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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
a0049d49-be99-4d08-9862-7819ef9de820 | cross-label-suppression-a-discriminative-and | 1705.02928 | null | http://arxiv.org/abs/1705.02928v1 | http://arxiv.org/pdf/1705.02928v1.pdf | Cross-label Suppression: A Discriminative and Fast Dictionary Learning with Group Regularization | This paper addresses image classification through learning a compact and
discriminative dictionary efficiently. Given a structured dictionary with each
atom (columns in the dictionary matrix) related to some label, we propose
cross-label suppression constraint to enlarge the difference among
representations for different classes. Meanwhile, we introduce group
regularization to enforce representations to preserve label properties of
original samples, meaning the representations for the same class are encouraged
to be similar. Upon the cross-label suppression, we don't resort to
frequently-used $\ell_0$-norm or $\ell_1$-norm for coding, and obtain
computational efficiency without losing the discriminative power for
categorization. Moreover, two simple classification schemes are also developed
to take full advantage of the learnt dictionary. Extensive experiments on six
data sets including face recognition, object categorization, scene
classification, texture recognition and sport action categorization are
conducted, and the results show that the proposed approach can outperform lots
of recently presented dictionary algorithms on both recognition accuracy and
computational efficiency. | ['Yuantao Gu', 'Xiudong Wang'] | 2017-05-08 | null | null | null | null | ['object-categorization'] | ['computer-vision'] | [ 3.98546487e-01 -2.12872982e-01 -4.55893964e-01 -5.27606726e-01
-2.72673637e-01 -1.53058350e-01 2.37398803e-01 2.89783895e-01
-3.73050392e-01 6.19133830e-01 1.49953932e-01 1.26457468e-01
-1.84314206e-01 -7.29511380e-01 -1.80183560e-01 -9.82635677e-01
6.72350675e-02 2.49535721e-02 -1.70147613e-01 2.98635103e-02
3.56674165e-01 4.06083614e-01 -1.83014333e+00 1.47591025e-01
8.48553896e-01 1.47981715e+00 9.07758176e-02 -3.29773873e-01
-8.59443173e-02 8.47044468e-01 -3.29571754e-01 -3.43404673e-02
4.01168019e-01 -5.98242164e-01 -3.87395263e-01 7.11064935e-01
3.73479038e-01 1.28432497e-01 -4.90480423e-01 1.46566939e+00
4.27729309e-01 4.72401679e-01 7.58412957e-01 -8.89374912e-01
-8.00082743e-01 1.49415463e-01 -9.16938961e-01 1.90908536e-01
2.30943203e-01 -3.23033094e-01 8.95886362e-01 -9.71985459e-01
5.11414349e-01 1.17129910e+00 4.51534986e-01 2.77307570e-01
-1.14150941e+00 -8.32343876e-01 4.02734816e-01 3.85888219e-01
-1.78840065e+00 -5.06445825e-01 1.00000751e+00 -4.39184219e-01
4.35706586e-01 3.66157681e-01 4.47453916e-01 6.16757035e-01
1.46445967e-02 6.19082630e-01 1.28883553e+00 -4.94069576e-01
1.62945926e-01 2.02430576e-01 3.65847766e-01 1.00398743e+00
1.68742269e-01 -1.05567031e-01 -4.91956472e-01 -6.36074319e-02
5.52269638e-01 1.70216694e-01 -4.46187317e-01 -5.14483631e-01
-1.03227174e+00 9.86664355e-01 2.41301462e-01 4.88155037e-01
-3.36708248e-01 -2.48359948e-01 5.30257106e-01 3.86319160e-01
3.29876781e-01 -6.82999566e-02 1.57054558e-01 2.77410716e-01
-6.37693703e-01 -2.42372066e-01 3.44622791e-01 1.01834047e+00
9.62335646e-01 2.54506677e-01 -1.30714566e-01 1.37607574e+00
2.82204807e-01 4.74897414e-01 6.94004357e-01 -7.37721205e-01
2.60280848e-01 8.01036298e-01 -2.72786498e-01 -1.75507069e+00
-2.14255154e-01 -4.93283957e-01 -1.23717046e+00 -8.73345509e-02
1.44845635e-01 3.38907838e-01 -7.79368520e-01 1.65218961e+00
2.49102995e-01 2.43814215e-01 -6.88821673e-02 1.04241371e+00
8.07160974e-01 6.91210628e-01 1.92149699e-01 -7.52930164e-01
1.24722588e+00 -8.23097289e-01 -8.02062988e-01 -1.26524225e-01
6.55714154e-01 -7.99172044e-01 9.60257649e-01 4.71572876e-01
-7.27211833e-01 -9.57512021e-01 -1.02180684e+00 4.13817391e-02
-1.35347649e-01 4.53198224e-01 7.91951597e-01 6.22560620e-01
-3.44023317e-01 2.89781451e-01 -4.52903569e-01 -2.09179789e-01
4.05385494e-01 4.12630588e-01 -5.32283366e-01 -4.96981055e-01
-9.79158282e-01 4.76456463e-01 5.02531707e-01 4.83664759e-02
-6.17240012e-01 -2.42641136e-01 -9.16773796e-01 -1.91908494e-01
2.64280438e-01 -3.77974063e-02 4.42873746e-01 -1.11000299e+00
-1.14578271e+00 1.15207219e+00 -2.39722669e-01 5.00977300e-02
-1.03912257e-01 1.92572013e-01 -7.60600209e-01 2.21271232e-01
2.23880291e-01 5.12350738e-01 9.90833819e-01 -1.17150247e+00
-5.71016848e-01 -3.98317963e-01 -8.60559568e-02 3.22323829e-01
-9.28578615e-01 -1.89882770e-01 -4.70429659e-01 -1.04791963e+00
6.56796336e-01 -7.55370200e-01 -8.83163884e-02 1.10192440e-01
-1.01837181e-01 -2.08568349e-01 9.79944050e-01 -6.38580620e-01
1.39695489e+00 -2.50484610e+00 3.87653112e-01 5.63269734e-01
-1.33124357e-02 1.30391419e-01 -2.21860901e-01 2.09048107e-01
-2.36477345e-01 -4.12147224e-01 -3.97933722e-01 -2.16794387e-01
-1.54558644e-01 6.61653280e-01 -2.62066036e-01 7.49923646e-01
-1.52565688e-01 1.27485245e-01 -7.22829998e-01 -6.63355768e-01
1.86905295e-01 2.42057309e-01 -5.23845017e-01 6.26569316e-02
2.46019587e-01 4.82796967e-01 -7.00078905e-01 7.69978523e-01
7.58839965e-01 -5.96983843e-02 3.32840770e-01 -5.10941148e-01
7.24579766e-02 -6.12084679e-02 -1.68366063e+00 1.77935827e+00
-3.92820209e-01 1.17539719e-01 2.24534109e-01 -1.81496501e+00
1.27503073e+00 4.55362611e-02 5.78748047e-01 -9.52029288e-01
2.60622352e-01 2.51141191e-01 -5.82947843e-02 -3.50181401e-01
2.41804644e-01 -2.21266672e-01 5.58514372e-02 3.76170605e-01
-4.62675816e-05 2.97376007e-01 3.36452425e-01 -1.46676734e-01
5.75005174e-01 -3.49509031e-01 2.91357428e-01 -7.35565364e-01
1.11239445e+00 -1.38017699e-01 1.02987599e+00 4.42686617e-01
-3.82025205e-02 3.49109441e-01 6.23881482e-02 -5.04773736e-01
-5.82915008e-01 -5.68562388e-01 -4.24819648e-01 1.19522619e+00
5.74937761e-01 -3.37080449e-01 -3.49712849e-01 -7.70195961e-01
5.59560210e-02 3.77794117e-01 -5.53046048e-01 -4.12472934e-01
-4.86733973e-01 -7.79454350e-01 2.01526955e-01 3.90858948e-01
7.35358775e-01 -9.08612907e-01 -1.46665230e-01 1.39049396e-01
-7.07068294e-02 -7.64651477e-01 -6.48307621e-01 1.55607536e-01
-8.38141739e-01 -1.03812027e+00 -6.17711425e-01 -1.34817076e+00
1.02163446e+00 5.83354712e-01 5.47209799e-01 2.93102860e-01
-3.30874264e-01 3.74030024e-01 -6.34447396e-01 1.65058807e-01
-2.18983516e-02 -2.21794441e-01 4.09514338e-01 5.01965642e-01
4.49657351e-01 -5.97961426e-01 -5.00969589e-01 5.75346947e-01
-9.59925234e-01 -2.07589954e-01 4.21096414e-01 1.11795092e+00
1.02255893e+00 3.42862278e-01 6.02754772e-01 -7.31997073e-01
4.52432305e-01 -3.75464708e-01 -3.55345041e-01 1.99182436e-01
-6.88450933e-01 -7.75375962e-02 6.99009895e-01 -6.10360086e-01
-7.64874935e-01 -2.26090103e-02 1.90490708e-02 -5.67200840e-01
6.48393109e-02 6.46694958e-01 -3.66986752e-01 -3.33870620e-01
4.61185157e-01 7.43478954e-01 1.04932390e-01 -6.27843261e-01
2.61793524e-01 7.40232050e-01 4.05273855e-01 -7.40546048e-01
5.75926304e-01 5.87540925e-01 9.98397693e-02 -8.68737757e-01
-8.32355261e-01 -4.70215231e-01 -5.77304721e-01 -8.46016183e-02
6.38910234e-01 -1.04564941e+00 -5.13000190e-01 3.75647247e-01
-6.44895792e-01 3.98590744e-01 -3.59600961e-01 7.27315009e-01
-3.07218820e-01 8.97208869e-01 -3.54045928e-01 -5.70067823e-01
-7.64798447e-02 -1.14661360e+00 7.87193239e-01 2.38905847e-01
5.95629029e-02 -8.75655591e-01 -2.45293915e-01 3.83305848e-01
1.12787418e-01 7.88136013e-03 1.06532001e+00 -5.33836782e-01
-1.75013497e-01 -2.77025223e-01 -2.85016060e-01 7.99154639e-01
5.30954719e-01 -5.17204225e-01 -6.78339958e-01 -7.33965695e-01
2.71526486e-01 -5.52891612e-01 9.39754903e-01 8.54133070e-02
1.56260693e+00 -3.26689512e-01 -3.13706964e-01 6.22561932e-01
1.17906678e+00 5.25907993e-01 5.68725348e-01 1.43992290e-01
8.11022878e-01 7.12368131e-01 7.52814829e-01 6.59980476e-01
-1.22340791e-01 6.18266225e-01 9.75094885e-02 -1.83848720e-02
-7.06685409e-02 -7.75946900e-02 2.44141370e-01 1.30721521e+00
-6.51107132e-02 7.30306283e-02 -4.52505261e-01 3.80823344e-01
-1.61006498e+00 -9.06347454e-01 1.32582132e-02 2.22800136e+00
7.36107528e-01 -1.06766380e-01 -1.28379703e-01 4.08734500e-01
8.90598238e-01 4.44686651e-01 -5.16672134e-01 -1.98146299e-01
-3.65002543e-01 2.74822444e-01 1.91938773e-01 2.80064821e-01
-1.19005597e+00 7.37483203e-01 5.61579370e+00 1.46203792e+00
-1.21502733e+00 4.59389500e-02 6.83047831e-01 2.63807833e-01
-1.51822522e-01 -4.69672233e-02 -6.43128157e-01 5.30539989e-01
2.52898872e-01 -3.43036093e-02 3.16703916e-01 9.43160295e-01
1.72812827e-02 -1.23335429e-01 -9.31896627e-01 1.49276721e+00
3.58837813e-01 -1.07959986e+00 3.47348928e-01 5.03855273e-02
7.05302954e-01 -6.93109274e-01 2.60939330e-01 3.74720275e-01
-2.76082456e-01 -9.40724134e-01 5.84240615e-01 4.24100637e-01
8.04029465e-01 -8.48188400e-01 4.86209810e-01 3.39985669e-01
-1.42774558e+00 -2.14468837e-01 -9.26089346e-01 -8.56982470e-02
-2.64376551e-01 7.26740777e-01 8.66479501e-02 6.96750343e-01
5.02559423e-01 1.27365446e+00 -4.73622769e-01 6.36291742e-01
8.93676430e-02 4.79081780e-01 -1.34483248e-01 2.56701738e-01
2.04257086e-01 -8.21489036e-01 4.32630837e-01 1.08996427e+00
3.34201753e-01 5.30719697e-01 7.75811493e-01 2.59309232e-01
1.17254211e-02 6.64667666e-01 -5.74662685e-01 3.63958664e-02
4.21017408e-01 1.07162511e+00 -7.04963863e-01 -4.27682161e-01
-5.40310442e-01 1.01376724e+00 -1.00134071e-02 1.19748443e-01
-6.85151041e-01 -4.63473141e-01 6.01938665e-01 -1.42408818e-01
2.30332106e-01 -2.19772726e-01 -1.98652148e-01 -1.17632139e+00
7.87435099e-02 -1.03303683e+00 6.72648191e-01 -3.28686208e-01
-1.40307403e+00 4.19476599e-01 -5.40881120e-02 -1.62842691e+00
2.78560787e-01 -5.95842540e-01 -2.37011313e-01 6.60969794e-01
-1.38738549e+00 -9.20066476e-01 -2.98283726e-01 9.98964190e-01
4.75600809e-01 -5.70682108e-01 7.82925248e-01 7.43568897e-01
-7.88969278e-01 8.27711701e-01 5.28104484e-01 1.11124873e-01
6.48096263e-01 -5.23486316e-01 -8.57226968e-01 5.29001236e-01
2.41110563e-01 7.81758964e-01 2.42269590e-01 -4.79686260e-01
-1.32446110e+00 -1.11146367e+00 5.89666784e-01 3.19610238e-01
3.77967328e-01 -1.10616475e-01 -9.99601603e-01 3.23914260e-01
-3.98274045e-04 2.71362513e-01 1.01399696e+00 3.87603156e-02
-5.70637286e-01 -4.99518931e-01 -1.09823537e+00 3.15890789e-01
1.22945166e+00 -5.84267855e-01 -5.99707782e-01 5.77969253e-01
1.27687573e-01 -1.24650180e-01 -9.73328888e-01 4.85830873e-01
4.52295661e-01 -8.22084665e-01 9.18282926e-01 -6.03746414e-01
1.58163726e-01 -5.57977319e-01 -7.17465937e-01 -9.20602381e-01
-6.34725034e-01 -2.35157013e-02 2.79592186e-01 1.23589242e+00
-1.88044295e-01 -5.58665752e-01 6.61076069e-01 2.88162753e-02
-2.77645320e-01 -6.65365577e-01 -1.20561385e+00 -9.99313891e-01
-1.95027664e-01 -1.35057792e-01 3.11142981e-01 1.37462914e+00
-1.96283534e-02 3.21235150e-01 -6.60710692e-01 8.38228539e-02
7.30361700e-01 5.11001527e-01 3.47159773e-01 -1.13729072e+00
-2.80339271e-01 -2.77787715e-01 -7.46389389e-01 -1.24601364e+00
3.34757209e-01 -1.10304952e+00 -2.45369181e-01 -1.03252208e+00
3.56451601e-01 -7.88934588e-01 -7.05916584e-01 6.88617945e-01
6.52900785e-02 6.27642512e-01 6.97876215e-02 5.29268563e-01
-5.73925316e-01 8.89396369e-01 1.27122879e+00 -5.12156129e-01
2.42050793e-02 -3.10693502e-01 -6.93785846e-01 7.31744826e-01
4.13089812e-01 -4.07109797e-01 -6.38423920e-01 -4.30554956e-01
-3.87548298e-01 -3.94608900e-02 4.46618870e-02 -1.21233118e+00
-2.11283844e-03 -3.00333828e-01 4.04811084e-01 -1.89130336e-01
3.74430746e-01 -9.11173284e-01 2.36585185e-01 5.73799670e-01
-4.43078667e-01 -2.31494725e-01 4.73566018e-02 7.32648969e-01
-5.50294280e-01 -3.61809671e-01 1.22119153e+00 9.08930227e-02
-1.00282979e+00 3.34641576e-01 -3.03330839e-01 -1.80963829e-01
1.13686752e+00 -5.36680818e-01 1.55408442e-01 -1.51608407e-01
-9.44465220e-01 6.61644861e-02 3.05206716e-01 3.88036609e-01
7.09495664e-01 -1.76869643e+00 -4.74306196e-01 4.95272279e-01
3.46520424e-01 -4.15676624e-01 3.54976743e-01 7.50802636e-01
-2.22482771e-01 2.36800835e-01 -4.50932324e-01 -6.14132881e-01
-1.25337040e+00 6.95835412e-01 5.39835580e-02 -7.31268823e-02
-5.21350086e-01 9.66953814e-01 4.21181768e-01 -1.48747623e-01
2.86562622e-01 4.71838377e-03 -4.22853619e-01 3.42579782e-01
4.71531630e-01 3.19547236e-01 6.87000435e-03 -1.08173501e+00
-4.32887733e-01 9.99021351e-01 -2.29354367e-01 3.88820887e-01
1.12989712e+00 -1.67814597e-01 -3.69824588e-01 2.81870514e-01
1.41373050e+00 1.09812856e-01 -7.94813395e-01 -5.28843462e-01
-2.75457501e-01 -8.08770895e-01 8.24658871e-02 -3.23345929e-01
-1.46187866e+00 7.67551661e-01 8.96000683e-01 -8.04878324e-02
1.46678686e+00 -3.40203315e-01 6.65439487e-01 5.02871692e-01
5.79492629e-01 -1.26322329e+00 1.70110062e-01 3.52996737e-01
8.22813272e-01 -1.04006970e+00 2.34925106e-01 -6.41114771e-01
-5.73350489e-01 1.08537221e+00 5.54666519e-01 -2.84316033e-01
7.64116585e-01 -3.22054803e-01 -7.12192953e-02 -1.36510402e-01
-9.95495096e-02 -2.39580736e-01 3.41252238e-01 4.80558783e-01
3.02862316e-01 -1.46039888e-01 -9.07505214e-01 3.82199913e-01
2.22826898e-01 -3.08330804e-01 8.41447562e-02 9.19969559e-01
-6.18599832e-01 -1.30616188e+00 -4.58118290e-01 4.95578438e-01
-1.33660823e-01 2.40698919e-01 -4.44611497e-02 6.55972600e-01
5.90328276e-01 9.51605558e-01 1.87875368e-02 -5.44625342e-01
3.50932032e-01 7.74339363e-02 4.60562408e-01 -7.55077064e-01
-1.21932209e-01 2.14494362e-01 -1.50100440e-01 -3.97865176e-01
-6.95806742e-01 -7.15883791e-01 -1.23820138e+00 -1.93626866e-01
-2.40836754e-01 4.59900141e-01 1.57173246e-01 7.48172164e-01
2.08527684e-01 1.54751375e-01 9.68577802e-01 -4.49047327e-01
-5.74632943e-01 -8.09406936e-01 -1.21891439e+00 6.76756680e-01
-1.77098528e-01 -1.13538122e+00 -2.46539190e-01 3.48905087e-01] | [12.387375831604004, 0.42809247970581055] |
64832fb9-9154-4871-a668-d69d1e55285f | mrfusion-a-deep-learning-architecture-to-fuse | 1806.11452 | null | http://arxiv.org/abs/1806.11452v1 | http://arxiv.org/pdf/1806.11452v1.pdf | MRFusion: A Deep Learning architecture to fuse PAN and MS imagery for land cover mapping | Nowadays, Earth Observation systems provide a multitude of heterogeneous
remote sensing data. How to manage such richness leveraging its complementarity
is a crucial chal- lenge in modern remote sensing analysis. Data Fusion
techniques deal with this point proposing method to combine and exploit
complementarity among the different data sensors. Considering optical Very High
Spatial Resolution (VHSR) images, satellites obtain both Multi Spectral (MS)
and panchro- matic (PAN) images at different spatial resolution. VHSR images
are extensively exploited to produce land cover maps to deal with agricultural,
ecological, and socioeconomic issues as well as assessing ecosystem status,
monitoring biodiversity and provid- ing inputs to conceive food risk monitoring
systems. Common techniques to produce land cover maps from such VHSR images
typically opt for a prior pansharpening of the multi-resolution source for a
full resolution processing. Here, we propose a new deep learning architecture
to jointly use PAN and MS imagery for a direct classification without any prior
image fusion or resampling process. By managing the spectral information at its
native spatial resolution, our method, named MRFusion, aims at avoiding the
possible infor- mation loss induced by pansharpening or any other hand-crafted
preprocessing. Moreover, the proposed architecture is suitably designed to
learn non-linear transformations of the sources with the explicit aim of taking
as much as possible advantage of the complementarity of PAN and MS imagery.
Experiments are carried out on two-real world scenarios depicting large areas
with different land cover characteristics. The characteristics of the proposed
scenarios underline the applicability and the generality of our method in
operational settings. | ['Kenji Ose', 'Remi Cresson', 'Raffaele Gaetano', 'Dino Ienco'] | 2018-06-29 | null | null | null | null | ['pansharpening'] | ['computer-vision'] | [ 6.33592188e-01 -2.02373460e-01 -6.13655187e-02 -3.25401783e-01
-6.69233739e-01 -4.54266816e-01 7.76191890e-01 1.63725689e-01
-5.57706177e-01 1.00981343e+00 8.02515373e-02 -3.51133317e-01
-5.09849429e-01 -1.52022874e+00 -5.90359509e-01 -1.01813102e+00
-2.29827508e-01 3.60006690e-02 -3.72928649e-01 -6.39140606e-01
-2.01879159e-01 8.30094814e-01 -1.96364939e+00 1.61523130e-02
1.25227273e+00 1.01899052e+00 7.06388772e-01 5.49685597e-01
-1.23346485e-01 2.54339844e-01 -7.89133385e-02 1.97251886e-01
7.41042256e-01 -1.13498852e-01 -3.24735641e-01 3.60895216e-01
3.56131345e-01 -3.37667227e-01 7.01419190e-02 1.48368442e+00
3.33279431e-01 6.28195554e-02 3.95707518e-01 -7.71566927e-01
-2.75186062e-01 6.29508138e-01 -9.11429465e-01 1.36798084e-01
-2.46955305e-01 8.75064582e-02 7.52416313e-01 -5.87097526e-01
2.64193088e-01 1.04209411e+00 5.33787906e-01 -3.45048010e-01
-1.28473926e+00 -5.43551922e-01 1.01445265e-01 4.60206792e-02
-1.68377686e+00 -4.90131438e-01 6.66050553e-01 -4.32313681e-01
4.54059631e-01 6.02425456e-01 7.54562080e-01 5.46836734e-01
3.93923342e-01 2.68542826e-01 1.43908262e+00 -3.16221952e-01
1.96090993e-03 -3.18267122e-02 -3.57509293e-02 8.74898955e-02
5.18897176e-01 2.77060628e-01 1.20400935e-02 7.41477013e-02
7.35556483e-01 4.33430731e-01 -5.29501021e-01 -3.01478989e-02
-1.18336987e+00 6.51683092e-01 9.20148849e-01 7.08907723e-01
-1.04220545e+00 -2.50078321e-01 -1.38709858e-01 5.64421535e-01
3.98378938e-01 3.02566350e-01 -2.28545070e-01 6.23274326e-01
-1.17373765e+00 1.29325911e-01 9.94180962e-02 6.93602979e-01
1.24418938e+00 2.47551724e-01 3.25356811e-01 6.23776674e-01
4.49880749e-01 9.74477232e-01 3.77714276e-01 -5.08513331e-01
4.39444035e-01 7.17519820e-01 2.83866614e-01 -1.15666962e+00
-5.37952244e-01 -7.22907901e-01 -1.54244733e+00 3.85612786e-01
-7.35607296e-02 -1.58268929e-01 -8.98176670e-01 1.55488455e+00
2.87780762e-01 1.12604700e-01 4.17400241e-01 9.21652675e-01
3.57727706e-01 9.38755691e-01 3.22556645e-01 -3.20858121e-01
1.42943203e+00 -1.79959014e-01 -7.06619620e-01 -3.20039064e-01
1.77149266e-01 -5.42053819e-01 6.13745809e-01 1.14797294e-01
-4.74296004e-01 -6.98189855e-01 -1.35168970e+00 3.62971038e-01
-7.12622404e-01 2.44628951e-01 6.22244596e-01 4.14088696e-01
-9.94901657e-01 5.69824040e-01 -6.71886683e-01 -3.30336779e-01
3.57575446e-01 1.98459253e-01 -5.31892121e-01 1.41631797e-01
-1.40026498e+00 9.53911245e-01 7.92938113e-01 7.70090759e-01
-5.33493638e-01 -5.80404699e-01 -7.54032314e-01 2.70167559e-01
1.86700478e-01 -3.45399946e-01 2.99945086e-01 -1.34147763e+00
-9.91434038e-01 6.31885052e-01 3.21224988e-01 -7.22495615e-01
3.82726461e-01 -1.47004306e-01 -8.53196204e-01 -9.49529782e-02
5.29282615e-02 3.64823788e-01 1.04777539e+00 -1.22115505e+00
-8.36771131e-01 -6.98822737e-01 -1.04247950e-01 5.44489861e-01
-2.52148569e-01 -3.35876167e-01 5.46972454e-01 -6.61737978e-01
4.72706825e-01 -6.31961584e-01 -4.03065264e-01 -1.47460938e-01
1.07695967e-01 7.15739846e-01 7.16782928e-01 -6.44819677e-01
8.50793123e-01 -2.15794230e+00 1.94982305e-01 3.10166359e-01
1.38650443e-02 4.31463599e-01 -3.43022555e-01 4.44866687e-01
-1.49542555e-01 7.50539228e-02 -7.63700724e-01 2.36488149e-01
-3.62207025e-01 2.17207655e-01 -5.48337162e-01 6.71550035e-01
2.89397866e-01 4.89812255e-01 -5.58434427e-01 -3.24666560e-01
7.14182079e-01 7.16719806e-01 2.21812516e-01 2.30264485e-01
-1.45843938e-01 5.70660830e-01 -5.38439989e-01 5.10752261e-01
1.31996822e+00 2.12272078e-01 2.39135176e-01 -3.75708133e-01
-6.46493375e-01 -3.86320591e-01 -1.49839389e+00 1.28897095e+00
-5.93088925e-01 3.41504306e-01 6.10142469e-01 -1.04521143e+00
1.20013559e+00 2.09562957e-01 4.76604164e-01 -7.16832340e-01
1.19904310e-01 3.62499684e-01 -1.06439032e-01 -2.58986264e-01
6.28982186e-01 -3.93124938e-01 1.26497969e-01 1.79837216e-02
-2.17914835e-01 -1.68976218e-01 -2.13451877e-01 -4.48946476e-01
3.72131526e-01 4.08572778e-02 7.78515637e-01 -4.33893532e-01
9.81268108e-01 1.45360291e-01 3.77328098e-01 5.71640670e-01
-4.10984159e-02 2.94315368e-01 -1.47891283e-01 -6.61836803e-01
-1.00111985e+00 -7.32812643e-01 -4.58036631e-01 7.46713102e-01
1.09602034e-01 5.05875766e-01 -1.93892360e-01 -7.30202347e-02
1.21806480e-01 5.76963127e-01 -5.87523222e-01 1.42870694e-01
-2.85672009e-01 -1.36193621e+00 4.64666188e-01 1.02334432e-02
1.05973351e+00 -8.57612967e-01 -9.88812089e-01 4.04821694e-01
-1.51018903e-01 -7.82226562e-01 5.03965437e-01 2.56891310e-01
-9.83665347e-01 -7.72055566e-01 -6.80327952e-01 -9.20217261e-02
1.23525478e-01 6.86987460e-01 7.45086670e-01 -1.66951463e-01
-3.03712748e-02 -1.86351463e-01 -4.28837806e-01 -1.80945456e-01
-4.29429293e-01 1.98755175e-01 -1.00251503e-01 5.80552459e-01
2.54412532e-01 -1.00776505e+00 -4.00485456e-01 4.91640866e-02
-1.28175056e+00 2.10947856e-01 1.11478496e+00 6.10654354e-01
5.94442129e-01 4.27046299e-01 7.00617433e-01 -4.53434825e-01
1.16001345e-01 -7.14358270e-01 -1.01546729e+00 4.78906006e-01
-3.98488760e-01 6.10155016e-02 3.62354815e-01 8.25179666e-02
-1.21794939e+00 2.80807406e-01 -1.65058300e-01 -3.95394303e-02
-5.87230444e-01 9.25326705e-01 -4.88252163e-01 -2.13573426e-01
5.81460893e-01 4.52999473e-01 -4.68795486e-02 -5.72290242e-01
3.94801736e-01 9.49965000e-01 6.50040448e-01 -1.17064260e-01
1.05516386e+00 7.24703968e-01 2.25199446e-01 -1.30121791e+00
-4.90854442e-01 -5.07330954e-01 -8.33151698e-01 -1.23472534e-01
8.05426538e-01 -1.32945418e+00 -2.64724791e-01 6.24699414e-01
-7.56339192e-01 1.22459173e-01 -1.47695228e-01 6.06007814e-01
-3.24172020e-01 3.07398826e-01 1.41496837e-01 -8.92523348e-01
-5.39982319e-01 -8.97358656e-01 8.03048432e-01 3.41541916e-01
5.08763373e-01 -6.80331588e-01 6.09013699e-02 1.10220954e-01
8.01703811e-01 5.80193102e-01 6.19500458e-01 -3.49601269e-01
-7.90432274e-01 -5.58406599e-02 -6.16804242e-01 2.90040433e-01
5.85471034e-01 -2.30711564e-01 -9.51165795e-01 -3.63360286e-01
2.42428079e-01 3.62272151e-02 1.10623705e+00 3.86339664e-01
6.71237767e-01 -2.85015762e-01 1.32219438e-02 8.48376334e-01
2.02198076e+00 -2.39373390e-02 7.21216023e-01 5.20914674e-01
6.17971659e-01 6.53245509e-01 7.80520082e-01 5.47809601e-01
1.66384548e-01 3.83132160e-01 9.39464033e-01 -2.95772225e-01
1.53454199e-01 1.38535827e-01 1.39307857e-01 1.65848255e-01
-3.59456807e-01 -4.33060266e-02 -8.89908612e-01 5.81276596e-01
-1.73312378e+00 -1.19684350e+00 -9.69760045e-02 2.35052180e+00
4.47719395e-01 -3.71833533e-01 -4.05611724e-01 1.94537506e-01
8.48255992e-01 7.67761588e-01 -3.90883088e-01 2.71472603e-01
-7.16171563e-01 1.87269688e-01 1.22029662e+00 5.35834730e-01
-1.49577487e+00 7.82092869e-01 4.38342381e+00 5.44063151e-01
-1.46124327e+00 -2.72215605e-02 1.17739163e-01 4.29148555e-01
-3.77063870e-01 -1.40087297e-02 -5.76665640e-01 8.88878033e-02
9.69481468e-01 -3.16301547e-02 4.19057995e-01 4.27100152e-01
3.88496011e-01 -2.96552658e-01 -2.61808723e-01 7.79860139e-01
-4.05721962e-01 -1.24113441e+00 1.87714428e-01 2.77044177e-01
5.90670645e-01 3.16075236e-01 -3.24538499e-02 -4.22960930e-02
3.76630098e-01 -8.24028611e-01 5.01562715e-01 7.61301994e-01
7.05725253e-01 -7.34474599e-01 8.50786686e-01 4.08828557e-01
-1.51854980e+00 -3.31247836e-01 -4.44833189e-01 -1.09157242e-01
1.25040770e-01 6.85419023e-01 -2.87947387e-01 1.27351189e+00
4.69388545e-01 8.94732654e-01 -4.72346157e-01 8.52956653e-01
-7.22031519e-02 1.70492485e-01 -5.32241404e-01 6.47748172e-01
4.65589046e-01 -5.18387973e-01 7.36088753e-01 8.47217143e-01
8.24744642e-01 3.53644699e-01 1.31359339e-01 7.41601944e-01
2.42725745e-01 1.74416751e-01 -7.43236601e-01 -1.11623459e-01
4.72867310e-01 1.33390307e+00 -4.18724567e-01 -2.85398841e-01
-1.88247189e-01 5.14848769e-01 -1.43791169e-01 3.74899864e-01
-4.71466690e-01 -1.78164124e-01 7.41764069e-01 8.99867192e-02
1.81010887e-01 -4.05556679e-01 -2.90718287e-01 -1.21662045e+00
-2.43223056e-01 -8.37811112e-01 2.14430198e-01 -6.07248664e-01
-8.45259666e-01 7.75444388e-01 2.09670737e-01 -1.62395978e+00
-4.32796702e-02 -2.48161033e-01 -2.79291570e-01 1.35503387e+00
-2.22718620e+00 -1.51605797e+00 -7.95155883e-01 4.86630261e-01
1.62987068e-01 -1.51682436e-01 7.69447684e-01 1.85128555e-01
-3.15878361e-01 -3.77524465e-01 3.51181686e-01 -3.07240844e-01
3.74699771e-01 -9.13580179e-01 1.00490667e-01 1.14612877e+00
-1.20493762e-01 1.70535117e-01 7.78095484e-01 -7.08325803e-01
-1.32405579e+00 -1.46355534e+00 5.61409354e-01 5.21454513e-01
5.63608944e-01 2.83043563e-01 -1.10143673e+00 6.60128057e-01
6.88421205e-02 4.59935963e-02 3.12718898e-01 -1.74285680e-01
-1.15828447e-01 -6.98380828e-01 -1.24767351e+00 2.18171492e-01
4.51961488e-01 -3.46332133e-01 -4.57508773e-01 1.41882807e-01
3.96449089e-01 1.03016175e-01 -1.03202212e+00 8.15132499e-01
6.13710642e-01 -1.10988021e+00 9.54337180e-01 -9.55167264e-02
2.73853481e-01 -4.77627784e-01 -7.96018302e-01 -1.38808978e+00
-4.09033924e-01 -4.50153947e-02 5.85794449e-01 1.03021657e+00
5.29442243e-02 -8.15380871e-01 1.56757072e-01 4.65909652e-02
8.59013870e-02 1.63899511e-01 -8.66629303e-01 -5.01493871e-01
6.30093068e-02 -1.81482345e-01 1.10336161e+00 1.06721413e+00
-7.18550324e-01 -1.02718107e-01 -4.92436469e-01 9.92920101e-01
8.20087492e-01 3.89409840e-01 7.14673579e-01 -1.64814830e+00
1.95638891e-02 -4.18674171e-01 -2.92072833e-01 -3.86825174e-01
-2.55846232e-02 -5.20409226e-01 -1.17332533e-01 -1.29167640e+00
-2.21598983e-01 -6.81448340e-01 -4.23509717e-01 5.71199000e-01
-5.98386079e-02 2.15603426e-01 9.84616578e-02 3.84475648e-01
3.64648998e-01 7.65240133e-01 8.49106312e-01 -2.90529519e-01
-4.04087752e-01 5.52556515e-02 -4.30093884e-01 5.88866889e-01
7.61039376e-01 -2.02242911e-01 -2.38575146e-01 -5.78097641e-01
2.76910484e-01 4.00001764e-01 6.38188064e-01 -1.35544491e+00
-5.75971305e-02 -5.15162468e-01 2.37859666e-01 -8.99960518e-01
1.36277810e-01 -1.33370006e+00 8.95477891e-01 3.10228080e-01
-5.27280010e-02 -3.11653882e-01 2.85655349e-01 4.06035513e-01
-4.32657003e-01 -1.84626393e-02 1.15852892e+00 -2.70638913e-01
-1.06025553e+00 2.76305169e-01 -4.04715985e-01 -9.03251112e-01
1.01411843e+00 -1.27185822e-01 -2.37077087e-01 -8.60998109e-02
-6.48382246e-01 2.21666515e-01 4.52869773e-01 3.93308610e-01
3.56347084e-01 -1.05321443e+00 -1.01129031e+00 6.71619773e-01
2.31233731e-01 -1.43806577e-01 7.18689322e-01 6.51557863e-01
-5.59039831e-01 3.28298301e-01 -8.51827502e-01 -4.44806635e-01
-1.01020563e+00 4.58249688e-01 5.79004645e-01 -2.57790536e-01
-4.92372394e-01 2.54304528e-01 -9.47608575e-02 -4.82388973e-01
-5.39520442e-01 -2.83593893e-01 -5.51581979e-01 5.40770948e-01
6.36144519e-01 1.29574835e-01 2.69032091e-01 -1.24144375e+00
-2.14663446e-01 6.02941573e-01 3.44868124e-01 -1.58697754e-01
1.56365108e+00 -6.27355158e-01 -3.04712176e-01 3.68721098e-01
8.23645175e-01 -2.63227761e-01 -1.09473264e+00 -5.70270419e-01
-8.24397355e-02 -6.81422710e-01 7.13042617e-01 -7.20771730e-01
-1.21192873e+00 9.16553855e-01 9.15273726e-01 2.61328429e-01
1.46929526e+00 -5.21702826e-01 3.03617746e-01 4.61518675e-01
4.63709682e-01 -7.56390095e-01 -1.02565610e+00 7.55530223e-02
1.01734746e+00 -1.43541098e+00 3.33819509e-01 -1.36780679e-01
-3.65272105e-01 1.07257533e+00 9.42390859e-02 -5.32813109e-02
7.64270842e-01 8.94992277e-02 9.91428792e-02 -7.44069517e-02
-1.94103375e-01 -7.46979594e-01 8.66245180e-02 4.84607279e-01
5.27443029e-02 5.68686128e-01 -1.44801691e-01 1.01085685e-01
1.34466924e-02 1.86237887e-01 5.06744742e-01 8.50973010e-01
-9.08866465e-01 -8.11392486e-01 -8.95107687e-01 4.61368829e-01
-6.60573244e-02 -1.35137022e-01 2.27818266e-01 6.86696053e-01
1.80671602e-01 8.11527312e-01 6.56349165e-03 -2.51586497e-01
8.59114602e-02 -2.16629073e-01 1.02156430e-01 -2.66457081e-01
-3.92189085e-01 2.31526822e-01 -2.59952039e-01 -3.29832524e-01
-9.24638808e-01 -6.73769772e-01 -7.01083601e-01 -3.61463398e-01
-1.93659127e-01 -5.34187956e-03 8.87414753e-01 1.07363009e+00
1.89141154e-01 3.16387713e-01 7.47300506e-01 -1.23901594e+00
-5.54250419e-01 -1.13023686e+00 -1.10636461e+00 -1.35272229e-02
9.22507524e-01 -7.20189631e-01 -2.83475220e-01 -1.86420336e-01] | [9.772173881530762, -1.7349746227264404] |
976c6a3a-a89d-46bc-8ba7-9c7dc830c43d | nestfuse-an-infrared-and-visible-image-fusion | 2007.00328 | null | https://arxiv.org/abs/2007.00328v2 | https://arxiv.org/pdf/2007.00328v2.pdf | NestFuse: An Infrared and Visible Image Fusion Architecture based on Nest Connection and Spatial/Channel Attention Models | In this paper we propose a novel method for infrared and visible image fusion where we develop nest connection-based network and spatial/channel attention models. The nest connection-based network can preserve significant amounts of information from input data in a multi-scale perspective. The approach comprises three key elements: encoder, fusion strategy and decoder respectively. In our proposed fusion strategy, spatial attention models and channel attention models are developed that describe the importance of each spatial position and of each channel with deep features. Firstly, the source images are fed into the encoder to extract multi-scale deep features. The novel fusion strategy is then developed to fuse these features for each scale. Finally, the fused image is reconstructed by the nest connection-based decoder. Experiments are performed on publicly available datasets. These exhibit that our proposed approach has better fusion performance than other state-of-the-art methods. This claim is justified through both subjective and objective evaluation. The code of our fusion method is available at https://github.com/hli1221/imagefusion-nestfuse | ['Xiao-Jun Wu', 'Tariq Durrani', 'Hui Li'] | 2020-07-01 | null | null | null | null | ['infrared-and-visible-image-fusion'] | ['computer-vision'] | [ 1.52015135e-01 -3.71169239e-01 1.22395800e-02 -1.99190587e-01
-8.63191307e-01 -2.08945096e-01 4.33757842e-01 1.72611419e-02
-2.81865358e-01 6.98949039e-01 4.47109193e-01 -8.35942402e-02
-1.96046960e-02 -8.13338697e-01 -8.23108852e-01 -7.46418774e-01
2.82201201e-01 -5.31893730e-01 -2.59447191e-02 -1.41957089e-01
1.33559018e-01 3.75915676e-01 -1.41158116e+00 3.31731021e-01
8.72224152e-01 1.56298959e+00 6.17942333e-01 8.10934305e-01
6.48149252e-02 8.40746343e-01 -1.64905548e-01 -3.30864906e-01
4.64161366e-01 -4.43766087e-01 -5.10496259e-01 -5.52303754e-02
1.67069569e-01 -4.81069475e-01 -6.03811800e-01 1.38494921e+00
7.51684606e-01 -1.50169795e-02 3.43730390e-01 -1.12292016e+00
-1.01611602e+00 3.82651836e-01 -7.83675015e-01 5.19939601e-01
2.85940588e-01 2.64055561e-02 8.06079865e-01 -9.60457563e-01
1.69649333e-01 1.07383609e+00 5.36591411e-01 1.05839305e-01
-8.05539370e-01 -7.91412473e-01 5.54032959e-02 3.29998761e-01
-1.46530139e+00 -4.52238172e-01 8.36160541e-01 -2.44865805e-01
7.13889420e-01 1.94265932e-01 6.11175001e-01 6.78888798e-01
5.69397092e-01 8.06815922e-01 1.11294067e+00 -4.60732430e-01
-1.14881821e-01 -4.99941558e-02 2.17001350e-03 6.41140580e-01
8.09561536e-02 3.51993680e-01 -5.49140573e-01 1.63522258e-01
9.29467380e-01 4.08773452e-01 -5.47357678e-01 1.18960440e-01
-1.24002934e+00 6.63838625e-01 1.02090704e+00 4.09114510e-01
-7.97967851e-01 2.94925749e-01 1.25795260e-01 3.24975282e-01
5.91297090e-01 -2.16606036e-01 -3.37992877e-01 3.20913792e-01
-7.71301985e-01 -1.33937925e-01 2.63570070e-01 8.71398628e-01
9.21751320e-01 -2.20475703e-01 -3.89631450e-01 6.69877172e-01
5.64633012e-01 5.94104826e-01 2.21541539e-01 -8.63147974e-01
5.12989104e-01 3.67287427e-01 -2.06515267e-02 -9.99592006e-01
-2.45173842e-01 -5.72009027e-01 -1.18061018e+00 2.98799723e-01
-3.20897758e-01 -3.80757332e-01 -9.71605301e-01 1.51251984e+00
1.59120113e-01 4.02708948e-01 2.32771173e-01 9.64980364e-01
1.22597265e+00 8.81173551e-01 9.18469876e-02 -7.05556944e-02
1.47626066e+00 -1.10840070e+00 -9.18718040e-01 8.91432539e-03
1.33075178e-01 -9.61841047e-01 4.26728129e-01 1.11428715e-01
-1.27377570e+00 -9.70239758e-01 -1.01563489e+00 -1.47624403e-01
-3.26811761e-01 5.55300713e-01 4.51151341e-01 2.93766677e-01
-1.33291757e+00 3.74544561e-01 -5.54967701e-01 -4.12188590e-01
4.79205638e-01 4.08636451e-01 -3.83074701e-01 -1.16697997e-01
-1.18729424e+00 7.58384764e-01 3.99495691e-01 2.75794983e-01
-6.63739562e-01 -2.89950073e-01 -9.21363175e-01 1.82595149e-01
2.88874190e-02 -1.07762253e+00 1.14774799e+00 -9.62086499e-01
-1.14818919e+00 3.47722501e-01 -2.45922551e-01 -3.82272065e-01
-5.82793448e-03 -1.80509701e-01 -4.90733087e-01 2.88309395e-01
4.26189564e-02 9.07537103e-01 8.50673318e-01 -1.36829317e+00
-9.99342203e-01 -2.42312908e-01 1.84740230e-01 6.18002057e-01
-2.93183625e-01 1.02444835e-01 -6.61314368e-01 -6.11194670e-01
1.03168637e-01 -4.71278697e-01 -3.93315032e-02 1.66355327e-01
-3.80028337e-01 5.25937229e-02 7.95457900e-01 -8.73686492e-01
1.10819173e+00 -2.34772539e+00 1.03494741e-01 -1.07658664e-02
3.82911056e-01 3.13769937e-01 -1.28177524e-01 4.32152450e-01
-1.89171016e-01 4.55107167e-02 -2.34210104e-01 -4.57888335e-01
-2.69700587e-01 -2.96935052e-01 -1.11282282e-01 5.32633960e-01
-1.83457360e-02 1.03301024e+00 -6.50910437e-01 -4.43277717e-01
7.70944417e-01 1.03247988e+00 -2.98886150e-01 3.85227174e-01
2.03005001e-01 6.88981950e-01 -3.66660565e-01 7.53504753e-01
9.98900950e-01 -2.06643999e-01 -3.66335809e-01 -8.06276858e-01
-3.38781327e-01 -6.57683760e-02 -8.58149946e-01 2.00609374e+00
-4.86457944e-01 5.15621185e-01 1.66730151e-01 -7.92121172e-01
8.59319150e-01 5.56909680e-01 3.63236517e-01 -8.69175971e-01
5.31787515e-01 -3.37247588e-02 -2.70730466e-01 -4.41584051e-01
5.66231668e-01 -1.02276914e-01 1.07849374e-01 7.79892802e-02
2.41193682e-01 1.82470486e-01 -1.53141677e-01 2.04887971e-01
8.89002979e-01 3.39623764e-02 3.03154409e-01 5.94245270e-02
7.76693046e-01 -4.38755006e-01 4.54413146e-01 4.81668979e-01
-3.11912954e-01 6.67188644e-01 -6.44514561e-02 -2.52825946e-01
-1.01925445e+00 -8.14786732e-01 1.07550219e-01 6.00853443e-01
4.92717981e-01 -1.81147963e-01 -4.83671963e-01 -4.04468238e-01
-1.53130978e-01 3.68416011e-01 -6.74173713e-01 -1.93179578e-01
2.99986601e-02 -4.48219687e-01 1.80549443e-01 6.50460899e-01
1.05613911e+00 -1.07215285e+00 -5.67764938e-01 2.89407186e-02
-4.82144505e-01 -1.00933862e+00 -4.66614902e-01 3.39541882e-02
-6.99521601e-01 -9.76261556e-01 -9.05332088e-01 -7.62137413e-01
5.45297027e-01 8.49368870e-01 7.04831541e-01 2.16887429e-01
-1.51202947e-01 2.22494215e-01 -5.76910734e-01 -2.99073786e-01
1.13958202e-01 -2.44494811e-01 -5.22940636e-01 3.56331468e-01
2.87394643e-01 -5.88289618e-01 -1.05393887e+00 -1.53499424e-01
-8.93793643e-01 3.39082450e-01 9.60016191e-01 6.94167018e-01
5.78159273e-01 2.57456750e-01 2.84619510e-01 -2.92215943e-01
5.89092374e-01 -6.66920304e-01 -5.33715963e-01 1.57943994e-01
-2.01256067e-01 -7.83821121e-02 4.05426294e-01 1.73814937e-01
-1.32722211e+00 1.92338929e-01 -2.85177588e-01 -5.64296186e-01
-1.98836371e-01 5.12423694e-01 -1.67948991e-01 -1.96335033e-01
2.29770795e-01 4.42749083e-01 -7.58581534e-02 -6.21608615e-01
4.89308953e-01 9.35592115e-01 7.16292679e-01 -1.10480279e-01
7.36214936e-01 6.61076307e-01 -4.25804794e-01 -4.77022082e-01
-6.91825509e-01 -5.10141969e-01 -6.32944822e-01 -2.80433357e-01
1.14967442e+00 -1.36400294e+00 -7.03592777e-01 6.65833950e-01
-1.27745891e+00 -8.21681041e-03 -1.18474804e-01 6.31016016e-01
-4.80464488e-01 2.91007817e-01 -6.54113233e-01 -7.65249789e-01
-7.04933345e-01 -1.24795938e+00 1.15050960e+00 5.45829475e-01
4.38080609e-01 -9.80288148e-01 -6.01490699e-02 3.09345573e-01
5.96073866e-01 1.23858877e-01 3.16924900e-01 -7.68834651e-02
-6.61535084e-01 -1.41333371e-01 -7.22962856e-01 4.66122001e-01
2.26417229e-01 -1.68904260e-01 -1.10177326e+00 -2.38153487e-01
1.00241475e-01 -1.54510317e-02 1.34103286e+00 6.59800589e-01
1.26122510e+00 -9.91115440e-03 -3.44387174e-01 9.64314401e-01
1.81281614e+00 2.90150017e-01 8.69743407e-01 2.88164943e-01
7.06895351e-01 1.09901778e-01 5.02570093e-01 3.83663595e-01
5.14985204e-01 5.03509820e-01 7.61849105e-01 -7.16946244e-01
-3.20132554e-01 -1.12591907e-01 2.32655704e-01 6.72512472e-01
-1.61506593e-01 -4.25913304e-01 -5.43153346e-01 5.43281555e-01
-1.98708665e+00 -1.06053090e+00 -6.08278625e-02 1.97126806e+00
3.99854273e-01 -2.93072820e-01 -1.98928699e-01 1.76493153e-01
8.07981968e-01 2.20247492e-01 -1.81435511e-01 -1.31685615e-01
-3.11830521e-01 1.33973062e-01 7.12812603e-01 5.88609576e-01
-1.30745852e+00 6.06598139e-01 5.66691971e+00 6.92255974e-01
-1.15781045e+00 3.94050002e-01 6.06576800e-01 1.73134759e-01
-2.85784543e-01 -4.38815802e-02 -4.95852947e-01 5.24442554e-01
8.22568297e-01 5.17720170e-02 3.38046163e-01 1.92199886e-01
2.10194007e-01 -2.87103444e-01 -4.32798654e-01 1.18285203e+00
2.21636742e-01 -1.33561409e+00 -1.83827549e-01 -8.86040647e-03
6.31521881e-01 2.31933981e-01 1.47780776e-01 -1.98489681e-01
2.83848792e-01 -9.22766507e-01 7.93382466e-01 9.08242762e-01
8.02133799e-01 -8.64310980e-01 1.02042508e+00 1.17423825e-01
-1.58755422e+00 -3.10193837e-01 -3.45013320e-01 1.28967360e-01
2.40844592e-01 6.47448421e-01 1.86506170e-03 1.13855553e+00
8.99190307e-01 1.01925445e+00 -5.03185689e-01 1.24766040e+00
-2.72419095e-01 1.70799226e-01 -1.02093212e-01 5.51890135e-01
1.03271887e-01 -5.85205890e-02 3.01895857e-01 1.08998120e+00
6.95662141e-01 2.97497600e-01 -1.13905117e-01 7.38504589e-01
-1.29722089e-01 -7.39615336e-02 -8.02673995e-01 3.77566159e-01
4.74559546e-01 1.54037309e+00 -4.88841385e-01 -4.55949783e-01
-8.57625484e-01 1.25976932e+00 8.38882178e-02 4.89541411e-01
-9.49927151e-01 -3.37611735e-01 4.27057952e-01 -3.19888324e-01
6.52797759e-01 -5.31583838e-02 -2.94832259e-01 -1.13411486e+00
-4.21446338e-02 -5.64195335e-01 3.42610925e-01 -1.38242185e+00
-1.09733391e+00 8.78858984e-01 -1.49124026e-01 -1.30648708e+00
4.90765065e-01 -2.45183036e-01 -5.85753560e-01 1.14230371e+00
-1.98424447e+00 -1.37798822e+00 -8.17427397e-01 8.68182898e-01
4.29493099e-01 -1.77697703e-01 5.56741834e-01 5.26512504e-01
-4.95613128e-01 2.93804914e-01 2.11955801e-01 6.13711821e-03
5.15082955e-01 -8.13780069e-01 2.72780329e-01 1.09658551e+00
5.85571900e-02 2.65293896e-01 3.37686807e-01 -6.22458577e-01
-1.20429921e+00 -1.26464677e+00 7.97317863e-01 1.57519147e-01
2.13269025e-01 -1.01942807e-01 -7.03976870e-01 6.17209315e-01
8.16878796e-01 3.62510026e-01 4.54745561e-01 -4.89505470e-01
-1.69380441e-01 -2.44550005e-01 -1.19569838e+00 1.40498251e-01
7.29285896e-01 -6.08636320e-01 -2.32590735e-01 2.51399517e-01
7.86332488e-01 -3.29566896e-01 -8.84687781e-01 4.31675673e-01
4.13953304e-01 -1.20676816e+00 9.55494642e-01 2.69905627e-01
5.41862667e-01 -6.49842620e-01 -4.16467130e-01 -1.23894775e+00
-6.66092396e-01 -2.71491855e-01 -1.12738334e-01 1.18711746e+00
2.52812058e-01 -6.52712464e-01 9.39823240e-02 1.91886827e-01
-1.48632437e-01 -6.57769978e-01 -6.64053917e-01 -3.94632369e-01
-3.32562983e-01 -2.44004041e-01 5.65135896e-01 6.15512550e-01
-2.57701755e-01 3.38231444e-01 -5.26108623e-01 3.77792209e-01
7.49041796e-01 1.92887317e-02 3.47093701e-01 -7.80495286e-01
-5.25058694e-02 -1.76054120e-01 -5.16215444e-01 -9.86124158e-01
-2.01428115e-01 -7.83715308e-01 -5.30072227e-02 -2.02316642e+00
4.70511943e-01 -1.38063222e-01 -6.37784958e-01 5.81271291e-01
-2.72070289e-01 5.60447395e-01 4.36695009e-01 3.40381891e-01
-5.79484999e-01 7.96835840e-01 1.06974638e+00 -1.23065233e-01
1.28917202e-01 -2.75140882e-01 -1.10181546e+00 4.10420567e-01
1.08976102e+00 -1.90533340e-01 -2.50770122e-01 -6.93587780e-01
-7.68982247e-02 1.65190846e-01 7.78809726e-01 -1.28171241e+00
5.61838388e-01 1.82269186e-01 6.86458409e-01 -7.95817375e-01
5.33394754e-01 -1.09417379e+00 3.36414963e-01 4.04606700e-01
-1.48177817e-01 6.77015334e-02 2.72200942e-01 5.82359374e-01
-4.56932187e-01 2.43298441e-01 7.49700725e-01 -2.22746246e-02
-8.21022391e-01 3.98904651e-01 3.29885557e-02 -6.39112890e-01
1.13696837e+00 -1.18293449e-01 -4.56442147e-01 -5.85844159e-01
-6.78972542e-01 1.11185230e-01 4.12747383e-01 3.30560178e-01
8.77751291e-01 -1.54431403e+00 -8.71075332e-01 4.34567839e-01
-9.34550725e-03 -2.28688985e-01 5.46232462e-01 1.08533430e+00
-3.89232665e-01 5.16580462e-01 -3.36188823e-01 -3.73915076e-01
-1.34224296e+00 6.67423785e-01 3.23903322e-01 1.79147050e-02
-5.85556090e-01 9.81803060e-01 4.92241234e-01 1.98211446e-02
6.51040971e-02 -3.16447258e-01 -3.68873149e-01 -2.19805956e-01
5.84861398e-01 7.54827037e-02 -6.03304394e-02 -1.02855277e+00
-3.35449934e-01 7.01979935e-01 1.18883558e-01 -5.37049472e-02
1.30856097e+00 -6.01208866e-01 -1.54767722e-01 1.09105222e-01
1.33131671e+00 -3.06553572e-01 -1.30916822e+00 -4.27618682e-01
-8.43279719e-01 -6.23419285e-01 5.89087367e-01 -9.01647389e-01
-1.61936498e+00 9.96755779e-01 1.06901264e+00 4.56699505e-02
1.74425709e+00 -3.91573124e-02 9.45701063e-01 -3.37390006e-01
7.27506578e-02 -6.94504559e-01 -2.34610528e-01 2.58399010e-01
9.48305070e-01 -1.33982480e+00 -2.80857496e-02 -4.67851639e-01
-4.67440128e-01 8.17334354e-01 4.10775185e-01 -3.56135666e-02
8.95027101e-01 3.36682796e-01 1.42631792e-02 -2.72624850e-01
-7.39020348e-01 -5.24661720e-01 2.95390278e-01 5.43216109e-01
5.55361271e-01 -3.56999263e-02 -2.73502320e-01 3.49516273e-01
1.61918744e-01 2.25990817e-01 1.33642524e-01 9.66415048e-01
-6.09636486e-01 -8.21045220e-01 -5.74536085e-01 3.34160477e-01
-4.68332529e-01 -4.49000180e-01 -1.47182643e-01 2.69048601e-01
3.33779395e-01 1.16822135e+00 3.43239233e-02 -7.40649819e-01
9.51935798e-02 -2.57407218e-01 3.59774917e-01 -2.76168078e-01
-5.74623466e-01 2.55938083e-01 -1.82329655e-01 -6.21433973e-01
-8.62938106e-01 -3.07003945e-01 -1.10921562e+00 -3.60594153e-01
-3.82777870e-01 5.46809658e-02 8.06895435e-01 6.88040435e-01
6.94156349e-01 7.85380006e-01 7.44958997e-01 -1.02963984e+00
-4.06599082e-02 -9.84494090e-01 -4.16304141e-01 2.17944548e-01
5.93141079e-01 -5.38572490e-01 -2.04471394e-01 1.81459457e-01] | [10.497836112976074, -1.8619657754898071] |
8992483f-3389-4b06-843b-136a5e42e5bf | inductive-linear-probing-for-few-shot-node | 2306.08192 | null | https://arxiv.org/abs/2306.08192v1 | https://arxiv.org/pdf/2306.08192v1.pdf | Inductive Linear Probing for Few-shot Node Classification | Meta-learning has emerged as a powerful training strategy for few-shot node classification, demonstrating its effectiveness in the transductive setting. However, the existing literature predominantly focuses on transductive few-shot node classification, neglecting the widely studied inductive setting in the broader few-shot learning community. This oversight limits our comprehensive understanding of the performance of meta-learning based methods on graph data. In this work, we conduct an empirical study to highlight the limitations of current frameworks in the inductive few-shot node classification setting. Additionally, we propose a simple yet competitive baseline approach specifically tailored for inductive few-shot node classification tasks. We hope our work can provide a new path forward to better understand how the meta-learning paradigm works in the graph domain. | ['Huan Liu', 'Nivedh Mudiam', 'Zhen Tan', 'Hirthik Mathavan'] | 2023-06-14 | null | null | null | null | ['node-classification', 'meta-learning', 'classification-1'] | ['graphs', 'methodology', 'methodology'] | [ 3.13145697e-01 4.13944095e-01 -1.00533962e+00 -1.28531262e-01
-6.30187273e-01 2.98371189e-03 8.37625206e-01 3.36673707e-01
-1.93804815e-01 4.25947994e-01 2.47850433e-01 -3.20199877e-01
-2.34146923e-01 -1.23962283e+00 -2.82143682e-01 -6.30201101e-01
-1.35338649e-01 3.35675210e-01 8.95010680e-02 -5.74045956e-01
1.51104763e-01 -2.15670802e-02 -1.45137584e+00 1.92331579e-02
5.80787420e-01 3.71789932e-01 -1.58912227e-01 8.55056942e-01
-6.34711310e-02 1.26042974e+00 -2.20979735e-01 -7.30932057e-01
-1.69264540e-01 -5.25661349e-01 -1.03945196e+00 -1.42704234e-01
2.78106242e-01 -3.39256048e-01 -7.17155039e-01 9.77282584e-01
4.75227594e-01 3.70900869e-01 8.93659651e-01 -1.71162510e+00
-7.18589067e-01 1.03232384e+00 -2.74494618e-01 3.34853560e-01
1.52470917e-01 -5.36342785e-02 1.64948845e+00 -8.71203303e-01
1.08061910e+00 1.00158048e+00 9.78338897e-01 8.10153961e-01
-1.19829309e+00 -3.15239400e-01 -5.70037356e-03 3.47263873e-01
-1.10731208e+00 -5.15400946e-01 9.07367408e-01 -2.86520034e-01
8.85506749e-01 1.92998629e-02 6.76818848e-01 1.34895194e+00
-5.22429049e-02 1.04875171e+00 7.98129857e-01 -7.33031392e-01
4.64620978e-01 6.64966628e-02 6.50028646e-01 8.78874540e-01
2.13680416e-01 9.42896754e-02 -3.84437650e-01 -3.60771567e-01
4.02685612e-01 5.25758564e-01 2.14391902e-01 -5.69282472e-01
-7.22322524e-01 1.48705530e+00 5.75694025e-01 8.33803296e-01
-1.27613246e-01 4.16200459e-01 7.98784971e-01 5.78709781e-01
1.02008927e+00 3.91432136e-01 -9.71352831e-02 -2.39895701e-01
-8.56989384e-01 1.34611968e-02 1.01820767e+00 1.07914901e+00
9.80664074e-01 1.07752956e-01 -5.01585305e-01 8.14321220e-01
2.97407538e-01 5.92502840e-02 2.39434242e-01 -8.53194475e-01
1.03867829e-01 4.68063772e-01 -4.07225490e-01 -8.20754707e-01
-3.36510807e-01 -4.31193382e-01 -6.18551493e-01 -1.07070871e-01
6.68469742e-02 -1.85582623e-01 -8.39066327e-01 1.55728185e+00
1.91370055e-01 6.75727069e-01 -6.34986162e-02 4.91504997e-01
1.29732907e+00 4.33479220e-01 5.74366987e-01 -3.46414745e-01
1.17591417e+00 -1.26289070e+00 -4.89721596e-01 -1.92412734e-01
1.39811671e+00 -2.31524944e-01 1.07081783e+00 -3.68913680e-01
-6.89780354e-01 -2.07772866e-01 -9.77103949e-01 -8.78253058e-02
-5.33481002e-01 -4.95301813e-01 1.44790745e+00 7.88595140e-01
-1.01590431e+00 5.70705354e-01 -5.26067972e-01 -1.06932032e+00
6.14454985e-01 -8.69231671e-02 -1.43430755e-01 -4.44239020e-01
-1.50970674e+00 7.80732632e-01 1.39739931e-01 -5.62998891e-01
-1.04828060e+00 -1.08273411e+00 -1.03616023e+00 2.56408840e-01
6.66059196e-01 -9.16531801e-01 1.40867770e+00 -5.30521631e-01
-1.20506728e+00 9.81721818e-01 -6.26842119e-03 -5.49908757e-01
1.34662256e-01 4.06281918e-01 -1.35259867e-01 1.61068216e-01
1.05687931e-01 3.41071010e-01 8.39221597e-01 -1.08325255e+00
-3.97045285e-01 -2.76683927e-01 4.50140983e-01 1.92590624e-01
-7.46802330e-01 -2.57302910e-01 -2.53034025e-01 -5.72823882e-01
-3.71045858e-01 -7.92555332e-01 -4.02961820e-01 -2.56073356e-01
-2.37067550e-01 -5.57696104e-01 5.89634180e-01 3.56341541e-01
1.30376720e+00 -1.94448555e+00 1.64362732e-02 -1.73260599e-01
6.94352388e-01 3.25550914e-01 -2.92960018e-01 1.07625401e+00
5.24954721e-02 2.51932442e-01 4.77009267e-02 -5.39949834e-01
2.80624423e-02 2.50106938e-02 -1.99722141e-01 3.58774185e-01
-1.34410784e-01 1.58038807e+00 -1.48592734e+00 -8.25026035e-01
4.17775214e-01 3.86042207e-01 -3.58974338e-01 9.42020118e-02
-3.44522186e-02 -4.59838063e-01 -6.43190324e-01 9.73086238e-01
2.18865320e-01 -4.93926346e-01 1.22688003e-01 -4.39017974e-02
2.83050120e-01 -9.83786136e-02 -3.19515318e-01 1.88606632e+00
-4.09060985e-01 7.52439439e-01 -1.42529964e-01 -1.36510861e+00
5.56248605e-01 3.28588933e-01 5.99524438e-01 -4.41739619e-01
4.67321187e-01 -2.31493697e-01 -1.36651501e-01 -4.34865981e-01
5.17114043e-01 -7.04388320e-01 -1.32089421e-01 6.89655364e-01
6.64934158e-01 -1.89265862e-01 2.74249583e-01 9.28400099e-01
1.46300507e+00 -2.92490721e-01 7.06192136e-01 -9.19183269e-02
-1.46119475e-01 2.88305014e-01 9.04244483e-02 1.35832083e+00
-6.36820674e-01 3.54779541e-01 4.33426976e-01 -2.39579082e-01
-7.95283914e-01 -8.90335202e-01 1.74529422e-02 1.98575997e+00
7.56595656e-02 -8.05756927e-01 -4.98245180e-01 -8.90641749e-01
7.96372592e-02 7.15861440e-01 -1.17476857e+00 -5.56256413e-01
1.42038008e-02 -9.00771976e-01 6.55766189e-01 5.77174842e-01
9.93294865e-02 -9.25870895e-01 -2.01708972e-01 9.86709297e-02
8.51891264e-02 -7.92518497e-01 -4.35223542e-02 2.56965548e-01
-1.14241791e+00 -1.18687284e+00 -6.82171345e-01 -8.59688878e-01
3.47117633e-01 9.80032086e-01 1.38664460e+00 4.79882032e-01
-4.45147574e-01 8.99091899e-01 -9.07555580e-01 -3.85431319e-01
-4.24187660e-01 5.68854153e-01 -2.11170077e-01 -4.43776429e-01
6.83231771e-01 -5.61041534e-01 -4.11045492e-01 -9.12948251e-02
-6.77379966e-01 -1.72596350e-01 3.56330872e-01 9.53126907e-01
-1.31794557e-01 -3.23807627e-01 8.50717485e-01 -1.61526322e+00
8.17914307e-01 -1.14257634e+00 2.33168513e-01 3.35110098e-01
-8.95484984e-01 -3.58204782e-01 4.33075190e-01 -3.63372445e-01
-9.26801860e-01 -5.30001521e-01 -1.07351288e-01 -2.69635320e-01
6.75421581e-02 8.34395051e-01 4.47020411e-01 -4.22139466e-01
8.71224105e-01 -1.09992055e-02 -1.25655130e-01 -2.17328519e-02
7.68937349e-01 4.79870021e-01 -1.10492976e-02 -2.45277733e-01
9.85098422e-01 5.60744107e-01 1.27758205e-01 -1.03020597e+00
-1.04001451e+00 -7.32988298e-01 -6.87642395e-01 -4.47951466e-01
5.42752326e-01 -9.00705278e-01 -4.94429737e-01 1.99631020e-01
-5.64802468e-01 -5.25962889e-01 -4.82797623e-01 1.55078858e-01
-5.56269586e-01 3.14364403e-01 -9.63248730e-01 -8.97451222e-01
-5.28648555e-01 -5.62649250e-01 8.33655059e-01 1.46622539e-01
-1.76263496e-01 -1.80098879e+00 4.67434168e-01 1.62176430e-01
5.01480699e-01 1.46944867e-03 9.16393936e-01 -9.12641227e-01
-1.52191743e-01 -6.39335215e-01 -5.46513125e-02 -3.73077482e-01
-6.71943650e-02 1.34747431e-01 -1.17966199e+00 -3.17377269e-01
-4.45494503e-01 -8.38486314e-01 1.37487888e+00 3.87327284e-01
4.88846421e-01 1.42407745e-01 -7.08921432e-01 2.77214110e-01
1.49791837e+00 -3.55676830e-01 5.11528969e-01 5.18139713e-02
8.42102826e-01 3.64087313e-01 8.30739439e-01 4.87798214e-01
6.20326042e-01 3.51533085e-01 2.81819314e-01 -9.61956307e-02
-3.10066015e-01 -4.53407884e-01 -5.37075773e-02 7.32935429e-01
-6.50561228e-02 -4.11428452e-01 -8.56899440e-01 6.34455204e-01
-2.02281618e+00 -1.47517669e+00 1.14669457e-01 1.66824591e+00
5.25710046e-01 8.49322751e-02 3.15069526e-01 1.80250868e-01
7.86137819e-01 7.41531610e-01 -2.68001884e-01 -2.74976760e-01
2.79739410e-01 1.82560265e-01 7.03577027e-02 3.30376089e-01
-1.34334838e+00 1.30005074e+00 7.50077629e+00 9.56067443e-01
-8.25278997e-01 4.27371591e-01 4.31340903e-01 -1.64253041e-02
-4.46700394e-01 2.93992192e-01 -5.48698664e-01 -6.18538521e-02
9.76090789e-01 -4.60850149e-01 1.21820070e-01 1.20746839e+00
-2.11332485e-01 9.83158201e-02 -1.32187128e+00 6.70085669e-01
3.24207127e-01 -1.71805108e+00 -2.30091125e-01 4.01634611e-02
8.65230858e-01 3.48383218e-01 -3.24702561e-02 1.06010735e+00
5.53082764e-01 -8.79445672e-01 1.80190261e-02 2.50185907e-01
6.79659486e-01 -5.28154314e-01 5.38537920e-01 2.98588037e-01
-1.22845566e+00 -8.66860226e-02 -4.55832779e-01 -4.54551339e-01
-2.00457815e-02 5.81750274e-01 -1.06518173e+00 3.10021549e-01
3.39774996e-01 1.17460465e+00 -6.11353815e-01 7.18866646e-01
3.24691683e-02 1.04704010e+00 2.26394862e-01 -5.00469506e-01
4.10149097e-01 8.57238248e-02 5.25678694e-01 1.24547875e+00
-1.93229556e-01 2.37520680e-01 2.74780512e-01 5.80101907e-01
-4.16326255e-01 2.39612281e-01 -1.25596690e+00 -4.41112697e-01
6.61032498e-01 1.72888684e+00 -9.35757697e-01 -5.55509627e-01
-6.67498887e-01 5.78310013e-01 6.52993083e-01 2.93760240e-01
-5.90589046e-01 -4.91229832e-01 3.24806064e-01 -7.77227879e-02
1.66811913e-01 1.05322406e-01 -3.86955827e-01 -1.34654105e+00
-7.74815738e-01 -2.60708749e-01 7.50148714e-01 -3.94374609e-01
-1.60733461e+00 1.57682076e-01 1.15232445e-01 -9.19183195e-01
-4.44174260e-01 -3.05839241e-01 -1.16902459e+00 8.71668085e-02
-1.34066236e+00 -1.53177226e+00 -4.23631668e-01 3.40076745e-01
7.49955773e-01 -1.16170011e-01 1.00239503e+00 5.63210016e-03
-5.81632793e-01 7.29327559e-01 1.32800937e-01 2.40700960e-01
6.80840254e-01 -1.16074157e+00 5.03731430e-01 5.19247830e-01
3.69506389e-01 7.19592690e-01 7.55944848e-01 -6.88773036e-01
-1.71670508e+00 -9.91488934e-01 6.90402687e-01 -4.87983108e-01
1.04321110e+00 -3.46527755e-01 -8.29196274e-01 8.23361814e-01
1.53083175e-01 3.89080971e-01 1.08190846e+00 7.86878407e-01
-4.96952474e-01 3.06879222e-01 -1.13674974e+00 7.35107481e-01
1.25594068e+00 -8.60342085e-01 -4.78658110e-01 4.01274174e-01
8.07670116e-01 3.58618677e-01 -1.14713037e+00 3.84943783e-01
5.58974385e-01 -7.54179299e-01 1.05796242e+00 -1.01385379e+00
6.93134546e-01 5.07741213e-01 -1.08918980e-01 -1.46766412e+00
-7.46121228e-01 -3.93792242e-01 -5.15380383e-01 1.33209395e+00
2.82398999e-01 -4.52215135e-01 1.06644320e+00 3.96816373e-01
-1.91956833e-02 -7.46372044e-01 -6.60203874e-01 -5.94170451e-01
2.25973472e-01 -5.09033382e-01 2.26170607e-02 1.30456436e+00
7.21357048e-01 7.04209805e-01 -5.91564417e-01 -6.11569941e-01
8.96885574e-01 1.54851481e-01 7.94836044e-01 -1.59912646e+00
-1.85473278e-01 -3.51443321e-01 -5.70910633e-01 -4.17612910e-01
7.28538871e-01 -1.30893862e+00 -6.11638883e-03 -1.65229678e+00
8.79403830e-01 -3.82082343e-01 -3.99260402e-01 3.89680833e-01
-4.61016327e-01 7.04436421e-01 2.12595046e-01 7.51458183e-02
-1.09480274e+00 5.94862461e-01 9.47072625e-01 -3.43646735e-01
4.68760282e-02 -2.88572945e-02 -9.51990902e-01 5.55427790e-01
8.29277515e-01 -5.62996864e-01 -7.28055060e-01 6.62900880e-02
3.94198969e-02 -1.91372544e-01 3.08207601e-01 -6.64226174e-01
3.71449590e-01 -2.73505926e-01 -1.05499990e-01 -1.09246917e-01
2.88846791e-01 -4.18728858e-01 -5.25543034e-01 4.79967028e-01
-5.15565097e-01 -5.74490607e-01 -3.68622065e-01 9.36612248e-01
1.67134747e-01 -6.87747478e-01 5.70572078e-01 -3.61949682e-01
-1.11609554e+00 6.99278653e-01 -5.33293784e-01 5.80074191e-01
1.24079442e+00 -2.48775840e-01 -5.09746850e-01 -5.85952461e-01
-6.12329125e-01 9.33675691e-02 6.63357973e-01 4.38010424e-01
5.78447878e-01 -1.42314184e+00 -6.02328897e-01 -2.21889600e-01
8.10662448e-01 -7.94234216e-01 3.10193688e-01 9.48149800e-01
7.44473636e-02 2.12825552e-01 1.64091259e-01 -2.84683198e-01
-1.14662707e+00 9.19181168e-01 1.89711433e-02 -2.29363218e-01
-7.63981283e-01 7.69416511e-01 -2.14327723e-01 -3.76172006e-01
2.03679651e-01 3.02599072e-01 -1.92881152e-01 3.84287506e-01
4.33055341e-01 7.68581271e-01 -2.27117255e-01 -3.98428470e-01
-3.65354508e-01 1.77426979e-01 -1.84046358e-01 9.70207378e-02
1.31647027e+00 -7.74758980e-02 6.32382855e-02 9.52896297e-01
1.47823071e+00 -4.62769479e-01 -6.34209216e-01 -3.98107558e-01
-1.52336480e-02 -1.95867434e-01 2.80700147e-01 -2.33247921e-01
-7.27230668e-01 8.60929549e-01 8.55207667e-02 5.33011436e-01
6.25253916e-01 3.94953787e-01 5.40586233e-01 6.44548416e-01
5.97323120e-01 -1.06651402e+00 2.72832125e-01 5.61152101e-01
9.64194313e-02 -1.65893495e+00 1.97567552e-01 -4.77839679e-01
-7.22693622e-01 9.77277458e-01 4.21049625e-01 -2.93344587e-01
1.07462633e+00 1.52102515e-01 -2.17456281e-01 -6.04959846e-01
-1.15889335e+00 -6.91482127e-01 6.09133504e-02 8.10217738e-01
7.61938334e-01 1.09381616e-01 -2.34486610e-01 3.10265571e-01
3.44658613e-01 1.51530281e-01 5.81168056e-01 1.13852167e+00
-7.28514016e-01 -9.08177733e-01 3.37897241e-01 7.98668325e-01
-1.90155059e-01 -3.02893311e-01 -6.33551300e-01 8.41724038e-01
-3.98638934e-01 1.24249446e+00 -1.15775570e-01 -6.12648487e-01
-5.76670691e-02 2.30433941e-01 6.06819332e-01 -1.12087274e+00
-5.65257549e-01 -2.69478202e-01 1.37681246e-01 -2.65905797e-01
-6.34560645e-01 -3.45917672e-01 -9.16822195e-01 -6.19213700e-01
-5.01541197e-01 2.66295195e-01 2.13938132e-01 1.13600338e+00
1.15883768e-01 4.21448171e-01 7.04197764e-01 -8.17083776e-01
-6.07534468e-01 -1.18157113e+00 -7.81847119e-01 5.04202664e-01
4.95353760e-03 -8.31284881e-01 -4.48311925e-01 -3.71954560e-01] | [9.924932479858398, 3.0778756141662598] |
a803864d-2d7d-4dee-962d-9485424e24a3 | rethinking-interactive-image-segmentation | 2101.04378 | null | https://arxiv.org/abs/2101.04378v3 | https://arxiv.org/pdf/2101.04378v3.pdf | Rethinking Interactive Image Segmentation: Feature Space Annotation | Despite the progress of interactive image segmentation methods, high-quality pixel-level annotation is still time-consuming and laborious - a bottleneck for several deep learning applications. We take a step back to propose interactive and simultaneous segment annotation from multiple images guided by feature space projection. This strategy is in stark contrast to existing interactive segmentation methodologies, which perform annotation in the image domain. We show that feature space annotation achieves competitive results with state-of-the-art methods in foreground segmentation datasets: iCoSeg, DAVIS, and Rooftop. Moreover, in the semantic segmentation context, it achieves 91.5% accuracy in the Cityscapes dataset, being 74.75 times faster than the original annotation procedure. Further, our contribution sheds light on a novel direction for interactive image annotation that can be integrated with existing methodologies. The supplementary material presents video demonstrations. Code available at https://github.com/LIDS-UNICAMP/rethinking-interactive-image-segmentation. | ['Alexandre X Falc{ã}o', 'Jord{ã}o Bragantini', 'Laurent Najman'] | 2021-01-12 | null | null | null | null | ['foreground-segmentation'] | ['computer-vision'] | [ 4.28877145e-01 4.18356508e-01 -2.07340255e-01 -3.78953427e-01
-1.01608181e+00 -9.80133832e-01 2.37495705e-01 -1.63127035e-02
-3.76349986e-01 5.60182452e-01 -2.66747296e-01 -4.24445778e-01
2.22886980e-01 -5.31859100e-01 -7.33238757e-01 -5.58702052e-01
2.05982074e-01 5.72893262e-01 7.06494033e-01 2.06819981e-01
2.09906399e-01 4.82853115e-01 -1.49444163e+00 2.85140961e-01
8.27417552e-01 7.89088786e-01 4.00090635e-01 7.68408239e-01
-3.24784309e-01 2.91431487e-01 -2.12441340e-01 -4.33712959e-01
3.25744927e-01 -4.04725403e-01 -1.30337334e+00 4.94628042e-01
6.16423011e-01 -3.74267012e-01 -1.86935195e-03 9.32499409e-01
1.88614756e-01 -3.33357090e-03 3.94917727e-01 -1.19130361e+00
-4.18338269e-01 4.43287492e-01 -6.81959093e-01 6.51032850e-02
2.98962682e-01 -2.77170353e-02 1.05186415e+00 -8.86094391e-01
9.30304289e-01 8.45495224e-01 6.64969087e-01 4.03802514e-01
-1.16621232e+00 -4.47661668e-01 4.55415010e-01 6.71823621e-02
-1.28378594e+00 -1.26287714e-01 4.17002261e-01 -6.18578672e-01
7.87905216e-01 4.09931719e-01 9.58207130e-01 7.72196531e-01
-4.05304432e-01 1.34190857e+00 1.00710595e+00 -4.92674440e-01
-3.24348994e-02 2.04455107e-02 2.79392719e-01 8.04828823e-01
9.67173949e-02 -3.96298558e-01 -2.78313428e-01 2.21864745e-01
9.56183672e-01 -1.78850405e-02 -2.07319856e-01 -4.29562718e-01
-1.21309733e+00 7.72752523e-01 2.73418635e-01 4.04438496e-01
-8.34732596e-03 2.64296532e-01 3.76696140e-01 -1.24713637e-01
5.86855888e-01 2.86183536e-01 -8.16136122e-01 -2.71325290e-01
-1.34922636e+00 1.24519348e-01 7.05984712e-01 9.85371232e-01
9.64272857e-01 -2.09403366e-01 7.75738358e-02 5.65821528e-01
2.03055814e-01 2.68172473e-01 4.95240325e-03 -1.40956008e+00
1.81247622e-01 6.49587691e-01 1.51832819e-01 -5.56892157e-01
-5.46996057e-01 -2.90985703e-01 -3.32305133e-01 1.65342778e-01
9.24734056e-01 -1.95636243e-01 -1.01587081e+00 1.15432358e+00
5.77156186e-01 1.95955232e-01 -3.89813513e-01 9.80245352e-01
9.17238414e-01 3.92490327e-01 7.15907365e-02 9.17824730e-02
1.39294934e+00 -1.41126823e+00 -6.16535723e-01 -1.04863770e-01
8.09602737e-01 -8.41338217e-01 1.43220913e+00 4.29758579e-01
-1.10535491e+00 -4.40119147e-01 -7.46073306e-01 -3.66931468e-01
-5.53036153e-01 3.20547134e-01 8.27874601e-01 7.64271975e-01
-1.10130239e+00 3.48754108e-01 -1.10074830e+00 -6.92069054e-01
7.32484698e-01 3.43177259e-01 -2.89288521e-01 2.05042556e-01
-6.70394003e-01 4.24916953e-01 5.33801317e-01 -1.38144597e-01
-6.16749525e-01 -8.16381216e-01 -6.40746176e-01 -1.73315659e-01
6.58242345e-01 -4.62749660e-01 1.44544995e+00 -1.30303490e+00
-1.52043474e+00 1.27388036e+00 -5.65889217e-02 -4.18694556e-01
9.78710890e-01 -6.78595781e-01 8.53526294e-02 4.73456770e-01
1.78887859e-01 1.30887234e+00 4.36794460e-01 -1.41957462e+00
-8.85260046e-01 -1.25805154e-01 1.86477348e-01 1.56435862e-01
7.75009617e-02 7.16010705e-02 -1.08726346e+00 -4.13873374e-01
2.04434007e-01 -1.02259803e+00 -4.59981769e-01 1.13474168e-01
-7.55754232e-01 -1.29447713e-01 1.03864908e+00 -7.02091813e-01
9.74393487e-01 -2.01543140e+00 1.55293534e-03 -4.41314243e-02
4.47952300e-02 3.85885864e-01 1.34027719e-01 3.79393548e-01
-1.07431412e-02 3.70192438e-01 -6.42278075e-01 -4.59919810e-01
6.04830384e-02 2.02634275e-01 -1.39173597e-01 4.68075901e-01
-5.46876304e-02 1.06201279e+00 -7.80012786e-01 -7.89323926e-01
4.65655625e-01 3.98585230e-01 -5.36622584e-01 -1.13090523e-01
-3.37284863e-01 8.68344188e-01 -3.77276063e-01 8.54801595e-01
6.87592387e-01 -3.44988436e-01 1.08741105e-01 -3.92460078e-02
-4.26415503e-01 -1.52940616e-01 -1.22582698e+00 2.08247066e+00
-8.11393559e-02 9.27230597e-01 2.29216293e-01 -1.09739530e+00
5.19640565e-01 1.02740139e-01 7.30388701e-01 -5.47791064e-01
1.68338612e-01 2.80532449e-01 -4.13144052e-01 -4.92632031e-01
5.30634463e-01 4.85433578e-01 -3.25693227e-02 2.81387120e-01
1.29920080e-01 -3.57460201e-01 4.77982163e-01 2.30541512e-01
7.28464484e-01 6.79486215e-01 2.31341884e-01 -3.62747431e-01
4.14872497e-01 2.73056120e-01 4.76982236e-01 8.06390762e-01
-3.62532020e-01 9.76965606e-01 6.17064536e-01 -4.58801031e-01
-9.75569189e-01 -9.93735194e-01 -3.51773798e-01 1.15662384e+00
2.83943355e-01 -3.92821670e-01 -1.45619249e+00 -9.93916452e-01
-2.51484722e-01 5.13255894e-01 -5.76405108e-01 7.30491161e-01
-5.49733341e-01 -5.28903127e-01 5.90182662e-01 6.27799630e-01
7.85795450e-01 -9.44556653e-01 -7.83955991e-01 -2.96614692e-02
-2.76841015e-01 -1.42614639e+00 -2.44799048e-01 1.26351237e-01
-8.53568673e-01 -1.25752330e+00 -8.36773932e-01 -8.68413746e-01
7.06420422e-01 1.08011529e-01 1.21191931e+00 1.96316391e-01
-5.37059247e-01 6.10295653e-01 -4.35179174e-01 -3.35019082e-01
-5.53431734e-02 4.74471062e-01 -5.87950885e-01 -2.75192529e-01
2.48944774e-01 -3.62163991e-01 -8.38943958e-01 4.26907033e-01
-8.51377904e-01 5.58780551e-01 3.63454014e-01 4.55471396e-01
9.50281024e-01 -3.14541370e-01 2.66204268e-01 -1.32217908e+00
-7.70399421e-02 -1.90652445e-01 -9.51541305e-01 1.07061304e-01
-3.31451595e-01 -5.21284640e-01 1.60166159e-01 4.29485291e-02
-1.09202385e+00 7.65760601e-01 -4.65474933e-01 -3.55760604e-02
-7.37869740e-01 9.12168622e-02 -2.17035711e-01 1.76243223e-02
4.35786575e-01 -4.68094870e-02 -2.93477058e-01 -5.37537515e-01
7.62800753e-01 4.29463536e-01 5.21924078e-01 -3.65158349e-01
5.63598573e-01 8.35204780e-01 -3.54331523e-01 -8.58645439e-01
-9.50888693e-01 -7.58701026e-01 -1.39631760e+00 -3.91337126e-01
1.32994318e+00 -7.63149559e-01 -3.12727064e-01 4.85754967e-01
-1.06948507e+00 -9.98034537e-01 -3.04488540e-01 1.18607469e-01
-7.36814678e-01 4.46032882e-01 -6.37597382e-01 -5.15806854e-01
-7.64714852e-02 -1.32361412e+00 1.38064444e+00 4.95000601e-01
-7.62453675e-02 -1.06146598e+00 -2.59734452e-01 8.38693082e-01
-1.20911285e-01 4.06554043e-01 3.85394543e-01 -6.89092755e-01
-8.35556984e-01 -1.34861052e-01 -5.32056332e-01 2.15336412e-01
-1.71295211e-01 2.83413827e-01 -1.08584464e+00 1.45355463e-01
-6.11609101e-01 -9.91824269e-02 8.83079767e-01 6.27331376e-01
1.43094110e+00 1.72335878e-01 -5.63765943e-01 7.46100605e-01
1.42595983e+00 5.21451868e-02 5.94796598e-01 5.07726312e-01
9.27743077e-01 5.45008957e-01 9.47632432e-01 3.06768358e-01
3.69325608e-01 6.85044706e-01 3.64006728e-01 -7.09284604e-01
-2.28099659e-01 6.12333752e-02 -1.52653277e-01 3.85973036e-01
-1.53415561e-01 -2.05409259e-01 -1.24989367e+00 6.74032152e-01
-2.07240653e+00 -6.17927909e-01 -7.49278247e-01 1.77974069e+00
5.91257989e-01 3.83468047e-02 3.53971571e-01 1.12765700e-01
6.47505224e-01 -9.43139642e-02 -5.23379385e-01 -2.47163907e-01
6.38000295e-02 -8.21710215e-04 7.86422372e-01 6.27076745e-01
-1.57398283e+00 1.36599624e+00 6.17641020e+00 8.86472046e-01
-8.15575957e-01 3.43202382e-01 9.06530499e-01 1.46835417e-01
3.03197876e-02 3.69941369e-02 -8.28343034e-01 2.59378523e-01
6.64591551e-01 4.51981008e-01 1.85946196e-01 9.05833602e-01
2.83355653e-01 -5.19998193e-01 -7.79135883e-01 7.22909391e-01
-1.52048364e-01 -1.43995070e+00 -4.21453327e-01 3.97013314e-02
9.81186569e-01 2.61521727e-01 2.43854448e-02 -3.65774683e-03
2.39657372e-01 -6.55581594e-01 7.86131263e-01 1.86604962e-01
8.01062226e-01 -6.07342482e-01 6.19444430e-01 2.02417463e-01
-1.28892875e+00 1.16264984e-01 -1.06496669e-01 1.61351666e-01
4.47095901e-01 5.66600502e-01 -7.68168867e-01 4.83118445e-01
9.28338110e-01 5.83311021e-01 -7.37232983e-01 1.15801144e+00
-4.08514231e-01 8.06800902e-01 -3.29165220e-01 4.04856801e-01
5.57525992e-01 -4.60630685e-01 2.87915528e-01 1.67328990e+00
1.01299345e-01 2.04506852e-02 4.78419632e-01 7.79546142e-01
-9.42014717e-03 2.17714086e-01 -5.03230870e-01 2.77150348e-02
2.35382572e-01 1.60720980e+00 -1.81289053e+00 -4.77616489e-01
-5.48043907e-01 1.38586271e+00 9.98822302e-02 5.38441300e-01
-1.21060717e+00 -3.38806748e-01 2.44040027e-01 1.16415992e-01
4.54092622e-01 -4.56625551e-01 -5.57284474e-01 -8.82000387e-01
-3.07047784e-01 -4.88068968e-01 3.07226121e-01 -6.66270614e-01
-6.51877582e-01 2.09203541e-01 7.21074194e-02 -9.37020123e-01
1.32366896e-01 -6.53422534e-01 -4.14817989e-01 2.62409598e-01
-1.32335377e+00 -1.51963365e+00 -4.71743077e-01 4.09044981e-01
1.00409293e+00 3.73174965e-01 6.56311393e-01 4.36474949e-01
-7.39377916e-01 3.32324326e-01 2.72795200e-01 3.72200578e-01
3.72378260e-01 -1.60974991e+00 5.58683157e-01 9.36255455e-01
5.23768067e-01 1.21846783e-03 4.49717849e-01 -5.11433184e-01
-8.88121724e-01 -1.13252962e+00 4.44490641e-01 -5.04860461e-01
6.60877168e-01 -5.75790465e-01 -7.55413353e-01 9.55530524e-01
4.10203665e-01 1.74284115e-01 7.54099786e-01 -2.13956498e-02
7.75765777e-02 3.94606411e-01 -9.60828364e-01 6.56492591e-01
1.25877881e+00 -3.76726359e-01 -1.10243373e-01 7.00137258e-01
8.40389311e-01 -7.42463708e-01 -7.07092106e-01 1.55196831e-01
4.73501354e-01 -1.19564426e+00 1.07726598e+00 -1.03088133e-01
1.79355234e-01 -3.64915907e-01 -5.43091632e-02 -7.10409105e-01
2.52454486e-02 -5.97571373e-01 2.46162608e-01 1.30641365e+00
4.94463563e-01 -3.90410513e-01 1.04397953e+00 5.88059247e-01
-5.01469374e-01 -8.07702065e-01 -6.75264120e-01 -6.62310660e-01
9.34771672e-02 -7.45620370e-01 2.37694845e-01 9.03876245e-01
-3.14403862e-01 -6.82724789e-02 1.35759383e-01 1.33603364e-01
4.89268631e-01 1.92952156e-01 8.35535944e-01 -1.11727250e+00
-1.59863174e-01 -5.72775543e-01 -1.43801853e-01 -1.03797042e+00
2.10715160e-01 -7.94754684e-01 5.46454601e-02 -1.97580242e+00
4.39351797e-02 -2.74774492e-01 1.47250816e-01 5.70522070e-01
-1.12471692e-01 9.74675536e-01 3.29691231e-01 1.45730600e-01
-1.15487611e+00 -6.33983314e-02 1.38599491e+00 1.48810044e-01
-2.74708688e-01 -3.05450913e-02 -3.78229141e-01 1.13810217e+00
1.05344188e+00 -1.85689092e-01 -3.33400518e-01 -3.86278570e-01
-9.46502760e-02 -1.94689408e-01 3.20293188e-01 -1.04368615e+00
1.53641164e-01 -6.74802586e-02 3.06565881e-01 -9.14409101e-01
2.76103199e-01 -8.01185727e-01 6.65208697e-02 3.30976754e-01
-1.17742151e-01 -2.09515050e-01 2.90967345e-01 4.25085723e-01
-8.90000537e-02 -3.65376443e-01 7.83863664e-01 -4.14681792e-01
-1.04430187e+00 2.04339013e-01 -6.42141283e-01 2.68930286e-01
1.27701437e+00 -6.86807930e-01 -1.74957693e-01 1.75214335e-02
-1.08934224e+00 1.86064109e-01 6.45323634e-01 6.67033046e-02
2.30831802e-01 -7.58914888e-01 -4.45317954e-01 -1.72364544e-02
-1.13421001e-01 3.48882675e-01 2.96596467e-01 1.17652404e+00
-1.03486967e+00 4.67419922e-01 -6.49059862e-02 -9.05308306e-01
-1.42337537e+00 1.71847269e-01 2.71714151e-01 -7.73989502e-03
-9.04298842e-01 9.64136899e-01 3.48113835e-01 -4.84886467e-01
2.29412019e-01 -3.83159518e-01 -2.40674755e-03 3.16567570e-01
1.97301388e-01 4.87045586e-01 -4.21754755e-02 -6.17374718e-01
-3.28734070e-01 7.68538296e-01 3.43180113e-02 -2.31525034e-01
1.10880733e+00 -4.45572376e-01 7.82392249e-02 4.60951746e-01
1.00508714e+00 -1.74098492e-01 -1.55520403e+00 1.25348434e-01
3.23297352e-01 -4.43857878e-01 1.73272826e-02 -9.39177513e-01
-1.21608806e+00 9.26594079e-01 7.30136514e-01 4.27147925e-01
1.04905272e+00 2.70260900e-01 6.86891437e-01 3.07165593e-01
1.36541113e-01 -1.49006176e+00 6.37573078e-02 5.04753292e-01
5.61481893e-01 -1.44214594e+00 4.36180383e-02 -8.76216352e-01
-6.74712241e-01 1.09081149e+00 6.49270296e-01 -9.54222083e-02
6.13905907e-01 2.94470787e-01 2.61811733e-01 -3.33359241e-01
-6.45029843e-02 -5.83898783e-01 2.69342631e-01 6.48266852e-01
6.68232560e-01 1.44043192e-01 -2.29640186e-01 4.00656253e-01
1.97497550e-02 -1.04516648e-01 5.02100289e-01 8.47210646e-01
-5.25947511e-01 -1.10153282e+00 -2.85248935e-01 2.65128046e-01
-6.43036425e-01 -2.73196287e-02 -4.28650349e-01 1.21163344e+00
2.44018793e-01 8.49801898e-01 2.39377424e-01 1.89856350e-01
6.02426864e-02 1.90857410e-01 3.18290293e-01 -6.85043097e-01
-5.46067476e-01 4.72358227e-01 1.62114099e-01 -8.46269906e-01
-5.67050576e-01 -9.94445503e-01 -1.71042967e+00 4.97642383e-02
-3.67251247e-01 -4.31503262e-03 7.33066499e-01 9.26454544e-01
2.65441120e-01 4.70175564e-01 -7.62205943e-02 -1.24327219e+00
6.38063431e-01 -5.88154376e-01 -4.07043755e-01 2.66184449e-01
-5.68714514e-02 -5.42230546e-01 -1.04682788e-01 5.67156315e-01] | [9.489652633666992, 0.2057841569185257] |
8e0d4233-21cd-4216-ad15-5dfd50e1359e | feature-fusion-for-robust-patch-matching-with | 1901.03547 | null | http://arxiv.org/abs/1901.03547v1 | http://arxiv.org/pdf/1901.03547v1.pdf | Feature Fusion for Robust Patch Matching With Compact Binary Descriptors | This work addresses the problem of learning compact yet discriminative patch
descriptors within a deep learning framework. We observe that features
extracted by convolutional layers in the pixel domain are largely complementary
to features extracted in a transformed domain. We propose a convolutional
network framework for learning binary patch descriptors where pixel domain
features are fused with features extracted from the transformed domain. In our
framework, while convolutional and transformed features are distinctly
extracted, they are fused and provided to a single classifier which thus
jointly operates on convolutional and transformed features. We experiment at
matching patches from three different datasets, showing that our feature fusion
approach outperforms multiple state-of-the-art approaches in terms of accuracy,
rate, and complexity. | ['Skjalg Lepsoy', 'Gianluca Francini', 'Andrea Migliorati', 'Riccardo Leonardi', 'Attilio Fiandrotti'] | 2019-01-11 | null | null | null | null | ['patch-matching'] | ['computer-vision'] | [ 5.77960730e-01 -1.93978563e-01 -3.24785143e-01 -4.51399803e-01
-1.11845219e+00 -5.17680168e-01 6.94585264e-01 3.04136276e-01
-5.44568777e-01 4.45344746e-01 4.21985164e-02 2.95581609e-01
-2.77061790e-01 -1.09555864e+00 -8.31743121e-01 -7.43750274e-01
-1.22522421e-01 -1.92491636e-01 3.23710889e-01 8.43640044e-03
1.26437828e-01 8.72907758e-01 -1.84506679e+00 6.24600172e-01
4.92244661e-01 1.62950063e+00 -1.81383416e-01 5.24566591e-01
2.02925861e-01 6.80269778e-01 -2.39291832e-01 -1.20916538e-01
6.93999887e-01 -9.18121040e-02 -6.36507034e-01 1.36156976e-01
1.10741556e+00 -3.25661212e-01 -7.48991787e-01 8.68143737e-01
2.84355551e-01 -2.18022466e-02 8.58154535e-01 -9.83956516e-01
-7.87554145e-01 -5.24117425e-02 -2.37168923e-01 7.54282847e-02
4.09557611e-01 -5.65461740e-02 1.32394040e+00 -9.89677429e-01
7.47549474e-01 8.37609589e-01 9.79446650e-01 -1.11942336e-01
-1.51845515e+00 -3.51428777e-01 -9.47149768e-02 1.18549921e-01
-1.61365199e+00 -4.59094226e-01 9.16676819e-01 -4.33171928e-01
1.18883145e+00 4.91327755e-02 5.66523015e-01 8.13427687e-01
3.12753558e-01 8.24598193e-01 9.62625146e-01 -4.01386410e-01
1.58619449e-01 -7.34931827e-02 8.61952752e-02 7.68516719e-01
1.02411385e-03 5.60959756e-01 -6.18230164e-01 -3.38577390e-01
7.41092503e-01 4.59338099e-01 -3.41422260e-02 -6.31731331e-01
-1.06708932e+00 1.10935187e+00 8.56541395e-01 4.93865788e-01
-8.60495090e-01 3.09306353e-01 1.81030139e-01 6.02771103e-01
6.52540863e-01 1.32831931e-01 -6.06905758e-01 8.75062570e-02
-1.21736419e+00 4.14125234e-01 6.87285900e-01 9.01865482e-01
1.39706779e+00 -4.08557177e-01 -4.92534965e-01 8.18414450e-01
1.93500090e-02 7.07215011e-01 1.79049641e-01 -8.28314781e-01
2.10909639e-02 8.78951609e-01 -1.66114271e-01 -1.24725997e+00
-2.20515206e-01 -4.29909229e-01 -7.44678736e-01 2.82414943e-01
1.46253929e-01 1.56806126e-01 -9.70890582e-01 1.56374490e+00
6.31477982e-02 1.26045957e-01 6.83529675e-02 5.98766088e-01
9.37759280e-01 4.21465546e-01 -1.12186931e-01 4.20589268e-01
1.14755142e+00 -9.77193952e-01 -1.55446650e-02 -8.38940144e-02
4.14133847e-01 -8.80905747e-01 3.92273664e-01 2.28733152e-01
-9.41648364e-01 -9.84638691e-01 -1.17963362e+00 -3.14915895e-01
-6.77774072e-01 3.49741995e-01 4.10178572e-01 2.40769953e-01
-1.40193021e+00 9.45024848e-01 -5.92570662e-01 -3.77688736e-01
8.29802096e-01 5.57266951e-01 -9.36555803e-01 -2.83271641e-01
-8.81235123e-01 7.26159513e-01 1.45600721e-01 -2.48952642e-01
-6.65960848e-01 -5.24475574e-01 -1.07392657e+00 3.02923709e-01
-3.35285574e-01 -4.22959328e-01 9.95981336e-01 -1.12295544e+00
-1.05367696e+00 8.89671743e-01 -1.77644908e-01 -3.56287301e-01
1.91979408e-01 -5.36130145e-02 -2.43784145e-01 5.76159179e-01
2.60295659e-01 8.32889020e-01 1.09138560e+00 -8.08657348e-01
-9.88636017e-01 -4.49486017e-01 1.31905958e-01 -1.79223716e-01
-2.57911503e-01 -2.46275783e-01 -9.09809321e-02 -5.83830237e-01
2.43026227e-01 -7.39326477e-01 -1.35138392e-01 6.79389119e-01
1.33917749e-01 -3.06370497e-01 1.05362773e+00 -5.65733731e-01
7.39027917e-01 -2.33524251e+00 5.21118715e-02 3.29479933e-01
3.57370615e-01 2.64248461e-01 -4.48209226e-01 5.54037869e-01
-1.06167495e-01 -2.85589337e-01 -2.39569306e-01 -2.52239913e-01
1.61327779e-01 1.70484036e-01 -3.15633208e-01 6.67893589e-01
8.21563244e-01 1.23227799e+00 -7.90625095e-01 -3.65055978e-01
3.98506492e-01 8.27198565e-01 -4.63107347e-01 1.50140107e-01
1.07553288e-01 1.06649455e-02 -4.64597940e-01 8.76854300e-01
8.76926422e-01 -1.29025489e-01 -2.30401218e-01 -3.20119590e-01
-2.29453295e-02 2.10638136e-01 -8.57101977e-01 1.85227740e+00
-6.04435503e-01 9.68056381e-01 -2.05539450e-01 -1.38952255e+00
1.20216417e+00 2.50202566e-01 6.15821958e-01 -9.03475285e-01
3.29664606e-03 2.97916532e-01 -3.96265358e-01 -2.00440347e-01
5.16474545e-01 8.95175263e-02 -1.57805249e-01 3.35163891e-01
6.30611539e-01 -2.85200309e-02 -5.98281063e-02 -2.61438429e-01
1.44715011e+00 2.11837247e-01 4.17270035e-01 -2.56114393e-01
5.55510223e-01 -1.99022945e-02 2.64221907e-01 9.00262356e-01
-2.07499281e-01 5.84281147e-01 3.94117892e-01 -8.16761672e-01
-1.08718586e+00 -1.12723374e+00 -5.45541108e-01 1.05464399e+00
1.44457459e-01 -4.75801080e-01 -4.20520246e-01 -8.64867806e-01
4.68383253e-01 -3.02412599e-01 -1.02479887e+00 -1.71946734e-01
-4.95677620e-01 2.52541397e-02 5.63198686e-01 7.86188722e-01
5.48407078e-01 -1.15723073e+00 -7.83586144e-01 4.03775424e-01
3.71288985e-01 -1.13151622e+00 -1.53559223e-01 5.31114340e-01
-6.61201119e-01 -1.20566010e+00 -7.00917840e-01 -1.05050468e+00
5.72751343e-01 4.22728300e-01 1.15132809e+00 1.92534581e-01
-5.46149433e-01 5.00877857e-01 -5.27856290e-01 -3.54776680e-02
1.61409727e-03 2.14429796e-01 -6.21896803e-01 1.55550122e-01
7.61453152e-01 -7.99086630e-01 -7.41817534e-01 -3.74417603e-02
-1.04254508e+00 -4.22784358e-01 8.92574549e-01 1.15155590e+00
6.90310299e-01 -2.54028052e-01 1.72336161e-01 -4.24984545e-01
3.15236866e-01 -2.48327985e-01 -5.26405096e-01 2.29946613e-01
-2.61022538e-01 2.31130078e-01 5.62297702e-01 -3.40107143e-01
-4.38303828e-01 5.12930632e-01 -1.19781189e-01 -2.83905715e-01
-5.76386809e-01 3.66450697e-01 1.85036480e-01 -7.23135471e-01
7.83524573e-01 4.38493729e-01 -8.22986811e-02 -4.45070416e-01
4.51997727e-01 7.24085927e-01 6.11126006e-01 -5.10082483e-01
9.08103228e-01 5.86084723e-01 1.47058606e-01 -7.11449862e-01
-6.22995496e-01 -5.71589530e-01 -1.15892994e+00 1.05477534e-01
6.83294773e-01 -1.05268550e+00 -3.04479510e-01 4.83973116e-01
-1.21684372e+00 1.16152868e-01 -6.33417904e-01 1.86210096e-01
-6.91075504e-01 2.50600219e-01 -4.82016444e-01 -9.94977802e-02
-2.94387043e-01 -1.09489298e+00 1.76292098e+00 2.56044328e-01
1.98164098e-02 -1.06062126e+00 3.15888584e-01 -2.53447652e-01
7.14556336e-01 5.13007820e-01 6.76736474e-01 -7.66190231e-01
-5.10941923e-01 -7.19784558e-01 -6.42141104e-01 5.43727696e-01
3.32748652e-01 -9.88925844e-02 -1.26162767e+00 -4.19609696e-01
-2.86462992e-01 -4.06714112e-01 1.28045392e+00 2.55612880e-01
1.03803492e+00 -2.00606063e-01 -3.39512587e-01 9.15913284e-01
1.69094181e+00 -1.25084370e-01 5.77614307e-01 2.57197618e-01
3.29234362e-01 2.72331715e-01 2.96517611e-01 4.05406177e-01
1.40845209e-01 6.63830638e-01 2.00396627e-01 -4.97542888e-01
-2.22381085e-01 -1.29011616e-01 1.13912456e-01 2.12883517e-01
1.72408238e-01 2.59542197e-01 -5.95825732e-01 7.76729047e-01
-1.92960107e+00 -9.91129279e-01 3.13573569e-01 1.76956844e+00
6.15680158e-01 -2.20829889e-01 -4.28089052e-02 1.87184244e-01
4.69315261e-01 4.23891753e-01 -4.05548424e-01 -2.79058784e-01
-3.68234605e-01 1.13479960e+00 4.87844646e-01 2.42493719e-01
-1.62580860e+00 8.02223206e-01 7.02393818e+00 7.10137844e-01
-1.33256567e+00 -5.56268282e-02 3.09079856e-01 2.88247377e-01
-1.46551132e-01 4.14075442e-02 -3.41354191e-01 1.51916698e-01
5.45729280e-01 -1.07443444e-01 2.84025282e-01 1.05268884e+00
-7.27112174e-01 5.87060174e-04 -1.37572956e+00 9.82760251e-01
1.21612169e-01 -1.44838238e+00 -8.36224779e-02 2.85191327e-01
9.14545536e-01 3.00736278e-01 2.38078535e-01 2.71886855e-01
1.35673329e-01 -1.21372092e+00 5.75864494e-01 6.43200397e-01
7.13076472e-01 -6.77896738e-01 8.41003895e-01 -1.76531821e-01
-1.46259046e+00 -6.25925437e-02 -6.80773199e-01 -2.22799435e-01
-4.79392678e-01 5.09445548e-01 -4.81313705e-01 7.33654380e-01
6.91091537e-01 1.27862513e+00 -7.55985618e-01 1.03043747e+00
-8.65851119e-02 1.68525055e-01 -2.14035302e-01 3.96576285e-01
5.60276568e-01 6.30806163e-02 1.80204045e-02 1.36165106e+00
3.83270502e-01 -1.85292855e-01 3.07744741e-01 1.02713466e+00
-1.50235847e-01 -3.02924234e-02 -9.22957957e-01 4.63388711e-02
3.50022644e-01 1.41302180e+00 -4.57617164e-01 -4.38907027e-01
-7.70505726e-01 1.06101716e+00 5.19923806e-01 2.52781302e-01
-5.23678362e-01 -7.79969335e-01 9.68844473e-01 -2.14838117e-01
1.12358272e+00 -2.71786842e-02 -1.74790412e-01 -1.06137085e+00
3.29730779e-01 -4.64141607e-01 1.35180637e-01 -3.15717101e-01
-1.59813166e+00 8.46050918e-01 -2.08491921e-01 -1.41816056e+00
-3.09858561e-01 -9.90444839e-01 -5.33188045e-01 1.00796545e+00
-1.93362701e+00 -1.44669342e+00 -5.71489751e-01 8.52727771e-01
1.55948400e-02 -2.84978688e-01 1.13065946e+00 2.42669553e-01
5.87238632e-02 6.35157883e-01 4.45567667e-01 4.00324851e-01
5.70019782e-01 -1.05424440e+00 6.11320734e-01 5.12326956e-01
4.71682042e-01 2.66770482e-01 -5.02820648e-02 -6.18062913e-02
-1.23417771e+00 -1.10637796e+00 8.53909492e-01 -6.85062706e-02
4.64617521e-01 -2.85978526e-01 -8.37880611e-01 3.44830275e-01
1.08262837e-01 8.45229387e-01 7.10002065e-01 -1.25220731e-01
-7.81182826e-01 -2.08893642e-01 -1.22094071e+00 -1.62402794e-01
8.95767152e-01 -1.25399017e+00 -5.36071539e-01 9.04808491e-02
4.28798497e-01 -8.08513686e-02 -1.22756445e+00 5.16659439e-01
8.30254257e-01 -1.17033172e+00 1.03888953e+00 -5.43471932e-01
4.64421988e-01 -1.12102576e-01 -6.49480820e-01 -1.08620727e+00
-6.10769153e-01 -6.09625243e-02 1.71465784e-01 7.25326538e-01
2.23612443e-01 -5.49426258e-01 7.63926446e-01 1.86524048e-01
1.39483169e-01 -8.44744384e-01 -1.21561527e+00 -7.78885424e-01
3.08036149e-01 -2.48345658e-01 7.16138542e-01 6.74751937e-01
-2.64375657e-01 9.19215474e-03 -4.50738333e-02 4.44503464e-02
5.18737137e-01 5.91175735e-01 6.26587570e-01 -1.56500161e+00
1.69613777e-04 -4.18273509e-01 -1.20050752e+00 -9.22887325e-01
3.55328470e-01 -9.68568921e-01 3.28659862e-02 -1.34737539e+00
3.09691757e-01 -2.35208556e-01 -7.37186134e-01 8.05287659e-01
4.81016934e-02 6.92138672e-01 1.25122115e-01 9.85256359e-02
-5.54336429e-01 5.06882906e-01 7.55926847e-01 -5.20538986e-01
8.10326114e-02 -3.84766281e-01 -5.73154807e-01 4.24381047e-01
5.04210174e-01 -5.61006546e-01 1.58003777e-01 -3.87529880e-01
-3.40072215e-01 -1.65287524e-01 7.88516104e-01 -1.38092542e+00
3.50677878e-01 2.49946430e-01 8.49445939e-01 -4.78593826e-01
3.78609180e-01 -1.02075553e+00 -1.34051129e-01 3.08605820e-01
-3.12938035e-01 -1.68392107e-01 1.43435344e-01 3.57892215e-01
-6.95351124e-01 1.00391932e-01 6.26616597e-01 7.16196820e-02
-8.37852597e-01 4.35343742e-01 -1.90003868e-02 -2.66985208e-01
9.72105801e-01 -4.64008689e-01 -2.50314027e-01 -1.78458571e-01
-6.94263339e-01 -3.39031488e-01 6.75372601e-01 3.61342698e-01
7.36617506e-01 -1.68384016e+00 -7.35146999e-01 6.32993221e-01
4.53667462e-01 -2.30560109e-01 -2.78880317e-02 5.00506163e-01
-3.47892880e-01 6.02880597e-01 -6.09078884e-01 -8.96493316e-01
-1.04295397e+00 5.46150625e-01 3.91236067e-01 -1.80369660e-01
-5.50669134e-01 6.20917618e-01 1.79883286e-01 -4.35481846e-01
2.22284481e-01 -4.72894609e-01 2.14454964e-01 1.25989199e-01
5.17018795e-01 9.80979800e-02 2.65512913e-01 -9.19120789e-01
-4.62357908e-01 1.01403546e+00 -1.04941987e-01 2.53465146e-01
1.74732721e+00 3.79473358e-01 -1.48391008e-01 -1.31969959e-01
1.95028031e+00 -3.27887446e-01 -1.38194644e+00 -9.04267192e-01
7.88541436e-02 -5.97336292e-01 1.71572685e-01 -5.39280951e-01
-1.21948957e+00 7.75024056e-01 8.68498623e-01 1.37673631e-01
1.34880352e+00 2.60474443e-01 7.10528851e-01 5.32191873e-01
2.81664640e-01 -8.24301898e-01 -6.09987341e-02 5.06360888e-01
7.59019494e-01 -1.28447068e+00 6.91681579e-02 -1.47877082e-01
-1.28020331e-01 1.33597469e+00 1.71136826e-01 -8.60806286e-01
1.01393819e+00 2.58818388e-01 -2.57752184e-02 -3.93560499e-01
-6.28957987e-01 -5.61644852e-01 7.42772996e-01 7.86905527e-01
4.01461482e-01 -8.81377533e-02 4.10087742e-02 3.43632489e-01
-2.12184638e-02 -2.66767796e-02 -2.64544368e-01 1.22358394e+00
-4.84108984e-01 -1.41662109e+00 -2.06344962e-01 5.22828281e-01
-3.30897331e-01 -9.75909233e-02 -6.33372068e-01 7.95556307e-01
4.62686211e-01 6.95970416e-01 2.72843033e-01 -5.74402750e-01
4.05304670e-01 -3.33438814e-02 6.23652697e-01 -3.66822839e-01
-7.28087902e-01 -1.15525708e-01 -3.11218143e-01 -8.23773801e-01
-7.37192392e-01 -4.87189233e-01 -5.22672176e-01 3.63849886e-02
-1.70754462e-01 -1.12676863e-02 3.84639025e-01 8.44575405e-01
4.87701803e-01 3.45881701e-01 9.86682653e-01 -1.19765341e+00
-4.91958737e-01 -8.39917421e-01 -6.70909882e-01 5.94982624e-01
8.10024738e-01 -6.29703045e-01 -2.21134588e-01 -2.32807584e-02] | [10.245447158813477, 0.11777088791131973] |
2ab448c9-884d-47bd-a4bf-dc4d5c8ffc56 | edge-aware-graph-representation-learning-and | 2007.11240 | null | https://arxiv.org/abs/2007.11240v1 | https://arxiv.org/pdf/2007.11240v1.pdf | Edge-aware Graph Representation Learning and Reasoning for Face Parsing | Face parsing infers a pixel-wise label to each facial component, which has drawn much attention recently. Previous methods have shown their efficiency in face parsing, which however overlook the correlation among different face regions. The correlation is a critical clue about the facial appearance, pose, expression etc., and should be taken into account for face parsing. To this end, we propose to model and reason the region-wise relations by learning graph representations, and leverage the edge information between regions for optimized abstraction. Specifically, we encode a facial image onto a global graph representation where a collection of pixels ("regions") with similar features are projected to each vertex. Our model learns and reasons over relations between the regions by propagating information across vertices on the graph. Furthermore, we incorporate the edge information to aggregate the pixel-wise features onto vertices, which emphasizes on the features around edges for fine segmentation along edges. The finally learned graph representation is projected back to pixel grids for parsing. Experiments demonstrate that our model outperforms state-of-the-art methods on the widely used Helen dataset, and also exhibits the superior performance on the large-scale CelebAMask-HQ and LaPa dataset. The code is available at https://github.com/tegusi/EAGRNet. | ['Hailin Shi', 'Gusi Te', 'Yinglu Liu', 'Wei Hu', 'Tao Mei'] | 2020-07-22 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/1543_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123570256.pdf | eccv-2020-8 | ['face-parsing'] | ['computer-vision'] | [-1.49958609e-02 3.77958924e-01 -2.46235684e-01 -8.86823833e-01
-5.87270975e-01 -3.47420663e-01 3.13411653e-01 -8.55102837e-02
1.66810662e-01 2.44137675e-01 2.85036713e-01 2.11818218e-01
1.13145910e-01 -9.63930547e-01 -7.40390778e-01 -6.39297247e-01
-1.03242680e-01 2.78232187e-01 -3.58540341e-02 -5.36673218e-02
-3.90351266e-02 6.48759842e-01 -1.38992441e+00 2.52182543e-01
3.99424076e-01 1.19050980e+00 -1.29581198e-01 4.59609270e-01
-2.86396414e-01 5.03030896e-01 -5.02672382e-02 -6.27662599e-01
1.13626681e-01 -4.03027773e-01 -6.93963766e-01 4.40111697e-01
7.16729462e-01 -3.19214165e-01 -2.37895906e-01 1.23208320e+00
2.24306181e-01 -1.55203447e-01 5.29730082e-01 -1.29211676e+00
-6.72554374e-01 3.63901258e-01 -1.17316210e+00 -1.84896663e-01
2.21369877e-01 -5.42892516e-02 1.21956694e+00 -9.33426142e-01
6.55756831e-01 1.57687163e+00 5.49045801e-01 6.82746291e-01
-8.95851672e-01 -8.08775127e-01 3.92784536e-01 7.50656128e-02
-1.28463221e+00 -5.21305323e-01 1.06830859e+00 -3.43629330e-01
4.07781601e-01 1.43637151e-01 6.70848727e-01 6.66934967e-01
6.48285225e-02 6.19067430e-01 8.42804074e-01 -1.61612675e-01
-3.36876884e-02 -1.65816665e-01 1.80166498e-01 1.52632475e+00
8.13789815e-02 -2.67217755e-01 -7.74821401e-01 -4.94526811e-02
7.89704382e-01 3.32943089e-02 -1.21749677e-01 -3.11360508e-01
-4.79645193e-01 9.15330231e-01 6.84328139e-01 -1.72433376e-01
-2.34399125e-01 3.82541150e-01 1.59489185e-01 -2.21045047e-01
6.70312822e-01 -2.78053582e-01 -4.79388207e-01 3.23253661e-01
-5.24651825e-01 -8.85895081e-03 6.17804170e-01 7.67968357e-01
1.28674304e+00 -3.03564936e-01 -1.29950643e-01 9.22882199e-01
7.25161731e-01 3.38982224e-01 -8.30053464e-02 -1.19253290e+00
4.73891109e-01 9.15887654e-01 -4.48714077e-01 -1.30064034e+00
-4.32024300e-01 -3.59606408e-02 -6.90582275e-01 2.58566946e-01
3.33321542e-01 -1.59796715e-01 -9.94911730e-01 1.88086498e+00
7.38454521e-01 3.74416560e-01 -1.90421745e-01 8.42438340e-01
1.12671149e+00 5.25116742e-01 2.91063011e-01 3.06176003e-02
1.77186382e+00 -9.81526494e-01 -5.32120466e-01 -4.06254441e-01
3.87052000e-01 -7.00459599e-01 7.66510963e-01 1.62095800e-01
-1.01280832e+00 -4.53451723e-01 -5.88683844e-01 -2.82242745e-01
-3.10019583e-01 4.88274574e-01 9.05088127e-01 5.96559346e-01
-1.14261413e+00 5.61915755e-01 -8.25205326e-01 -3.60573739e-01
8.95187259e-01 5.41032612e-01 -5.87502301e-01 -1.09903462e-01
-8.05028319e-01 2.38122389e-01 2.01705378e-02 5.30897379e-01
-6.26434803e-01 -5.44946551e-01 -1.19966137e+00 5.29096387e-02
3.27323139e-01 -5.93743145e-01 8.98959875e-01 -1.03949463e+00
-1.46332109e+00 1.11944878e+00 -3.86941344e-01 2.15422302e-01
1.31003469e-01 -4.08480726e-02 -1.92446530e-01 2.86343038e-01
9.25293118e-02 1.07684910e+00 8.73319149e-01 -1.37955344e+00
-5.06126881e-01 -7.81625092e-01 3.86687927e-02 1.31789118e-01
-2.68681884e-01 2.27988377e-01 -1.10581195e+00 -3.39208931e-01
4.65393990e-01 -9.55405235e-01 -8.31144676e-02 2.42317602e-01
-5.85806906e-01 -4.01608944e-01 8.08591783e-01 -8.07453692e-01
1.01407254e+00 -2.17631006e+00 1.98710546e-01 3.55962992e-01
4.25710201e-01 -1.61863953e-01 -2.43456602e-01 9.57482755e-02
-5.13923019e-02 1.19464949e-01 -2.90801227e-01 -6.28296494e-01
-1.82495177e-01 2.13078424e-01 1.32352104e-02 6.37054205e-01
4.12445158e-01 9.35438573e-01 -6.82469010e-01 -7.64456868e-01
7.06360713e-02 7.39764154e-01 -7.68290102e-01 2.12805316e-01
-2.94977903e-01 5.26787817e-01 -8.94830644e-01 9.28727269e-01
9.30690348e-01 -3.32758933e-01 2.66502917e-01 -3.85283500e-01
2.27302909e-01 -1.62599698e-01 -9.69986200e-01 1.64450753e+00
-1.49471462e-01 3.68545502e-01 3.92395318e-01 -7.74412036e-01
1.00220215e+00 5.16911224e-03 5.25284827e-01 -3.84531140e-01
3.41508448e-01 -2.19615608e-01 -3.40177178e-01 -5.54623008e-01
1.19071648e-01 1.55651614e-01 2.36037597e-02 3.04148138e-01
1.00408494e-01 3.31919789e-02 -8.36127102e-02 2.74890900e-01
7.04563260e-01 3.76739144e-01 1.17843308e-01 -2.20972016e-01
5.71990252e-01 -3.55875760e-01 8.16151381e-01 -4.13555130e-02
-6.83550909e-02 7.99637020e-01 9.89989519e-01 -4.49072868e-01
-3.99615973e-01 -8.42837155e-01 -6.08594976e-02 1.17367744e+00
2.77754992e-01 -7.32273340e-01 -1.20651948e+00 -8.11155260e-01
3.14679779e-02 4.61167796e-03 -1.06282163e+00 1.47483155e-01
-5.99230409e-01 -8.18525493e-01 1.33991078e-01 5.94389439e-01
4.27789032e-01 -1.16702223e+00 -1.45886987e-01 -2.13529542e-01
3.75070199e-02 -1.28453863e+00 -5.21835208e-01 -2.07962111e-01
-7.08853424e-01 -1.17782533e+00 -2.64615715e-01 -9.23009872e-01
1.20998561e+00 -7.98503682e-03 1.01903582e+00 4.27021414e-01
-4.90051955e-01 5.13941646e-01 -1.83123887e-01 -9.63773429e-02
1.80188686e-01 -1.04778543e-01 -3.62762570e-01 5.41235566e-01
1.53834835e-01 -4.05094296e-01 -7.42236853e-01 1.72511995e-01
-5.33450007e-01 2.25380376e-01 3.67093027e-01 5.83621204e-01
9.76017535e-01 -1.16601080e-01 1.21337764e-01 -1.48520732e+00
1.85523048e-01 -5.72349012e-01 -6.16731703e-01 3.47676396e-01
-1.49725288e-01 1.00045025e-01 2.84299672e-01 7.98596367e-02
-1.14342165e+00 3.70599240e-01 -1.95912629e-01 -4.34255600e-01
-1.30754888e-01 1.96919709e-01 -6.78861201e-01 -2.15807006e-01
-2.41934787e-02 -2.66378403e-01 -4.45677228e-02 -3.44950348e-01
5.20512700e-01 2.98770785e-01 4.26312387e-01 -8.74773681e-01
5.56595445e-01 6.86392725e-01 2.85270333e-01 -7.25808203e-01
-9.44044590e-01 -1.98743075e-01 -6.73427641e-01 -3.11030269e-01
1.15060067e+00 -7.96514332e-01 -7.73565888e-01 3.72642666e-01
-1.30243623e+00 -3.57251287e-01 1.38032317e-01 8.50233436e-02
-4.35942203e-01 3.43405932e-01 -7.37227261e-01 -6.24332428e-01
-1.78224817e-01 -1.23557127e+00 1.50373077e+00 7.54424930e-01
-6.76882919e-03 -1.09171116e+00 -2.29043365e-01 4.82152224e-01
-1.78429872e-01 3.68611872e-01 8.12423110e-01 -9.42870975e-02
-7.07239866e-01 6.64691851e-02 -6.32947206e-01 9.63280722e-02
1.64119422e-01 3.94167155e-01 -1.22235501e+00 4.00510840e-02
-3.16027164e-01 -1.99276701e-01 9.05352890e-01 5.17580092e-01
1.72852242e+00 -1.22571610e-01 -4.61881548e-01 1.05132151e+00
1.25998163e+00 -1.92355990e-01 4.63942111e-01 -2.09627718e-01
1.19873536e+00 1.03340971e+00 4.79920000e-01 4.58463192e-01
6.35916591e-01 5.02167583e-01 6.40965581e-01 -2.56282538e-01
-3.29908669e-01 -4.63614613e-01 2.56067336e-01 6.13053977e-01
-9.75666568e-02 -1.11841962e-01 -7.33402729e-01 2.44826511e-01
-1.72960317e+00 -4.41152215e-01 -1.62296772e-01 1.65475690e+00
5.40318608e-01 -1.91420093e-01 -1.00697875e-01 -4.50779974e-01
8.20353448e-01 3.49775076e-01 -4.69992250e-01 -3.68621707e-01
1.70661837e-01 3.53974253e-01 2.82034308e-01 6.65690362e-01
-1.00631320e+00 1.29214549e+00 5.38907528e+00 6.16005719e-01
-1.06854224e+00 9.78491753e-02 1.33856809e+00 2.71609664e-01
-3.13328236e-01 6.50166050e-02 -1.00438488e+00 2.52467245e-01
4.03793722e-01 1.78058118e-01 4.72310960e-01 8.80767763e-01
3.44993221e-03 -2.26506758e-02 -9.97880757e-01 9.77837145e-01
2.68646061e-01 -1.04807162e+00 1.05468100e-02 1.72844782e-01
6.85210466e-01 -2.48864621e-01 1.26387984e-01 2.81675681e-02
2.90427595e-01 -1.22148871e+00 5.44311881e-01 4.94518936e-01
8.33358169e-01 -6.66556060e-01 4.13842112e-01 -8.79319310e-02
-1.62569094e+00 2.14482322e-01 -5.00009775e-01 1.76304504e-01
-1.01074882e-01 3.70908767e-01 -7.00089216e-01 3.52763623e-01
8.06214869e-01 7.62436986e-01 -6.70449615e-01 3.57292354e-01
-5.96909285e-01 4.67819661e-01 -1.83894962e-01 2.65370548e-01
5.85334636e-02 -5.99737048e-01 -6.06562048e-02 9.44163978e-01
2.76705950e-01 4.44695801e-01 1.04353331e-01 9.16850984e-01
-5.14254928e-01 2.62559235e-01 -4.27933425e-01 2.11380161e-02
3.01356435e-01 1.83906102e+00 -1.03917682e+00 -1.23300277e-01
-6.67210698e-01 9.97070074e-01 7.48104274e-01 3.30790281e-01
-8.78705740e-01 7.28409588e-02 9.83206391e-01 9.77363139e-02
3.42132688e-01 -9.65216290e-03 -3.15205693e-01 -8.36811006e-01
-8.20652489e-03 -4.01319027e-01 5.52292228e-01 -7.84716010e-01
-1.33437300e+00 7.48409748e-01 -1.68041348e-01 -7.93700814e-01
1.25283793e-01 -9.24527824e-01 -8.66886675e-01 7.48306990e-01
-1.64042592e+00 -1.40702558e+00 -5.51633000e-01 6.58572137e-01
4.07776654e-01 -4.20543849e-02 7.40321040e-01 4.24976945e-02
-9.30152714e-01 6.35850191e-01 -5.73496819e-01 5.50451279e-01
3.89885426e-01 -9.96690094e-01 3.85021418e-01 6.11182272e-01
2.92316765e-01 5.92481077e-01 2.47624457e-01 -6.79301381e-01
-1.54663694e+00 -1.36062455e+00 7.44686544e-01 -3.24859828e-01
4.85643864e-01 -5.00128567e-01 -8.99126947e-01 7.84168482e-01
4.32424098e-02 6.97003663e-01 5.84621310e-01 3.09456497e-01
-4.81133610e-01 -3.37169677e-01 -1.17133284e+00 5.11627555e-01
1.26122415e+00 -4.82471496e-01 -2.90480331e-02 2.84237176e-01
5.29952526e-01 -4.79488283e-01 -7.91230798e-01 2.17787534e-01
6.04526520e-01 -1.06148827e+00 8.93336475e-01 -4.73818898e-01
5.68438947e-01 -1.45980865e-01 -8.01673755e-02 -1.09232330e+00
-2.17644393e-01 -4.09416765e-01 1.98388174e-01 1.59989727e+00
3.40038210e-01 -6.07754767e-01 1.22428131e+00 7.31321514e-01
1.57910705e-01 -1.07330644e+00 -7.77107835e-01 -1.98162608e-02
-1.55484825e-01 -3.28644961e-01 8.96346807e-01 7.55924940e-01
-2.54904240e-01 2.46307641e-01 -1.96469784e-01 4.96227115e-01
7.86317945e-01 3.54321659e-01 6.96398497e-01 -1.25968575e+00
-5.96738271e-02 -4.59745437e-01 -4.82324600e-01 -8.67651343e-01
6.17339611e-01 -9.47478414e-01 -8.73148516e-02 -1.34572387e+00
4.27606523e-01 -5.52551627e-01 -5.31504825e-02 7.24932969e-01
-3.83774430e-01 6.60001099e-01 7.55222812e-02 -1.54859528e-01
-5.22611022e-01 4.50358510e-01 1.49016094e+00 3.06634754e-02
1.06825545e-01 -2.77222544e-01 -9.07148540e-01 9.90636587e-01
8.67069960e-01 -4.16650444e-01 -2.77391464e-01 -5.87739408e-01
8.90816376e-02 1.50143340e-01 3.34368169e-01 -6.65300369e-01
1.76551282e-01 -2.24723265e-01 6.13179445e-01 -2.98221618e-01
5.49155176e-01 -6.95953250e-01 7.99226835e-02 6.70794994e-02
-1.01741657e-01 -8.36674348e-02 4.16505150e-02 4.71368551e-01
-2.14380428e-01 1.26862396e-02 7.66199470e-01 -1.09445877e-01
-8.75531673e-01 1.00585723e+00 5.35304248e-01 3.41321044e-02
1.05102599e+00 -6.96533397e-02 -1.56692132e-01 -2.07757831e-01
-8.19162846e-01 3.21483910e-01 4.42771494e-01 3.35859805e-01
6.57107294e-01 -1.33073795e+00 -6.70820415e-01 6.58556044e-01
9.20950919e-02 1.95761815e-01 3.57640415e-01 5.17828286e-01
-6.47326946e-01 -6.08221255e-03 -2.68785268e-01 -7.46121228e-01
-1.65410340e+00 1.08712338e-01 3.55710864e-01 9.23986733e-02
-5.29714465e-01 1.31629848e+00 7.83929467e-01 -3.39284092e-01
-1.07459305e-03 -3.11133415e-01 -4.18536991e-01 1.29545033e-01
3.01777095e-01 1.18001476e-02 -1.44469604e-01 -1.02865386e+00
-4.54369277e-01 1.29398882e+00 3.21041308e-02 1.78829208e-01
1.37477469e+00 -1.14413381e-01 -4.37443554e-01 -2.79282313e-02
1.44770384e+00 2.56181061e-01 -1.46094799e+00 -7.38717541e-02
-1.95024535e-01 -5.65491676e-01 2.72496305e-02 -2.66841024e-01
-1.92798078e+00 1.05537045e+00 3.58293772e-01 -1.36451080e-01
1.25994623e+00 5.77206016e-01 6.34244978e-01 -2.95034915e-01
2.31408983e-01 -7.22677827e-01 1.67624647e-04 2.25939035e-01
7.64249444e-01 -1.30344892e+00 1.14936247e-01 -1.21948183e+00
-7.29364514e-01 1.23866236e+00 7.84657836e-01 -2.00016111e-01
7.75672615e-01 3.18194181e-01 5.52192144e-02 -5.62450469e-01
-5.27936876e-01 -1.31051123e-01 3.94618839e-01 4.56988841e-01
5.34807384e-01 4.28419620e-01 5.33072017e-02 8.28766644e-01
-2.74279386e-01 -3.03309560e-01 -6.09200895e-02 4.61507380e-01
-3.14544082e-01 -1.25294805e+00 -2.88810313e-01 2.84294367e-01
-3.84873569e-01 1.28286898e-01 -4.98168528e-01 6.42444432e-01
4.84685719e-01 6.79411232e-01 2.58619279e-01 -2.17635155e-01
1.32599190e-01 -1.24214463e-01 6.44053638e-01 -8.34658623e-01
-3.63474429e-01 5.80253974e-02 -1.31387174e-01 -8.88295531e-01
-1.54349610e-01 -7.95437753e-01 -1.60755837e+00 -2.97186017e-01
9.01673734e-02 -2.49702949e-02 6.46282971e-01 6.81944430e-01
3.49366337e-01 4.35456008e-01 6.74675167e-01 -8.79643202e-01
1.30275980e-01 -5.72271824e-01 -7.12379634e-01 5.11328995e-01
1.31360874e-01 -7.48615146e-01 -2.45373175e-01 1.86047539e-01] | [13.429743766784668, 0.6541008353233337] |
eceabe2e-c250-4c53-afaf-9365629ef701 | point-cloud-classification-using-content | 2303.04599 | null | https://arxiv.org/abs/2303.04599v1 | https://arxiv.org/pdf/2303.04599v1.pdf | Point Cloud Classification Using Content-based Transformer via Clustering in Feature Space | Recently, there have been some attempts of Transformer in 3D point cloud classification. In order to reduce computations, most existing methods focus on local spatial attention, but ignore their content and fail to establish relationships between distant but relevant points. To overcome the limitation of local spatial attention, we propose a point content-based Transformer architecture, called PointConT for short. It exploits the locality of points in the feature space (content-based), which clusters the sampled points with similar features into the same class and computes the self-attention within each class, thus enabling an effective trade-off between capturing long-range dependencies and computational complexity. We further introduce an Inception feature aggregator for point cloud classification, which uses parallel structures to aggregate high-frequency and low-frequency information in each branch separately. Extensive experiments show that our PointConT model achieves a remarkable performance on point cloud shape classification. Especially, our method exhibits 90.3% Top-1 accuracy on the hardest setting of ScanObjectNN. Source code of this paper is available at https://github.com/yahuiliu99/PointConT. | ['FeiYue Wang', 'Lingxi Li', 'Yisheng Lv', 'Bin Tian', 'Yahui Liu'] | 2023-03-08 | null | null | null | null | ['3d-point-cloud-classification', 'point-cloud-classification'] | ['computer-vision', 'computer-vision'] | [-4.04962271e-01 -5.08496284e-01 -2.14554101e-01 -4.26859587e-01
-9.40692782e-01 -4.19410110e-01 4.45286036e-01 4.82814699e-01
1.34213507e-01 1.51748568e-01 1.06730074e-01 -3.33864957e-01
-4.27125424e-01 -1.16885293e+00 -9.13011074e-01 -6.78234637e-01
-4.84346375e-02 6.26857996e-01 5.21572888e-01 -9.86818597e-02
4.48520482e-01 9.34619784e-01 -1.79822493e+00 3.79812509e-01
8.53300750e-01 1.33592510e+00 2.59858221e-01 2.01670378e-01
-5.77739835e-01 3.46561372e-01 -3.66762638e-01 2.21265834e-02
1.70962423e-01 6.54864982e-02 -8.04636955e-01 -2.73460180e-01
6.17413461e-01 -1.28755227e-01 -4.15037125e-01 1.04298449e+00
4.91480291e-01 -9.00269225e-02 4.19774950e-01 -1.41385019e+00
-8.64097118e-01 4.78092611e-01 -7.74636388e-01 4.88024533e-01
1.54622987e-01 2.69217044e-02 1.16965747e+00 -1.40965259e+00
1.96126342e-01 1.21742237e+00 7.23398685e-01 -3.36774364e-02
-1.00534499e+00 -8.97147655e-01 2.71807849e-01 3.78312618e-01
-1.71768153e+00 -5.03609516e-02 1.04705656e+00 -3.23805302e-01
1.01123989e+00 3.73601496e-01 9.62019026e-01 5.00849009e-01
1.29375324e-01 7.44036376e-01 7.52715051e-01 -1.15864493e-01
4.48177941e-02 -3.66034001e-01 3.65054369e-01 4.58525896e-01
9.64448527e-02 4.28525731e-02 -5.78736365e-01 -4.04001325e-01
8.65172029e-01 5.68998933e-01 -2.60563940e-01 -4.64662313e-01
-1.24620497e+00 8.63129973e-01 1.10665464e+00 3.83774847e-01
-3.03063303e-01 2.48751536e-01 2.30031863e-01 1.95145577e-01
7.58408368e-01 3.34750488e-02 -3.07075709e-01 4.67569046e-02
-5.92433751e-01 2.60591298e-01 3.34475011e-01 1.35559034e+00
1.08015943e+00 -3.99118006e-01 -1.63105652e-01 8.52456093e-01
2.81400800e-01 7.40562975e-01 2.33985081e-01 -4.47771877e-01
5.36306977e-01 9.64627683e-01 -3.05312514e-01 -1.18077135e+00
-2.28273526e-01 -5.39628625e-01 -7.01986134e-01 1.34870991e-01
-1.43834621e-01 4.75530088e-01 -8.70869935e-01 1.19960856e+00
6.14506721e-01 4.70611572e-01 -5.43859482e-01 9.09101844e-01
1.00945020e+00 8.00969064e-01 -5.60410880e-02 1.88622713e-01
1.30773938e+00 -7.90970027e-01 -1.45213082e-01 2.90931374e-01
3.23244244e-01 -7.28822947e-01 1.29512811e+00 1.71926282e-02
-1.17061841e+00 -7.41406143e-01 -8.08154941e-01 -3.40595931e-01
-4.09523726e-01 -2.59563297e-01 6.57568514e-01 6.49481341e-02
-9.22161818e-01 6.58516943e-01 -9.05861497e-01 -2.04527110e-01
7.69471943e-01 4.57580954e-01 -1.18487418e-01 -2.12312620e-02
-7.21906483e-01 3.22326958e-01 -8.06168765e-02 -4.48743403e-02
-4.64080006e-01 -1.11723924e+00 -6.08804345e-01 2.05638453e-01
8.35336596e-02 -5.59775829e-01 1.05341578e+00 -4.26999420e-01
-1.00105262e+00 7.66396761e-01 -4.79129374e-01 -2.52648294e-02
1.46843508e-01 -3.26950192e-01 -1.76228926e-01 -3.02195246e-03
4.74218935e-01 6.51722908e-01 6.65694475e-01 -1.21481645e+00
-9.10189152e-01 -7.29085624e-01 -6.61897957e-02 2.35192686e-01
-2.83378720e-01 -4.08028886e-02 -8.55513215e-01 -5.90407133e-01
6.44470751e-01 -7.84312546e-01 9.44651961e-02 1.42964676e-01
-2.94251531e-01 -8.00553322e-01 1.14266753e+00 6.37365431e-02
1.03548813e+00 -2.31646752e+00 -1.67472735e-01 4.08286452e-01
4.68374401e-01 2.50906730e-03 -3.50806527e-02 5.17960608e-01
-2.32992217e-01 1.48235843e-01 -2.11426899e-01 -2.57603854e-01
-1.54184923e-01 1.69267982e-01 -7.11585402e-01 4.59764987e-01
2.58539706e-01 1.06494129e+00 -8.47393572e-01 -5.40078044e-01
3.73331487e-01 5.82140446e-01 -6.61585450e-01 5.90884089e-02
-8.86371583e-02 2.27535501e-01 -7.89920688e-01 1.02873707e+00
1.08921969e+00 -4.91416067e-01 -4.83553320e-01 -2.36470714e-01
-2.79872268e-01 5.39673448e-01 -7.86379635e-01 1.69554996e+00
-4.27484661e-01 3.69229704e-01 -2.46912882e-01 -6.68386102e-01
9.67398286e-01 -5.30551970e-02 8.94827247e-01 -6.56956971e-01
1.66690873e-03 2.67478019e-01 -2.57100940e-01 -1.61779851e-01
4.25031453e-01 1.74195662e-01 5.04543334e-02 1.85310990e-01
-2.39081547e-01 -3.15593392e-01 -4.41412479e-01 -6.02196902e-02
1.06774545e+00 2.98817083e-02 1.26440391e-01 -5.01969635e-01
3.74290556e-01 1.87708095e-01 3.04022074e-01 5.71961761e-01
4.52495087e-03 9.35248733e-01 1.60909593e-01 -7.09862173e-01
-8.79031003e-01 -1.13053989e+00 -5.28801799e-01 9.41713810e-01
6.33293748e-01 -5.55773973e-01 -2.56112337e-01 -6.40880823e-01
5.74700773e-01 3.24806482e-01 -5.88563442e-01 -2.23393030e-02
-6.93445504e-01 -4.58535641e-01 1.35180339e-01 6.82183862e-01
4.20944840e-01 -9.40363109e-01 -4.36089486e-01 2.59507019e-02
-9.37016830e-02 -6.45151973e-01 -4.85672981e-01 1.27236336e-01
-1.06799781e+00 -9.06613231e-01 -4.88233894e-01 -8.37664723e-01
6.45532310e-01 8.71325552e-01 1.28475010e+00 4.81245011e-01
-3.59681733e-02 1.54894426e-01 -5.39005756e-01 -5.52903473e-01
4.42186803e-01 4.04448092e-01 -3.24384928e-01 -4.10715006e-02
8.44223142e-01 -8.44423950e-01 -6.18485570e-01 5.02853036e-01
-5.43914318e-01 -1.17848195e-01 5.14048874e-01 8.10258806e-01
9.99876082e-01 -1.72529832e-01 2.26000816e-01 -6.39907956e-01
3.42224360e-01 -7.64260352e-01 -5.57859182e-01 -8.92960802e-02
-3.50110799e-01 -2.09194586e-01 3.29869002e-01 -1.88954458e-01
-3.93067002e-01 6.69609848e-03 -3.51522952e-01 -9.68400598e-01
-1.43856406e-01 4.93666559e-01 -1.27215967e-01 -3.57831597e-01
3.45436096e-01 4.17632073e-01 -2.33352467e-01 -7.20511556e-01
8.91975239e-02 6.40637755e-01 2.81625837e-01 -8.48051250e-01
8.94936621e-01 7.97093272e-01 -5.89215197e-02 -7.48384714e-01
-8.15701246e-01 -6.92066252e-01 -7.60610282e-01 -5.82604483e-03
5.73518813e-01 -8.86033058e-01 -9.29260254e-01 2.33580709e-01
-1.31432569e+00 3.13465782e-02 -5.85985124e-01 1.93920866e-01
-3.88607383e-01 1.56402171e-01 -3.76693666e-01 -5.26036620e-01
-3.97860497e-01 -1.13606930e+00 1.60790324e+00 1.00181609e-01
1.40672937e-01 -5.21023810e-01 2.78011500e-03 -6.51149526e-02
3.88400227e-01 6.71675354e-02 1.00028825e+00 -4.03991997e-01
-1.15747607e+00 -8.49791467e-02 -5.36017239e-01 -1.82448119e-01
1.30363747e-01 -5.97919710e-02 -9.40934002e-01 -2.54038632e-01
-4.57830466e-02 -2.68676765e-02 8.26280534e-01 3.41112882e-01
1.77370071e+00 1.00483455e-01 -6.82337821e-01 9.55165565e-01
1.46801603e+00 1.27623677e-01 5.59195161e-01 1.09324381e-01
9.87074196e-01 2.01934844e-01 5.73935330e-01 4.49326038e-01
6.00125253e-01 8.36696446e-01 6.63824260e-01 -2.35537589e-01
-1.04176596e-01 -3.55707914e-01 -1.96151033e-01 9.54592943e-01
-2.76659817e-01 -9.00995657e-02 -1.17165637e+00 6.93267465e-01
-1.83521175e+00 -8.40174556e-01 -3.58961165e-01 2.15871787e+00
3.95239264e-01 9.04489085e-02 9.97490436e-02 5.85149676e-02
7.24415123e-01 2.41482660e-01 -4.85243767e-01 1.57256439e-01
-2.46011987e-02 3.51756245e-01 4.80596900e-01 3.67494553e-01
-1.24222541e+00 7.98897982e-01 5.65739536e+00 1.04568923e+00
-1.21501124e+00 1.56254113e-01 4.70417202e-01 -1.87955037e-01
-4.22885120e-01 -1.27807213e-02 -8.05836678e-01 6.84635997e-01
4.02825207e-01 -1.51953295e-01 8.66935626e-02 1.09726536e+00
-1.07391454e-01 2.81159937e-01 -9.39555049e-01 1.17390370e+00
-6.81612417e-02 -1.44744074e+00 2.70117909e-01 2.42666557e-01
3.95729542e-01 5.66657543e-01 3.25080939e-02 1.17452137e-01
1.26093738e-02 -8.75645757e-01 8.34190309e-01 3.98452371e-01
7.56877601e-01 -8.88857186e-01 6.55315936e-01 2.35496849e-01
-1.74421108e+00 6.82970881e-02 -6.69179082e-01 3.93171720e-02
-8.25114846e-02 7.71290004e-01 -4.49265987e-01 6.64071679e-01
1.13094962e+00 1.09681737e+00 -4.12147820e-01 1.32060361e+00
1.04152650e-01 4.18024868e-01 -6.72072589e-01 -7.86733702e-02
2.60993868e-01 -1.31035700e-01 5.92070341e-01 9.72877562e-01
6.36760414e-01 2.38738820e-01 3.88221949e-01 8.41131866e-01
-9.65754036e-03 1.25331253e-01 -7.36896753e-01 4.79301572e-01
8.48920524e-01 9.47219431e-01 -7.89470255e-01 -2.89449781e-01
-5.84582329e-01 6.88017786e-01 5.53668201e-01 6.82425871e-02
-9.21716630e-01 -4.83545989e-01 1.02925885e+00 4.48996365e-01
5.49153328e-01 -3.51073980e-01 -7.51506031e-01 -9.71451998e-01
2.85159737e-01 -4.72858906e-01 4.15173829e-01 -8.28903079e-01
-1.40683532e+00 7.80488372e-01 6.25176579e-02 -1.75580609e+00
5.25163710e-01 -4.42883432e-01 -6.81609690e-01 8.78635526e-01
-1.62794292e+00 -1.26704562e+00 -7.10507989e-01 7.20404446e-01
5.91105998e-01 6.28838018e-02 4.75057244e-01 5.26988328e-01
-1.66941404e-01 4.46899951e-01 -6.43401071e-02 -4.90134535e-03
2.66067713e-01 -1.12198007e+00 7.32806683e-01 4.34882909e-01
2.92062640e-01 7.60304749e-01 7.74149224e-02 -7.43100643e-01
-1.46052015e+00 -1.14665258e+00 9.42329943e-01 -6.53707743e-01
4.98828709e-01 -4.79999334e-01 -1.14000273e+00 5.62769592e-01
-7.39118010e-02 4.04424787e-01 4.43569243e-01 2.35241726e-01
-5.14072120e-01 -3.40601653e-01 -8.97125006e-01 4.19238359e-01
1.57362032e+00 -6.11898482e-01 -5.37117600e-01 4.65252846e-01
9.37868476e-01 -5.74957669e-01 -9.45808053e-01 4.99175847e-01
1.14915043e-01 -9.32522714e-01 1.29629064e+00 -3.37385833e-01
4.39752698e-01 -6.05485201e-01 -2.20843375e-01 -1.06918645e+00
-8.86662364e-01 -2.12097466e-01 3.41141075e-02 1.08455801e+00
3.15708548e-01 -7.94054627e-01 9.02333021e-01 1.75116718e-01
-4.65476066e-01 -1.01276445e+00 -1.06267166e+00 -7.85441279e-01
1.16805419e-01 -6.28923178e-01 1.28639495e+00 9.86462295e-01
-1.48474559e-01 2.36200884e-01 1.31679937e-01 4.89615440e-01
4.82389182e-01 9.51479793e-01 7.70981252e-01 -1.48608601e+00
9.30952653e-02 -6.31976962e-01 -7.14434862e-01 -1.46718907e+00
-1.08912595e-01 -1.23286176e+00 -1.66902795e-01 -1.38746166e+00
1.90928623e-01 -1.13971722e+00 -4.13515121e-01 6.58169687e-01
-8.02323744e-02 3.48133922e-01 2.13709697e-01 7.50824213e-01
-4.66749161e-01 7.69289255e-01 1.15744734e+00 -2.34117851e-01
-5.26516475e-02 7.89483786e-02 -6.65122569e-01 5.04403234e-01
7.09533751e-01 -6.57650650e-01 -2.68482149e-01 -8.90749216e-01
-6.96443170e-02 -2.73520947e-01 5.86730719e-01 -1.26961613e+00
4.93427306e-01 -1.81784526e-01 3.98723841e-01 -1.33325911e+00
4.14052695e-01 -1.09094238e+00 1.65423051e-01 2.36716866e-01
4.72970977e-02 4.85680878e-01 1.72584429e-01 5.14340878e-01
-4.14391160e-01 1.80367991e-01 6.78822339e-01 -6.52801991e-03
-5.99826872e-01 8.86225045e-01 3.08272451e-01 -1.48051783e-01
1.07351756e+00 -2.28904620e-01 -3.16063792e-01 -7.91553259e-02
-4.19786930e-01 2.84705997e-01 5.90492189e-01 4.80731636e-01
8.51388335e-01 -1.63621187e+00 -6.50248945e-01 5.28170884e-01
3.38331401e-01 4.69725013e-01 2.89600253e-01 6.09015644e-01
-6.13480806e-01 3.50172758e-01 -1.30747810e-01 -1.25460804e+00
-1.06347048e+00 6.81062400e-01 2.83965945e-01 1.80083364e-01
-1.12042081e+00 1.02019572e+00 3.89343530e-01 -2.82551110e-01
-1.08856512e-02 -5.83154023e-01 8.87401775e-02 -1.25757888e-01
4.48754698e-01 1.46988615e-01 3.93957585e-01 -6.61391795e-01
-6.55775964e-01 1.21139061e+00 -7.63923004e-02 2.83298194e-01
1.48445880e+00 1.81749195e-01 -2.83062905e-01 3.62033069e-01
1.37132847e+00 1.41434863e-01 -1.02848589e+00 -4.35218692e-01
-1.42090738e-01 -1.04491651e+00 1.63207039e-01 -2.57505476e-01
-1.32879138e+00 9.04029846e-01 4.52716321e-01 5.27188897e-01
1.13945949e+00 4.51873660e-01 7.88376629e-01 1.76336318e-01
6.78496718e-01 -5.42894423e-01 -3.08895111e-01 6.85567558e-01
1.16263556e+00 -1.12535608e+00 -1.98461278e-03 -8.42174113e-01
-1.92679167e-01 8.60630751e-01 6.31585419e-01 -4.60929155e-01
1.10210037e+00 3.16023529e-01 -1.31702319e-01 -7.12154210e-01
-6.86246514e-01 -2.84075022e-01 2.61351794e-01 4.92926687e-01
2.63572752e-01 8.60324875e-02 1.95663739e-02 2.74840653e-01
-3.24432015e-01 -2.51228571e-01 -2.51644224e-01 9.89085317e-01
-3.69491577e-01 -8.64871502e-01 -3.77990901e-01 5.78847289e-01
3.24589288e-04 -2.11821243e-01 -2.45592684e-01 7.27444828e-01
2.21998528e-01 5.55473268e-01 5.85108757e-01 -5.45638859e-01
6.61546111e-01 -2.25215122e-01 9.80418250e-02 -5.50843179e-01
-6.14401221e-01 1.14612088e-01 -4.57963318e-01 -7.38464415e-01
-1.77033052e-01 -7.07351923e-01 -1.24421132e+00 -4.95226443e-01
-2.49045283e-01 2.51146436e-01 5.09424269e-01 4.92185712e-01
9.06616986e-01 4.58748996e-01 8.11172724e-01 -1.09766459e+00
-2.79932678e-01 -6.88751101e-01 -3.20722163e-01 3.47730696e-01
4.65347230e-01 -9.11808074e-01 -1.33483976e-01 -3.44555706e-01] | [7.8443145751953125, -3.4241747856140137] |
9808f6ca-9ae1-4927-8fd7-9da3e4becbf7 | gnn-encoder-learning-a-dual-encoder | 2204.08241 | null | https://arxiv.org/abs/2204.08241v2 | https://arxiv.org/pdf/2204.08241v2.pdf | GNN-encoder: Learning a Dual-encoder Architecture via Graph Neural Networks for Dense Passage Retrieval | Recently, retrieval models based on dense representations are dominant in passage retrieval tasks, due to their outstanding ability in terms of capturing semantics of input text compared to the traditional sparse vector space models. A common practice of dense retrieval models is to exploit a dual-encoder architecture to represent a query and a passage independently. Though efficient, such a structure loses interaction between the query-passage pair, resulting in inferior accuracy. To enhance the performance of dense retrieval models without loss of efficiency, we propose a GNN-encoder model in which query (passage) information is fused into passage (query) representations via graph neural networks that are constructed by queries and their top retrieved passages. By this means, we maintain a dual-encoder structure, and retain some interaction information between query-passage pairs in their representations, which enables us to achieve both efficiency and efficacy in passage retrieval. Evaluation results indicate that our method significantly outperforms the existing models on MSMARCO, Natural Questions and TriviaQA datasets, and achieves the new state-of-the-art on these datasets. | ['Rui Yan', 'Dongyan Zhao', 'Wei Wu', 'Jingang Wang', 'Yang Yang', 'Jiahao Liu', 'Jiduan Liu'] | 2022-04-18 | null | null | null | null | ['natural-questions', 'triviaqa', 'passage-retrieval'] | ['miscellaneous', 'miscellaneous', 'natural-language-processing'] | [-2.61692077e-01 -3.51682514e-01 -4.23712611e-01 2.34553944e-02
-9.92680252e-01 -4.66223001e-01 9.24773395e-01 4.47435915e-01
-3.12773228e-01 5.08795559e-01 8.17175806e-01 -1.83477059e-01
-4.05209094e-01 -1.19752026e+00 -6.28786683e-01 -3.12395841e-01
-8.37194398e-02 5.62576890e-01 2.93291628e-01 -7.68692017e-01
2.45672762e-01 1.25706822e-01 -1.31046367e+00 3.29334557e-01
8.59042287e-01 7.84630835e-01 1.68233335e-01 4.80389029e-01
-5.21797299e-01 1.17334807e+00 -4.26294088e-01 -4.82197970e-01
1.72220692e-02 -4.58711088e-01 -1.18189991e+00 -2.99206614e-01
2.66575664e-01 -6.09815121e-01 -1.32672119e+00 8.49901736e-01
4.08411652e-01 5.24503827e-01 7.77611375e-01 -7.83502817e-01
-1.41001844e+00 6.76799893e-01 -3.92196864e-01 3.11948061e-01
7.81902373e-01 -4.73319143e-01 1.74484158e+00 -7.94800162e-01
6.14476323e-01 1.22104168e+00 2.51922548e-01 2.04539642e-01
-1.04328036e+00 -3.18308532e-01 -2.11860333e-02 4.17579800e-01
-1.66123545e+00 -3.58476937e-01 7.73325503e-01 8.32452402e-02
1.23011410e+00 2.96012461e-01 6.35987699e-01 9.32684183e-01
2.28884384e-01 1.14211297e+00 9.19629782e-02 -4.05269384e-01
-1.22721441e-01 -1.25890568e-01 3.96568060e-01 6.77911520e-01
1.81976050e-01 -1.52990028e-01 -4.03283358e-01 -3.94193649e-01
7.64827609e-01 5.39421856e-01 -6.29084229e-01 -4.32521403e-01
-9.54837084e-01 1.15711570e+00 1.01990151e+00 5.94067752e-01
-5.69813371e-01 6.96111545e-02 4.41229165e-01 5.87300599e-01
3.98274988e-01 4.87353772e-01 5.64765409e-02 -2.47972868e-02
-8.92885625e-01 4.60293829e-01 9.29478824e-01 1.19499540e+00
6.98738396e-01 -2.67715037e-01 -7.06112981e-01 1.14228630e+00
4.73204345e-01 3.50732774e-01 7.59364605e-01 -4.24809366e-01
6.36293709e-01 9.55650210e-01 -6.41982118e-03 -1.35512900e+00
-8.23876634e-02 -6.34119570e-01 -1.07497478e+00 -8.83356214e-01
-3.23671281e-01 3.77798200e-01 -7.78870046e-01 1.44195879e+00
-1.34995252e-01 -6.32292405e-02 2.75018096e-01 9.70101535e-01
1.23270392e+00 1.09752464e+00 -4.78135832e-02 3.62780392e-02
1.21304226e+00 -1.18903816e+00 -8.02813113e-01 -5.67939021e-02
8.40892315e-01 -7.34902740e-01 9.98412967e-01 -3.71131927e-01
-1.06011939e+00 -5.03529191e-01 -8.43093812e-01 -6.32358789e-01
-4.52356547e-01 -1.29682114e-02 7.13962972e-01 5.75573929e-02
-1.23415935e+00 4.05252844e-01 -4.57033992e-01 -4.68613416e-01
1.87221706e-01 1.00990832e-01 -2.91959316e-01 -6.36314094e-01
-1.69226956e+00 6.27304435e-01 3.74409795e-01 4.05738391e-02
-6.79783821e-01 -5.79380155e-01 -1.18831336e+00 6.97346985e-01
7.38873184e-02 -1.04361224e+00 1.14002085e+00 -3.24925393e-01
-1.19083750e+00 4.15362090e-01 -2.34992549e-01 -3.59620661e-01
-1.45136341e-01 -2.85545319e-01 -5.19974172e-01 6.02701783e-01
1.46477431e-01 5.10920525e-01 7.21040606e-01 -1.01772881e+00
-2.52322614e-01 -1.85674414e-01 5.75875759e-01 4.71869022e-01
-6.30867720e-01 -2.43706286e-01 -1.09391582e+00 -4.53600049e-01
1.39789596e-01 -6.46373034e-01 -1.81336716e-01 -2.60642827e-01
-2.78965741e-01 -6.27655625e-01 6.98035896e-01 -5.05219162e-01
1.66771305e+00 -2.05100584e+00 2.27108717e-01 8.65384787e-02
4.72906858e-01 3.80632222e-01 -5.88424206e-01 1.19890511e+00
4.25213784e-01 9.41864699e-02 1.09764837e-01 -4.05948132e-01
1.84496753e-02 2.49418378e-01 -6.39519751e-01 2.56014049e-01
1.30500257e-01 1.42544067e+00 -1.16260946e+00 -6.90343201e-01
4.11592424e-02 7.62494147e-01 -6.22937620e-01 5.08394897e-01
-1.41962945e-01 -1.91469356e-01 -1.01965344e+00 5.77081144e-01
3.06454927e-01 -7.82436132e-01 -1.57593619e-02 -1.17606431e-01
5.08401990e-01 6.14690661e-01 -5.44724464e-01 1.99519610e+00
-4.88481551e-01 4.15608615e-01 -2.57524818e-01 -7.80748069e-01
9.39974368e-01 5.04788399e-01 4.64376152e-01 -1.31212890e+00
-1.75987989e-01 7.78106749e-02 -4.76377606e-01 -4.34809655e-01
1.25072479e+00 8.79805684e-02 -1.28129557e-01 5.96057951e-01
2.41402179e-01 -1.19092450e-01 4.33200121e-01 9.70729470e-01
1.21578443e+00 -1.48715556e-01 9.81778577e-02 1.90700330e-02
5.73997796e-01 -7.36343116e-02 7.40953675e-03 1.04334319e+00
2.03434885e-01 7.54154265e-01 3.40975285e-01 -1.48298636e-01
-8.67279351e-01 -9.64303374e-01 9.25930142e-02 1.13106072e+00
3.78606141e-01 -7.75154948e-01 -8.57541710e-02 -4.61966544e-01
1.48545578e-01 4.90440011e-01 -6.41580522e-01 -6.69535458e-01
-5.37920535e-01 -1.87901959e-01 4.43933725e-01 5.09408116e-01
4.58255380e-01 -1.04621553e+00 9.12724286e-02 3.52396369e-01
-5.60466170e-01 -7.97322392e-01 -6.81180418e-01 -3.07645112e-01
-9.53065693e-01 -9.49338138e-01 -1.11268330e+00 -9.26317871e-01
4.31369364e-01 1.01524436e+00 1.57309473e+00 5.23911953e-01
-2.85425503e-02 6.24919116e-01 -1.00238478e+00 1.30982324e-01
3.12176626e-03 4.36824977e-01 -3.45662683e-01 -1.75697163e-01
5.97191691e-01 -3.88537645e-01 -9.50496078e-01 -3.18430960e-02
-1.41824806e+00 -4.44179833e-01 5.74823141e-01 1.08904397e+00
4.92509514e-01 -2.44032368e-01 7.00979412e-01 -6.70516908e-01
1.11918342e+00 -9.41238105e-01 -2.73600638e-01 6.71233058e-01
-5.53805351e-01 1.21536039e-01 5.42189062e-01 -4.21589538e-02
-8.41022193e-01 -6.92281663e-01 -2.15357855e-01 -6.18479013e-01
4.58901346e-01 1.13106656e+00 4.33757246e-01 3.08080260e-02
5.20039082e-01 6.24831915e-01 -5.36263585e-02 -5.89589655e-01
5.18487692e-01 7.32714117e-01 4.34146225e-02 -4.13859457e-01
5.39984763e-01 1.54399946e-01 -2.24783987e-01 -7.48840630e-01
-9.55977559e-01 -1.20357573e+00 -3.33105415e-01 1.55074373e-01
4.52981770e-01 -1.13848591e+00 -3.63371462e-01 -8.54037851e-02
-1.02202380e+00 4.05448914e-01 -3.42035115e-01 4.46019381e-01
-2.59197623e-01 6.32630110e-01 -9.82138872e-01 -3.88190359e-01
-6.27493203e-01 -8.80894363e-01 1.34181786e+00 2.94422030e-01
6.03710711e-02 -1.25169230e+00 2.96529651e-01 3.31415296e-01
7.92437732e-01 -4.94876325e-01 1.13463390e+00 -7.05148101e-01
-6.22058749e-01 -6.53359711e-01 -5.67237914e-01 1.16032297e-02
2.20158815e-01 -4.19624627e-01 -5.12946367e-01 -6.98659956e-01
-2.25649670e-01 -7.85645187e-01 1.17651296e+00 -1.51275620e-02
9.95818794e-01 -3.45651418e-01 -2.89633662e-01 3.78655285e-01
1.61208618e+00 -2.34460458e-01 8.28008831e-01 1.91066235e-01
6.04586899e-01 2.16081411e-01 2.53946930e-01 3.68356824e-01
6.16087437e-01 5.25585413e-01 3.48299503e-01 3.35700363e-02
-2.18108729e-01 -7.45506704e-01 6.99177980e-02 1.36341655e+00
1.64685935e-01 -6.03620827e-01 -6.72197163e-01 8.17650557e-01
-1.84881091e+00 -1.02169812e+00 2.04519063e-01 1.96039593e+00
6.59409523e-01 -3.27879429e-01 -1.42776668e-01 -1.95004195e-01
2.47748256e-01 6.39103591e-01 -1.09666608e-01 -1.22760378e-01
-1.24359213e-01 2.14117885e-01 3.65817137e-02 3.90261948e-01
-7.38014698e-01 9.01064038e-01 6.51973963e+00 8.23249698e-01
-7.48479664e-01 -2.01826870e-01 9.77907702e-02 -1.38515532e-01
-7.96697795e-01 -1.79325178e-01 -7.25006819e-01 1.10700928e-01
8.75539362e-01 -5.78191638e-01 3.03061157e-01 5.10517180e-01
-2.56989479e-01 3.06639761e-01 -1.13043642e+00 9.26526546e-01
4.34216410e-01 -1.59572673e+00 8.67966712e-01 -1.55614212e-01
6.34253025e-01 1.66758269e-01 -1.20923836e-02 1.09125638e+00
1.81933895e-01 -1.00216460e+00 3.32163125e-02 6.30535722e-01
4.82434392e-01 -5.90420723e-01 9.47386563e-01 3.01341057e-01
-1.34608805e+00 -1.56439818e-03 -9.04881299e-01 3.59402522e-02
9.25999209e-02 1.70092136e-01 -5.45055151e-01 1.18087363e+00
4.34098303e-01 1.10779369e+00 -5.34430206e-01 1.23433745e+00
-5.76998331e-02 5.61940491e-01 -1.01038165e-01 -3.21172684e-01
7.33526826e-01 -3.71937007e-01 3.78811300e-01 1.23930335e+00
2.52211004e-01 1.53356701e-01 2.55050272e-01 8.15328777e-01
-5.42660177e-01 4.16113436e-01 -8.99756968e-01 -3.79238635e-01
5.24327397e-01 9.52619314e-01 -1.09785095e-01 -4.61044252e-01
-8.21019530e-01 8.96222949e-01 7.26256907e-01 8.34596694e-01
-3.22609842e-01 -7.23018408e-01 4.93365079e-01 -1.04028277e-01
5.36688447e-01 -1.90007482e-02 4.20468658e-01 -1.49952257e+00
2.34007254e-01 -7.79424012e-01 6.91886544e-01 -5.35721183e-01
-1.56045306e+00 7.02296972e-01 -1.13977291e-01 -1.45560586e+00
-5.31495392e-01 -2.95855589e-02 -3.91270936e-01 1.02231073e+00
-2.05433416e+00 -1.27745032e+00 1.58223473e-02 9.87329245e-01
4.72652555e-01 -5.32971285e-02 1.11836600e+00 5.13376057e-01
-2.05004275e-01 5.97988427e-01 5.05992293e-01 4.53605801e-01
4.17792857e-01 -9.75994408e-01 2.20996737e-01 3.83046448e-01
6.68397009e-01 9.92832780e-01 1.10602222e-01 -3.93227309e-01
-1.87825274e+00 -9.87887859e-01 1.27551484e+00 -2.27655381e-01
7.49956667e-01 -1.26748607e-02 -1.33480155e+00 6.35928333e-01
3.89549196e-01 4.41808067e-02 8.87440085e-01 5.75473368e-01
-5.68252385e-01 5.06337509e-02 -5.86138189e-01 6.26856983e-01
7.44315088e-01 -1.12556362e+00 -1.19030511e+00 4.73441094e-01
1.13265967e+00 -1.39735386e-01 -1.05264962e+00 2.39595294e-01
3.44308525e-01 -7.00453043e-01 1.19653034e+00 -8.02794576e-01
5.23971617e-01 4.10565697e-02 -2.55282253e-01 -1.26949525e+00
-6.68065727e-01 -3.28710020e-01 -7.48645067e-01 1.02186465e+00
3.10815126e-01 -3.40659440e-01 5.70942879e-01 3.59411031e-01
1.10732531e-02 -9.65535462e-01 -7.88962305e-01 -7.11249590e-01
2.98695862e-01 -1.86826824e-03 7.34929919e-01 8.07536423e-01
2.89780587e-01 7.85251856e-01 -2.51465321e-01 -3.15557607e-02
2.65292495e-01 6.40619695e-01 4.41866308e-01 -1.09935737e+00
-1.67740628e-01 -4.50260133e-01 -3.91313136e-01 -1.79451311e+00
3.18140686e-01 -1.09483552e+00 -1.83856890e-01 -2.02877498e+00
4.41311955e-01 -2.67410815e-01 -7.78096735e-01 2.23090366e-01
-4.36574131e-01 1.64421156e-01 1.15686134e-02 6.51612639e-01
-1.15652657e+00 1.23831606e+00 1.45551085e+00 -5.03743708e-01
-1.51341990e-01 -1.50869861e-01 -7.94549644e-01 -3.75988074e-02
3.35194498e-01 -3.23814273e-01 -7.64711201e-01 -1.02435994e+00
4.03536201e-01 4.87321556e-01 1.86292663e-01 -6.14636123e-01
5.86999059e-01 3.01114798e-01 2.72562951e-01 -6.37882173e-01
4.21606243e-01 -7.03395188e-01 -4.78654712e-01 1.47481337e-01
-6.50540769e-01 1.58061802e-01 -5.95304556e-02 8.99955630e-01
-1.06883860e+00 -2.17989206e-01 1.06334396e-01 -1.72199175e-01
-6.95296943e-01 7.35537231e-01 -1.04585074e-01 1.27724767e-01
3.53389680e-01 1.57188386e-01 -4.28741038e-01 -8.86050940e-01
-2.46619672e-01 6.11950338e-01 2.39630282e-01 7.84094453e-01
1.00149310e+00 -1.55742633e+00 -8.99518728e-01 1.86417833e-01
5.82388163e-01 4.37926594e-03 3.27705383e-01 4.45341647e-01
-3.67384523e-01 1.13125288e+00 1.10705823e-01 -4.04850215e-01
-8.63312185e-01 7.29345798e-01 -2.72697899e-02 -8.01154077e-01
-7.89741993e-01 8.55371892e-01 6.06760159e-02 -3.46607566e-01
3.03291976e-01 4.39609475e-02 -7.11008012e-01 -3.11308308e-03
7.97905684e-01 5.78186363e-02 5.35886437e-02 -5.73883712e-01
-4.34226617e-02 4.02021378e-01 -6.86272681e-01 2.17900604e-01
1.24749291e+00 -3.22143078e-01 -3.44549417e-01 1.97736904e-01
1.82424355e+00 -1.98020220e-01 -5.24559617e-01 -7.84178972e-01
-8.90008062e-02 -6.18359268e-01 2.45298564e-01 -3.62394899e-01
-9.90870357e-01 8.57740164e-01 -4.97971289e-03 4.65010524e-01
1.08864534e+00 1.79952353e-01 1.17604125e+00 9.79760528e-01
4.51601952e-01 -6.01113796e-01 2.00662971e-01 9.04998899e-01
1.00329363e+00 -1.22046804e+00 5.72120920e-02 -9.45637003e-02
-2.94279993e-01 1.00343049e+00 3.17373902e-01 -3.07965189e-01
6.52597249e-01 -5.85286200e-01 -1.41024590e-01 -5.53450108e-01
-9.40071881e-01 -3.38001102e-01 7.98360288e-01 2.80675739e-01
7.20734477e-01 -1.88062489e-01 -3.65034252e-01 2.02969819e-01
3.66345346e-02 8.04061536e-03 1.10597074e-01 1.15995169e+00
-3.61295462e-01 -1.11970210e+00 1.45694450e-01 7.74952650e-01
-4.48038042e-01 -4.90637749e-01 -3.48496556e-01 6.73343599e-01
-9.39472079e-01 1.06836545e+00 1.30013615e-01 -2.76952773e-01
2.64822632e-01 -5.72628826e-02 2.28716001e-01 -7.83623695e-01
-6.31186724e-01 -1.53871819e-01 -1.93524398e-02 -6.77092195e-01
-2.70276397e-01 -1.34241700e-01 -8.67718697e-01 -3.09047192e-01
-5.26961088e-01 8.81880999e-01 2.25733563e-01 8.04968059e-01
6.50591552e-01 5.79181612e-01 7.96190917e-01 -3.25709134e-01
-8.61348152e-01 -1.09958434e+00 -8.43261361e-01 6.50532424e-01
4.94545817e-01 -4.21314597e-01 -3.00963789e-01 -4.94480073e-01] | [11.32787799835205, 7.807013511657715] |
0df8496b-18c4-4116-afd3-db37408a29d8 | generalization-bounds-and-algorithms-for-1 | 2205.14692 | null | https://arxiv.org/abs/2205.14692v1 | https://arxiv.org/pdf/2205.14692v1.pdf | Generalization bounds and algorithms for estimating conditional average treatment effect of dosage | We investigate the task of estimating the conditional average causal effect of treatment-dosage pairs from a combination of observational data and assumptions on the causal relationships in the underlying system. This has been a longstanding challenge for fields of study such as epidemiology or economics that require a treatment-dosage pair to make decisions but may not be able to run randomized trials to precisely quantify their effect and heterogeneity across individuals. In this paper, we extend (Shalit et al, 2017) to give new bounds on the counterfactual generalization error in the context of a continuous dosage parameter which relies on a different approach to defining counterfactuals and assignment bias adjustment. This result then guides the definition of new learning objectives that can be used to train representation learning algorithms for which we show empirically new state-of-the-art performance results across several benchmark datasets for this problem, including in comparison to doubly-robust estimation methods. | ['Giulia Prando', 'Anish Dhir', 'Alexis Bellot'] | 2022-05-29 | null | null | null | null | ['epidemiology'] | ['medical'] | [ 6.52213573e-01 1.50966629e-01 -1.04845524e+00 -3.84196252e-01
-1.01796699e+00 -4.84112680e-01 8.24498832e-01 4.12179768e-01
-4.76411760e-01 1.26041186e+00 5.96681356e-01 -6.93859100e-01
-7.04172373e-01 -5.32936335e-01 -9.38047349e-01 -6.68223739e-01
-4.80008066e-01 6.70226276e-01 -5.73356688e-01 2.11066023e-01
1.83974102e-01 3.27033371e-01 -1.11723495e+00 6.37607574e-02
7.53670692e-01 5.45622408e-01 -3.55177343e-01 3.77493501e-01
2.60430247e-01 3.95858347e-01 -1.85347199e-01 -2.22908363e-01
2.21070185e-01 -4.96753722e-01 -5.68554044e-01 -2.19485164e-01
4.87766087e-01 -2.47857675e-01 -1.28193498e-01 8.29426706e-01
7.49716461e-01 3.59582268e-02 1.31294751e+00 -1.14312971e+00
-5.37085533e-01 1.10600507e+00 -8.44644964e-01 2.00918272e-01
1.64571881e-01 2.30895758e-01 9.83709514e-01 -2.67338544e-01
4.22420561e-01 1.42918754e+00 6.38397753e-01 4.30757731e-01
-1.67677534e+00 -7.99225211e-01 2.58263677e-01 1.25047648e-02
-1.00893497e+00 -4.95369434e-01 4.28370744e-01 -6.89795554e-01
5.16547441e-01 2.44938418e-01 2.35958114e-01 1.50366509e+00
4.36004132e-01 3.42131019e-01 1.42978549e+00 -4.48158175e-01
4.90752190e-01 -2.30275273e-01 1.08959638e-01 2.29630932e-01
6.26184225e-01 6.28266335e-01 -1.76332369e-01 -6.44513905e-01
7.23436892e-01 3.39410231e-02 -3.90723079e-01 -6.12743139e-01
-1.23533046e+00 1.42178905e+00 4.44102615e-01 4.36803773e-02
-7.70391822e-01 4.61092532e-01 5.51127076e-01 9.40305144e-02
6.27674103e-01 6.30914152e-01 -6.08177066e-01 1.91314399e-01
-8.68299305e-01 6.43606961e-01 5.39451897e-01 3.13696086e-01
9.10482481e-02 -1.65056571e-01 -6.03197694e-01 6.67420506e-01
3.00377980e-03 6.06382787e-01 2.93616533e-01 -8.75534117e-01
5.17444611e-01 3.10755014e-01 4.49677736e-01 -3.36398900e-01
-6.88305616e-01 -3.55384856e-01 -1.10289764e+00 -2.39043217e-02
6.51096821e-01 -4.52313006e-01 -7.62305140e-01 2.11391449e+00
4.64018941e-01 6.07890368e-01 -1.74166128e-01 6.38800979e-01
4.07707959e-01 3.60636324e-01 4.58655328e-01 -7.98834860e-01
1.28325021e+00 -2.05980927e-01 -7.26903498e-01 -7.56683350e-02
7.59578109e-01 -4.31427330e-01 7.94708908e-01 9.35554802e-02
-1.10288095e+00 4.42780145e-02 -8.77793193e-01 2.18237847e-01
-2.63879746e-01 -9.27386060e-03 1.09569466e+00 7.64769852e-01
-5.59031546e-01 7.95510709e-01 -5.57939649e-01 -1.64239347e-01
7.96344340e-01 5.57990909e-01 -1.17793657e-01 -7.05578551e-02
-1.29190767e+00 9.63386834e-01 1.71535671e-01 -1.28594935e-01
-1.01484501e+00 -1.73941660e+00 -7.40042984e-01 2.27862760e-01
5.29962540e-01 -1.18190753e+00 1.34476364e+00 -7.67877579e-01
-1.24431157e+00 6.58882499e-01 9.87216234e-02 -7.82826543e-01
7.43590951e-01 1.74535200e-01 -9.68139619e-02 -3.49999487e-01
2.85416186e-01 3.52433085e-01 6.78910553e-01 -7.74599433e-01
-5.46299636e-01 -6.76804423e-01 -5.20322844e-03 3.30880247e-02
1.59636214e-01 2.79233884e-02 3.56543183e-01 -7.29791701e-01
-4.52041507e-01 -8.12467456e-01 -4.68178689e-01 -3.30679595e-01
-4.52957481e-01 -3.90629053e-01 1.29988998e-01 -5.11895657e-01
1.24030471e+00 -1.74257803e+00 2.09446192e-01 -3.05863209e-02
8.93063173e-02 -1.74299970e-01 -8.64984095e-02 3.91971648e-01
-6.85817719e-01 2.29477435e-01 -6.14660859e-01 -4.46938835e-02
1.25849396e-01 -1.85465649e-01 -5.02759635e-01 9.42344129e-01
1.25464574e-01 1.00511789e+00 -8.32479060e-01 -1.29871219e-01
1.30807370e-01 2.49699041e-01 -5.71871519e-01 1.20694600e-01
-1.73700318e-01 3.07562590e-01 -3.62642407e-01 1.09412491e-01
7.02591598e-01 -2.23607317e-01 5.70390344e-01 4.93793860e-02
7.41103664e-02 4.58264977e-01 -1.27449048e+00 1.22386301e+00
-8.07517707e-01 1.84341177e-01 -2.45059207e-01 -1.48874032e+00
1.76717788e-01 5.17244756e-01 6.17814362e-01 -3.62537652e-01
2.70454437e-01 9.68397483e-02 2.75585651e-01 -4.73147690e-01
-2.30802760e-01 -8.34451258e-01 -3.55629444e-01 7.07675636e-01
-2.37239793e-01 -7.86052365e-03 -3.39829475e-02 -1.92100435e-01
8.63229513e-01 -1.57723755e-01 7.91562557e-01 -6.59255922e-01
2.68974334e-01 -3.46163809e-01 3.93526793e-01 1.04384637e+00
1.91712663e-01 3.43375564e-01 8.92738104e-01 -4.00049508e-01
-1.02684546e+00 -1.10791576e+00 -7.29755998e-01 8.02301705e-01
-6.08904779e-01 3.45397830e-01 -4.07324344e-01 -8.59669566e-01
5.36266029e-01 1.03072596e+00 -1.26621222e+00 -3.57804656e-01
-2.59137869e-01 -1.50709605e+00 4.13054287e-01 6.51122689e-01
2.96599027e-02 -5.87700963e-01 -4.32885021e-01 1.92168519e-01
5.54736182e-02 -4.90596116e-01 -5.15702248e-01 1.13200560e-01
-9.36612606e-01 -1.29176390e+00 -9.27682757e-01 -6.72188774e-02
2.63266414e-01 -2.30367720e-01 1.14651167e+00 -4.04742867e-01
-8.33528042e-02 2.00691745e-01 2.45190114e-01 -8.19181859e-01
-3.47841442e-01 -1.39057666e-01 2.78527737e-01 -1.04201965e-01
2.19080850e-01 -5.38532972e-01 -7.34713316e-01 3.23567949e-02
-8.97914588e-01 -3.33989292e-01 4.61135864e-01 1.04507458e+00
3.21924627e-01 -2.00382382e-01 1.14394319e+00 -1.35886860e+00
6.33888721e-01 -9.38748121e-01 -9.72978711e-01 2.97248662e-01
-7.27551877e-01 4.03166741e-01 4.41404045e-01 -6.74494684e-01
-1.13707232e+00 -1.97604835e-01 4.00504380e-01 2.07611024e-02
1.99505836e-01 7.09044218e-01 -1.22204788e-01 5.09353161e-01
7.25790381e-01 -3.33507538e-01 -1.15119427e-01 -4.05761510e-01
4.98812079e-01 7.28604496e-01 2.61160970e-01 -6.21479690e-01
1.86423331e-01 6.01623297e-01 5.72098255e-01 -3.58867317e-01
-1.00549984e+00 -1.98144019e-01 -2.05889672e-01 4.30789411e-01
6.92879140e-01 -1.01257074e+00 -1.02646959e+00 6.56707399e-03
-8.47197652e-01 -5.61272144e-01 -2.96843797e-01 8.78013849e-01
-9.34442699e-01 -4.07192707e-02 -9.43362489e-02 -9.32744563e-01
-2.11051833e-02 -9.92791533e-01 1.04352713e+00 -2.15752855e-01
-2.69111961e-01 -1.29583943e+00 5.02085805e-01 1.07472099e-01
2.80482054e-01 4.71325368e-01 1.28101921e+00 -3.57024908e-01
-1.05272785e-01 -5.13877757e-02 -1.93423703e-01 -9.41807851e-02
4.28953946e-01 -2.49201700e-01 -8.89121532e-01 -3.91636133e-01
-1.03004957e-02 -1.77761137e-01 1.14504910e+00 1.44924021e+00
1.46050489e+00 -6.81199431e-01 -5.92850506e-01 3.29472214e-01
1.41518760e+00 6.49580806e-02 4.52905476e-01 -1.54548720e-01
3.16115230e-01 7.03290939e-01 2.99782544e-01 5.58309972e-01
2.60164648e-01 1.03853810e+00 1.21673830e-01 1.38644148e-02
1.66471168e-01 -2.66517907e-01 5.71474843e-02 -4.70609456e-01
5.78779504e-02 -2.19116762e-01 -5.29949903e-01 7.56172657e-01
-1.88842607e+00 -1.05656171e+00 -2.46025041e-01 2.78742003e+00
1.18936586e+00 -1.46559134e-01 5.98766506e-01 -1.91850781e-01
7.22765505e-01 -5.94671257e-02 -7.08412468e-01 -5.67046881e-01
5.78318499e-02 3.29437226e-01 9.75585103e-01 6.10042155e-01
-1.17699802e+00 1.76984161e-01 7.22701788e+00 6.33812070e-01
-1.00359321e+00 2.16610059e-01 9.84877288e-01 -2.13513955e-01
-3.64692301e-01 -4.02396098e-02 -5.68629742e-01 6.81637466e-01
1.43169403e+00 -4.46053475e-01 3.69039387e-01 2.19313517e-01
6.80156171e-01 -1.36310816e-01 -1.61990523e+00 6.14998341e-01
-3.41688097e-01 -1.45948529e+00 -1.71197429e-01 2.75382191e-01
1.15686715e+00 -8.95737633e-02 2.44582027e-01 1.51620939e-01
7.68626451e-01 -1.47388351e+00 2.33669937e-01 5.25611043e-01
9.28614736e-01 -5.86505055e-01 8.53190720e-01 1.55099407e-01
-4.18277621e-01 -3.81403536e-01 -2.99277514e-01 -2.96911538e-01
6.36055544e-02 8.99302542e-01 -9.18770909e-01 6.61916077e-01
3.50853473e-01 6.41671181e-01 -7.68411085e-02 1.15122461e+00
-1.67197213e-01 9.49478090e-01 -3.82066250e-01 1.22806109e-01
6.19749874e-02 -2.52501410e-03 7.51496851e-02 9.48453605e-01
3.40517163e-01 -1.82131380e-02 -3.67044449e-01 8.64873469e-01
-4.32254612e-01 2.87438594e-02 -8.31640244e-01 -2.37448104e-02
3.55245143e-01 6.64366663e-01 -3.33929360e-01 -3.15858424e-01
-2.79753298e-01 1.75799951e-01 1.78771421e-01 3.24249506e-01
-8.14788342e-01 3.22054327e-01 6.26027465e-01 2.19845265e-01
2.14551404e-01 4.55185860e-01 -4.21666950e-01 -9.99817252e-01
-2.32546687e-01 -9.08601224e-01 9.84947562e-01 -2.87674189e-01
-1.59987009e+00 -4.94135827e-01 7.10911036e-01 -8.31109583e-01
-5.82160652e-01 -5.28870523e-01 -6.52867734e-01 1.01697600e+00
-1.38252461e+00 -8.55011821e-01 5.20777047e-01 1.51822880e-01
3.40128630e-01 2.13308752e-01 7.24215925e-01 1.21500783e-01
-6.61333621e-01 3.88759285e-01 4.81365234e-01 -3.80122781e-01
7.07708359e-01 -1.43154347e+00 -1.07918777e-01 3.80396128e-01
-1.26825884e-01 4.73572075e-01 8.13174486e-01 -6.46759808e-01
-1.24499285e+00 -9.48190093e-01 7.99264133e-01 -6.79388106e-01
6.98467255e-01 -2.91190684e-01 -5.64361811e-01 9.53664184e-01
1.24313816e-01 -1.99921325e-01 6.58968925e-01 6.68256521e-01
-4.00851369e-01 -1.27605453e-01 -1.14451003e+00 5.31725407e-01
8.33151698e-01 -1.70521334e-01 -6.48384571e-01 6.17113888e-01
4.14372563e-01 -1.20813921e-01 -7.74590313e-01 5.03252566e-01
6.79379046e-01 -5.55612803e-01 1.15160537e+00 -1.34881306e+00
6.62557721e-01 5.29855452e-02 5.14818430e-02 -1.54380941e+00
-2.11867049e-01 -4.58435297e-01 1.77413836e-01 1.00140524e+00
5.30378222e-01 -8.08177650e-01 3.72038364e-01 5.17879188e-01
5.16955376e-01 -7.27209210e-01 -1.19991040e+00 -7.25313962e-01
7.50467896e-01 -2.84503043e-01 8.62840950e-01 1.09598279e+00
3.55934612e-02 4.12440419e-01 -5.10806561e-01 2.75972277e-01
8.28304887e-01 3.85451734e-01 4.52958852e-01 -1.19872451e+00
-4.13747370e-01 -7.09954560e-01 -1.33424804e-01 -3.89737785e-01
4.54930007e-01 -8.33978832e-01 -2.18278244e-01 -1.31788933e+00
6.47242427e-01 -3.64245385e-01 -3.71285319e-01 2.32773542e-01
-5.08446813e-01 -1.39970407e-01 -2.30137363e-01 -4.03942138e-01
1.52701372e-02 5.72734773e-01 7.94235885e-01 -4.38855410e-01
-1.85086668e-01 4.16958034e-01 -9.87567127e-01 4.06281859e-01
6.27542377e-01 -8.23989451e-01 -6.04176700e-01 5.50421402e-02
2.81339407e-01 3.25605094e-01 6.28759921e-01 -2.64564842e-01
-3.44308347e-01 -6.44834757e-01 3.89250278e-01 5.24797896e-03
-1.08690687e-01 -6.15295589e-01 3.03948790e-01 7.74908602e-01
-8.42041850e-01 -3.31915058e-02 2.88470954e-01 6.88494503e-01
3.84873927e-01 -2.45123893e-01 6.24560356e-01 1.13913417e-02
5.37849255e-02 2.12703869e-01 -1.60320222e-01 3.00219178e-01
8.50255311e-01 5.25888324e-01 -4.17387873e-01 -5.13889015e-01
-5.50666451e-01 4.21376944e-01 5.65595888e-02 2.49224335e-01
1.78535089e-01 -1.36915886e+00 -1.27624869e+00 -3.01987916e-01
5.98809533e-02 -5.67977667e-01 5.48236132e-01 1.13687050e+00
1.37018129e-01 6.29333317e-01 5.45299007e-03 -2.60135889e-01
-8.73973191e-01 9.91688073e-01 4.16063428e-01 -5.92552543e-01
-2.64652997e-01 3.14793617e-01 6.89004123e-01 -3.94800693e-01
-9.08178687e-02 -3.64819676e-01 -4.91316579e-02 2.07073152e-01
6.64028585e-01 5.09285688e-01 -7.66640902e-02 -9.36332420e-02
-4.25210446e-01 2.30974823e-01 9.90639478e-02 -1.21950962e-01
1.58110380e+00 1.52350785e-05 8.63749385e-02 5.45214593e-01
1.19991076e+00 -2.57677644e-01 -1.25761712e+00 -6.45107329e-02
-2.93101035e-02 -4.81287658e-01 3.14207166e-01 -1.15028763e+00
-8.08870256e-01 6.95423365e-01 8.07059586e-01 1.94208071e-01
9.19967890e-01 1.78197339e-01 -1.78177506e-01 8.53233691e-03
-1.10705663e-02 -8.28388512e-01 -5.41390598e-01 -2.91316628e-01
1.07973146e+00 -1.31239665e+00 4.95027751e-01 -1.66466922e-01
-3.73643816e-01 5.65442741e-01 -1.18496418e-01 -1.67407945e-01
7.13380694e-01 2.20457502e-02 -3.57491791e-01 -8.02334398e-02
-8.89911771e-01 -8.26348364e-02 5.42112291e-01 6.66264594e-01
6.42567158e-01 4.88730073e-01 -9.33170378e-01 5.49428403e-01
-3.07624824e-02 2.99354643e-01 7.31371105e-01 3.38105887e-01
3.40487242e-01 -1.06953299e+00 -2.89202660e-01 1.01982760e+00
-6.70182228e-01 -9.39753279e-02 -1.23169474e-01 1.04288542e+00
-8.92161056e-02 7.58715034e-01 1.21376246e-01 5.11183202e-01
4.35236424e-01 1.20674193e-01 6.60890281e-01 -5.51863849e-01
-2.13697404e-01 -6.96724653e-02 1.11920670e-01 -4.44224745e-01
-7.56009698e-01 -9.89318371e-01 -7.12306678e-01 -2.78205335e-01
-1.94121465e-01 1.40385240e-01 4.33226109e-01 1.07380307e+00
2.49886408e-01 5.70100307e-01 7.27950692e-01 -5.22245288e-01
-1.19460666e+00 -9.79846239e-01 -6.83986545e-01 3.09385598e-01
5.69358885e-01 -9.49550033e-01 -4.51737046e-01 -3.67652506e-01] | [8.06107234954834, 5.374167442321777] |
57e3a7b1-5f73-4cd5-99f7-1f3351933e03 | task-independent-capsule-based-agents-for | null | null | https://www.researchgate.net/publication/357764898_Task_Independent_Capsule-Based_Agents_for_Deep_Q-Learning | https://link.springer.com/chapter/10.1007/978-3-030-93842-0_4 | Task Independent Capsule-Based Agents for Deep Q-Learning | In recent years, Capsule Networks (CapsNets) have achieved promising results in tasks such as object recognition thanks to their invariance characteristics towards pose and lighting. They have been proposed as an alternative to relational insensitive and translation invariant Convolutional Neural Networks (CNN). It has been empirically proven that CapsNets are capable of achieving competitive performance while requiring significantly fewer parameters. This is a desirable characteristic for Deep reinforcement learning which is known to be sample-inefficient during training. In this paper, we propose DCapsQN, a task-independent CapsNets-based architecture in the deep reinforcement learning setting. We experiment in the model-free reinforcement learning setting, more specifically in Deep Q-Learning using the Atari suite as the testbed of our analysis. To the best of our knowledge, this work constitutes the first CapsNets-based deep reinforcement learning architecture to learn state-action value functions without the need for task-specific adaptation. Our results show that, in this setting, DCapsQN requires 92% fewer parameters than the baseline. Moreover, despite their smaller size, the DCapsQN provides significant boosts in performance (score), ranging between 10%–77% while further stabilizing the Deep Q-Learning. This is supported by our empirical results which shows that DCapsQN agents outperform the benchmark Double-DQN agent, with Prioritized experience replay, in eight out of the nine selected environments. | ['Steven Latre ́', 'Jose ́ Oramas', 'Peter Hellinckx', 'Kevin Mets', 'Tom De Schepper', 'Akash Singh'] | 2022-01-11 | null | null | null | benelux-conference-on-artificial-intelligence | ['object-recognition'] | ['computer-vision'] | [-2.45335922e-01 -1.40978366e-01 -1.45155266e-01 -8.70978087e-02
-5.16229033e-01 -6.30557954e-01 7.40580261e-01 -2.84356147e-01
-7.59344339e-01 8.00093055e-01 -3.82424481e-02 -1.28882334e-01
-3.58689815e-01 -6.24177873e-01 -1.04803658e+00 -7.97015011e-01
-4.98152107e-01 5.25036156e-01 2.13965371e-01 -5.87363362e-01
-4.53469940e-02 6.69659078e-01 -1.57864380e+00 -2.21247047e-01
5.05625308e-01 1.10101402e+00 -4.75695878e-02 6.32795393e-01
4.90462869e-01 8.15359414e-01 -6.96884096e-01 -3.51615071e-01
6.84847116e-01 -1.00789048e-01 -5.33519506e-01 -1.68831095e-01
6.46251500e-01 -5.54689050e-01 -4.64562953e-01 6.89784229e-01
7.48600006e-01 2.84597933e-01 5.25081679e-02 -1.31222034e+00
-5.05008876e-01 5.23924828e-01 -8.44264179e-02 1.69782639e-01
-5.56106232e-02 5.86607814e-01 1.04151130e+00 -5.46984136e-01
6.02287114e-01 1.26428378e+00 6.26310825e-01 6.78186834e-01
-1.14067578e+00 -6.65614665e-01 1.37377247e-01 9.62369815e-02
-9.35594678e-01 -2.67682165e-01 6.57825053e-01 1.11314051e-01
1.31583560e+00 -8.10191855e-02 7.47514546e-01 1.54228246e+00
4.52212125e-01 8.75453293e-01 1.17414486e+00 -1.78416431e-01
6.12055898e-01 6.12453697e-03 -4.25115496e-01 4.48777616e-01
1.66534752e-01 5.76137781e-01 -5.70595622e-01 1.02818161e-01
8.79761040e-01 -1.51502237e-01 4.01613638e-02 -9.32402790e-01
-1.12265635e+00 8.95632625e-01 8.93229663e-01 3.13884646e-01
-4.27816391e-01 9.72683609e-01 7.20996261e-01 7.18077838e-01
1.13660008e-01 8.08482111e-01 -8.59726846e-01 -4.75777358e-01
-4.41066116e-01 6.16622806e-01 8.76778603e-01 9.22449410e-01
3.75521690e-01 5.97332597e-01 -1.50410905e-01 5.46093464e-01
2.54257545e-02 7.16684163e-01 5.21708190e-01 -1.08667886e+00
2.83752382e-01 3.60144258e-01 5.59930503e-02 -7.08127975e-01
-6.95951581e-01 -1.07015955e+00 -5.91631591e-01 5.35248756e-01
2.89914072e-01 -2.82529980e-01 -9.27947998e-01 2.03005505e+00
1.33400887e-01 2.32967064e-02 4.19934660e-01 1.01884818e+00
4.20420527e-01 2.81153172e-01 1.20714806e-01 1.38222113e-01
1.18053317e+00 -1.00960636e+00 -3.40841174e-01 -1.03843529e-02
4.95791495e-01 -4.59506810e-01 1.15214336e+00 6.54663861e-01
-8.88244927e-01 -5.54074585e-01 -1.21857107e+00 4.06589329e-01
-3.87413681e-01 -1.48660079e-01 8.92089128e-01 7.44800746e-01
-1.31408644e+00 7.80553162e-01 -1.00019765e+00 -5.83336174e-01
5.12049198e-01 7.71233737e-01 -2.15597630e-01 1.01400040e-01
-1.18083072e+00 9.34791625e-01 4.15483266e-01 -7.46977404e-02
-1.43281519e+00 -7.57640719e-01 -6.09429359e-01 1.26997143e-01
7.18615115e-01 -4.97008890e-01 1.58926392e+00 -1.18463266e+00
-1.91657948e+00 3.00639808e-01 5.84250808e-01 -1.11603212e+00
7.37024963e-01 -4.02036577e-01 -2.20373973e-01 2.07791641e-01
-2.24628925e-01 1.03800189e+00 8.90283644e-01 -1.01933384e+00
-4.25457120e-01 -1.75996810e-01 6.08036637e-01 2.26334974e-01
-3.68229657e-01 -1.62872806e-01 -2.21177429e-01 -5.58975935e-01
-4.15700078e-01 -1.27292418e+00 -3.01645368e-01 3.32956575e-02
4.65376899e-02 -4.40491468e-01 8.60858858e-01 -3.79143581e-02
3.14952433e-01 -2.05026007e+00 4.45237271e-02 5.21535240e-02
8.03772062e-02 4.73298699e-01 -4.74608243e-01 3.97235811e-01
-2.61958446e-02 -2.33536094e-01 -1.56237110e-01 -3.08780998e-01
3.73645246e-01 5.46155810e-01 -2.79820889e-01 5.47154725e-01
3.43261719e-01 1.22456050e+00 -9.29906428e-01 5.32063609e-03
3.64226907e-01 4.61281896e-01 -6.50048018e-01 5.98364100e-02
-5.29868543e-01 2.76722640e-01 -2.91891605e-01 5.11229277e-01
5.31904995e-01 1.24209255e-01 1.66016743e-01 4.49161343e-02
-1.74974483e-02 4.13383208e-02 -9.75318968e-01 1.87613344e+00
-5.46572626e-01 6.43553257e-01 -1.36490092e-01 -1.17728651e+00
9.30453777e-01 4.21988487e-01 6.40138865e-01 -1.48934877e+00
9.53058153e-02 2.58959293e-01 2.44542927e-01 -1.15607686e-01
4.56764519e-01 -1.35958642e-01 -2.27807611e-01 1.76278040e-01
3.89500678e-01 -1.50539681e-01 3.97700161e-01 -6.66582864e-03
1.29467404e+00 5.23308873e-01 -6.51841909e-02 -4.14916426e-01
3.56534243e-01 -3.01855467e-02 4.96477008e-01 8.77009273e-01
-4.37830210e-01 3.91644776e-01 5.30063391e-01 -7.82340050e-01
-1.12548161e+00 -1.12237215e+00 -1.39982447e-01 1.15124643e+00
1.03532828e-01 -1.77690044e-01 -6.82692230e-01 -6.54008090e-01
-2.94941268e-03 4.31895107e-01 -6.64116442e-01 -3.79156917e-01
-8.77668858e-01 -5.28845191e-01 7.37643540e-01 8.72640610e-01
8.87777805e-01 -1.46923959e+00 -9.78924632e-01 4.04110879e-01
4.98193592e-01 -1.15907264e+00 9.26345773e-03 6.59825981e-01
-8.54762197e-01 -1.01346338e+00 -7.87417054e-01 -5.28161645e-01
2.51883537e-01 2.34222761e-03 1.25158215e+00 8.29318315e-02
4.51456904e-02 6.08007312e-01 -4.47291404e-01 -3.63200188e-01
-2.65740186e-01 4.99213606e-01 2.19008967e-01 -4.31488752e-01
1.32868245e-01 -6.54568791e-01 -7.95298338e-01 4.46711808e-01
-1.07174873e+00 -4.97334987e-01 7.47407615e-01 9.18205261e-01
3.86094719e-01 -1.17525354e-01 6.62242293e-01 -4.62776214e-01
6.12309039e-01 -1.59127533e-01 -9.25639391e-01 -2.31978092e-02
-5.86547792e-01 2.80535281e-01 1.00027537e+00 -4.19591337e-01
-7.61783779e-01 9.13330168e-02 -7.90107548e-02 -4.30260658e-01
-1.91868469e-01 2.89686590e-01 1.44329295e-01 -3.79167318e-01
8.67177188e-01 3.84087041e-02 1.13015942e-01 -1.51159838e-01
1.72802091e-01 7.18578771e-02 4.36265230e-01 -7.12018669e-01
9.63135660e-01 4.55856472e-01 2.97617346e-01 -5.08271813e-01
-5.78668535e-01 -2.22742856e-01 -3.66604179e-01 -1.12041585e-01
6.96726263e-01 -9.63787138e-01 -1.12829673e+00 3.72921497e-01
-7.20614851e-01 -7.80008852e-01 -5.18617213e-01 4.67113554e-01
-9.20752585e-01 1.21089153e-03 -5.39612532e-01 -5.02203465e-01
-4.04825717e-01 -1.23909533e+00 8.06889892e-01 3.02960068e-01
1.36825621e-01 -9.73956883e-01 3.32093030e-01 -5.44259138e-02
8.98347676e-01 3.10785651e-01 3.40561092e-01 -6.28742933e-01
-5.88017762e-01 3.91631462e-02 -5.21156266e-02 5.22815645e-01
-1.13770939e-01 -1.86210185e-01 -1.04424131e+00 -8.57665241e-01
-2.51801968e-01 -7.44953632e-01 8.77498806e-01 2.30418146e-01
9.07754719e-01 -1.38028404e-02 2.70556867e-01 7.01335311e-01
1.51745689e+00 1.56703353e-01 6.47138357e-01 8.85718763e-01
4.08094943e-01 1.15880243e-01 5.42401314e-01 3.66928279e-01
1.71679839e-01 8.91782224e-01 1.27502382e+00 -1.77726641e-01
-1.46006227e-01 -3.07168141e-02 7.91929007e-01 3.04691255e-01
-8.90049264e-02 -1.75576910e-01 -6.65105820e-01 5.10189772e-01
-1.83257639e+00 -7.48795927e-01 3.16050470e-01 1.88707495e+00
5.36206245e-01 3.82145762e-01 3.94609064e-01 -1.03112936e-01
1.00249164e-01 1.02327600e-01 -9.09851313e-01 -7.15941846e-01
-1.56101182e-01 6.79188728e-01 6.98588967e-01 6.09842353e-02
-1.32742441e+00 1.06153512e+00 5.88159561e+00 5.94872653e-01
-1.36532879e+00 3.37154674e-03 2.43519202e-01 -2.23145589e-01
1.65461853e-01 -2.70308554e-01 -4.14657652e-01 1.60421669e-01
1.20032191e+00 1.97763383e-01 6.71569824e-01 1.14147949e+00
-3.17798071e-02 6.06079251e-02 -1.01692367e+00 6.74183190e-01
-9.59249511e-02 -1.14081883e+00 -2.29143783e-01 1.35176480e-01
8.06559801e-01 6.07117474e-01 4.32475626e-01 9.01084542e-01
6.50751889e-01 -1.16614437e+00 8.62890482e-01 1.55341670e-01
3.78343523e-01 -1.05522513e+00 9.21575367e-01 3.95790301e-02
-9.15230751e-01 -3.63733619e-01 -5.71404815e-01 -1.27820581e-01
-3.07497799e-01 8.01722482e-02 -7.17922151e-01 6.10873938e-01
9.12330687e-01 7.43893027e-01 -5.69154024e-01 1.09831727e+00
-2.09166914e-01 6.48431599e-01 -3.55022490e-01 -7.64787644e-02
7.38401592e-01 1.87883303e-01 5.20884991e-01 9.53384340e-01
1.20266028e-01 -4.78774816e-01 1.26428142e-01 7.58299232e-01
-3.35082471e-01 -2.03182712e-01 -5.11289358e-01 1.17822833e-01
1.37446940e-01 1.33966279e+00 -6.54690564e-01 -1.39787318e-02
-1.96176678e-01 6.35184765e-01 3.81790847e-01 2.31131449e-01
-1.01156580e+00 -1.37836337e-01 8.37724805e-01 -2.56829411e-01
8.37146163e-01 -4.07928735e-01 1.93288341e-01 -9.38545942e-01
6.05392307e-02 -1.10229170e+00 1.97911635e-01 -5.59629500e-01
-1.05392528e+00 7.29897141e-01 -4.98454496e-02 -1.33720660e+00
-3.92339170e-01 -1.07632577e+00 -3.04327548e-01 2.87331164e-01
-1.86882257e+00 -1.01186585e+00 -1.85899913e-01 7.00798273e-01
4.11713034e-01 -3.34566206e-01 9.18677092e-01 1.51885152e-01
-5.39161682e-01 7.45058119e-01 2.52332628e-01 9.79560614e-02
6.64175868e-01 -1.54542315e+00 3.72110456e-01 6.42377615e-01
1.93634436e-01 5.51457465e-01 8.60725939e-01 -1.92506701e-01
-1.74252141e+00 -1.09422553e+00 -2.41068266e-02 -2.87638664e-01
6.67241514e-01 -4.46674198e-01 -5.54289579e-01 4.91542459e-01
5.26629269e-01 3.13432902e-01 1.97809860e-01 7.14284331e-02
-5.04054904e-01 -5.86732507e-01 -1.08700156e+00 6.41650200e-01
9.40901875e-01 -2.26638570e-01 -2.40960464e-01 3.66635472e-01
8.85927618e-01 -6.57475471e-01 -1.00701511e+00 3.38116676e-01
4.58663166e-01 -1.12887704e+00 9.55742121e-01 -4.67994452e-01
3.88786197e-01 -1.48313046e-01 -2.49528691e-01 -1.55651605e+00
-1.03146687e-01 -7.07319140e-01 -5.68541773e-02 7.08424151e-01
1.52561888e-01 -7.15036213e-01 9.52200115e-01 1.83997959e-01
-2.88649261e-01 -7.64455974e-01 -1.17467761e+00 -1.25935745e+00
3.58262420e-01 -3.84708434e-01 6.48859203e-01 6.27875447e-01
-5.04415154e-01 1.30350294e-03 -3.77511531e-01 -1.59521535e-01
4.62331802e-01 -1.55115917e-01 9.48162556e-01 -1.02730107e+00
-4.29547250e-01 -4.29677397e-01 -5.58478713e-01 -7.97542334e-01
1.62375838e-01 -6.17764354e-01 1.74974501e-02 -1.12969148e+00
-4.28841971e-02 -5.58341920e-01 -5.96165001e-01 6.31362319e-01
3.76088709e-01 4.01489317e-01 5.11294007e-01 -8.96346569e-02
-9.70003307e-01 7.70088255e-01 1.31978834e+00 -1.13245189e-01
-7.48884752e-02 -4.71907705e-02 -5.06500661e-01 3.96899313e-01
1.26428032e+00 -4.39252645e-01 -4.91559565e-01 -4.20625895e-01
3.43140423e-01 -3.23292285e-01 5.85395932e-01 -1.35932219e+00
-2.84770075e-02 -4.10711020e-02 2.97717214e-01 -1.96145490e-01
4.01122898e-01 -1.04369962e+00 -1.32617801e-01 8.07113469e-01
-2.39475027e-01 3.05722058e-01 5.17614603e-01 4.21641648e-01
4.65556160e-02 4.06888090e-02 6.59194231e-01 -1.09008066e-01
-8.24431717e-01 2.58776695e-01 -3.28831077e-01 -2.68569309e-02
9.50986385e-01 -2.31622070e-01 -4.35284823e-01 -3.59849155e-01
-2.81051308e-01 2.06223562e-01 3.57305944e-01 6.75102413e-01
3.22980434e-01 -1.30563891e+00 -5.75974345e-01 5.32241575e-02
7.56633282e-02 7.86227267e-03 7.03998804e-02 7.75641680e-01
-6.03300333e-01 7.17658997e-01 -6.42933786e-01 -7.83072174e-01
-8.20810676e-01 5.32379448e-01 7.09013164e-01 -5.54248333e-01
-6.00657940e-01 4.97190416e-01 -3.12606432e-02 -7.08190680e-01
4.37282145e-01 -5.62912464e-01 2.73470562e-02 -2.78288484e-01
1.01827994e-01 2.05022499e-01 4.76669312e-01 -2.25221679e-01
-4.14778829e-01 2.43219942e-01 -2.47627124e-01 -5.95519841e-02
1.67541933e+00 4.44960564e-01 2.38231048e-01 -7.56560788e-02
1.08657885e+00 -4.09988374e-01 -1.84917450e+00 -1.84395954e-01
-2.66819056e-02 1.03924135e-02 1.32027209e-01 -1.01395333e+00
-1.44070387e+00 6.04826212e-01 8.94323945e-01 1.42632008e-01
1.09321404e+00 -2.18577802e-01 6.22692883e-01 8.08865845e-01
5.53707004e-01 -1.16222119e+00 5.36185980e-01 6.51576221e-01
1.00889575e+00 -1.18322766e+00 -3.73267792e-02 4.37250823e-01
-5.40544748e-01 1.05748487e+00 8.66374493e-01 -6.09141171e-01
2.22271547e-01 2.63527662e-01 7.91134387e-02 -1.78384960e-01
-9.54107225e-01 -3.22172999e-01 -4.51220647e-02 7.46916771e-01
1.40011713e-01 4.25894652e-03 6.55206144e-02 6.96817189e-02
-3.44295233e-01 -1.63930431e-01 3.04931849e-01 1.08446789e+00
-1.59535363e-01 -1.22084689e+00 -7.78777823e-02 -4.92762104e-02
-3.31198782e-01 1.68609530e-01 -2.00410560e-01 1.27785492e+00
-9.62185338e-02 7.33259261e-01 7.13860542e-02 -1.70493290e-01
6.33728981e-01 -4.54468310e-01 6.43666208e-01 -1.94118053e-01
-1.18944871e+00 -2.16151416e-01 -8.52079615e-02 -9.92317617e-01
-5.92379451e-01 -6.12809181e-01 -1.31452954e+00 -3.44359100e-01
-1.20017804e-01 -1.00346208e-01 7.48947561e-01 8.86707604e-01
4.16370720e-01 7.99075305e-01 5.79708040e-01 -8.36650908e-01
-1.01565146e+00 -7.63280571e-01 -2.33297080e-01 4.31849629e-01
4.06208664e-01 -9.05539989e-01 -9.84843820e-02 -5.20539999e-01] | [4.077277183532715, 1.607572078704834] |
1a6ddd6d-7af4-41d2-8545-1d07f973bd71 | scene-graph-as-pivoting-inference-time-image | 2305.12256 | null | https://arxiv.org/abs/2305.12256v2 | https://arxiv.org/pdf/2305.12256v2.pdf | Scene Graph as Pivoting: Inference-time Image-free Unsupervised Multimodal Machine Translation with Visual Scene Hallucination | In this work, we investigate a more realistic unsupervised multimodal machine translation (UMMT) setup, inference-time image-free UMMT, where the model is trained with source-text image pairs, and tested with only source-text inputs. First, we represent the input images and texts with the visual and language scene graphs (SG), where such fine-grained vision-language features ensure a holistic understanding of the semantics. To enable pure-text input during inference, we devise a visual scene hallucination mechanism that dynamically generates pseudo visual SG from the given textual SG. Several SG-pivoting based learning objectives are introduced for unsupervised translation training. On the benchmark Multi30K data, our SG-based method outperforms the best-performing baseline by significant BLEU scores on the task and setup, helping yield translations with better completeness, relevance and fluency without relying on paired images. Further in-depth analyses reveal how our model advances in the task setting. | ['Tat-Seng Chua', 'Min Zhang', 'Meishan Zhang', 'Qian Liu', 'Hao Fei'] | 2023-05-20 | null | null | null | null | ['multimodal-machine-translation'] | ['natural-language-processing'] | [ 4.98496503e-01 2.02640533e-01 -9.14392844e-02 -2.56642491e-01
-9.83165085e-01 -7.53925979e-01 1.21972895e+00 -3.42020839e-01
-2.57491797e-01 5.15708745e-01 3.99211735e-01 -4.70300019e-01
5.85149765e-01 -3.43077630e-01 -1.10778105e+00 -4.43764240e-01
7.00345576e-01 7.83287764e-01 -1.39828309e-01 -1.49900302e-01
5.26744053e-02 8.54048133e-02 -1.18468821e+00 8.25662017e-01
9.00705755e-01 6.02569878e-01 6.85311913e-01 7.72283018e-01
-5.41847833e-02 7.52790868e-01 -4.31188464e-01 -9.02719975e-01
1.63834006e-01 -7.64220536e-01 -9.17026043e-01 5.73582828e-01
7.54215956e-01 -3.86600077e-01 -3.75232011e-01 8.69414628e-01
4.03817385e-01 -1.10956289e-01 7.55847037e-01 -1.18388975e+00
-9.87780452e-01 4.80388701e-01 -3.91147643e-01 -3.00297141e-01
6.18100703e-01 5.92323422e-01 9.07005191e-01 -1.31520414e+00
1.16238046e+00 1.35091579e+00 6.93503395e-02 4.42792803e-01
-1.66350186e+00 -1.85427502e-01 -6.01629242e-02 1.94095805e-01
-1.07522631e+00 -7.17652798e-01 5.79437971e-01 -3.89080614e-01
9.18339133e-01 2.76183099e-01 6.27157986e-01 1.72916031e+00
1.64884076e-01 9.97112036e-01 1.29502392e+00 -8.42235565e-01
1.07079551e-01 2.52801061e-01 -4.78204548e-01 7.93091118e-01
-1.89665016e-02 2.99269557e-02 -8.65221202e-01 3.40260386e-01
7.74025798e-01 -1.22967318e-01 -3.36811393e-01 -8.13373208e-01
-1.94969165e+00 6.07378185e-01 3.80424768e-01 2.19148219e-01
-2.65360326e-01 1.28720522e-01 1.23553991e-01 4.65919584e-01
2.65655249e-01 5.04722953e-01 -7.57720545e-02 -5.29810339e-02
-1.13498366e+00 2.56464053e-02 5.03164411e-01 1.24711573e+00
7.90611923e-01 1.36591241e-01 -6.94052339e-01 5.76262236e-01
2.10713059e-01 9.15119529e-01 3.12726229e-01 -1.11898446e+00
8.41424108e-01 5.02625346e-01 5.33172376e-02 -6.72169209e-01
1.09043375e-01 -3.84945989e-01 -6.40034735e-01 3.86605486e-02
4.68402177e-01 2.23494247e-01 -1.22306740e+00 1.66091025e+00
6.48659542e-02 -2.81915724e-01 3.44261378e-01 9.74479496e-01
8.28794301e-01 6.44613802e-01 -1.23533592e-01 -1.79128170e-01
1.34512031e+00 -1.37004447e+00 -7.61125386e-01 -4.20022815e-01
4.71191555e-01 -8.12341332e-01 1.73496616e+00 1.33262649e-01
-1.24991417e+00 -5.38215458e-01 -8.90232265e-01 -5.28164387e-01
-2.73598760e-01 4.85724747e-01 2.10658461e-01 2.22230062e-01
-1.47230208e+00 2.17196807e-01 -7.37994254e-01 -7.15631247e-01
3.17189813e-01 -4.01072614e-02 -4.85217214e-01 -3.63394052e-01
-8.07781994e-01 1.08638871e+00 1.86415434e-01 -6.18577981e-03
-1.42654812e+00 -3.22029293e-01 -1.05017745e+00 -1.70246571e-01
2.72794783e-01 -1.09090519e+00 1.41837490e+00 -1.36201036e+00
-1.58781695e+00 1.29599714e+00 -4.93824661e-01 -2.05885828e-01
8.97405565e-01 1.20623641e-01 1.75201278e-02 4.91160423e-01
1.19962193e-01 1.18563426e+00 1.20399809e+00 -1.73219907e+00
-3.78332026e-02 -2.77038276e-01 9.11796764e-02 5.62735379e-01
-1.40693128e-01 -1.06705405e-01 -8.15278828e-01 -5.79553008e-01
-1.04076579e-01 -9.26912427e-01 -4.67792451e-02 -2.42584571e-02
-6.00463688e-01 3.84658277e-01 5.65760076e-01 -8.73498917e-01
6.44574106e-01 -2.02077031e+00 5.61974525e-01 -1.81140646e-01
1.93041876e-01 -8.70709717e-02 -4.44372177e-01 7.19072640e-01
2.25007609e-01 -1.89980760e-01 -2.79629111e-01 -9.33667719e-01
6.02828935e-02 2.87269264e-01 -4.78680015e-01 1.02248341e-01
1.32378712e-01 1.72950256e+00 -9.64041054e-01 -6.81486785e-01
2.86260158e-01 2.78798103e-01 -3.61519217e-01 3.88752013e-01
-5.37942290e-01 7.23206639e-01 -1.25826359e-01 6.75069571e-01
3.57022047e-01 -4.92845386e-01 3.60129565e-01 -3.98831487e-01
1.10176660e-01 3.45094837e-02 -2.43247554e-01 2.35352850e+00
-5.67674637e-01 9.37338948e-01 -8.82807299e-02 -7.84063280e-01
5.09825647e-01 2.47689798e-01 -1.86031103e-01 -1.07552862e+00
-8.40069130e-02 1.02155872e-01 -1.93305731e-01 -5.62610269e-01
6.65847600e-01 1.51564479e-01 -4.17931527e-02 5.86027503e-01
3.51212025e-01 -4.09838319e-01 5.15280962e-02 6.80302322e-01
6.82620168e-01 5.71623921e-01 2.68056132e-02 -7.35606253e-02
1.11619979e-01 3.54197413e-01 -1.93728819e-01 7.88339138e-01
5.19372411e-02 9.61409330e-01 4.50648338e-01 -4.25021425e-02
-1.38213301e+00 -1.42178357e+00 3.31663907e-01 1.10115635e+00
1.49056092e-01 -4.10214782e-01 -9.22812581e-01 -7.74738252e-01
-3.78737539e-01 1.05604780e+00 -5.40084660e-01 4.42880467e-02
-3.21289957e-01 -2.04893574e-01 4.47542548e-01 2.50011146e-01
4.15423542e-01 -1.13379300e+00 -5.45676291e-01 -2.36492202e-01
-7.51727939e-01 -1.55731988e+00 -5.66760182e-01 -1.92389816e-01
-7.32940912e-01 -6.18679106e-01 -1.06788206e+00 -9.36825931e-01
1.05916369e+00 3.75208080e-01 1.32242060e+00 -2.48218954e-01
-8.63662362e-02 6.68952405e-01 -2.39231229e-01 -5.58509342e-02
-8.62605870e-01 -2.62845993e-01 -2.72514194e-01 9.92493033e-02
-1.78674623e-01 -5.41021943e-01 -5.70401967e-01 3.72078381e-02
-8.50144446e-01 1.11951542e+00 8.95613074e-01 8.92406106e-01
5.55599272e-01 -9.27391648e-01 3.75327468e-02 -5.19010127e-01
5.10659337e-01 -2.42264718e-02 -3.09878916e-01 6.07307374e-01
-4.32566702e-01 1.07584596e-01 5.88483155e-01 -5.20522058e-01
-1.02465355e+00 -2.25116834e-02 3.88545334e-01 -7.59777844e-01
-8.07462335e-02 2.80889273e-01 -3.47707301e-01 1.65704191e-01
6.69340074e-01 7.99999833e-01 4.54194732e-02 -2.28845417e-01
9.84244943e-01 6.97288692e-01 8.39527786e-01 -7.16051698e-01
8.06445420e-01 5.88271856e-01 -1.90196261e-01 -6.26078606e-01
-6.11572266e-01 -6.77555278e-02 -8.71010780e-01 -2.17584223e-01
1.06526065e+00 -1.05736196e+00 -2.94153929e-01 2.92224914e-01
-1.37451327e+00 -7.25274801e-01 -9.40670669e-02 2.78705865e-01
-9.66507018e-01 4.46799368e-01 -5.19907534e-01 -5.05668283e-01
-2.50277579e-01 -1.48646855e+00 1.63599944e+00 -2.39690140e-01
-9.43762660e-02 -9.26163018e-01 -1.33982003e-01 7.69308090e-01
2.08993301e-01 6.84060976e-02 1.00663543e+00 -3.60957831e-01
-9.60791469e-01 4.03281897e-01 -4.71863031e-01 4.93400283e-02
-4.86927815e-02 -2.10881457e-01 -1.07863450e+00 -3.47727895e-01
-3.13258231e-01 -6.41529679e-01 7.98898101e-01 1.00667708e-01
7.66145527e-01 -3.68351132e-01 -2.29462147e-01 6.74382746e-01
1.22506428e+00 -1.83311880e-01 5.62600970e-01 2.55531639e-01
1.01872790e+00 6.07271552e-01 3.79848540e-01 -1.28835320e-01
6.77728176e-01 8.60156536e-01 3.82893324e-01 -5.23004293e-01
-5.25644481e-01 -6.96292222e-01 7.87090123e-01 8.04636061e-01
-2.18736567e-03 -4.74932224e-01 -8.48888397e-01 6.72415614e-01
-1.93457413e+00 -8.17814112e-01 2.72976551e-02 2.06661510e+00
9.96461928e-01 -1.83612928e-01 -1.24048978e-01 -4.84859288e-01
5.04019737e-01 1.65127352e-01 -4.15677309e-01 -4.33603168e-01
-3.38158518e-01 -2.55184509e-02 1.13572031e-01 5.83130360e-01
-6.93934202e-01 1.35511398e+00 6.12648726e+00 7.93723404e-01
-1.08830428e+00 2.48395428e-01 6.85922027e-01 -1.75510406e-01
-6.78163707e-01 1.05745137e-01 -2.04113841e-01 2.05160722e-01
7.76842713e-01 2.45608445e-02 7.80611217e-01 3.75536710e-01
4.31084305e-01 -1.19961843e-01 -1.42003191e+00 1.18007791e+00
4.33105946e-01 -1.52397323e+00 5.62588751e-01 1.28151923e-01
9.46909130e-01 4.99635972e-02 2.60374248e-01 2.28803143e-01
3.27987403e-01 -1.02556229e+00 1.06580067e+00 5.09122312e-01
1.43010914e+00 -2.87372500e-01 1.96056604e-01 3.18969578e-01
-6.66208148e-01 3.63590568e-01 -7.27351755e-02 1.67461529e-01
3.68489355e-01 2.22134903e-01 -9.67803419e-01 8.62812936e-01
2.97673911e-01 6.78402722e-01 -8.21778119e-01 3.08947295e-01
-5.17667949e-01 4.92615759e-01 9.67989638e-02 1.77237108e-01
3.19723666e-01 -3.10201943e-01 5.32183886e-01 1.30260146e+00
2.95281380e-01 -3.67196351e-01 2.39535332e-01 1.19646025e+00
-2.84221411e-01 9.01389495e-02 -8.56436253e-01 -2.72660822e-01
1.28641605e-01 1.30975485e+00 -5.99300504e-01 -6.42617464e-01
-3.68148416e-01 1.78297806e+00 5.40470004e-01 9.18723106e-01
-7.67698467e-01 2.25267693e-01 3.32684398e-01 8.56572092e-02
1.70400962e-01 -3.48306656e-01 -3.48038733e-01 -1.60149646e+00
5.40566966e-02 -1.11550820e+00 -9.37211812e-02 -1.54055619e+00
-7.59234607e-01 5.34518063e-01 2.11719275e-02 -1.13832116e+00
-5.10187626e-01 -6.05283916e-01 -4.26679730e-01 8.13216865e-01
-1.28881979e+00 -1.79143119e+00 -3.11996341e-01 7.02315450e-01
8.59776735e-01 -7.53219649e-02 7.50236154e-01 -2.24291176e-01
-2.63631105e-01 6.62605643e-01 5.76212257e-02 1.38991043e-01
8.50860298e-01 -1.11932564e+00 6.29096806e-01 1.10959899e+00
7.35718250e-01 5.61163545e-01 6.30042493e-01 -5.76010764e-01
-1.77428651e+00 -1.13609052e+00 9.40674901e-01 -8.02564085e-01
5.62266052e-01 -8.70081186e-01 -4.69866753e-01 8.12217832e-01
8.90237510e-01 -3.66583318e-01 1.37432545e-01 -3.37954789e-01
-4.40337241e-01 2.44566113e-01 -8.08483720e-01 1.22520232e+00
1.30408430e+00 -8.82358909e-01 -7.71552324e-01 5.48723519e-01
9.66847956e-01 -4.35415506e-01 -4.87656981e-01 9.04638320e-02
4.72814232e-01 -8.55433047e-01 1.00313389e+00 -5.35529256e-01
1.02719903e+00 -3.74344617e-01 -2.59535283e-01 -1.49854016e+00
-3.31986509e-02 -7.99585521e-01 -3.27884629e-02 1.01054251e+00
7.08744526e-01 -2.63339013e-01 2.76075602e-01 1.89981312e-01
-2.34295383e-01 -5.14555335e-01 -6.47737920e-01 -5.84569931e-01
-9.27341953e-02 -3.15933824e-01 2.34477475e-01 9.42544103e-01
6.82739243e-02 6.45800889e-01 -4.75819528e-01 -1.88278958e-01
6.38112962e-01 4.40603137e-01 1.03179693e+00 -5.65111816e-01
-4.81108457e-01 -3.46601665e-01 -3.14438716e-02 -1.14459336e+00
1.55945823e-01 -1.17375767e+00 1.43781947e-02 -1.84840763e+00
7.03120947e-01 3.65229756e-01 1.97214261e-01 5.25331080e-01
-1.70306548e-01 5.18403411e-01 3.86667252e-01 4.59420472e-01
-7.59753525e-01 6.84749663e-01 1.80752707e+00 -3.78631115e-01
5.80941141e-02 -6.72848880e-01 -5.28299093e-01 3.19603354e-01
4.04585004e-01 -9.80285183e-03 -5.70263028e-01 -9.87442553e-01
1.73004970e-01 2.76551157e-01 7.06636071e-01 -4.93914723e-01
2.13906504e-02 -2.54247487e-01 5.94675779e-01 -5.54296494e-01
5.02469718e-01 -4.27376777e-01 1.15899004e-01 1.98478594e-01
-5.40430069e-01 2.34562755e-01 -1.93680804e-02 4.19622719e-01
-9.87042636e-02 3.01731139e-01 4.38930482e-01 -2.35781476e-01
-6.26955211e-01 1.84234977e-01 -1.94117174e-01 6.40863925e-02
7.66855717e-01 -4.29841846e-01 -5.44539332e-01 -6.98655903e-01
-7.07917035e-01 1.36844233e-01 1.15676475e+00 4.87313062e-01
7.53984571e-01 -1.45300758e+00 -8.99390996e-01 1.19150423e-01
5.99796414e-01 -1.36873260e-01 -1.36885345e-02 1.04882383e+00
-4.56450909e-01 4.12512690e-01 -1.48641035e-01 -9.47906792e-01
-1.10708559e+00 6.67924166e-01 1.24087878e-01 -1.88822433e-01
-6.63071513e-01 4.85641241e-01 7.25361526e-01 -5.42604148e-01
-6.63684309e-02 -2.19482809e-01 2.65400559e-01 -2.93022543e-01
3.05582792e-01 -9.33837891e-02 -2.86397282e-02 -7.77383327e-01
-8.80839750e-02 3.92182112e-01 3.23076509e-02 -9.54125643e-01
8.23799670e-01 -4.32199478e-01 2.67933635e-03 4.88900632e-01
1.03142941e+00 1.80068333e-02 -1.35242712e+00 -2.30459690e-01
-2.67790139e-01 -4.21723545e-01 -2.15337425e-01 -1.19115019e+00
-6.16279006e-01 1.14244294e+00 2.77772486e-01 -4.28762585e-01
9.92415667e-01 1.64328262e-01 5.03811419e-01 4.84690428e-01
3.48736733e-01 -7.84325361e-01 4.31929380e-01 5.05408943e-01
1.16822505e+00 -1.27362597e+00 -3.06425095e-01 -1.22735977e-01
-1.05816865e+00 8.73593390e-01 5.29885530e-01 3.44198465e-01
-3.35689843e-01 -9.06588137e-02 2.83189356e-01 -3.36342901e-02
-8.36443126e-01 -4.23510581e-01 6.39844239e-01 8.09175074e-01
2.03510121e-01 1.37190387e-01 4.74262200e-02 1.33848250e-01
-3.67118716e-01 -2.69223154e-01 4.16658759e-01 5.29542625e-01
-1.55333638e-01 -9.67045844e-01 -3.21252912e-01 -8.18822309e-02
-4.06546460e-04 -5.27849495e-01 -8.24777663e-01 7.56715894e-01
-2.70489156e-01 9.73300338e-01 6.17909767e-02 -2.77060688e-01
1.19131655e-01 2.76739538e-01 9.72848594e-01 -5.55730641e-01
-2.52297699e-01 1.78608894e-01 1.17666587e-01 -7.81312585e-01
-3.10699910e-01 -5.12170255e-01 -1.03216302e+00 -4.33681086e-02
8.49445313e-02 -2.25305721e-01 7.76665628e-01 1.10072267e+00
5.77684224e-01 3.75208586e-01 4.22970414e-01 -1.09960151e+00
-2.50588208e-01 -1.05958831e+00 1.09524608e-01 7.08920896e-01
3.51817608e-01 -2.50361800e-01 -7.54929259e-02 6.17716730e-01] | [11.339550018310547, 1.4018806219100952] |
ae391e9e-b5d8-4542-86bc-217051238991 | pure-exploration-in-multi-armed-bandits-with-1 | 2306.15856 | null | https://arxiv.org/abs/2306.15856v1 | https://arxiv.org/pdf/2306.15856v1.pdf | Pure exploration in multi-armed bandits with low rank structure using oblivious sampler | In this paper, we consider the low rank structure of the reward sequence of the pure exploration problems. Firstly, we propose the separated setting in pure exploration problem, where the exploration strategy cannot receive the feedback of its explorations. Due to this separation, it requires that the exploration strategy to sample the arms obliviously. By involving the kernel information of the reward vectors, we provide efficient algorithms for both time-varying and fixed cases with regret bound $O(d\sqrt{(\ln N)/n})$. Then, we show the lower bound to the pure exploration in multi-armed bandits with low rank sequence. There is an $O(\sqrt{\ln N})$ gap between our upper bound and the lower bound. | ['Eiji Takimoto', 'Kohei Hatano', 'Atsuyoshi Nakamura', 'Yaxiong Liu'] | 2023-06-28 | null | null | null | null | ['multi-armed-bandits'] | ['miscellaneous'] | [ 7.96061307e-02 3.58701259e-01 -7.16940463e-01 1.17415050e-02
-9.43209469e-01 -1.16481137e+00 -1.51743710e-01 -1.17998511e-01
-7.61953115e-01 9.34007883e-01 7.57849123e-03 -8.48409891e-01
-8.85948539e-01 -8.93694520e-01 -9.51767921e-01 -9.92650867e-01
-5.96407533e-01 7.78270304e-01 9.64103173e-03 -1.06562495e-01
3.63038361e-01 5.55842578e-01 -1.00692284e+00 -9.62300375e-02
6.05779052e-01 1.46746159e+00 3.61241214e-02 7.80673265e-01
-1.45021111e-01 5.57481229e-01 -1.47003710e-01 -3.46032739e-01
9.61495876e-01 -5.24112940e-01 -1.07944584e+00 7.74918031e-03
-2.37766847e-01 -5.43141425e-01 -5.83672643e-01 1.31733310e+00
2.61867046e-01 2.49146357e-01 2.71647453e-01 -1.09631205e+00
-2.13744953e-01 1.20283258e+00 -1.23775995e+00 3.70519817e-01
1.64586678e-01 -1.29817113e-01 1.35052562e+00 -1.51163101e-01
5.36282599e-01 1.19604588e+00 -8.64929259e-02 2.25918233e-01
-1.39331377e+00 -8.49357307e-01 4.98743087e-01 1.30907884e-02
-9.11590219e-01 -2.52752602e-01 5.30709743e-01 -6.63108751e-02
2.78696746e-01 6.61787033e-01 5.29408634e-01 6.79188311e-01
-5.58753192e-01 1.14598823e+00 1.40001118e+00 -3.96705151e-01
4.73141432e-01 -5.64801060e-02 4.84227568e-01 3.64938974e-01
6.98755085e-01 5.76892495e-01 -3.12686026e-01 -3.34820241e-01
7.60101855e-01 9.64488462e-02 -3.00970525e-01 -5.43174505e-01
-7.59385467e-01 9.74714458e-01 2.57544607e-01 -1.71133384e-01
-2.40156412e-01 6.05509281e-01 1.53976321e-01 7.87226975e-01
8.83919969e-02 2.77747482e-01 -3.33542854e-01 -3.81924629e-01
-8.76157224e-01 1.79192618e-01 9.13189054e-01 1.16969228e+00
7.26404727e-01 -2.75709420e-01 -3.02713007e-01 3.27828139e-01
6.27516583e-02 6.62741184e-01 -2.78486172e-03 -1.20366657e+00
1.21250355e+00 1.33787617e-01 8.26240659e-01 -6.10859275e-01
-3.14440280e-01 -7.13757575e-01 -6.91519260e-01 1.77254543e-01
7.48453259e-01 -3.67498904e-01 -5.01100659e-01 1.84493887e+00
2.84429967e-01 -4.57835823e-01 -2.96850689e-02 8.95410776e-01
-1.61864355e-01 4.66304928e-01 -7.44054019e-01 -7.96639800e-01
9.96504188e-01 -9.62948382e-01 -4.69209462e-01 -3.12257737e-01
6.10383928e-01 -3.99746776e-01 8.81190598e-01 6.30209327e-01
-1.61414051e+00 1.34235486e-01 -9.20420289e-01 5.19627035e-01
2.04523444e-01 -1.00912407e-01 7.85363913e-01 8.03531826e-01
-6.88648641e-01 6.61755741e-01 -7.66421139e-01 3.65269870e-01
4.92543370e-01 6.79009676e-01 -3.64254462e-03 -1.90247908e-01
-6.89764082e-01 1.14613451e-01 4.16320711e-01 1.16181411e-01
-9.66885507e-01 -1.36101022e-01 -2.60832369e-01 1.95747539e-01
1.21008658e+00 -1.49121091e-01 1.39532709e+00 -5.47231019e-01
-1.41560757e+00 2.95110732e-01 -6.58635125e-02 -4.66792494e-01
8.27839434e-01 -3.62432867e-01 3.85909438e-01 1.82187945e-01
5.11731505e-02 -1.97705984e-01 4.75916594e-01 -8.96773040e-01
-8.64286184e-01 -7.09222436e-01 5.43696404e-01 3.05812508e-02
-1.95732817e-01 -1.91604793e-01 -3.91587049e-01 -3.26221377e-01
4.26718503e-01 -1.18116379e+00 -5.19394934e-01 -3.22896004e-01
-5.20089447e-01 5.82783483e-02 -3.01757101e-02 -5.96704893e-02
1.19908178e+00 -2.08535409e+00 2.14475855e-01 8.55652332e-01
2.04796139e-02 -1.19787648e-01 -2.67410338e-01 5.14679313e-01
1.60470217e-01 2.09355846e-01 2.10571572e-01 -1.80297773e-02
2.26917848e-01 2.89509714e-01 -9.17645335e-01 5.90985656e-01
-7.67115474e-01 8.28730345e-01 -8.85898530e-01 9.92161222e-03
-5.06250679e-01 -6.73306644e-01 -5.76048732e-01 1.39979020e-01
-2.11691096e-01 3.07892989e-02 -9.57932472e-01 4.61019695e-01
1.06025624e+00 -4.92014050e-01 5.18623412e-01 3.18572313e-01
2.11259965e-02 2.69242316e-01 -1.73115444e+00 1.47684252e+00
-2.68972963e-01 4.06731851e-02 5.73107898e-01 -1.00343513e+00
2.69392073e-01 -3.51586118e-02 3.82967323e-01 -7.08037972e-01
1.79752186e-01 5.35023093e-01 -2.47207331e-03 -1.68232933e-01
4.55508173e-01 -1.86175168e-01 -1.33876324e-01 1.20677960e+00
-4.70858812e-01 5.75594306e-01 3.58127922e-01 2.66804427e-01
1.35155880e+00 -1.10865220e-01 4.88920137e-02 -9.46374387e-02
9.91359800e-02 -1.99783832e-01 5.39202809e-01 1.47832572e+00
1.23730525e-01 -2.40473971e-02 1.26482046e+00 -1.56169355e-01
-8.10889900e-01 -1.03483832e+00 2.63768554e-01 1.39745963e+00
3.04924369e-01 -1.79754689e-01 -4.23065186e-01 -7.38577902e-01
2.77764916e-01 5.12981832e-01 -9.94981706e-01 1.67897552e-01
-3.92695159e-01 -8.44990313e-01 2.58993655e-01 4.58462209e-01
3.51321906e-01 -3.98354024e-01 -7.39272773e-01 2.06920356e-01
7.09674805e-02 -7.29104638e-01 -5.96346915e-01 6.75312161e-01
-1.13050663e+00 -1.02694809e+00 -6.56097710e-01 -1.42922461e-01
7.06516623e-01 5.33197105e-01 5.68122268e-01 -3.39203566e-01
9.32188928e-02 1.22059278e-01 -3.27278286e-01 -2.62918025e-01
2.02003643e-01 2.70668089e-01 -9.67927836e-03 -2.73398727e-01
1.63000859e-02 -6.26097262e-01 -9.42110300e-01 5.38860857e-01
-8.86795044e-01 -1.25510201e-01 8.60200644e-01 8.01324785e-01
7.66895652e-01 1.34808021e-02 3.98667395e-01 -8.75610530e-01
6.02929652e-01 -4.95748580e-01 -1.46043825e+00 2.69245833e-01
-6.43221796e-01 5.61336160e-01 5.79468668e-01 -6.77908957e-01
-6.61286652e-01 1.65998172e-02 4.69149917e-01 -2.84076244e-01
5.77054679e-01 3.64068627e-01 9.62590352e-02 -8.05192441e-02
4.03992146e-01 2.56529301e-01 -5.35858385e-02 -8.47583055e-01
4.00429010e-01 4.40558851e-01 3.46260130e-01 -9.11469281e-01
9.31445360e-01 6.37228072e-01 4.50406522e-01 -7.66897425e-02
-1.19484568e+00 -2.50252634e-01 1.72148988e-01 3.26997548e-01
-1.14615597e-01 -5.40195882e-01 -1.64609766e+00 -3.52946013e-01
-6.72622681e-01 -3.88501704e-01 -5.97730756e-01 6.46983981e-01
-9.63275850e-01 4.69407767e-01 -3.49444658e-01 -1.66879892e+00
-1.80024236e-01 -9.91508961e-01 3.47722113e-01 2.22009495e-01
2.77682960e-01 -2.29511216e-01 8.18659887e-02 2.66857088e-01
2.53975868e-01 -5.45620844e-02 7.06636846e-01 -7.36403525e-01
-1.08219123e+00 -1.85994685e-01 -4.38058674e-01 -1.21022515e-01
-9.96984020e-02 -8.70982885e-01 -3.46164793e-01 -6.30268157e-01
9.25007313e-02 -5.36094129e-01 1.04815888e+00 3.52754653e-01
1.42446792e+00 -9.09540415e-01 -4.47750628e-01 6.67900622e-01
1.29186153e+00 3.11302722e-01 4.45900828e-01 5.20880580e-01
-3.37289982e-02 3.67189080e-01 1.06112432e+00 9.57328916e-01
-2.44633451e-01 5.17430365e-01 6.44394279e-01 4.26672250e-01
6.93556070e-01 -2.22222075e-01 2.53116667e-01 -1.28593817e-01
-2.26519242e-01 -2.72620529e-01 -3.91467988e-01 5.07051945e-01
-2.15644050e+00 -9.07548904e-01 2.49191433e-01 2.71994591e+00
1.07417405e+00 5.03327966e-01 4.47414547e-01 7.42518157e-02
3.82622033e-01 2.61639133e-02 -9.68815506e-01 -6.82411015e-01
-7.12802112e-02 2.91122615e-01 1.29735482e+00 5.02370596e-01
-6.53859496e-01 6.18373215e-01 6.39064646e+00 1.27605796e+00
-4.96800482e-01 -1.98831526e-03 6.59614265e-01 -1.13492501e+00
-4.78433788e-01 2.33361840e-01 -9.12332475e-01 4.73747730e-01
8.41724455e-01 -3.51211011e-01 1.08344162e+00 1.03551328e+00
-4.33080867e-02 -5.01602829e-01 -1.30126953e+00 7.95245707e-01
-5.10676324e-01 -1.22664964e+00 -4.03596103e-01 7.03595102e-01
8.32135558e-01 -2.00948820e-01 4.27928746e-01 4.94643040e-02
8.87364268e-01 -1.03098321e+00 5.40935099e-01 4.30949964e-02
7.46822298e-01 -1.13773167e+00 4.94003832e-01 6.23549044e-01
-8.65956366e-01 -7.17154801e-01 -5.01596630e-01 -7.44993240e-02
1.95238367e-01 6.11212611e-01 -4.37699854e-01 6.19468868e-01
6.07522130e-01 -1.36032179e-01 1.95557371e-01 1.13756573e+00
-1.24502644e-01 4.37351912e-01 -7.92554796e-01 -2.56674558e-01
6.63250804e-01 -5.66363037e-01 6.53239369e-01 7.00785041e-01
3.85209620e-01 3.37118655e-01 2.07919493e-01 6.75451815e-01
-2.26718783e-01 2.91079953e-02 -1.36993140e-01 -2.53509104e-01
3.96301985e-01 9.09531593e-01 -7.07781553e-01 4.60933074e-02
3.35943103e-02 7.40517318e-01 4.77641046e-01 4.25166935e-01
-6.79603398e-01 -6.15464747e-01 6.03149414e-01 -6.01402670e-02
8.92519295e-01 -1.47957489e-01 -3.40846628e-01 -6.43686652e-01
3.65340322e-01 -4.28179294e-01 5.78401208e-01 1.74210258e-02
-9.69517112e-01 3.01323652e-01 1.18015446e-01 -9.41952109e-01
-3.42042416e-01 -4.21417803e-01 -2.53071696e-01 8.99459541e-01
-1.50545835e+00 -4.51340258e-01 4.59395409e-01 3.61370325e-01
-9.29812249e-03 -3.88566926e-02 4.71088260e-01 -1.41291291e-01
-4.22124505e-01 9.20702577e-01 8.30110550e-01 -2.06904635e-01
1.55119926e-01 -1.17285812e+00 -1.42102078e-01 5.52300274e-01
-8.56812671e-02 5.35984576e-01 6.96697295e-01 -4.76855129e-01
-1.72681069e+00 -6.35193527e-01 1.67293385e-01 9.60621014e-02
9.06570256e-01 -1.93117231e-01 -2.73177266e-01 7.64197230e-01
-1.18862659e-01 -6.64264038e-02 8.37102175e-01 3.88398498e-01
-3.88514519e-01 -3.64139110e-01 -1.06883991e+00 6.04319155e-01
1.24804461e+00 -1.53381601e-01 -1.28731772e-01 2.62470722e-01
6.64518595e-01 -4.79855299e-01 -6.10921860e-01 1.52326941e-01
7.64384747e-01 -8.39448750e-01 8.70983958e-01 -9.61920738e-01
2.01333210e-01 1.11556232e-01 -4.31578130e-01 -8.82296205e-01
-3.52158159e-01 -1.31303000e+00 -7.81738937e-01 6.41913712e-01
6.70713484e-01 -6.47266150e-01 1.21466899e+00 6.31358862e-01
4.32787925e-01 -1.09764910e+00 -1.33728373e+00 -1.15324473e+00
3.22861522e-01 -2.78454214e-01 6.59904599e-01 1.56691849e-01
2.74479747e-01 5.17599322e-02 -6.73576891e-01 -5.60471695e-03
8.28640699e-01 8.97230685e-01 6.50339127e-01 -8.53575110e-01
-1.14918089e+00 -4.66346174e-01 3.94389331e-01 -1.65162683e+00
-2.14847133e-01 -7.31840611e-01 -3.61764491e-01 -1.16207051e+00
7.83289969e-01 -7.78012455e-01 -6.19531393e-01 4.53221530e-01
1.09702617e-01 -2.83015251e-01 3.31026882e-01 2.72640288e-01
-1.01047206e+00 4.59333926e-01 1.38215637e+00 -3.44218798e-02
-3.87612551e-01 5.66057444e-01 -1.16642940e+00 4.51053619e-01
7.90630758e-01 -7.62585580e-01 -4.42916483e-01 -3.57634902e-01
8.52158189e-01 6.33748293e-01 -8.33854824e-03 -3.46366823e-01
1.20644964e-01 -6.92503810e-01 2.28910353e-02 -9.86642301e-01
2.77862191e-01 -7.54715741e-01 1.53960288e-02 7.94987559e-01
-8.83095264e-01 -2.19386786e-01 -8.06925222e-02 1.04153955e+00
4.54798639e-01 -5.44658899e-01 5.13270795e-01 -1.12757511e-01
3.86789888e-01 5.39427817e-01 -1.89473748e-01 1.98252261e-01
1.00414920e+00 -1.82234153e-01 -4.02731270e-01 -5.74645221e-01
-8.14823806e-01 7.20197916e-01 1.53488323e-01 -3.30959231e-01
2.61743665e-01 -1.24559140e+00 -3.18106771e-01 -1.12263255e-01
-2.38690481e-01 5.05535454e-02 3.24028105e-01 1.03282523e+00
1.49903834e-01 6.43972158e-01 -8.58874023e-02 -1.10644124e-01
-9.95348692e-01 9.36624825e-01 -4.94805425e-02 -8.25724542e-01
-9.06306952e-02 9.11820054e-01 2.24485904e-01 1.71533957e-01
5.61461449e-01 -9.14269835e-02 2.74221480e-01 1.53294364e-02
7.08384812e-01 8.27965379e-01 -4.10838157e-01 3.18241060e-01
-1.67132858e-02 9.22527611e-02 -6.46854460e-01 -7.55056679e-01
1.37065184e+00 -2.78821707e-01 -6.14289194e-02 1.85487792e-01
1.18148780e+00 1.68368354e-01 -1.32908249e+00 -5.93637288e-01
3.41978148e-02 -9.29843485e-01 -7.12443143e-02 -6.54849172e-01
-1.10416842e+00 5.39380312e-01 6.10830605e-01 4.65743244e-01
1.08199513e+00 1.55131742e-01 6.73038542e-01 8.59487593e-01
8.30224156e-01 -1.34340930e+00 -4.58502918e-02 4.73246992e-01
7.18048513e-01 -8.23345423e-01 1.05707616e-01 -1.61094427e-01
-4.11805451e-01 9.22876656e-01 1.65726826e-01 -1.79009229e-01
5.30216694e-01 1.40949056e-01 -8.83649409e-01 9.88302082e-02
-7.76202917e-01 -3.31043512e-01 -2.01308459e-01 5.24256155e-02
-9.05948952e-02 2.59244025e-01 -6.69297338e-01 1.00728595e+00
-2.73846090e-01 3.42687219e-02 4.85239089e-01 1.01825142e+00
-6.70602441e-01 -1.60123146e+00 -5.09114385e-01 7.41293013e-01
-7.83465147e-01 1.07842408e-01 -1.77886069e-01 4.53892708e-01
-5.79426110e-01 9.80228662e-01 -1.59627780e-01 -1.65842220e-01
1.22947827e-01 -3.87237996e-01 6.43757701e-01 -1.90352723e-01
-2.21866444e-01 3.43135238e-01 6.96608722e-02 -7.93065608e-01
1.45728849e-02 -4.61687207e-01 -9.83314633e-01 -6.10926390e-01
-4.23753232e-01 7.23587394e-01 3.51527542e-01 6.90918624e-01
2.96171695e-01 8.51843655e-02 1.24435925e+00 -2.06012636e-01
-1.47290945e+00 -7.28181064e-01 -1.14231575e+00 -1.16876408e-01
4.20920253e-01 -5.79240322e-01 -4.83370870e-01 -9.50963974e-01] | [4.577956199645996, 3.3597092628479004] |
95361383-a7dc-466a-8fe6-419018f87d68 | generating-safe-diversity-in-nlg-via | 2004.14364 | null | https://arxiv.org/abs/2004.14364v2 | https://arxiv.org/pdf/2004.14364v2.pdf | Informed Sampling for Diversity in Concept-to-Text NLG | Deep-learning models for language generation tasks tend to produce repetitive output. Various methods have been proposed to encourage lexical diversity during decoding, but this often comes at a cost to the perceived fluency and adequacy of the output. In this work, we propose to ameliorate this cost by using an Imitation Learning approach to explore the level of diversity that a language generation model can reliably produce. Specifically, we augment the decoding process with a meta-classifier trained to distinguish which words at any given timestep will lead to high-quality output. We focus our experiments on concept-to-text generation where models are sensitive to the inclusion of irrelevant words due to the strict relation between input and output. Our analysis shows that previous methods for diversity underperform in this setting, while human evaluation suggests that our proposed method achieves a high level of diversity with minimal effect to the output's fluency and adequacy. | ['Giulio Zhou', 'Gerasimos Lampouras'] | 2020-04-29 | null | https://aclanthology.org/2021.findings-emnlp.213 | https://aclanthology.org/2021.findings-emnlp.213.pdf | findings-emnlp-2021-11 | ['concept-to-text-generation'] | ['natural-language-processing'] | [ 1.55971721e-01 4.17495936e-01 -1.13016121e-01 -1.12029940e-01
-6.11531854e-01 -4.85068768e-01 1.01956344e+00 1.45028383e-01
-2.28545755e-01 8.49930644e-01 5.85438013e-01 -1.93597600e-01
3.37631732e-01 -7.36840010e-01 -5.01487255e-01 -3.49796861e-01
2.50591248e-01 3.96592647e-01 -2.67543674e-01 -3.53251159e-01
5.53421676e-01 7.06882402e-02 -1.66441858e+00 2.43993863e-01
1.38906002e+00 3.33570987e-01 4.15777832e-01 7.84866512e-01
-2.44616792e-01 8.83085430e-01 -9.94263291e-01 -4.54332858e-01
1.00390486e-01 -1.08241343e+00 -8.01164806e-01 6.21041395e-02
1.21545896e-01 -3.71805847e-01 3.10485605e-02 9.07693863e-01
6.41391873e-01 1.26485839e-01 1.05567431e+00 -1.05338776e+00
-8.18315327e-01 1.01165485e+00 4.58894782e-02 1.04121923e-01
5.35352468e-01 4.49751705e-01 8.15465868e-01 -7.21502841e-01
7.84032762e-01 1.24797344e+00 4.01257813e-01 7.85272300e-01
-1.31921434e+00 -6.03079796e-01 2.07747389e-02 -3.23863953e-01
-1.14343214e+00 -6.60687983e-01 7.36713350e-01 -5.05917132e-01
1.16338515e+00 7.45918229e-03 7.31162488e-01 1.48643684e+00
4.39175129e-01 7.52687752e-01 1.19513440e+00 -5.71221709e-01
5.00502706e-01 2.93450475e-01 -3.07851374e-01 4.54788297e-01
1.64596722e-01 1.60952643e-01 -7.19992161e-01 6.48667216e-02
7.06093669e-01 -7.20545590e-01 -1.00944690e-01 -3.01577300e-02
-1.03320026e+00 9.70149696e-01 1.95680812e-01 5.09063363e-01
-5.14637649e-01 1.49703547e-01 3.73683453e-01 5.62288463e-01
6.00773215e-01 8.84481788e-01 -1.35440558e-01 -5.47239602e-01
-9.57663953e-01 3.96466374e-01 8.76314461e-01 8.30388427e-01
3.87767524e-01 2.20235005e-01 -5.72018802e-01 8.63458514e-01
1.82389826e-01 2.46460706e-01 9.14388955e-01 -8.57345402e-01
4.26495939e-01 4.21675444e-01 1.94923744e-01 -7.23158956e-01
-1.12994656e-01 -5.41087747e-01 -5.36438406e-01 2.16325521e-01
2.94102222e-01 -3.48557979e-01 -9.27486122e-01 2.00982428e+00
-4.09604311e-02 -3.48006845e-01 2.25461975e-01 6.63529456e-01
5.22607803e-01 6.20260000e-01 3.38297129e-01 -3.79649311e-01
7.33353198e-01 -8.40471923e-01 -6.53298378e-01 -2.29813606e-01
8.36655855e-01 -8.13883066e-01 1.37328053e+00 2.79188126e-01
-1.38236248e+00 -5.81227183e-01 -9.38653946e-01 1.09114870e-01
-3.77973542e-02 3.86748835e-02 3.79105717e-01 7.43654132e-01
-1.19573629e+00 7.81366408e-01 -5.87411284e-01 -4.32783842e-01
1.71195552e-01 5.70228994e-02 1.08115584e-01 3.05819869e-01
-1.19014359e+00 1.33460450e+00 6.42139971e-01 -2.34886721e-01
-7.99417019e-01 -2.50601351e-01 -6.09856486e-01 -3.64010595e-02
-5.07207401e-02 -9.07625854e-01 1.45288444e+00 -1.49028277e+00
-1.75952339e+00 5.92256546e-01 -6.13147765e-02 -2.63368577e-01
7.22460628e-01 -5.57953380e-02 -1.68040782e-01 -1.55373290e-01
1.78860381e-01 1.10419941e+00 8.12552452e-01 -1.16161180e+00
-5.06251454e-01 1.29219303e-02 9.51363822e-04 5.23443580e-01
-3.79855633e-01 -1.98835731e-01 7.92772397e-02 -8.64457965e-01
-1.59798980e-01 -8.86785090e-01 -1.46836162e-01 -5.24735808e-01
-2.44482443e-01 -4.67934936e-01 9.08265710e-02 -5.23712933e-01
1.28191161e+00 -1.58585584e+00 4.17428911e-01 -9.28632766e-02
-1.38867214e-01 1.39265612e-01 -3.25020313e-01 7.29268253e-01
4.00221407e-01 3.54791880e-01 -1.47091761e-01 -3.80126953e-01
-4.79559191e-02 6.87667131e-02 -2.48850405e-01 -3.83380949e-02
3.21261466e-01 9.43000376e-01 -1.14381063e+00 -3.57821167e-01
-2.08626091e-01 3.16895962e-01 -7.35949218e-01 3.31186444e-01
-5.12418866e-01 4.04961526e-01 -2.88911432e-01 3.37916493e-01
1.45820320e-01 -1.01476222e-01 2.08030149e-01 4.74839240e-01
-1.47284150e-01 7.49933124e-01 -8.56052458e-01 1.63581693e+00
-7.09714055e-01 4.82591897e-01 -3.62274349e-01 -5.06784141e-01
9.04571891e-01 3.49306673e-01 -1.31558850e-01 -9.49877441e-01
1.65317819e-01 3.00192595e-01 4.90423560e-01 -3.40536535e-01
9.95388329e-01 -3.57423782e-01 -4.93400320e-02 8.58134210e-01
-1.88227706e-02 -4.59924668e-01 2.64630139e-01 1.42500356e-01
8.03363919e-01 3.49823296e-01 2.98069954e-01 -5.19818187e-01
1.21015713e-01 1.08401040e-02 2.93060869e-01 8.87466967e-01
-1.25793561e-01 2.98439592e-01 4.74215090e-01 6.97514638e-02
-1.37308335e+00 -8.91441107e-01 2.63033152e-01 1.15479803e+00
-8.11314508e-02 -3.17318320e-01 -9.19085562e-01 -4.13062692e-01
-4.02356148e-01 1.32909334e+00 -3.60888630e-01 -4.67463642e-01
-5.16264200e-01 -4.35684115e-01 6.89004898e-01 4.32166100e-01
1.11846320e-01 -1.50364017e+00 -9.06906307e-01 5.37367463e-01
-3.93284917e-01 -7.50096858e-01 -5.31210005e-01 6.51884973e-02
-1.06633532e+00 -4.14983422e-01 -7.85930395e-01 -7.94937491e-01
6.64820313e-01 -2.05545574e-01 1.18601155e+00 1.46086141e-01
2.13576064e-01 3.53640690e-02 -4.53011423e-01 -3.45284551e-01
-1.20318794e+00 4.31301922e-01 6.11543544e-02 -5.11659622e-01
1.28965363e-01 -4.82347608e-01 -4.85707045e-01 -2.82659531e-02
-9.09783542e-01 5.48199356e-01 6.29462361e-01 9.72319007e-01
1.20318167e-01 -1.27610445e-01 9.99571204e-01 -5.25804818e-01
1.49527919e+00 -5.03722906e-01 -1.71592414e-01 1.49441734e-01
-9.31121469e-01 4.16741192e-01 8.38535368e-01 -7.03097641e-01
-9.97746348e-01 -1.78356126e-01 -1.24165587e-01 6.48302808e-02
-1.33079097e-01 5.28326154e-01 2.11043715e-01 4.35366511e-01
9.35878634e-01 4.25393790e-01 1.37787610e-01 -8.24657083e-02
5.79224646e-01 8.45487595e-01 5.98847158e-02 -7.05988586e-01
3.66717994e-01 -2.25158900e-01 -4.14617807e-01 -6.41305566e-01
-4.46848035e-01 2.55020469e-01 -3.58295798e-01 -2.72069067e-01
5.09944260e-01 -8.95300210e-01 -1.91966489e-01 3.86529565e-01
-1.16817427e+00 -7.74000108e-01 -2.32035130e-01 3.95575434e-01
-8.24342132e-01 1.65130660e-01 -6.68555200e-01 -9.86180961e-01
-5.27631283e-01 -1.21074665e+00 9.28384483e-01 1.61725163e-01
-7.28175461e-01 -8.51304293e-01 1.53746814e-01 1.45181781e-03
6.06596291e-01 -7.92908575e-03 1.06915474e+00 -5.55761933e-01
-3.44536692e-01 3.06950480e-01 1.98957384e-01 2.88607955e-01
1.34327918e-01 -4.46246006e-02 -7.15173364e-01 -2.08122596e-01
-6.55568019e-02 -5.70515692e-01 6.29279554e-01 3.24574262e-01
6.23767555e-01 -5.06865382e-01 3.37423384e-02 7.74469301e-02
1.06850684e+00 2.49537244e-01 5.62821031e-01 3.01779211e-01
3.66750479e-01 7.26008534e-01 5.26163816e-01 4.13588613e-01
4.00509268e-01 5.84131181e-01 1.64355803e-02 2.08103716e-01
-2.70983279e-01 -6.44088566e-01 6.40722275e-01 9.92674589e-01
8.25318098e-02 -6.79547906e-01 -8.02210450e-01 7.06522584e-01
-1.70720065e+00 -1.00850463e+00 1.33061454e-01 2.18803167e+00
1.14899588e+00 3.50303531e-01 2.58343190e-01 2.89204959e-02
6.07390225e-01 1.50560047e-02 -3.20193201e-01 -1.03353262e+00
2.00264901e-02 1.71703279e-01 7.60165751e-02 5.35036027e-01
-3.55583161e-01 1.27196646e+00 7.24509954e+00 7.89056182e-01
-1.15852618e+00 -2.43953373e-02 7.10204244e-01 -2.53424406e-01
-7.46532142e-01 -1.77881241e-01 -5.45488179e-01 6.90957665e-01
1.16495860e+00 -4.29860532e-01 5.88333964e-01 5.30149221e-01
3.30949992e-01 -9.89418253e-02 -1.20729041e+00 5.56925237e-01
3.15114148e-02 -1.02398872e+00 2.79751182e-01 1.75248981e-02
9.58506882e-01 -2.14698106e-01 1.08857125e-01 5.22983074e-01
3.05232137e-01 -1.35228038e+00 9.72379386e-01 3.98367882e-01
6.60753906e-01 -8.97601962e-01 4.19013470e-01 7.13206708e-01
-6.42664254e-01 -1.89338867e-02 -1.82216629e-01 -5.04134655e-01
1.90248549e-01 3.17015320e-01 -1.26434350e+00 1.05835078e-02
-3.56195420e-02 8.71949941e-02 -5.35742521e-01 5.63729942e-01
-4.70238090e-01 6.12203062e-01 -1.13241605e-01 -6.05540395e-01
2.62804449e-01 -2.94784270e-02 5.08401930e-01 1.21448350e+00
6.63928747e-01 -9.06965360e-02 1.44272521e-01 1.06890869e+00
-1.51462495e-01 3.66990238e-01 -8.51021588e-01 -4.11280543e-01
5.31759739e-01 6.65125310e-01 -6.78753436e-01 -4.91577744e-01
8.44363272e-02 1.26828504e+00 5.06465316e-01 1.51578024e-01
-5.35138309e-01 -1.05161861e-01 5.20856380e-01 5.08597121e-02
1.35514408e-01 -3.55067164e-01 -5.67708135e-01 -9.11926448e-01
-1.01814099e-01 -1.10606098e+00 -1.06691971e-01 -7.38376677e-01
-1.00066507e+00 6.43678427e-01 -9.60083678e-02 -9.74399507e-01
-9.43656147e-01 -8.73719975e-02 -5.39043784e-01 1.07390940e+00
-1.17688382e+00 -9.40333843e-01 1.24622494e-01 7.63687715e-02
9.67921972e-01 -2.61557817e-01 7.13344276e-01 -2.58736014e-01
-8.91928896e-02 7.24869311e-01 -6.55233487e-02 -2.08149478e-01
5.40122211e-01 -1.14471173e+00 6.35314107e-01 8.97451699e-01
2.37004943e-02 6.56630695e-01 1.11199999e+00 -9.27491963e-01
-1.09562457e+00 -8.51759315e-01 1.36445463e+00 -4.18747544e-01
3.80710036e-01 -1.95585147e-01 -6.89896047e-01 3.87688965e-01
3.79325956e-01 -8.03767145e-01 4.18298751e-01 1.30392417e-01
-7.07668886e-02 4.89556342e-01 -1.12895405e+00 1.05785990e+00
1.22756159e+00 -6.01573825e-01 -8.27786207e-01 1.30001515e-01
7.64154136e-01 -3.25305134e-01 -6.20867372e-01 1.58751696e-01
5.23183048e-01 -9.10743177e-01 4.77522522e-01 -4.32845086e-01
9.55283344e-01 -3.17240283e-02 1.44961372e-01 -1.90757155e+00
-4.16819274e-01 -5.58974087e-01 -7.14461282e-02 1.20459974e+00
5.10709643e-01 -3.80243033e-01 4.37457323e-01 4.21741992e-01
-1.93761975e-01 -5.92239916e-01 -7.75116324e-01 -7.73283958e-01
5.58857977e-01 -2.13637561e-01 5.69943964e-01 8.07530284e-01
2.93005347e-01 6.83969736e-01 -4.31544483e-01 -2.82496005e-01
2.45711386e-01 -4.75382665e-03 6.16578102e-01 -9.53918636e-01
-3.68186265e-01 -9.25274014e-01 -3.04019004e-02 -8.97631586e-01
2.44707406e-01 -1.03183937e+00 3.15360516e-01 -1.53325629e+00
1.24947429e-01 -3.46306801e-01 2.27181241e-02 1.45588949e-01
-3.97095978e-01 1.09716713e-01 2.36581653e-01 2.57998437e-01
-2.62243539e-01 5.77697933e-01 1.38960099e+00 6.81586564e-02
-6.65492356e-01 -2.33426943e-01 -9.65675712e-01 3.18549305e-01
1.09678054e+00 -4.35505390e-01 -6.29817843e-01 -4.99402493e-01
2.67675638e-01 8.08560029e-02 -8.91410187e-02 -7.48881817e-01
-1.55745046e-02 -2.78114140e-01 2.75004119e-01 -1.10541195e-01
7.44311213e-02 -1.73957944e-01 2.24392712e-01 6.52618110e-01
-7.66171753e-01 2.70209074e-01 1.00163758e-01 2.27935642e-01
2.84352619e-02 -4.07089442e-01 6.77516699e-01 -1.71806619e-01
-3.79028827e-01 -2.54373968e-01 -8.11419547e-01 1.65817514e-01
7.28814781e-01 -2.90128738e-01 -1.59141108e-01 -5.70501685e-01
-3.84813964e-01 6.41212538e-02 7.53958225e-01 7.40159631e-01
3.88471246e-01 -1.35372305e+00 -9.24841106e-01 1.16374575e-01
9.22092944e-02 -4.79265451e-01 -2.26806059e-01 4.18405890e-01
-3.22046846e-01 3.99684638e-01 -2.96392471e-01 -3.69357586e-01
-9.79414880e-01 3.07681918e-01 3.29521686e-01 -2.34944850e-01
-5.89412630e-01 6.30534053e-01 -1.90247178e-01 -1.82468057e-01
2.27033511e-01 -3.65954190e-01 -1.09190181e-01 1.37254804e-01
2.07723796e-01 2.90771276e-01 2.78471448e-02 -5.90485394e-01
5.16555011e-02 1.03447355e-01 -1.72206014e-01 -5.54353595e-01
7.34071553e-01 2.80970875e-02 2.70490468e-01 5.16177773e-01
8.13474000e-01 -1.01385139e-01 -1.17131138e+00 2.42748663e-01
8.03064033e-02 -3.32370728e-01 -3.35510401e-03 -1.12524152e+00
-3.72054875e-01 6.85997427e-01 4.25440550e-01 2.27983370e-01
8.62352550e-01 -1.97743297e-01 7.92272747e-01 2.06476152e-01
4.07776147e-01 -1.52036881e+00 3.08538496e-01 6.70268357e-01
9.09077525e-01 -1.12031054e+00 -3.19843680e-01 1.80298343e-01
-9.68474805e-01 9.11768734e-01 6.81503475e-01 1.01588197e-01
-1.40592540e-02 5.59018590e-02 1.48770347e-01 1.81431249e-01
-1.10778165e+00 -1.83240741e-01 9.71030146e-02 6.23159349e-01
9.64639187e-01 2.44804516e-01 -9.61956680e-01 1.73078433e-01
-6.78822100e-01 1.09196767e-01 6.49589002e-01 8.60593140e-01
-6.40388429e-01 -1.23282969e+00 -1.34711415e-01 3.92753601e-01
-3.59362483e-01 -2.68316209e-01 -5.87269664e-01 5.04567504e-01
7.01747462e-02 1.17179453e+00 2.22447723e-01 -2.35969543e-01
8.35062340e-02 3.53739500e-01 6.89008057e-01 -7.60566890e-01
-8.95255625e-01 1.06236868e-01 2.53562897e-01 -2.21733928e-01
-1.03397280e-01 -5.66946447e-01 -1.28762138e+00 -2.65918285e-01
-2.54433244e-01 1.00273140e-01 6.59003854e-01 8.72270882e-01
3.88997316e-01 5.10147035e-01 5.28037965e-01 -6.97245836e-01
-8.40221465e-01 -1.25828326e+00 -3.03863883e-01 5.01754582e-01
1.74928941e-02 -3.65021020e-01 -2.75399417e-01 -1.05072983e-01] | [11.720465660095215, 9.148380279541016] |
6b112117-e6da-4f0a-b015-30d45aadf00c | learning-multi-attention-convolutional-neural | null | null | http://openaccess.thecvf.com/content_iccv_2017/html/Zheng_Learning_Multi-Attention_Convolutional_ICCV_2017_paper.html | http://openaccess.thecvf.com/content_ICCV_2017/papers/Zheng_Learning_Multi-Attention_Convolutional_ICCV_2017_paper.pdf | Learning Multi-Attention Convolutional Neural Network for Fine-Grained Image Recognition | Recognizing fine-grained categories (e.g., bird species) highly relies on discriminative part localization and part-based fine-grained feature learning. Existing approaches predominantly solve these challenges independently, while neglecting the fact that part localization (e.g., head of a bird) and fine-grained feature learning (e.g., head shape) are mutually correlated. In this paper, we propose a novel part learning approach by a multi-attention convolutional neural network (MA-CNN), where part generation and feature learning can reinforce each other. MA-CNN consists of convolution, channel grouping and part classification sub-networks. The channel grouping network takes as input feature channels from convolutional layers, and generates multiple parts by clustering, weighting and pooling from spatially-correlated channels. The part classification network further classifies an image by each individual part, through which more discriminative fine-grained features can be learned. Two losses are proposed to guide the multi-task learning of channel grouping and part classification, which encourages MA-CNN to generate more discriminative parts from feature channels and learn better fine-grained features from parts in a mutual reinforced way. MA-CNN does not need bounding box/part annotation and can be trained end-to-end. We incorporate the learned parts from MA-CNN with part-CNN for recognition, and show the best performances on three challenging published fine-grained datasets, e.g., CUB-Birds, FGVC-Aircraft and Stanford-Cars.
| ['Jiebo Luo', 'Jianlong Fu', 'Heliang Zheng', 'Tao Mei'] | 2017-10-01 | null | null | null | iccv-2017-10 | ['fine-grained-image-recognition'] | ['computer-vision'] | [-5.11884764e-02 -2.29128689e-01 -1.92652941e-02 -7.43321002e-01
-5.98024368e-01 -6.32924020e-01 4.23785597e-01 -1.49398282e-01
-2.42229432e-01 6.02323651e-01 2.35015944e-01 3.75676006e-01
-8.37463364e-02 -9.27919984e-01 -1.15343976e+00 -6.60283685e-01
-6.03167526e-02 2.04274520e-01 2.60426044e-01 4.48828787e-02
-3.86247877e-03 7.25182414e-01 -1.77931404e+00 5.57310343e-01
6.04963124e-01 1.69809222e+00 1.21435225e-01 5.25078833e-01
-2.97449738e-01 5.57100058e-01 -5.46964467e-01 -2.34152108e-01
4.69895214e-01 -1.45559549e-01 -6.89861894e-01 4.87065881e-01
6.67412221e-01 -2.70872712e-01 -2.06629142e-01 8.81189585e-01
2.68096298e-01 1.29419431e-01 7.89493620e-01 -1.25479090e+00
-7.42201030e-01 6.48492157e-01 -5.24603605e-01 -2.02517230e-02
-1.54143482e-01 1.76538602e-01 1.11188447e+00 -8.79073083e-01
9.47859585e-02 1.51309359e+00 7.56567538e-01 5.36971867e-01
-1.27799892e+00 -9.63817835e-01 4.84425724e-01 -7.02230912e-03
-1.58604097e+00 -9.73657966e-02 7.19188690e-01 -4.95137036e-01
9.37920153e-01 4.98610772e-02 7.71847188e-01 6.22031331e-01
1.44498676e-01 8.70618939e-01 1.02942228e+00 -3.57715450e-02
2.33471449e-02 -1.27177045e-01 1.36105403e-01 9.73470867e-01
2.17039123e-01 3.42622310e-01 -3.20114821e-01 -2.48445775e-02
9.27008331e-01 8.14065874e-01 -1.49543270e-01 -5.17136812e-01
-1.14932108e+00 8.45805824e-01 1.13218820e+00 1.71424210e-01
-5.07291913e-01 3.72112304e-01 2.28071928e-01 2.02603459e-01
2.83814549e-01 2.56675273e-01 -8.59233975e-01 2.18764216e-01
-8.49693775e-01 1.08195700e-01 5.97212434e-01 1.05785716e+00
1.35985470e+00 -1.03912145e-01 -6.38005197e-01 1.10614455e+00
5.63960731e-01 4.84740525e-01 5.74642420e-01 -3.73756170e-01
1.48161143e-01 9.74846065e-01 -1.99752510e-01 -4.24759746e-01
-3.99222016e-01 -5.58826506e-01 -1.02263093e+00 2.56153673e-01
2.31548086e-01 -1.50513366e-01 -1.39721906e+00 1.68244684e+00
2.01351658e-01 8.64285082e-02 -3.02827030e-01 9.60794687e-01
1.06695294e+00 5.63636005e-01 2.54619628e-01 4.68974143e-01
1.50767326e+00 -1.42843950e+00 8.51290673e-02 -2.07744181e-01
2.10413262e-01 -7.02713072e-01 6.60751760e-01 -1.44180132e-03
-7.66205609e-01 -1.16439426e+00 -9.00138676e-01 8.94301534e-02
-6.64066494e-01 3.63002777e-01 6.35588527e-01 1.95156425e-01
-7.82467008e-01 6.63581192e-01 -5.73196113e-01 -6.73876405e-02
8.88351142e-01 4.96212631e-01 -6.27063513e-01 -1.38701856e-01
-7.74310470e-01 3.35021436e-01 4.53710616e-01 1.51286885e-01
-1.11830831e+00 -6.51344478e-01 -1.14720297e+00 3.77699852e-01
-4.10490446e-02 -7.90044367e-01 1.11151552e+00 -1.25018966e+00
-1.24356639e+00 8.71472955e-01 1.18152209e-01 -3.68104905e-01
-1.16225563e-01 -8.84985849e-02 -3.76396418e-01 -6.90251216e-03
2.42592037e-01 1.22879553e+00 1.45628965e+00 -1.09899986e+00
-9.66215253e-01 -2.73080021e-01 1.47572175e-01 -1.25463217e-01
-1.85191110e-02 -2.15617314e-01 -3.90273362e-01 -9.11023736e-01
-1.73903510e-01 -7.80891240e-01 -3.53046685e-01 2.09928572e-01
-3.94696176e-01 -3.66607070e-01 7.11939871e-01 -1.11598805e-01
9.83926415e-01 -2.27715135e+00 -2.96859127e-02 1.27518654e-01
3.83277684e-01 1.97294086e-01 -5.50785899e-01 9.55544859e-02
-2.07712024e-01 -4.54380214e-02 -1.63712665e-01 -6.22583032e-02
2.03745648e-01 1.61993697e-01 -9.22103822e-02 3.76067430e-01
7.10466743e-01 1.31493437e+00 -5.21021903e-01 -3.22256953e-01
3.57112080e-01 5.01205027e-01 -6.83344424e-01 5.34825981e-01
-1.11852594e-01 1.49523035e-01 -5.77248633e-01 9.90437329e-01
8.37272108e-01 -2.99878508e-01 -5.02293229e-01 -4.80862379e-01
-1.46751687e-01 -1.94301292e-01 -1.08294141e+00 1.52299356e+00
-7.23322272e-01 8.64511430e-02 1.83540270e-01 -8.38585138e-01
9.58317161e-01 -1.02342479e-02 -2.89502516e-02 -6.68808520e-01
2.47558340e-01 1.63511679e-01 2.06111282e-01 -9.65721160e-02
1.92770660e-01 -3.16203356e-01 -4.78257209e-01 2.80609846e-01
6.43255711e-01 -9.57024768e-02 1.29282340e-01 -2.74391294e-01
8.46232712e-01 7.71412477e-02 3.61113459e-01 -1.00811861e-01
4.44551498e-01 -7.65862986e-02 5.69533527e-01 6.31543398e-01
-2.78218925e-01 6.12946033e-01 -2.00767983e-02 -6.25897706e-01
-7.66012490e-01 -9.08589005e-01 -2.92505562e-01 1.58979094e+00
2.30481192e-01 -1.09677508e-01 -5.71163952e-01 -1.09091544e+00
5.29299021e-01 -1.36525119e-02 -1.15464747e+00 -3.44994634e-01
-3.18808556e-01 -3.86608392e-01 4.28358495e-01 9.95210707e-01
6.66754842e-01 -1.55480063e+00 -5.50338686e-01 4.40094978e-01
2.55448729e-01 -7.63061702e-01 -8.08509588e-01 8.09601247e-01
-7.01698780e-01 -1.03702915e+00 -9.85659540e-01 -9.45972919e-01
8.72915447e-01 4.67807531e-01 1.29571986e+00 1.04777917e-01
-6.22260273e-01 3.53485852e-01 -5.35527289e-01 -2.40422919e-01
2.91426063e-01 -9.73051488e-02 -2.98179537e-01 4.09151465e-01
3.67643476e-01 -4.48695064e-01 -7.79884517e-01 5.98921716e-01
-8.10362101e-01 -2.94749379e-01 1.45486772e+00 1.35064971e+00
1.05833638e+00 -1.30964480e-02 3.57400060e-01 -6.89018250e-01
1.69324040e-01 -3.89126807e-01 -4.94648397e-01 2.66038209e-01
-5.54336607e-02 1.67688563e-01 9.28142667e-01 -4.67408925e-01
-7.21139491e-01 3.81081015e-01 -3.43107074e-01 -7.54101396e-01
-8.30550849e-01 8.96000490e-02 -4.09071177e-01 -5.12460232e-01
3.79713237e-01 3.84657681e-01 -4.31995869e-01 -6.79501653e-01
5.36654115e-01 4.64819312e-01 4.44431722e-01 -4.66499388e-01
9.30852354e-01 2.36252978e-01 -1.58905804e-01 -6.06187880e-01
-1.01973546e+00 -8.70173037e-01 -8.83084536e-01 1.42740935e-01
9.84377801e-01 -1.28530097e+00 -5.94104052e-01 6.59865499e-01
-8.46914649e-01 -4.43442643e-01 -5.73368192e-01 3.30743760e-01
-3.81994307e-01 -9.95583758e-02 -4.19524252e-01 -2.24446684e-01
-4.42795932e-01 -8.94074678e-01 1.62761319e+00 6.78224027e-01
1.52136296e-01 -7.07179964e-01 -1.09310091e-01 8.19553155e-03
4.83715177e-01 8.06477740e-02 5.04231930e-01 -6.13153517e-01
-7.11667418e-01 -4.24774468e-01 -6.46680057e-01 4.18581486e-01
1.17179818e-01 -1.65896028e-01 -1.04562008e+00 -3.71176481e-01
-6.77739680e-01 -6.73494339e-01 1.32837296e+00 4.00955498e-01
1.64823222e+00 -4.14075643e-01 -3.86672318e-01 9.74112689e-01
1.43441653e+00 1.60652399e-02 2.76725858e-01 -1.01123504e-01
8.59520733e-01 3.11764210e-01 6.01093292e-01 6.33985043e-01
2.48709977e-01 4.61935312e-01 5.21489501e-01 -2.97815204e-01
-5.03693998e-01 -3.29290986e-01 9.24453735e-02 3.77943367e-01
9.80620533e-02 4.74334098e-02 -2.39520222e-01 5.81487417e-01
-1.65048134e+00 -1.05139637e+00 2.29568183e-01 1.74691772e+00
5.98401845e-01 3.98511291e-02 3.46014082e-01 -2.13980958e-01
6.66244149e-01 1.77619740e-01 -6.11732543e-01 -2.48917997e-01
-8.80316272e-02 5.79781115e-01 4.31466341e-01 3.75153460e-02
-1.62209690e+00 1.10140729e+00 4.87281132e+00 9.82128680e-01
-1.29502773e+00 1.73385628e-02 6.77920699e-01 1.69545248e-01
7.79887438e-02 -3.30541015e-01 -1.04790926e+00 3.26737285e-01
3.59729141e-01 4.27440226e-01 4.27941054e-01 1.28343189e+00
-4.59520936e-01 1.44189581e-01 -1.04454803e+00 1.06186616e+00
9.08701308e-03 -1.29743803e+00 1.02233760e-01 -3.51513848e-02
8.10375154e-01 1.71636820e-01 -2.03498662e-01 7.39617825e-01
3.35719883e-01 -1.19758332e+00 1.04467213e+00 6.24284983e-01
9.33409631e-01 -7.53646731e-01 8.34653497e-01 2.12843999e-01
-1.85935104e+00 -3.45655411e-01 -7.36696005e-01 6.29544398e-03
-2.13456094e-01 5.87163806e-01 -3.70748281e-01 3.04238468e-01
9.08314526e-01 7.28621483e-01 -7.38205433e-01 1.42143428e+00
-2.48583347e-01 5.87190807e-01 -1.80850610e-01 -1.26978010e-01
4.70988244e-01 1.90157399e-01 -9.95922834e-02 1.53148508e+00
7.27866143e-02 4.78668697e-02 6.14618659e-01 1.00465989e+00
-4.43627596e-01 -2.22140014e-01 -2.23943219e-01 -1.40528962e-01
1.88226342e-01 1.93669879e+00 -6.60366476e-01 -4.48984295e-01
-6.45070374e-01 1.09723639e+00 5.69782853e-01 1.61854908e-01
-7.06976473e-01 -7.81398535e-01 9.73080397e-01 -5.63085079e-02
1.18932688e+00 1.63270712e-01 -3.92937437e-02 -1.05733192e+00
-1.65919438e-01 -5.15992105e-01 4.16568518e-01 -3.84135783e-01
-1.78710914e+00 9.14919972e-01 -3.16375583e-01 -1.33537996e+00
-2.53051929e-02 -7.88652658e-01 -8.76493096e-01 9.75938380e-01
-1.71704316e+00 -1.67525911e+00 -6.26706719e-01 9.47158694e-01
5.54855406e-01 -7.90562853e-02 8.51427972e-01 2.99439937e-01
-4.53173965e-01 8.43237579e-01 -3.38656783e-01 4.68913943e-01
5.46342075e-01 -1.44958830e+00 3.21654022e-01 5.59147298e-01
1.52766496e-01 4.87252533e-01 -7.68098608e-02 -4.34248269e-01
-1.12625957e+00 -1.86462379e+00 6.13269508e-01 -5.74547872e-02
4.68812227e-01 -5.53494513e-01 -5.60300350e-01 3.69225085e-01
-1.76239073e-01 9.14616227e-01 7.44165957e-01 1.07703462e-01
-7.86443710e-01 -3.45575333e-01 -1.19865203e+00 -1.38839126e-01
9.73326027e-01 -5.97452223e-01 -6.01728261e-01 2.32796613e-02
6.20886087e-01 -7.39112347e-02 -8.53345811e-01 4.44736153e-01
6.38042271e-01 -8.65503430e-01 9.30116355e-01 -6.62574232e-01
2.81904995e-01 -6.55325472e-01 -2.84144521e-01 -1.54803562e+00
-1.13420439e+00 -6.03180565e-02 -7.00365379e-02 1.14314389e+00
3.37413728e-01 -3.64096493e-01 6.16732121e-01 -6.94570411e-03
-5.77023864e-01 -9.33602214e-01 -6.65104508e-01 -7.09152400e-01
1.74436476e-02 -2.28644878e-01 1.00166965e+00 3.94192308e-01
-6.00136459e-01 3.87248963e-01 -1.77146360e-01 1.54290736e-01
5.53147972e-01 1.01299071e+00 7.79841483e-01 -1.41930819e+00
-4.72731888e-01 -7.07226217e-01 -6.18880153e-01 -1.23118258e+00
1.10677682e-01 -8.87924969e-01 5.15696824e-01 -1.42848241e+00
4.22023207e-01 -5.58054686e-01 -4.90878940e-01 1.04630589e+00
-2.28119999e-01 7.88395822e-01 1.89033955e-01 -2.81258319e-02
-1.00207305e+00 7.32013941e-01 1.50531042e+00 -4.72229302e-01
6.43571690e-02 1.94139391e-01 -9.07241940e-01 5.88471830e-01
5.20731568e-01 -3.38093936e-01 1.44809172e-01 4.92070941e-03
-5.07064342e-01 -3.52226406e-01 7.25064814e-01 -1.21776640e+00
-3.43683288e-02 8.98426101e-02 1.19418716e+00 -6.04046226e-01
2.35758066e-01 -1.00201404e+00 -9.14613530e-02 4.61220980e-01
1.57832894e-02 -4.22175676e-01 3.52104276e-01 4.79598910e-01
-5.17113268e-01 4.22444381e-02 1.05303574e+00 -2.90978789e-01
-1.07310212e+00 9.62173939e-01 -4.29435857e-02 1.17482906e-02
9.19292271e-01 -3.72683495e-01 -1.18199468e-01 1.87705517e-01
-8.54224682e-01 2.14474469e-01 1.98217779e-01 5.36352456e-01
6.02429986e-01 -1.58132696e+00 -6.27142012e-01 6.41038597e-01
5.60935438e-01 6.09971732e-02 6.55245006e-01 5.72786093e-01
-6.64949492e-02 5.87095261e-01 -6.36688709e-01 -5.38348198e-01
-1.02393997e+00 7.42222905e-01 3.68228823e-01 -2.75385797e-01
-3.15149635e-01 1.42630661e+00 8.45836580e-01 -6.61992967e-01
2.50950396e-01 -6.68660104e-01 -1.50777653e-01 -1.39222741e-02
5.91597795e-01 -3.29106152e-02 -4.25840914e-03 -8.01807463e-01
-4.45985496e-01 9.51799572e-01 -1.97695553e-01 6.16045713e-01
1.44312203e+00 1.14137702e-01 -1.67879171e-03 -4.82975096e-02
1.44412136e+00 -2.29836125e-02 -1.70761204e+00 -3.46994042e-01
-5.26500046e-01 -3.53919894e-01 -7.75902346e-02 -1.08814096e+00
-1.33679724e+00 1.02074194e+00 6.40087724e-01 1.37820855e-01
1.26090908e+00 4.18216646e-01 8.95190895e-01 1.99728101e-01
2.95086414e-01 -7.51874864e-01 1.43103138e-01 8.26745510e-01
1.10178161e+00 -1.27422154e+00 -3.71020228e-01 -2.27891564e-01
-4.62122202e-01 1.13915575e+00 9.51972127e-01 -4.33035254e-01
1.12599778e+00 1.92891538e-01 1.18331739e-03 -2.27383256e-01
-6.54656112e-01 -7.76422441e-01 9.23570096e-01 5.73478520e-01
3.94964427e-01 5.55201411e-01 2.71675825e-01 1.28698158e+00
-4.58326042e-02 -1.89035028e-01 -3.78242254e-01 7.49993265e-01
-7.13960707e-01 -8.91931057e-01 -4.26598281e-01 8.92690241e-01
-2.71049052e-01 -1.23224542e-01 -4.95500565e-01 5.58715343e-01
1.01984715e+00 5.94803691e-01 3.90567124e-01 -6.18342102e-01
3.82834703e-01 -1.73361808e-01 5.34513891e-01 -8.34772825e-01
-9.28285897e-01 1.70132220e-01 -2.46503592e-01 -8.47846508e-01
-2.88273513e-01 -3.02622944e-01 -1.07795632e+00 -1.46528408e-02
-3.98513556e-01 1.73459962e-01 2.92100042e-01 8.51867974e-01
4.53198493e-01 7.79048145e-01 8.28664780e-01 -1.36924660e+00
-3.14054459e-01 -1.03386998e+00 -9.38213944e-01 4.02075917e-01
4.49403763e-01 -8.33813488e-01 -2.00305134e-01 -1.15858100e-01] | [9.582928657531738, 1.986059308052063] |
94fd0e8c-1ed7-426e-b109-511d4a62bf16 | intrinsic-decomposition-of-document-images-in | 2011.14447 | null | https://arxiv.org/abs/2011.14447v1 | https://arxiv.org/pdf/2011.14447v1.pdf | Intrinsic Decomposition of Document Images In-the-Wild | Automatic document content processing is affected by artifacts caused by the shape of the paper, non-uniform and diverse color of lighting conditions. Fully-supervised methods on real data are impossible due to the large amount of data needed. Hence, the current state of the art deep learning models are trained on fully or partially synthetic images. However, document shadow or shading removal results still suffer because: (a) prior methods rely on uniformity of local color statistics, which limit their application on real-scenarios with complex document shapes and textures and; (b) synthetic or hybrid datasets with non-realistic, simulated lighting conditions are used to train the models. In this paper we tackle these problems with our two main contributions. First, a physically constrained learning-based method that directly estimates document reflectance based on intrinsic image formation which generalizes to challenging illumination conditions. Second, a new dataset that clearly improves previous synthetic ones, by adding a large range of realistic shading and diverse multi-illuminant conditions, uniquely customized to deal with documents in-the-wild. The proposed architecture works in a self-supervised manner where only the synthetic texture is used as a weak training signal (obviating the need for very costly ground truth with disentangled versions of shading and reflectance). The proposed approach leads to a significant generalization of document reflectance estimation in real scenes with challenging illumination. We extensively evaluate on the real benchmark datasets available for intrinsic image decomposition and document shadow removal tasks. Our reflectance estimation scheme, when used as a pre-processing step of an OCR pipeline, shows a 26% improvement of character error rate (CER), thus, proving the practical applicability. | ['Dimitris Samaras', 'Maria Vanrell', 'Ramon Baldrich', 'Ke Ma', 'Hassan Ahmed Sial', 'Sagnik Das'] | 2020-11-29 | null | null | null | null | ['shadow-removal', 'intrinsic-image-decomposition'] | ['computer-vision', 'computer-vision'] | [ 7.33546674e-01 -2.96572804e-01 5.89963496e-01 -1.62815839e-01
-5.78626633e-01 -7.03598976e-01 8.97962987e-01 -6.90247267e-02
-3.42999756e-01 8.61926258e-01 -1.09515302e-01 4.36181528e-03
-6.22159429e-02 -6.72874391e-01 -7.13434279e-01 -9.65296328e-01
3.04719836e-01 6.29849434e-01 3.74006368e-02 -2.37236112e-01
1.03664838e-01 7.68559933e-01 -2.11097336e+00 4.51206982e-01
9.59980190e-01 7.60402620e-01 4.84463632e-01 8.24776590e-01
-2.15638787e-01 3.60491782e-01 -7.84871876e-01 -1.54492691e-01
3.81309122e-01 -3.25121313e-01 -1.47027686e-01 3.60129654e-01
6.53816760e-01 -4.53046948e-01 2.14624122e-01 7.42416978e-01
4.61723655e-01 -2.03947145e-02 6.12878025e-01 -7.00191855e-01
-4.83287841e-01 1.10829927e-01 -5.43267608e-01 -5.14052451e-01
3.74521345e-01 3.08681518e-01 6.18193269e-01 -7.50369370e-01
5.95721662e-01 1.00869119e+00 6.82133973e-01 4.96258259e-01
-1.58845258e+00 -1.12805225e-01 -2.70790681e-02 2.53059044e-02
-1.12025630e+00 -3.91861469e-01 9.43192065e-01 -3.71055543e-01
6.47974491e-01 5.64688861e-01 5.41549563e-01 1.69256532e+00
-2.70988941e-01 4.97555584e-01 1.92932093e+00 -7.54092753e-01
2.88628638e-01 5.03617048e-01 -3.21259946e-02 4.55707073e-01
4.75676894e-01 2.41259970e-02 -4.07106608e-01 8.14958364e-02
3.71654123e-01 -2.48923659e-01 -6.27290130e-01 -4.79042113e-01
-1.05299151e+00 2.41618395e-01 2.39004642e-02 4.26907629e-01
-2.05183476e-01 -6.93043228e-03 1.69712216e-01 1.61623567e-01
4.06812459e-01 2.40429357e-01 -4.76347119e-01 4.84048203e-02
-1.12431419e+00 1.63556457e-01 9.21042085e-01 4.87557322e-01
7.40813494e-01 1.49975911e-01 -3.70052867e-02 1.00814831e+00
2.95759916e-01 9.80371296e-01 3.13484967e-01 -4.67398196e-01
2.34708056e-01 4.67664510e-01 3.89382094e-01 -1.03462040e+00
-4.33004290e-01 -5.57860553e-01 -7.84662247e-01 6.74388468e-01
6.83415055e-01 1.11305967e-01 -9.48001027e-01 1.50156915e+00
3.10694635e-01 -1.62677065e-01 1.87973306e-01 1.09244823e+00
4.94815379e-01 4.49495733e-01 -5.23140609e-01 -2.34528750e-01
1.30917084e+00 -6.84897304e-01 -6.64217234e-01 -8.15385282e-02
2.75897771e-01 -1.16749930e+00 1.62695944e+00 1.15000367e+00
-8.05196762e-01 -4.61962253e-01 -1.15213490e+00 -1.06128398e-02
-6.32812858e-01 6.14272654e-01 4.22658473e-01 1.20397580e+00
-9.87023473e-01 5.45615435e-01 -5.61142921e-01 -3.59489083e-01
2.15692282e-01 1.41002581e-01 -3.25706929e-01 -4.04414773e-01
-8.05572987e-01 7.76383638e-01 9.59303826e-02 4.61966127e-01
-7.60136187e-01 -5.21687567e-01 -4.21787143e-01 2.68246569e-02
2.96320647e-01 -3.08870405e-01 5.14424622e-01 -1.31529343e+00
-1.98443162e+00 9.58703399e-01 1.31576672e-01 -1.08645767e-01
1.14547884e+00 -4.34324890e-01 -2.85282403e-01 4.69644442e-02
-5.09744704e-01 1.28795579e-01 1.14287531e+00 -2.10861897e+00
3.35340291e-01 -3.51121157e-01 -9.07212272e-02 -2.07170695e-02
-4.24840838e-01 -2.55901963e-01 -4.42716658e-01 -5.22405386e-01
-1.96020976e-01 -8.90249729e-01 1.74358577e-01 1.29945695e-01
-5.93146563e-01 7.42053866e-01 8.50991249e-01 -7.81222522e-01
4.46331084e-01 -1.89074504e+00 1.29510403e-01 1.34845302e-01
-3.85707200e-01 5.34624755e-01 -2.94832528e-01 5.63219666e-01
-7.53293037e-02 -2.69287407e-01 -5.23986340e-01 -8.09715807e-01
8.36500898e-02 1.52043596e-01 -3.55497628e-01 5.11715531e-01
1.01732336e-01 3.54934186e-01 -5.01328051e-01 -3.61319512e-01
5.79025805e-01 1.05337667e+00 -2.04717785e-01 1.09016024e-01
-4.65772867e-01 5.80435753e-01 -2.17053983e-02 3.22169870e-01
1.15904093e+00 2.46058047e-01 2.77488112e-01 -3.63192141e-01
-1.85842723e-01 -2.02655405e-01 -1.46130061e+00 1.93833268e+00
-9.66414928e-01 6.90743148e-01 4.19630766e-01 -8.48079085e-01
1.27721500e+00 6.59507960e-02 2.51663595e-01 -7.75272846e-01
1.68572187e-01 3.76877487e-01 -3.81302685e-01 -5.01395464e-01
3.46489370e-01 -4.14007017e-03 5.75460672e-01 4.83867288e-01
-2.61701792e-01 -3.45018983e-01 1.72183052e-01 -9.71250609e-02
8.51286709e-01 8.06763053e-01 -4.17704880e-01 -3.93354297e-01
9.10335481e-01 -2.90165305e-01 1.05799682e-01 6.50795698e-01
4.53203231e-01 1.11566663e+00 4.94047314e-01 -3.74724865e-01
-1.05848205e+00 -8.36691380e-01 -1.62219912e-01 6.09269559e-01
-8.03631991e-02 5.49816713e-02 -1.05488634e+00 -3.91177565e-01
-1.55960754e-01 8.26463521e-01 -6.28033161e-01 1.51606590e-01
-5.45924664e-01 -1.15240788e+00 4.41207975e-01 -7.04439133e-02
4.81787622e-01 -9.20468152e-01 -7.65508831e-01 -1.12680249e-01
3.93221378e-02 -1.37695372e+00 2.57603168e-01 1.84480742e-01
-6.42074764e-01 -1.16575229e+00 -8.39387178e-01 -1.00693256e-01
5.86047411e-01 2.63067544e-01 1.09904993e+00 1.45716414e-01
-8.06080878e-01 5.13347507e-01 -3.65363240e-01 -3.57289940e-01
-5.20755827e-01 -3.51892769e-01 -3.17050636e-01 5.77891231e-01
-2.60234084e-02 -4.82248217e-01 -6.63434029e-01 3.41691822e-01
-1.25954187e+00 2.57294297e-01 5.71958244e-01 7.79152453e-01
3.38955849e-01 -1.76861763e-01 1.58982009e-01 -1.09106827e+00
4.37022716e-01 1.79582372e-01 -9.79214013e-01 4.19636816e-01
-5.21994472e-01 7.63352737e-02 9.52016652e-01 -3.61261219e-01
-1.59285855e+00 1.01911187e-01 1.27245441e-01 5.13779595e-02
-6.32528543e-01 -1.08560942e-01 -4.41215187e-01 -4.90781143e-02
9.24387157e-01 2.23775685e-01 -6.27379641e-02 -6.64060831e-01
3.47571254e-01 4.79635894e-01 2.41819531e-01 -7.39691198e-01
1.04863262e+00 9.93697464e-01 3.99804026e-01 -1.19774795e+00
-5.03756166e-01 -5.18281758e-02 -7.77926266e-01 -3.04080337e-01
7.61805654e-01 -5.50227821e-01 -6.71631396e-01 7.75968790e-01
-1.19255829e+00 -6.59002066e-01 -1.53304979e-01 4.26318467e-01
-3.91045213e-01 8.84990335e-01 -3.28271657e-01 -1.11677659e+00
-3.31544787e-01 -1.15699303e+00 1.21666384e+00 1.16816871e-02
3.13223183e-01 -8.14991891e-01 1.89069659e-01 6.37412012e-01
4.84450191e-01 5.50482750e-01 7.89839864e-01 3.22212726e-02
-6.17914438e-01 -1.06305048e-01 -5.27871728e-01 7.56052434e-01
9.23575982e-02 3.01676214e-01 -1.64773631e+00 -2.00791329e-01
3.21903941e-03 -3.80336404e-01 1.00842917e+00 1.55436657e-02
1.01693499e+00 1.66549847e-01 2.71502495e-01 5.45110166e-01
1.91646361e+00 -4.35875595e-01 7.50127971e-01 3.46216768e-01
7.70167947e-01 1.03262961e+00 4.10986036e-01 5.54348111e-01
-1.50272325e-01 8.92051816e-01 7.34342158e-01 -5.30514777e-01
-6.50508523e-01 3.28740627e-01 3.70297134e-01 3.35879892e-01
-3.66018295e-01 -6.26064897e-01 -8.54097128e-01 2.24905998e-01
-1.59082258e+00 -6.81735098e-01 -7.68112481e-01 2.58645916e+00
6.54978156e-01 1.89020336e-02 -3.16310316e-01 5.15253246e-01
4.35225278e-01 1.43272787e-01 -2.11188138e-01 -4.05745894e-01
-6.86096013e-01 4.09854859e-01 4.94120568e-01 5.92111588e-01
-6.48676217e-01 6.26645565e-01 4.50081968e+00 5.44101179e-01
-1.27423429e+00 8.80148038e-02 2.82994509e-01 -9.57084596e-02
-5.88808835e-01 -9.38402265e-02 -3.88869047e-01 2.94463038e-01
6.48557842e-01 9.27006602e-01 6.76939368e-01 3.67301524e-01
4.79831219e-01 -3.72428119e-01 -8.68456483e-01 1.07161272e+00
5.99392354e-01 -7.40650356e-01 5.89082623e-03 5.17061464e-02
7.10556507e-01 -9.14578736e-02 2.97686011e-01 -2.83381224e-01
-8.02254304e-02 -9.51286018e-01 5.77193260e-01 8.10781300e-01
7.93981671e-01 -4.63528126e-01 4.73929107e-01 2.14508981e-01
-4.33842421e-01 9.17015150e-02 -3.76536369e-01 2.35457882e-01
4.47746888e-02 1.14775872e+00 -6.93712890e-01 6.59142315e-01
4.84839797e-01 2.20915452e-01 -6.74656451e-01 6.64611518e-01
-3.08663100e-01 4.93274957e-01 -5.31016707e-01 7.14764073e-02
-2.16817483e-03 -5.30260742e-01 4.61414635e-01 1.63275445e+00
1.81305751e-01 -4.23498541e-01 -3.54753077e-01 1.13313854e+00
2.92655885e-01 2.11651817e-01 -3.94713998e-01 1.64740428e-01
-3.08589220e-01 1.60719109e+00 -9.72091496e-01 2.17943430e-01
-3.65681469e-01 1.25077188e+00 -2.44857091e-02 6.55647337e-01
-7.06261694e-01 -2.35652164e-01 2.32297674e-01 1.66584656e-01
1.96197227e-01 -4.33535129e-01 -4.27610874e-01 -1.31210828e+00
2.85233617e-01 -9.48800027e-01 -3.72166127e-01 -1.11749625e+00
-9.00746226e-01 4.77668524e-01 -3.29889715e-01 -8.93426776e-01
1.94645733e-01 -1.20932639e+00 -4.53445911e-01 1.04827237e+00
-1.94630754e+00 -1.47540236e+00 -8.54662716e-01 6.83987498e-01
4.45309639e-01 -5.96235953e-02 1.15414119e+00 3.63688022e-01
-4.74007994e-01 2.28970408e-01 6.19931400e-01 -3.38188320e-01
8.56741786e-01 -1.30563319e+00 -4.34837304e-02 9.54052925e-01
2.85162896e-01 2.83881456e-01 9.42448735e-01 -1.65609539e-01
-1.77745438e+00 -5.08443296e-01 3.72978240e-01 -3.32439125e-01
3.21506351e-01 -9.60635841e-01 -9.25238788e-01 -4.64767627e-02
1.36604816e-01 -4.75000665e-02 5.33907413e-01 -5.55834640e-03
-5.73203623e-01 -4.12497967e-01 -1.17199349e+00 5.16091347e-01
6.72804832e-01 -4.29208100e-01 -3.02028023e-02 7.21790254e-01
5.76363876e-02 -9.50489342e-02 -4.45954949e-01 1.23112611e-01
7.37010598e-01 -1.65555453e+00 9.78209138e-01 7.53134638e-02
4.99220282e-01 -4.33312237e-01 -2.28291690e-01 -9.79147077e-01
4.42826986e-01 -6.00923240e-01 1.62284866e-01 1.40155780e+00
3.05937231e-01 -6.19896472e-01 6.98585093e-01 4.12606895e-01
1.66181866e-02 -3.17839324e-01 -6.28861427e-01 -5.64883292e-01
-3.23958427e-01 -4.37866688e-01 4.39745903e-01 7.58514225e-01
-7.88244069e-01 -8.44859928e-02 -4.00793165e-01 1.03462592e-01
8.25504899e-01 2.96238691e-01 1.15474868e+00 -1.41150904e+00
-8.04512978e-01 -2.46777162e-01 -9.61127505e-02 -5.96338868e-01
1.46986395e-01 -4.18312252e-01 1.88903399e-02 -1.38094246e+00
8.01989138e-02 -5.52568853e-01 2.59466209e-02 1.76267266e-01
1.58849284e-01 5.08921623e-01 2.03527093e-01 7.73706064e-02
-1.69298828e-01 4.62282687e-01 9.64289904e-01 -1.31639093e-01
-2.42916252e-02 -1.40409380e-01 -1.45821840e-01 6.98699117e-01
5.97798109e-01 -1.53886542e-01 -5.33975661e-01 -6.92501009e-01
5.89601994e-01 -4.92581725e-01 6.28365874e-01 -9.73992705e-01
-3.56754869e-01 1.30836323e-01 6.17748737e-01 -4.07219440e-01
6.10230625e-01 -1.17288315e+00 1.46978617e-01 3.79841119e-01
-6.78067058e-02 -4.55087960e-01 1.79298013e-01 5.97436190e-01
1.71899557e-01 -5.23890257e-01 8.18376362e-01 -8.61473978e-02
-1.83372736e-01 -3.85011673e-01 -1.31706409e-02 -3.01978499e-01
6.98072970e-01 -4.55125064e-01 -3.95314902e-01 -2.56924808e-01
-4.03904855e-01 -5.98634839e-01 8.39055955e-01 1.29185811e-01
2.54456669e-01 -6.11885786e-01 -7.28453040e-01 1.92213833e-01
-3.69000025e-02 -5.88316051e-03 4.06112850e-01 7.71249473e-01
-9.16369379e-01 1.07182771e-01 -1.68746203e-01 -6.53512776e-01
-1.39161420e+00 5.50082207e-01 3.05788308e-01 -2.29160890e-01
-5.69845140e-01 4.16751146e-01 1.30485699e-01 -4.47717428e-01
2.97273714e-02 -2.37678766e-01 -6.52232841e-02 4.35688421e-02
2.68705577e-01 4.05678868e-01 6.89249933e-01 -5.38671851e-01
-7.70835429e-02 8.71903181e-01 2.52777398e-01 -2.21628547e-01
1.53247476e+00 2.53402162e-02 -1.39547259e-01 4.39570308e-01
1.04563463e+00 4.36383426e-01 -1.38708794e+00 6.85705105e-03
-2.94236302e-01 -6.66522861e-01 1.53002888e-01 -1.22509742e+00
-1.01128614e+00 1.15945327e+00 8.01190019e-01 2.44160276e-02
1.30749369e+00 -7.40976036e-01 3.44466418e-01 4.28273857e-01
2.38076136e-01 -1.20977175e+00 1.35222331e-01 9.32527632e-02
1.08865738e+00 -1.15718400e+00 2.77312368e-01 -3.55175495e-01
-4.81715262e-01 1.56071925e+00 1.28201261e-01 3.98239531e-02
2.18942076e-01 4.45225537e-01 3.14528942e-01 4.76470916e-03
-2.08178625e-01 -2.32319221e-01 1.43355131e-01 8.61728668e-01
5.54998398e-01 -2.52810627e-01 -3.75946254e-01 1.00250952e-01
7.39629492e-02 -2.00980037e-01 6.31047726e-01 5.92491806e-01
2.27548629e-02 -1.30031240e+00 -7.56967425e-01 -3.82783525e-02
-2.59490371e-01 9.77566466e-03 -5.85651278e-01 7.80830145e-01
2.04797417e-01 8.59494507e-01 -3.20353180e-01 2.19627097e-01
2.92101145e-01 5.71101792e-02 8.05198908e-01 -3.45562011e-01
-5.99527299e-01 1.80145219e-01 3.86534669e-02 -5.69877982e-01
-6.37403369e-01 -7.44442165e-01 -8.76295447e-01 1.16722710e-01
-5.08125424e-01 -4.71615613e-01 1.48175156e+00 7.94096172e-01
1.18919395e-01 5.48823774e-01 4.34492081e-01 -1.05388784e+00
-4.56268579e-01 -8.60707998e-01 -5.68121314e-01 7.50316262e-01
3.28450382e-01 -5.72234929e-01 -4.66119617e-01 1.77612826e-01] | [10.062957763671875, -2.809771776199341] |
e4d6be64-d349-4728-b0b3-6d4037dc3dab | offline-policy-evaluation-for-reinforcement | 2306.14063 | null | https://arxiv.org/abs/2306.14063v1 | https://arxiv.org/pdf/2306.14063v1.pdf | Offline Policy Evaluation for Reinforcement Learning with Adaptively Collected Data | Developing theoretical guarantees on the sample complexity of offline RL methods is an important step towards making data-hungry RL algorithms practically viable. Currently, most results hinge on unrealistic assumptions about the data distribution -- namely that it comprises a set of i.i.d. trajectories collected by a single logging policy. We consider a more general setting where the dataset may have been gathered adaptively. We develop theory for the TMIS Offline Policy Evaluation (OPE) estimator in this generalized setting for tabular MDPs, deriving high-probability, instance-dependent bounds on its estimation error. We also recover minimax-optimal offline learning in the adaptive setting. Finally, we conduct simulations to empirically analyze the behavior of these estimators under adaptive and non-adaptive regimes. | ['Yu-Xiang Wang', 'Ming Yin', 'Dan Xiao', 'Sunil Madhow'] | 2023-06-24 | null | null | null | null | ['offline-rl'] | ['playing-games'] | [-2.02242389e-01 1.18155353e-01 -9.44169700e-01 -1.85049295e-01
-1.31586933e+00 -1.01025164e+00 3.37062091e-01 1.75838277e-01
-8.44048977e-01 1.13807416e+00 8.52208287e-02 -7.42043853e-01
-3.31575572e-01 -4.32471842e-01 -9.15106714e-01 -7.74173796e-01
-6.02332711e-01 8.98604453e-01 -9.62321162e-02 3.34284753e-01
1.32619679e-01 5.61700881e-01 -1.03594160e+00 -4.51627433e-01
6.46813333e-01 1.09548783e+00 5.15145669e-03 9.12680805e-01
2.45573446e-01 6.29411042e-01 -6.32076323e-01 -2.71477662e-02
5.37190259e-01 -4.05851930e-01 -4.57860351e-01 7.01192021e-02
1.51454052e-02 -8.85021210e-01 -3.59884113e-01 1.01073432e+00
5.14723420e-01 3.18202227e-01 6.76214039e-01 -1.26163864e+00
9.60125700e-02 8.85684788e-01 -5.09060919e-01 3.08521181e-01
1.40290618e-01 2.75197595e-01 9.27195787e-01 -7.94749781e-02
5.46775818e-01 1.25609815e+00 2.79042959e-01 5.28370380e-01
-1.32014132e+00 -6.22486830e-01 4.30474728e-01 2.24062055e-02
-8.47948253e-01 -5.81916928e-01 2.43041322e-01 -4.52141874e-02
4.67931986e-01 4.61844243e-02 5.10588050e-01 1.10006154e+00
1.46674782e-01 1.13725138e+00 1.14870214e+00 -8.01568031e-02
7.11023688e-01 3.19855213e-02 -1.03026126e-02 4.26830351e-01
8.02261233e-01 2.23914653e-01 -3.63350928e-01 -5.96642673e-01
6.83493078e-01 -5.87421469e-02 -1.32403150e-01 -4.58039045e-01
-8.26423347e-01 7.82128394e-01 -2.14283347e-01 -2.46129140e-01
-5.23765504e-01 5.33583403e-01 4.14715081e-01 6.69607639e-01
4.31369871e-01 1.26253724e-01 -6.53127193e-01 -7.09022045e-01
-7.37449884e-01 6.78016782e-01 1.11021841e+00 1.10959888e+00
3.00237268e-01 5.34766726e-02 -2.67293662e-01 2.17102453e-01
-2.19152570e-02 9.73940492e-01 1.07803471e-01 -1.41031420e+00
8.82962584e-01 -2.84681410e-01 1.19702280e+00 -3.68659586e-01
-4.34742063e-01 -3.98088306e-01 -3.19988281e-01 -1.43686831e-01
9.08858240e-01 -7.46463358e-01 -3.60931933e-01 1.96893919e+00
4.71402973e-01 -1.16424197e-02 5.60320355e-02 8.13465893e-01
-4.72913742e-01 5.89283884e-01 -8.96494538e-02 -8.81650329e-01
6.06430054e-01 -3.51449609e-01 -7.23038971e-01 4.65626419e-02
6.94279194e-01 -1.09416775e-01 1.12801218e+00 5.71839571e-01
-1.34195316e+00 2.29714245e-01 -6.34401083e-01 3.48736554e-01
1.39712006e-01 -1.79917715e-03 4.84039873e-01 5.99228084e-01
-7.75870681e-01 5.77609003e-01 -9.28566158e-01 -2.75594473e-01
4.77794945e-01 3.25865716e-01 3.18353057e-01 7.29773790e-02
-5.39831340e-01 5.27262211e-01 3.92735869e-01 -2.33961955e-01
-1.43908572e+00 -5.61254323e-01 -2.34229267e-01 5.40088601e-02
1.00830114e+00 -3.69676590e-01 1.94614351e+00 -6.89956605e-01
-1.67224741e+00 1.71067938e-01 -2.44510174e-01 -9.74403977e-01
9.24205959e-01 -2.48265594e-01 6.40455931e-02 2.65383601e-01
-1.30072804e-02 -8.18156600e-02 8.08081090e-01 -1.19776499e+00
-8.90217125e-01 -4.32346582e-01 2.25263998e-01 4.01648283e-01
-2.01150119e-01 -3.26787084e-01 1.26415655e-01 -2.12964058e-01
-3.28360111e-01 -1.09957707e+00 -4.08653349e-01 -1.55395046e-01
-2.63576835e-01 -1.49623692e-01 6.50543392e-01 -3.86766523e-01
1.13251638e+00 -1.73684275e+00 -2.46804386e-01 3.01764876e-01
-1.17117546e-01 -1.48106351e-01 -3.62242684e-02 7.44372010e-01
7.36077368e-01 7.43623972e-02 1.22433901e-02 -3.69327098e-01
2.80836493e-01 3.60527039e-01 -8.07660222e-01 8.98271263e-01
-5.02902269e-01 4.64452684e-01 -1.07575250e+00 -2.47264430e-01
8.74326080e-02 -4.63400185e-01 -7.18840778e-01 2.68551677e-01
-6.54841244e-01 5.15497029e-01 -8.69192123e-01 4.32489902e-01
5.00884712e-01 -2.00719133e-01 6.63462341e-01 4.19505268e-01
-1.87912151e-01 2.35519111e-01 -1.07142794e+00 1.26221907e+00
-5.58046162e-01 4.17065918e-01 4.69370484e-01 -1.04766011e+00
1.49505496e-01 1.23889573e-01 7.48086810e-01 -5.24712265e-01
4.04301077e-01 1.96803868e-01 -1.54830813e-01 -3.21824610e-01
4.41386312e-01 -1.41097575e-01 -3.23146909e-01 8.38250875e-01
-2.12812021e-01 2.85807580e-01 8.05835798e-02 8.78078938e-02
1.17561531e+00 7.85299987e-02 4.44758862e-01 -3.01543385e-01
-4.13897960e-03 1.25711024e-01 3.74248445e-01 1.37490535e+00
-4.32080120e-01 -2.23930791e-01 6.91536665e-01 -3.77491824e-02
-1.03692925e+00 -9.97548878e-01 -5.69010973e-02 1.01280117e+00
1.65399507e-01 -4.85894755e-02 -7.23501027e-01 -7.11231470e-01
5.30252039e-01 9.59513843e-01 -4.74855661e-01 6.52183518e-02
-4.39856619e-01 -6.17911577e-01 2.80432045e-01 2.44980082e-01
3.37440491e-01 -4.96217728e-01 -9.73010421e-01 3.95126373e-01
-1.52974755e-01 -1.18658304e+00 -5.16892910e-01 1.22596666e-01
-1.03650510e+00 -1.13537288e+00 -4.37528014e-01 7.18848482e-02
5.30646086e-01 4.12157655e-01 7.60537446e-01 -6.37024522e-01
3.96862388e-01 1.16350257e+00 -8.39723200e-02 -5.53700089e-01
-2.61597157e-01 1.89389333e-01 4.10453409e-01 -7.60032982e-02
2.74716578e-02 -3.14436764e-01 -7.79277503e-01 1.51257455e-01
-6.50704563e-01 -6.67266071e-01 3.18807155e-01 4.27784860e-01
8.08633983e-01 1.96510423e-02 6.93684399e-01 -7.85538971e-01
1.06033206e+00 -9.06477869e-01 -1.18155277e+00 3.36478621e-01
-5.54610908e-01 2.66254872e-01 1.04496264e+00 -8.05469990e-01
-9.44795012e-01 -8.17204043e-02 3.54444593e-01 -5.34798145e-01
5.14239073e-02 1.64935410e-01 -1.38665825e-01 8.70008618e-02
2.82083571e-01 3.70459080e-01 8.26867372e-02 -5.33461094e-01
3.39074999e-01 5.98340511e-01 4.20014083e-01 -1.25718606e+00
6.62783146e-01 5.95923603e-01 1.69829488e-01 -8.74461114e-01
-9.96398509e-01 -4.09568429e-01 -2.53561046e-02 -2.20113307e-01
1.81587473e-01 -7.84959972e-01 -1.59838831e+00 -8.97875428e-03
-6.85447395e-01 -1.11870408e+00 -5.63957334e-01 5.19886196e-01
-1.24582160e+00 2.38903776e-01 -3.41968536e-01 -1.67173302e+00
-3.24651487e-02 -6.45165563e-01 7.79331088e-01 1.34662732e-01
1.17629103e-01 -1.04667485e+00 3.05916637e-01 -9.53282118e-02
1.67718217e-01 1.02400579e-01 5.44414461e-01 -7.55480111e-01
-6.29009247e-01 -5.69610484e-02 1.02992229e-01 -3.49604189e-02
-1.59400582e-01 -2.60455459e-01 -6.15325034e-01 -7.06289768e-01
5.11825224e-03 -5.59560478e-01 5.00612438e-01 7.92727053e-01
1.56651974e+00 -1.20190203e+00 -3.68556589e-01 3.13970059e-01
1.48188603e+00 3.23758602e-01 -3.15720290e-02 2.61098564e-01
8.39481726e-02 2.85828233e-01 1.24661171e+00 1.31746149e+00
5.14542401e-01 4.13986415e-01 4.26281482e-01 6.08301401e-01
7.54498303e-01 -6.54631197e-01 7.49735653e-01 2.47149020e-01
2.26768047e-01 -6.74374580e-01 -5.15808880e-01 5.61043799e-01
-2.08635831e+00 -1.08146644e+00 4.13562894e-01 2.82429934e+00
9.34990644e-01 1.69848293e-01 9.29469347e-01 -1.17397293e-01
3.91999662e-01 2.72449013e-02 -1.23142660e+00 -3.01318675e-01
1.61217675e-01 1.01022854e-01 1.40927804e+00 5.14913142e-01
-8.31787586e-01 8.17965508e-01 7.06370783e+00 8.39715362e-01
-8.32716703e-01 2.44496241e-01 5.06759882e-01 -5.45135200e-01
-1.57300755e-01 -4.26615514e-02 -8.73789012e-01 7.19364345e-01
1.63349342e+00 -6.29978359e-01 8.72983038e-01 1.09518206e+00
7.01382697e-01 -4.97299761e-01 -1.10602140e+00 9.16838467e-01
-6.32600307e-01 -1.08118832e+00 -4.00664717e-01 5.74358582e-01
7.35838950e-01 4.22750749e-02 2.36276731e-01 4.18678015e-01
9.55815136e-01 -4.88961875e-01 7.34020054e-01 4.04513210e-01
1.00749600e+00 -1.10748017e+00 3.07215482e-01 8.13347459e-01
-9.11773860e-01 -6.79705799e-01 -4.66537267e-01 6.17059022e-02
3.26667950e-02 2.88317055e-01 -7.47795284e-01 8.53515640e-02
3.23562115e-01 3.95459741e-01 2.98078746e-01 1.06788480e+00
1.26675174e-01 9.42509472e-01 -8.18636358e-01 -2.59931564e-01
4.24957484e-01 -3.21508974e-01 6.05499625e-01 9.37362850e-01
3.09539557e-01 2.60911137e-01 3.86566103e-01 3.94099444e-01
-1.57024205e-01 -1.69936404e-01 -7.00654566e-01 -2.62052536e-01
1.00031567e+00 8.78399491e-01 -3.60281706e-01 -2.88630724e-01
9.12684798e-02 5.81151366e-01 2.59740204e-01 5.67403257e-01
-9.69417751e-01 4.03065281e-03 9.34747696e-01 8.81447569e-02
4.16273326e-01 -6.51586175e-01 1.16921226e-02 -1.03266668e+00
-4.50593494e-02 -8.37437928e-01 6.28736973e-01 1.81916375e-02
-1.14507985e+00 -4.44218874e-01 3.85840863e-01 -1.05723548e+00
-5.26741028e-01 -2.57852882e-01 -3.09941590e-01 2.31088564e-01
-1.35880923e+00 -2.88515508e-01 4.02540028e-01 5.70359170e-01
4.95655626e-01 2.52975464e-01 2.92801201e-01 -5.80989048e-02
-7.17288911e-01 7.56172061e-01 9.12137270e-01 -3.38117987e-01
3.10845047e-01 -1.17585111e+00 -1.31880701e-01 6.50684237e-01
-1.97897896e-01 3.18350554e-01 9.02625263e-01 -5.28761506e-01
-1.99741387e+00 -1.00709414e+00 -2.92220898e-02 -2.35304236e-01
9.32862759e-01 -1.88473567e-01 -4.11433131e-01 9.26553428e-01
-3.14311504e-01 -1.28835201e-01 3.23009402e-01 -2.37232000e-02
9.61478204e-02 -2.53138781e-01 -1.44471335e+00 6.50524557e-01
1.12404239e+00 -3.50970179e-01 -1.13225281e-02 3.97161067e-01
5.20372570e-01 -4.29785371e-01 -8.35904002e-01 -2.48122346e-02
5.51903427e-01 -4.36886042e-01 5.54530382e-01 -9.37602580e-01
-3.67085904e-01 2.51506329e-01 -2.75728792e-01 -1.16188931e+00
2.95075774e-01 -1.27142763e+00 -7.96231329e-01 7.24143207e-01
5.47043532e-02 -1.03381741e+00 7.32363820e-01 6.07733548e-01
4.59897906e-01 -6.77739918e-01 -1.14348090e+00 -1.30460703e+00
2.30817899e-01 -4.40372765e-01 4.27530497e-01 4.61197615e-01
7.78271630e-02 -1.18664794e-01 -4.75418180e-01 1.41250163e-01
1.15499806e+00 2.20680416e-01 9.86785710e-01 -7.15115607e-01
-4.76986438e-01 -2.42871255e-01 5.36659539e-01 -1.61864626e+00
4.21322763e-01 -3.04155141e-01 1.66956067e-01 -1.05640054e+00
1.96565107e-01 -7.71032155e-01 -2.26343304e-01 1.53330415e-01
2.67166048e-01 -6.84370458e-01 1.73076853e-01 3.41281861e-01
-1.16938901e+00 7.88452327e-01 9.96143937e-01 3.89338285e-01
-3.06909204e-01 5.03824413e-01 -4.18882310e-01 4.69757438e-01
1.22471380e+00 -7.19806254e-01 -7.90872931e-01 -2.55574547e-02
4.83356416e-02 8.14067662e-01 6.11938424e-02 -6.20405316e-01
3.22665833e-02 -8.78170013e-01 -1.60210967e-01 -8.60777378e-01
1.38748839e-01 -9.20785427e-01 -2.44579509e-01 6.31903172e-01
-8.16739202e-01 5.10222046e-04 2.74696033e-02 1.16916978e+00
6.35924816e-01 9.60525591e-03 7.66368628e-01 -4.78775315e-02
-2.56577437e-03 7.18410075e-01 -5.72462261e-01 7.63403654e-01
1.06150830e+00 9.35261026e-02 -3.11741918e-01 -8.22357833e-01
-4.84635502e-01 5.36992848e-01 4.33730185e-01 -2.24246174e-01
3.41973513e-01 -8.57867002e-01 -2.73256153e-01 -2.80903518e-01
-4.25758809e-01 -2.75899947e-01 -1.27815917e-01 8.95687282e-01
2.26826277e-02 5.75492740e-01 9.96824056e-02 -2.35381007e-01
-8.46651077e-01 5.92526615e-01 4.10870105e-01 -4.10400301e-01
-3.41391802e-01 1.44846410e-01 -2.14220971e-01 -1.70371667e-01
5.94293773e-01 -3.95011455e-01 4.83036131e-01 -5.36872335e-02
4.84807581e-01 8.75147164e-01 -5.33264220e-01 1.32331192e-01
-7.23804086e-02 -2.09194161e-02 5.97870536e-02 -8.24764431e-01
1.15940249e+00 -4.07820791e-01 4.65241581e-01 7.51919270e-01
1.02403784e+00 5.87073117e-02 -1.74698579e+00 -4.00534868e-01
1.80532426e-01 -5.62330961e-01 -4.59318049e-02 -6.15628481e-01
-7.20587134e-01 5.20878732e-01 3.99067402e-01 2.85870999e-01
8.21063817e-01 -2.10670024e-01 8.16184402e-01 7.92932570e-01
1.03691375e+00 -1.34488583e+00 -2.60911137e-01 3.01976353e-01
4.14702803e-01 -9.51590300e-01 1.32237330e-01 4.37991589e-01
-5.65114558e-01 9.56013143e-01 1.57660007e-01 -2.79148370e-01
5.14992952e-01 1.43980756e-01 -6.62933946e-01 3.01630527e-01
-1.16127038e+00 -2.77737141e-01 -3.01759303e-01 3.80935580e-01
-2.81887412e-01 4.94949847e-01 -5.39234757e-01 4.78985846e-01
-2.08911449e-02 1.66427583e-01 7.00634718e-01 9.92452025e-01
-7.36711204e-01 -9.19742167e-01 -2.16296494e-01 5.30027151e-01
-6.16178334e-01 4.26931381e-01 -3.04970652e-01 9.37023163e-01
-7.10597217e-01 9.56013501e-01 8.45337510e-02 -7.48699345e-03
-2.21945420e-02 -2.42949516e-01 8.29113424e-01 -1.60050616e-01
-2.14025602e-01 4.52042520e-02 2.08514765e-01 -8.83793652e-01
-1.26577184e-01 -8.97145092e-01 -1.20138156e+00 -9.02813077e-01
4.25887220e-02 3.04565489e-01 6.60310924e-01 8.25653195e-01
3.86754721e-01 -3.33796859e-01 1.09092426e+00 -6.44693255e-01
-1.75176466e+00 -5.95258772e-01 -6.14155531e-01 -1.82306319e-01
6.31946027e-01 -5.78869641e-01 -6.74711108e-01 -6.28944814e-01] | [4.295804500579834, 2.7747855186462402] |
714d36d3-79ca-4643-bb26-174fbf316ee6 | frame-interpolation-transformer-and | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Plack_Frame_Interpolation_Transformer_and_Uncertainty_Guidance_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Plack_Frame_Interpolation_Transformer_and_Uncertainty_Guidance_CVPR_2023_paper.pdf | Frame Interpolation Transformer and Uncertainty Guidance | Video frame interpolation has seen important progress in recent years, thanks to developments in several directions. Some works leverage better optical flow methods with improved splatting strategies or additional cues from depth, while others have investigated alternative approaches through direct predictions or transformers. Still, the problem remains unsolved in more challenging conditions such as complex lighting or large motion. In this work, we are bridging the gap towards video production with a novel transformer-based interpolation network architecture capable of estimating the expected error together with the interpolated frame. This offers several advantages that are of key importance for frame interpolation usage: First, we obtained improved visual quality over several datasets. The improvement in terms of quality is also clearly demonstrated through a user study. Second, our method estimates error maps for the interpolated frame, which are essential for real-life applications on longer video sequences where problematic frames need to be flagged. Finally, for rendered content a partial rendering pass of the intermediate frame, guided by the predicted error, can be utilized during the interpolation to generate a new frame of superior quality. Through this error estimation, our method can produce even higher-quality intermediate frames using only a fraction of the time compared to a full rendering. | ['Christopher Schroers', 'Markus Gross', 'Matthias B. Hullin', 'Abdelaziz Djelouah', 'Karlis Martins Briedis', 'Markus Plack'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['video-frame-interpolation'] | ['computer-vision'] | [ 2.68983722e-01 -3.91559377e-02 -1.81192048e-02 -2.29729697e-01
-4.36208040e-01 -3.22252423e-01 4.63592201e-01 1.20753676e-01
-2.66309440e-01 9.60693955e-01 2.30248004e-01 -2.92710904e-02
1.24656238e-01 -7.81867862e-01 -7.46811628e-01 -5.50361276e-01
-1.79803491e-01 1.29078664e-02 5.39238751e-01 -2.51285821e-01
3.55248153e-01 5.63982189e-01 -1.82590103e+00 3.52158487e-01
9.00218666e-01 1.20086288e+00 3.59302938e-01 8.06665003e-01
-4.43331935e-02 9.04653013e-01 -4.73923713e-01 -2.35871434e-01
4.06932473e-01 -5.73812068e-01 -5.62547028e-01 2.43976414e-02
6.79901958e-01 -7.43293166e-01 -2.74107400e-02 7.19257653e-01
3.68300885e-01 2.11330384e-01 1.02809742e-01 -1.11717653e+00
1.74305530e-03 2.13799804e-01 -2.55000412e-01 4.21588905e-02
8.61268878e-01 3.64794165e-01 7.67494142e-01 -8.63462687e-01
8.47761154e-01 1.18764639e+00 6.69176161e-01 4.41206902e-01
-1.42355859e+00 -5.69273293e-01 5.71891591e-02 5.66159904e-01
-9.72788513e-01 -5.45577645e-01 7.19933152e-01 -4.93638545e-01
6.27006471e-01 3.31171960e-01 9.55451310e-01 8.78442049e-01
6.58769310e-02 4.79550898e-01 9.50439751e-01 -3.59040350e-01
2.36579165e-01 1.60735503e-01 -4.32083666e-01 4.99509335e-01
-1.31303683e-01 4.29856360e-01 -6.05144441e-01 1.44912332e-01
1.03178215e+00 -3.21990937e-01 -8.09655726e-01 -3.98001701e-01
-1.33826625e+00 5.39759398e-01 3.95996004e-01 3.16605836e-01
-4.14595217e-01 8.61776993e-02 3.73869181e-01 1.95873335e-01
6.04540825e-01 3.91205400e-01 -1.74954981e-01 -5.47007263e-01
-1.51262128e+00 4.04589593e-01 7.29886889e-01 6.76560521e-01
9.16798890e-01 1.29880130e-01 -3.11323166e-01 6.46129906e-01
2.12095186e-01 2.39587754e-01 -2.40751766e-02 -1.49172473e+00
4.84965056e-01 2.45490044e-01 4.71437633e-01 -1.17339480e+00
-2.37008899e-01 -2.60988265e-01 -8.43485534e-01 9.26753461e-01
7.61495650e-01 -3.57347503e-02 -5.32421947e-01 1.44800878e+00
3.99502188e-01 5.84920049e-01 -2.81869531e-01 1.10514581e+00
3.92313868e-01 7.16374636e-01 -2.03185439e-01 -3.50871712e-01
1.19481862e+00 -8.25799048e-01 -8.32375050e-01 8.63079652e-02
3.37896556e-01 -9.96888101e-01 8.52348447e-01 7.69149840e-01
-1.42324960e+00 -7.90853739e-01 -1.10808897e+00 -1.70661390e-01
9.18808877e-02 -2.05222890e-01 4.13857162e-01 6.14827871e-01
-1.38074517e+00 9.48345006e-01 -7.78160036e-01 -1.15760423e-01
3.99515837e-01 2.21655205e-01 -2.20761448e-01 -1.13504209e-01
-1.08724427e+00 7.14135468e-01 1.24802977e-01 1.71631724e-01
-5.75649500e-01 -1.08206594e+00 -9.60035026e-01 1.38168484e-01
2.42273062e-01 -8.20941806e-01 9.85414088e-01 -1.13786161e+00
-1.79241037e+00 3.37880135e-01 -3.61019999e-01 -5.27528644e-01
1.01306903e+00 -3.62493724e-01 -1.14197001e-01 2.68047512e-01
-1.70578197e-01 8.59207571e-01 7.96132445e-01 -1.32081938e+00
-8.64528775e-01 5.38940430e-02 4.48035151e-01 2.37044096e-02
-6.00021891e-02 -1.83353350e-01 -3.67284805e-01 -6.85923457e-01
-1.42593667e-01 -7.82175303e-01 -2.05715865e-01 6.02786958e-01
-7.34812068e-03 1.95654765e-01 8.30425560e-01 -7.76221454e-01
1.26984847e+00 -1.99835992e+00 1.89851463e-01 9.98977479e-03
1.59387499e-01 4.88849342e-01 8.09999835e-03 3.96776438e-01
1.12010604e-02 1.13771474e-02 -8.51973817e-02 -5.17717004e-01
-3.25755119e-01 -4.69283164e-02 -2.72023112e-01 1.82500482e-01
2.88154751e-01 4.54039633e-01 -1.15241194e+00 -4.20998901e-01
7.13355660e-01 9.68113303e-01 -9.30713654e-01 2.74805158e-01
-2.06805050e-01 1.00238657e+00 4.02697027e-02 2.83251554e-01
7.97739506e-01 6.13274463e-02 -6.10686168e-02 -4.77421373e-01
-3.96831006e-01 2.51037806e-01 -1.44870913e+00 1.86197734e+00
-6.26718462e-01 1.03736925e+00 1.64914392e-02 -6.11121774e-01
7.35034645e-01 4.48872209e-01 6.37384772e-01 -7.55543709e-01
-3.32910158e-02 1.31238535e-01 -1.73579991e-01 -4.47418630e-01
8.04935992e-01 1.84400201e-01 7.62480021e-01 1.53414428e-01
-2.36046344e-01 -1.54660344e-01 4.52261090e-01 9.48079862e-03
7.97378123e-01 7.39795268e-01 3.48345004e-02 -1.23568341e-01
7.52264619e-01 -2.15207666e-01 5.65070033e-01 5.88456035e-01
-1.33819982e-01 9.48475242e-01 5.39032757e-01 -5.46133399e-01
-1.25220978e+00 -8.41442883e-01 -1.49158493e-01 4.45263326e-01
1.88975126e-01 -6.30781114e-01 -6.79345787e-01 -2.21251994e-01
-4.36578482e-01 5.94944119e-01 -3.45013291e-01 1.69782177e-01
-9.23054278e-01 -1.64024428e-01 2.38087058e-01 3.09342593e-01
6.17968440e-01 -9.59006488e-01 -1.03070199e+00 4.58423942e-01
-3.76678586e-01 -1.27661908e+00 -3.38781297e-01 -4.46522385e-01
-1.09464478e+00 -9.81651127e-01 -9.26511526e-01 -4.30533051e-01
4.50370193e-01 2.21227244e-01 1.23014092e+00 3.57693970e-01
-1.72958091e-01 5.70519790e-02 -3.16447049e-01 9.52176079e-02
-5.74274123e-01 -3.16526622e-01 -2.86008686e-01 1.21007435e-01
-3.85207564e-01 -7.29924738e-01 -9.98734355e-01 4.94516671e-01
-8.53701532e-01 4.72835779e-01 1.42925968e-02 8.68534088e-01
3.69132400e-01 -2.71512508e-01 2.39592180e-01 -5.97612500e-01
3.40982378e-01 -2.09847555e-01 -8.28006744e-01 -1.55205980e-01
-4.95494574e-01 5.73919602e-02 8.61564457e-01 -2.74672210e-01
-1.21717811e+00 -1.07246764e-01 -3.90643954e-01 -3.79371494e-01
-3.70046496e-02 6.16451576e-02 1.37687430e-01 -1.64207015e-02
5.73142827e-01 2.75152158e-02 6.01185113e-02 -2.66072541e-01
1.06819980e-01 3.21903169e-01 3.79811496e-01 -4.46942389e-01
5.74230015e-01 6.45726264e-01 2.26129159e-01 -7.18996823e-01
-3.63799632e-01 -2.01228499e-01 -5.41469455e-01 -6.31972432e-01
6.85906291e-01 -7.87542164e-01 -1.07557428e+00 4.44392502e-01
-1.35632062e+00 -3.87708277e-01 -2.64003724e-01 6.27656698e-01
-5.25400519e-01 5.68123400e-01 -6.24902904e-01 -7.91881859e-01
-4.17299308e-02 -1.58844507e+00 1.01438785e+00 2.21573547e-01
-2.01170996e-01 -1.03737426e+00 -9.47538689e-02 1.94327608e-01
7.07875788e-01 3.11965615e-01 5.64782679e-01 3.33751082e-01
-1.01121986e+00 8.54915604e-02 -2.65922159e-01 3.93328160e-01
2.85466671e-01 3.67296487e-01 -9.85375702e-01 -3.11220556e-01
-1.20593108e-01 3.33434790e-02 5.98456502e-01 3.77065480e-01
1.08892536e+00 -2.50370838e-02 -8.22409149e-03 7.14915633e-01
1.43531835e+00 2.18175843e-01 1.01377809e+00 3.98430258e-01
6.10329747e-01 4.62743104e-01 7.59852409e-01 6.43752754e-01
3.43380094e-01 1.04375672e+00 5.02558291e-01 -1.20982803e-01
-5.87638617e-01 -1.33974105e-01 3.22032273e-01 3.65446895e-01
-6.58672154e-01 -2.41275862e-01 -6.09195292e-01 3.53834510e-01
-1.72294641e+00 -1.16216850e+00 -3.26372415e-01 2.58902168e+00
7.01035321e-01 1.86652824e-01 7.78535828e-02 3.92728657e-01
5.96200347e-01 2.40474775e-01 -3.07076782e-01 -3.10775369e-01
-4.18351823e-03 1.69860691e-01 2.53360033e-01 7.92380095e-01
-5.66927731e-01 5.61370432e-01 6.44506025e+00 6.07901752e-01
-1.41746294e+00 -2.30733007e-01 6.70361161e-01 -4.96351644e-02
-4.90673929e-01 2.83050358e-01 -5.75789094e-01 6.52036011e-01
8.26962233e-01 1.34936005e-01 4.95644927e-01 4.42224681e-01
8.83632243e-01 -6.04102671e-01 -1.05874872e+00 1.01100826e+00
-1.55245988e-02 -1.61042202e+00 -1.61537245e-01 9.69756022e-02
5.37421465e-01 -3.53889406e-01 -1.16067767e-01 -1.55232742e-01
-1.75545558e-01 -7.22157478e-01 8.59292269e-01 6.21004224e-01
8.01287293e-01 -5.87145090e-01 5.75122595e-01 2.32503250e-01
-1.14042151e+00 2.33300343e-01 -2.20214233e-01 -3.91283751e-01
5.63408375e-01 9.00827348e-01 -7.65856028e-01 6.54376149e-01
6.96990967e-01 8.36090744e-01 -4.21965539e-01 1.32367098e+00
-2.64464080e-01 5.64996719e-01 -2.62744308e-01 3.95207047e-01
-9.67812464e-02 -3.88432533e-01 6.46102190e-01 1.06168163e+00
5.10655999e-01 -2.23605365e-01 -5.64551540e-02 8.60975087e-01
2.29700044e-01 1.10043980e-01 -4.57586318e-01 4.91056502e-01
2.62471884e-01 1.20706904e+00 -5.78016222e-01 -3.98092866e-01
-2.70227969e-01 1.09273636e+00 9.46954042e-02 4.21304047e-01
-1.00239527e+00 -1.21016644e-01 8.00025165e-01 3.41526508e-01
2.71674305e-01 -3.64089608e-01 -1.93752438e-01 -1.21569347e+00
1.35841653e-01 -5.82145929e-01 -7.32729957e-02 -1.02694845e+00
-5.93957901e-01 8.05756509e-01 -1.30141452e-01 -1.62550390e+00
-6.63187861e-01 -3.12076002e-01 -3.43440980e-01 8.59863043e-01
-1.81356263e+00 -7.17475414e-01 -6.69918895e-01 4.73387271e-01
7.09650636e-01 3.46793175e-01 6.82276011e-01 7.31558204e-01
-2.45930165e-01 3.80408585e-01 -1.74194619e-01 -2.34577671e-01
8.14583182e-01 -1.02225459e+00 2.85798579e-01 1.00263071e+00
-9.40172374e-02 2.15560168e-01 1.00226665e+00 -3.85719389e-01
-1.09508491e+00 -6.10210896e-01 9.13226247e-01 -1.44663572e-01
3.40052068e-01 -2.82669067e-01 -9.21906650e-01 2.83706844e-01
3.64608705e-01 3.01191092e-01 2.06130311e-01 -2.87324101e-01
-6.97482675e-02 -2.77579516e-01 -1.24045992e+00 6.20881319e-01
1.01158309e+00 -1.93206847e-01 8.19343477e-02 -1.11546330e-01
6.82081342e-01 -6.77692533e-01 -9.04894352e-01 3.07012469e-01
8.19311202e-01 -1.91336584e+00 9.66938198e-01 6.33259043e-02
8.21037233e-01 -5.30767381e-01 9.55977365e-02 -1.32485044e+00
-6.00196160e-02 -7.24648476e-01 -1.28774777e-01 1.28970706e+00
2.44116038e-02 -4.28466082e-01 8.05979192e-01 6.11411572e-01
-7.12970719e-02 -6.99891150e-01 -8.66106331e-01 -5.71427882e-01
-3.84263158e-01 -6.76656246e-01 5.10779619e-01 6.40777647e-01
-2.10153237e-01 -1.70959800e-01 -7.29801297e-01 -1.04765728e-01
5.39560974e-01 -2.45578233e-02 8.83471549e-01 -1.01428020e+00
-2.71899581e-01 -2.41995260e-01 -6.03600621e-01 -1.42402089e+00
-1.32085204e-01 -3.28561574e-01 8.35904703e-02 -1.43703640e+00
-3.66841525e-01 -5.87073445e-01 4.32816669e-02 5.71118295e-02
-1.44022346e-01 4.83182281e-01 5.61372995e-01 1.21146210e-01
-1.90299258e-01 3.37051719e-01 1.50064635e+00 1.47626147e-01
-2.34410629e-01 3.88188288e-02 -1.72901139e-01 6.39448464e-01
5.91034710e-01 -1.42621055e-01 -4.58048612e-01 -5.02828479e-01
3.33202094e-01 5.61365783e-01 5.35433471e-01 -1.35253012e+00
1.22963466e-01 -4.04421911e-02 3.47761691e-01 -5.21417737e-01
5.70498288e-01 -9.90366161e-01 4.37905610e-01 4.80998158e-01
-1.47803426e-01 1.40753999e-01 9.45735499e-02 3.12420636e-01
-3.30312580e-01 -1.32347837e-01 8.57723594e-01 6.26057684e-02
-7.46614933e-01 7.32678622e-02 -2.27014005e-01 -3.02370578e-01
8.80727589e-01 -6.99335337e-01 3.47827300e-02 -6.60288513e-01
-6.21690691e-01 -1.03101552e-01 6.53556526e-01 3.03565830e-01
6.87783062e-01 -1.27324021e+00 -7.24274874e-01 3.41588080e-01
-1.60293803e-01 -1.38956681e-02 1.96174532e-01 1.02500927e+00
-8.40282619e-01 1.38063088e-01 -3.90785426e-01 -9.53725815e-01
-1.16819990e+00 3.89119297e-01 2.66364276e-01 -1.61343619e-01
-7.71625996e-01 4.74408060e-01 4.39089239e-02 3.20874006e-01
1.11951694e-01 -6.57423079e-01 -2.22916991e-01 2.00147748e-01
7.96807885e-01 5.94201684e-01 1.36013299e-01 -6.14113390e-01
-7.75110796e-02 5.74243248e-01 3.04189712e-01 -1.35197461e-01
1.21770465e+00 -3.72240245e-01 3.91452685e-02 3.44758451e-01
9.83628929e-01 2.74707168e-01 -1.90063596e+00 5.37709072e-02
-3.27431053e-01 -1.07614708e+00 6.30979687e-02 -6.10401690e-01
-1.34661198e+00 7.67013848e-01 6.70405447e-01 1.61099091e-01
1.32889509e+00 -4.72420514e-01 9.07811761e-01 -2.54425585e-01
6.35918856e-01 -7.10308433e-01 -1.92443043e-01 2.80496866e-01
7.92450309e-01 -1.19138324e+00 -6.08995110e-02 -6.74740195e-01
-2.24653319e-01 1.43838561e+00 3.25288266e-01 -1.09777175e-01
4.25765902e-01 3.61955822e-01 6.52761310e-02 2.46015757e-01
-7.32123017e-01 -4.33633886e-02 1.80841833e-01 6.20832384e-01
6.96873128e-01 -3.87193114e-01 -3.94962132e-01 -2.97717929e-01
-9.33396965e-02 4.47685033e-01 8.31165075e-01 6.06142581e-01
-3.09360325e-01 -1.22579563e+00 -4.54684615e-01 1.53283969e-01
-2.88375825e-01 -1.95514888e-01 3.99808228e-01 5.60386062e-01
1.67595446e-01 1.02376270e+00 1.05471127e-01 -8.15720484e-02
3.49873513e-01 -3.89422506e-01 5.96004426e-01 -1.66469678e-01
-6.02783740e-01 1.72543868e-01 1.15147889e-01 -1.09356225e+00
-7.93361366e-01 -5.58646381e-01 -1.18168843e+00 -4.77945477e-01
-2.19485104e-01 6.03695065e-02 7.33289778e-01 7.49088168e-01
4.54242408e-01 6.24592006e-01 5.28670490e-01 -1.46262848e+00
1.62437439e-01 -5.86974084e-01 -2.01981768e-01 5.45031488e-01
5.16442835e-01 -5.78963697e-01 -3.92412454e-01 3.44445258e-01] | [10.765586853027344, -1.413009524345398] |
86ab0393-5693-4bfd-8dcc-036014b551cd | brent-bidirectional-retrieval-enhanced | 2304.09649 | null | https://arxiv.org/abs/2304.09649v1 | https://arxiv.org/pdf/2304.09649v1.pdf | BRENT: Bidirectional Retrieval Enhanced Norwegian Transformer | Retrieval-based language models are increasingly employed in question-answering tasks. These models search in a corpus of documents for relevant information instead of having all factual knowledge stored in its parameters, thereby enhancing efficiency, transparency, and adaptability. We develop the first Norwegian retrieval-based model by adapting the REALM framework and evaluating it on various tasks. After training, we also separate the language model, which we call the reader, from the retriever components, and show that this can be fine-tuned on a range of downstream tasks. Results show that retrieval augmented language modeling improves the reader's performance on extractive question-answering, suggesting that this type of training improves language models' general ability to use context and that this does not happen at the expense of other abilities such as part-of-speech tagging, dependency parsing, named entity recognition, and lemmatization. Code, trained models, and data are made publicly available. | ['Egil Rønningstad', 'David Samuel', 'Sondre Wold', 'Lucas Georges Gabriel Charpentier'] | 2023-04-19 | null | null | null | null | ['lemmatization', 'dependency-parsing', 'part-of-speech-tagging'] | ['natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [-9.42651927e-02 4.15007710e-01 -6.41588941e-02 -3.19376230e-01
-1.45351136e+00 -9.86498594e-01 6.19416535e-01 5.02989709e-01
-9.12834883e-01 4.88474190e-01 6.30895734e-01 -7.35997498e-01
-2.35875487e-01 -5.02565920e-01 -4.16964591e-01 -1.03673013e-02
3.01458418e-01 5.80366313e-01 5.15178978e-01 -4.88424212e-01
4.09286410e-01 4.73238081e-01 -9.54261720e-01 4.89893436e-01
8.65795851e-01 2.45560497e-01 4.92871523e-01 6.33272588e-01
-5.59532404e-01 1.13076091e+00 -7.63742566e-01 -7.09637880e-01
-3.03709596e-01 -8.20480064e-02 -1.27932978e+00 -3.64088744e-01
1.20767422e-01 -4.15557027e-01 -1.43265456e-01 4.96660471e-01
4.14409518e-01 2.42452249e-01 4.67755139e-01 -4.68288779e-01
-9.21995997e-01 6.64901733e-01 4.45977896e-02 5.08846462e-01
4.30101871e-01 -1.39114216e-01 1.37303376e+00 -8.33099544e-01
9.70627010e-01 1.10306823e+00 4.70387131e-01 7.23283350e-01
-9.39303100e-01 -2.21106902e-01 2.73017108e-01 -2.29537800e-01
-9.79955554e-01 -7.49830663e-01 4.00547922e-01 -3.66821736e-01
1.66415763e+00 2.07192406e-01 2.72722393e-02 6.82658851e-01
1.62324458e-01 8.21672022e-01 7.85083592e-01 -8.72668266e-01
-9.01445150e-02 4.05520707e-01 4.92966771e-01 4.17905420e-01
2.96427786e-01 -1.36375114e-01 -3.78893942e-01 -2.48721465e-01
3.91026825e-01 -3.62184793e-01 -2.86452830e-01 1.10236481e-02
-8.91387582e-01 7.86608815e-01 9.35779065e-02 6.36664689e-01
-2.16613352e-01 6.49979152e-03 3.47094506e-01 4.88617033e-01
4.84203368e-01 1.03267157e+00 -9.88265336e-01 -5.32518290e-02
-7.01651156e-01 7.18131736e-02 1.13245857e+00 8.99364531e-01
6.21837676e-01 -4.50616330e-01 -2.63998449e-01 1.17010963e+00
4.38964754e-01 5.55881739e-01 6.53296888e-01 -1.09977198e+00
7.04233468e-01 7.49525607e-01 1.53967500e-01 -3.24174285e-01
-3.96461278e-01 -5.27847886e-01 1.25662193e-01 -4.31521475e-01
5.72817445e-01 -2.95858860e-01 -8.06256056e-01 1.85347736e+00
2.12086633e-01 -7.28451967e-01 4.42501426e-01 4.66467530e-01
8.74096096e-01 7.12564766e-01 6.48940504e-01 -3.21873575e-02
1.61923873e+00 -8.41928482e-01 -6.81610107e-01 -7.84124076e-01
1.32716537e+00 -9.03754771e-01 9.77258325e-01 -1.66030630e-01
-1.34787333e+00 -3.33993137e-01 -4.83137816e-01 -7.06371248e-01
-7.23828375e-01 1.39961466e-01 6.05017424e-01 5.52907646e-01
-1.22751522e+00 3.40486139e-01 -6.43973947e-01 -6.09569073e-01
-7.67426342e-02 1.43416747e-01 -3.54316801e-01 -2.05841929e-01
-1.36154342e+00 1.23043907e+00 1.12368658e-01 -1.93401556e-02
-2.65035987e-01 -3.44115227e-01 -9.25610602e-01 2.55192429e-01
2.82721788e-01 -7.34998167e-01 1.79896641e+00 -6.93435490e-01
-1.26241267e+00 9.23533618e-01 -3.49574298e-01 -8.89763087e-02
-1.83858693e-01 -4.10058469e-01 -3.71066421e-01 2.58429736e-01
2.05150113e-01 4.62765843e-01 5.09894431e-01 -1.13207042e+00
-7.52270579e-01 -5.21675527e-01 3.55523795e-01 4.75791425e-01
-2.47060061e-01 6.04926527e-01 -5.52768230e-01 -3.06517571e-01
5.64821213e-02 -6.14517808e-01 -3.82837504e-02 -4.66639847e-01
1.53038591e-01 -5.50606728e-01 3.22678059e-01 -1.14809012e+00
1.48093450e+00 -1.86563969e+00 -2.30411977e-01 -5.23122549e-02
-1.48406699e-01 4.12935287e-01 -4.50861901e-01 8.72828007e-01
-2.49396334e-03 7.40404546e-01 3.22049931e-02 -2.26250276e-01
9.20022093e-03 3.05277050e-01 -2.13222995e-01 -1.34431899e-01
4.78537887e-01 1.21715975e+00 -8.04940641e-01 -4.53839898e-01
-1.62575290e-01 2.38018036e-01 -5.19482195e-01 9.99147594e-02
-2.63017952e-01 7.64270201e-02 -7.30009198e-01 4.53853965e-01
1.06402263e-01 -2.39788443e-01 3.42069209e-01 4.63267416e-01
-1.27089903e-01 1.26187801e+00 -7.58700728e-01 1.56201220e+00
-7.73314536e-01 5.55254281e-01 2.91172981e-01 -5.30388355e-01
5.79811513e-01 6.76993668e-01 -1.69316277e-01 -1.20262349e+00
-2.16919065e-01 3.11232120e-01 3.11635852e-01 -8.21665943e-01
7.04861462e-01 -5.38966581e-02 -1.61164686e-01 8.25779498e-01
1.28087163e-01 -1.12808041e-01 2.87203968e-01 6.46536648e-01
1.18686604e+00 2.07315832e-01 1.92183837e-01 -2.00668916e-01
5.26991010e-01 4.34141725e-01 1.17092729e-01 8.29681218e-01
2.11986959e-01 9.49735641e-02 3.40424538e-01 3.34634513e-01
-8.98511350e-01 -7.36870706e-01 -2.08451897e-01 1.75845778e+00
-3.51152122e-01 -4.68784750e-01 -7.35848129e-01 -7.13327229e-01
-2.63927162e-01 1.22671843e+00 -2.84964770e-01 -1.53496459e-01
-1.01841688e+00 -3.01358730e-01 4.13172990e-01 5.11057317e-01
9.94475931e-03 -1.45471025e+00 -4.96866286e-01 4.13519174e-01
-2.33974800e-01 -9.24383104e-01 -4.44252253e-01 2.55215377e-01
-1.10244346e+00 -8.40379775e-01 -5.16323745e-01 -8.58043969e-01
4.92646247e-01 1.14490382e-01 1.54248440e+00 5.62579334e-01
3.86346072e-01 9.44493711e-01 -6.59709513e-01 -3.28791291e-01
-7.24445283e-01 6.95746064e-01 -8.77130806e-01 -7.12096274e-01
5.01849949e-01 -2.40054652e-01 -3.27976644e-01 -3.46057229e-02
-1.12027574e+00 -3.57639819e-01 8.59284103e-01 7.33022690e-01
7.07790852e-02 -5.45453370e-01 7.54195809e-01 -1.23336089e+00
1.03943884e+00 -3.78792495e-01 -4.47114348e-01 6.27732038e-01
-4.69347268e-01 4.43262160e-01 2.66434729e-01 -1.33672625e-01
-1.38078344e+00 -2.44766489e-01 -5.29977381e-01 4.09585357e-01
-2.26839170e-01 9.30999398e-01 -1.15823664e-01 1.57133281e-01
7.25294650e-01 -3.91083695e-02 -1.00781627e-01 -8.75206411e-01
6.11371636e-01 7.19138384e-01 2.59304047e-01 -7.97053993e-01
7.28339076e-01 -1.07567295e-01 -6.75306380e-01 -6.86366439e-01
-1.00053978e+00 -1.01000595e+00 -4.89241064e-01 1.90632910e-01
6.94730341e-01 -9.08085823e-01 -4.19492990e-01 4.08104658e-02
-1.34867501e+00 -4.34213161e-01 -4.89654630e-01 3.33066106e-01
-6.55980706e-02 2.83817112e-01 -9.17462528e-01 -8.58578920e-01
-5.24169803e-01 -7.88899422e-01 1.10314059e+00 1.23220444e-01
-2.94876963e-01 -1.33422816e+00 2.28496611e-01 7.16807544e-01
4.55266148e-01 -6.89701676e-01 1.39109337e+00 -1.49231291e+00
-6.15654528e-01 -3.34006220e-01 -1.43727437e-01 2.96659201e-01
-2.97677755e-01 -3.15672338e-01 -9.04256701e-01 -9.30463150e-03
1.29145294e-01 -2.97563523e-01 7.82466352e-01 -1.39571493e-02
4.18949693e-01 -4.30587769e-01 -2.85611749e-01 3.47932801e-02
1.30417907e+00 2.03151792e-01 6.42023504e-01 4.93612468e-01
4.00874138e-01 1.15557611e+00 3.79586905e-01 -3.60712200e-01
8.39307368e-01 4.29847062e-01 -1.08140670e-02 9.64349136e-02
-1.83315024e-01 -3.76336128e-01 2.82384068e-01 9.44028437e-01
3.31175208e-01 -4.10864562e-01 -1.15940201e+00 8.72012138e-01
-1.76557088e+00 -7.88835406e-01 2.25399416e-02 2.13496661e+00
1.07791436e+00 -5.41730225e-02 -2.45697841e-01 -3.51319015e-01
4.33442771e-01 3.90862394e-03 -2.22160667e-01 -7.50528872e-01
2.69264996e-01 6.31643713e-01 4.06413168e-01 9.20047224e-01
-6.92839563e-01 1.44306469e+00 6.81719828e+00 4.41978484e-01
-7.22069919e-01 3.90976787e-01 3.74069452e-01 4.39518511e-01
-6.96076095e-01 3.87585580e-01 -1.04474878e+00 -8.48820060e-02
1.34096456e+00 2.86306366e-02 2.07744822e-01 6.22239947e-01
2.14758478e-02 -3.36276442e-01 -1.01634729e+00 1.24441825e-01
-6.16466142e-02 -1.02506793e+00 9.24830586e-02 -9.74741504e-02
1.67053878e-01 8.47242922e-02 -3.08482975e-01 7.71275520e-01
4.12769556e-01 -7.68284380e-01 7.36862898e-01 4.44810361e-01
2.43021891e-01 -4.26798910e-01 8.31830204e-01 5.41993558e-01
-7.75957108e-01 -9.76850167e-02 -2.56164223e-01 2.68056076e-02
2.46661469e-01 1.14798903e-01 -7.46082008e-01 4.48562443e-01
3.25357020e-01 1.16989374e-01 -9.41378117e-01 9.20067489e-01
-6.26756966e-01 7.29056001e-01 -2.50851125e-01 -2.62394845e-01
1.13399796e-01 3.54573391e-02 4.17649329e-01 1.34781444e+00
2.41267905e-02 3.67575049e-01 -1.66979581e-01 4.84857619e-01
-2.20897347e-01 5.79914927e-01 -3.10317457e-01 -2.72715002e-01
5.68640649e-01 1.23067379e+00 -4.67027634e-01 -3.55982453e-01
-6.56320035e-01 6.97093189e-01 6.18394017e-01 5.02423525e-01
-1.59960568e-01 -4.06840324e-01 1.10922270e-01 1.86499402e-01
3.47702384e-01 -4.10943955e-01 -7.93129802e-02 -1.13651371e+00
1.36459202e-01 -1.08587444e+00 6.36249304e-01 -1.04801762e+00
-1.12229407e+00 3.81643981e-01 4.08632420e-02 -3.54952812e-01
-6.66099191e-01 -7.51838923e-01 -2.56877989e-01 1.15603971e+00
-1.87353981e+00 -1.13186646e+00 4.07443166e-01 3.38542759e-01
4.22577322e-01 3.24592888e-01 8.75183940e-01 1.53904989e-01
-4.60870385e-01 3.48247766e-01 1.13939837e-01 5.44153929e-01
9.04012680e-01 -1.12281239e+00 4.49629903e-01 8.50592971e-01
3.81456017e-01 1.25580430e+00 2.43444338e-01 -5.53180397e-01
-1.32296503e+00 -7.16803610e-01 1.85502708e+00 -9.73295629e-01
8.63377333e-01 -3.02175999e-01 -1.18096709e+00 8.69420052e-01
3.41187805e-01 -4.21752602e-01 7.92255342e-01 6.10956907e-01
-3.88642788e-01 7.00183734e-02 -9.18383121e-01 4.70293432e-01
9.62121725e-01 -8.73372614e-01 -1.29919231e+00 5.01793921e-01
8.34484816e-01 -2.38351807e-01 -8.19801092e-01 -2.30199769e-02
3.41532886e-01 -5.12852848e-01 8.54626715e-01 -9.17750061e-01
1.05251342e-01 1.06244653e-01 -1.06239147e-01 -1.00171471e+00
-2.99787998e-01 -6.05915904e-01 1.87770024e-01 1.44582701e+00
1.07249618e+00 -9.39742684e-01 1.54652178e-01 1.17632997e+00
-3.75536025e-01 -3.65402907e-01 -8.30855846e-01 -4.53222632e-01
5.15279949e-01 -3.67967427e-01 3.74557674e-01 6.90344036e-01
1.71613142e-01 1.03777075e+00 4.95039493e-01 1.45522371e-01
-4.67068935e-03 -3.49160731e-02 4.28413808e-01 -1.24947500e+00
-2.86311954e-01 -2.18932360e-01 3.59558433e-01 -1.16869628e+00
2.95126110e-01 -1.06935513e+00 2.22324416e-01 -2.02892876e+00
-1.49407193e-01 -5.77165067e-01 -4.00981344e-02 6.40175760e-01
-4.27306920e-01 -2.71854222e-01 1.64056361e-01 4.75576133e-01
-7.07695603e-01 8.59641284e-02 1.12002289e+00 3.90679806e-01
-2.97373593e-01 -5.38869128e-02 -1.19008589e+00 5.92587113e-01
6.33872807e-01 -7.51747191e-01 -3.29022884e-01 -9.16410208e-01
4.34674740e-01 3.40801865e-01 1.74133673e-01 -4.83738959e-01
4.00168091e-01 2.68408302e-02 2.32577294e-01 -1.52326122e-01
1.56827435e-01 -7.47956216e-01 -4.11459059e-01 1.69999495e-01
-6.27285540e-01 3.32118750e-01 4.58966970e-01 1.78406999e-01
-4.82033752e-02 -8.52371931e-01 3.29942942e-01 -3.50715637e-01
-6.39392138e-01 -1.43182084e-01 -6.78534329e-01 5.32611668e-01
3.31299096e-01 1.47836097e-02 -5.31709731e-01 -4.92921025e-01
-5.97402632e-01 1.76906422e-01 2.84153283e-01 3.21262240e-01
6.76711425e-02 -5.89548349e-01 -4.83202040e-01 -1.12034462e-01
1.47687629e-01 -2.04824820e-01 -1.67610332e-01 5.67620158e-01
-3.30402285e-01 9.18074310e-01 4.95489866e-01 -4.87498529e-02
-8.88853431e-01 3.93248826e-01 3.02301496e-01 -7.98959851e-01
-1.70580283e-01 8.06724310e-01 1.18343588e-02 -6.82702482e-01
-3.77773978e-02 -2.09030360e-01 -6.50082469e-01 1.85040593e-01
3.04166049e-01 -9.28121358e-02 2.69721895e-01 -5.41184843e-01
-4.14308846e-01 3.61964107e-01 -4.36287165e-01 -5.16082823e-01
1.08671451e+00 -4.02974039e-01 -2.61614799e-01 2.95382649e-01
9.62437451e-01 4.54700768e-01 -5.69886208e-01 -3.66881877e-01
6.92768872e-01 2.72666607e-02 8.98742750e-02 -1.11731505e+00
-5.51396012e-01 9.19659376e-01 1.16194978e-01 2.87730426e-01
9.25424397e-01 3.29421103e-01 7.82097518e-01 9.36784208e-01
7.00717866e-02 -1.27244186e+00 -2.19079435e-01 8.63368273e-01
6.83496833e-01 -1.02398276e+00 -9.73668769e-02 -2.76178390e-01
-4.96999264e-01 9.62899625e-01 3.78913105e-01 1.59564123e-01
5.76724470e-01 1.64942458e-01 3.77423435e-01 -3.12452286e-01
-9.91950572e-01 -6.34468913e-01 4.03696686e-01 5.15942752e-01
1.00052774e+00 -4.28677559e-01 -6.46586239e-01 6.70696735e-01
-1.71733752e-01 -3.15912932e-01 1.25563070e-01 1.26420629e+00
-6.18863165e-01 -1.57645011e+00 -9.76648331e-02 3.98699433e-01
-9.09259200e-01 -7.44912207e-01 -3.92971575e-01 1.11613321e+00
-3.40145290e-01 1.27726316e+00 1.06196806e-01 3.16327780e-01
4.56358969e-01 7.37768590e-01 2.63464957e-01 -1.25738549e+00
-1.14125395e+00 1.88559666e-01 6.67685568e-01 -3.27293336e-01
-3.84754807e-01 -6.46197081e-01 -1.04955661e+00 1.64985806e-01
-7.65918851e-01 6.39360666e-01 8.32951069e-01 1.16914928e+00
6.96147978e-01 1.37818053e-01 1.40170276e-01 1.08098174e-02
-7.73899853e-01 -9.99079168e-01 -8.26716870e-02 1.21230133e-01
1.50748342e-01 -7.25828297e-03 -4.17907387e-01 1.11067770e-02] | [11.20077133178711, 8.039229393005371] |
b5651fcb-c090-4aad-b361-a03748e34d77 | task-oriented-hand-motion-retargeting-for | 1810.01845 | null | http://arxiv.org/abs/1810.01845v1 | http://arxiv.org/pdf/1810.01845v1.pdf | Task-Oriented Hand Motion Retargeting for Dexterous Manipulation Imitation | Human hand actions are quite complex, especially when they involve object
manipulation, mainly due to the high dimensionality of the hand and the vast
action space that entails. Imitating those actions with dexterous hand models
involves different important and challenging steps: acquiring human hand
information, retargeting it to a hand model, and learning a policy from
acquired data. In this work, we capture the hand information by using a
state-of-the-art hand pose estimator. We tackle the retargeting problem from
the hand pose to a 29 DoF hand model by combining inverse kinematics and PSO
with a task objective optimisation. This objective encourages the virtual hand
to accomplish the manipulation task, relieving the effect of the estimator's
noise and the domain gap. Our approach leads to a better success rate in the
grasping task compared to our inverse kinematics baseline, allowing us to
record successful human demonstrations. Furthermore, we used these
demonstrations to learn a policy network using generative adversarial imitation
learning (GAIL) that is able to autonomously grasp an object in the virtual
space. | ['Tae-Kyun Kim', 'Guillermo Garcia-Hernando', 'Dafni Antotsiou'] | 2018-10-03 | null | null | null | null | ['motion-retargeting'] | ['computer-vision'] | [ 1.05617821e-01 2.25157589e-01 -1.66963506e-02 3.01571518e-01
-3.91686440e-01 -7.80895174e-01 5.50304174e-01 -8.30610693e-01
-6.23655260e-01 8.06445658e-01 2.05560066e-02 6.59894720e-02
-2.51526117e-01 -2.41033062e-01 -1.01894641e+00 -7.31650889e-01
-2.38485299e-02 1.03043151e+00 4.41280678e-02 -2.81933039e-01
2.60961324e-01 8.70740533e-01 -1.38442528e+00 -2.01376721e-01
6.52989447e-01 5.52204490e-01 7.51511991e-01 8.66545618e-01
4.40131605e-01 6.81924284e-01 -5.86151481e-01 1.61225125e-02
6.89791918e-01 -3.19693208e-01 -7.76581526e-01 -4.37477464e-03
5.53369261e-02 -6.42286062e-01 -5.73036790e-01 8.34763765e-01
7.85818398e-01 4.65396166e-01 7.52743185e-01 -1.39159584e+00
-4.07604992e-01 6.56041741e-01 -3.61792892e-01 -2.57916421e-01
5.17969608e-01 7.54695594e-01 6.60273612e-01 -4.70398813e-01
1.06316638e+00 1.50899947e+00 3.60191226e-01 9.21279490e-01
-1.29331315e+00 -5.49758077e-01 -5.28318137e-02 2.59507000e-01
-1.18636763e+00 -8.70047957e-02 6.40097022e-01 -6.19868279e-01
8.90032530e-01 -5.03791645e-02 8.92504692e-01 1.83551419e+00
1.74905360e-01 9.41950858e-01 1.00895500e+00 -4.30788755e-01
1.64735690e-01 -1.85512558e-01 -5.63829660e-01 5.55331469e-01
-2.95568496e-01 6.08516634e-01 -2.05703154e-01 9.63816419e-02
1.20420325e+00 1.28964752e-01 -3.71745020e-01 -9.13630426e-01
-1.51950216e+00 4.00980026e-01 6.87684357e-01 1.66522995e-01
-1.00882566e+00 4.27582711e-01 9.80032831e-02 3.66491437e-01
-4.32138056e-01 1.02738774e+00 -6.52222514e-01 -5.22809744e-01
-2.65689135e-01 7.97713041e-01 1.08911085e+00 1.13577020e+00
1.76550224e-01 -8.89605209e-02 -4.04534996e-01 2.74138302e-01
2.30252817e-01 7.15272725e-01 1.08569324e-01 -1.32953477e+00
5.68045318e-01 2.47275382e-01 4.61054057e-01 -4.75914925e-01
-3.90644699e-01 -2.20968664e-01 -3.83237451e-01 9.00887728e-01
7.51586616e-01 -1.81836650e-01 -1.04072118e+00 1.89268780e+00
3.07947189e-01 -3.13472480e-01 -4.52572033e-02 1.40202582e+00
-7.54616708e-02 2.23359391e-01 1.13044880e-01 2.09010229e-01
1.02088571e+00 -9.28235829e-01 -5.51602304e-01 -2.24078536e-01
5.56739047e-02 -5.30242443e-01 1.29867494e+00 5.25998354e-01
-9.30233717e-01 -4.62549955e-01 -7.85125613e-01 1.77299157e-01
-4.07844894e-02 2.13873923e-01 4.60732043e-01 -8.84113461e-03
-5.20802855e-01 1.28013325e+00 -1.15720165e+00 -2.51483768e-01
3.63228142e-01 5.25174737e-01 -3.00073445e-01 1.43227652e-01
-8.63759339e-01 1.36371911e+00 5.82971990e-01 2.13609915e-03
-1.23186600e+00 -6.10692143e-01 -4.18371081e-01 -9.38197896e-02
8.31729054e-01 -7.12655485e-01 1.12744272e+00 -6.17344201e-01
-2.15766931e+00 3.87920767e-01 5.01732409e-01 -8.97441953e-02
1.10125256e+00 -4.45916176e-01 3.98293912e-01 1.30252372e-02
-6.71085566e-02 7.27089703e-01 1.46844006e+00 -1.36757600e+00
-1.75733835e-01 -5.26964486e-01 2.86627393e-02 1.93411112e-01
1.04992643e-01 -5.31269789e-01 -3.24588329e-01 -9.39666688e-01
1.63259692e-02 -1.45512879e+00 -1.27203956e-01 1.74876764e-01
-4.66404974e-01 -2.14724213e-01 6.74702227e-01 -8.84827435e-01
4.26733762e-01 -1.82651532e+00 1.38085163e+00 2.70110697e-01
1.22123249e-01 2.19667122e-01 -4.10511166e-01 4.38561589e-01
6.92835599e-02 -4.31205064e-01 4.18156385e-02 -1.19538106e-01
3.18896055e-01 3.54869723e-01 -3.91720265e-01 3.41715693e-01
-1.46941096e-02 1.21589994e+00 -1.10244989e+00 -1.01115257e-01
2.20219329e-01 5.08786023e-01 -6.73264742e-01 6.90710187e-01
-6.73293591e-01 1.19024897e+00 -6.49191976e-01 5.09487569e-01
1.34954527e-01 2.43092358e-01 1.49859741e-01 -2.65466213e-01
6.40552789e-02 -1.40504554e-01 -1.10758257e+00 2.09951615e+00
-5.90145886e-01 1.23353846e-01 3.53097051e-01 -6.61218643e-01
6.69784606e-01 2.12609753e-01 4.56769466e-01 -3.16503048e-01
5.89597225e-01 3.14737529e-01 3.12736958e-01 -7.95864582e-01
9.01424587e-02 4.53014784e-02 4.01700884e-01 5.99695683e-01
3.09276074e-01 -8.26961517e-01 -2.02450797e-01 -2.36739621e-01
1.04610491e+00 9.53240097e-01 -1.37017056e-01 1.68626513e-02
1.06883526e-01 -1.39223253e-02 -6.20410554e-02 6.80137336e-01
-5.37266843e-02 3.83482546e-01 3.34213614e-01 -1.26505882e-01
-1.39828646e+00 -1.13034606e+00 5.32455683e-01 8.86564314e-01
-1.80522189e-01 3.44358623e-01 -7.69208729e-01 -5.76236963e-01
4.60483253e-01 8.22262645e-01 -7.45640755e-01 -2.39492774e-01
-9.21130300e-01 7.69093707e-02 2.95524925e-01 7.39139020e-01
2.01674893e-01 -1.59117520e+00 -1.04643869e+00 2.29301766e-01
6.32155165e-02 -9.52553093e-01 -5.87904871e-01 1.62978023e-01
-6.30478919e-01 -1.17820668e+00 -1.30435848e+00 -5.96312881e-01
4.70838815e-01 -4.05981094e-01 5.88212669e-01 -3.99689347e-01
-7.38061249e-01 6.86328173e-01 -3.65475476e-01 -4.82151419e-01
-7.79781759e-01 2.72308230e-01 3.82115960e-01 -4.11837846e-01
-2.16264412e-01 -7.56599665e-01 -4.76508111e-01 2.82930493e-01
-5.26814461e-01 -2.17196032e-01 9.87703383e-01 1.07540524e+00
4.19153124e-01 -3.73721093e-01 3.09858531e-01 2.18167007e-01
7.33989060e-01 4.37190942e-03 -6.58532619e-01 2.11997807e-01
-4.13754642e-01 4.86714542e-01 5.50871432e-01 -1.14208210e+00
-8.79184365e-01 4.41356331e-01 -5.08968793e-02 -9.19594467e-01
4.00416143e-02 -1.95668668e-01 -8.30878615e-02 -1.68844894e-01
8.38294864e-01 1.31565765e-01 7.29516864e-01 -6.23880386e-01
7.60106564e-01 5.03472745e-01 6.32860124e-01 -7.80066073e-01
8.84550035e-01 2.49012798e-01 2.99683332e-01 -4.51289266e-01
-4.10452187e-01 -1.86828431e-02 -1.10341561e+00 -2.55874038e-01
8.77376676e-01 -3.79492730e-01 -1.42695963e+00 5.38789034e-01
-1.26760769e+00 -9.16632175e-01 -5.72134554e-01 7.30092049e-01
-1.31518877e+00 2.96579838e-01 -3.64751130e-01 -6.43315911e-01
-2.64017761e-01 -1.41088986e+00 1.08133256e+00 -1.25968829e-01
-4.01523173e-01 -4.35461700e-01 2.39023212e-02 -1.70065984e-02
4.10252213e-01 3.88600975e-01 6.45715356e-01 -2.95016944e-01
-5.38859606e-01 -2.75527090e-01 6.58209845e-02 4.18446273e-01
6.11444712e-02 -4.38099623e-01 -6.02964580e-01 -7.69519389e-01
-2.32068952e-02 -6.48606181e-01 4.38843906e-01 1.58912167e-01
1.01914346e+00 -2.38102257e-01 -4.05500889e-01 4.23140705e-01
8.71518612e-01 1.47218853e-01 3.09601277e-01 4.39457446e-01
7.43948281e-01 6.68059587e-01 6.85908973e-01 4.49391842e-01
-1.02643140e-01 1.01950240e+00 6.39285505e-01 4.88292158e-01
-2.14909613e-01 -4.61634189e-01 1.21047847e-01 1.37781471e-01
-7.01983631e-01 4.78409044e-03 -7.91918635e-01 2.86963612e-01
-1.81312025e+00 -6.87698483e-01 4.33289349e-01 1.97668028e+00
8.02523434e-01 -2.05632970e-01 2.53916234e-01 4.72287759e-02
3.05692941e-01 -3.33658755e-01 -9.97316241e-01 1.31313786e-01
3.89631331e-01 3.16134363e-01 4.64963257e-01 4.04467016e-01
-7.79950619e-01 1.05539477e+00 5.71512032e+00 5.28554142e-01
-1.08920169e+00 -1.79138750e-01 -5.78195333e-01 -2.56997585e-01
4.22413111e-01 -2.99391717e-01 -3.15101087e-01 3.16043586e-01
2.02129066e-01 7.99760744e-02 1.40923846e+00 9.41475332e-01
-1.13629557e-01 -7.52265193e-03 -1.46907091e+00 9.61036801e-01
-7.77092427e-02 -6.97053134e-01 -8.11248943e-02 2.09227093e-02
4.62218732e-01 -3.59807312e-02 3.72013748e-02 4.20984745e-01
3.20208788e-01 -1.03021538e+00 9.60033953e-01 7.93260753e-01
6.76619112e-01 -4.48510677e-01 2.59914339e-01 7.39571095e-01
-5.71814716e-01 -3.79582167e-01 -2.04793867e-02 -1.00418232e-01
2.96735525e-01 -3.22534949e-01 -1.02354443e+00 2.29898602e-01
5.96364558e-01 3.48788708e-01 1.04866475e-01 4.94326174e-01
-5.63742697e-01 -1.55757278e-01 -4.36358929e-01 -2.95278460e-01
-9.10088420e-03 -1.56347588e-01 1.22041667e+00 4.89791423e-01
1.00225791e-01 6.69731721e-02 3.06368172e-01 1.41842747e+00
1.26661703e-01 -3.48245651e-01 -5.50062895e-01 -2.97134638e-01
1.55954853e-01 8.71761143e-01 -3.60009968e-01 -1.67747691e-01
4.89871562e-01 1.41652644e+00 4.18038726e-01 3.87114912e-01
-5.93432486e-01 -5.38508296e-01 6.20973647e-01 -2.19127774e-01
3.89317393e-01 -5.36456347e-01 1.31108731e-01 -1.06345212e+00
3.09947908e-01 -1.11041903e+00 -3.95606667e-01 -7.84821689e-01
-1.07277989e+00 4.44555312e-01 8.04445967e-02 -9.73290086e-01
-7.66954124e-01 -8.23337734e-01 -1.06663942e-01 1.12815726e+00
-9.68153775e-01 -8.52359831e-01 -4.84569103e-01 5.81213355e-01
4.57304955e-01 -2.66845345e-01 7.74929225e-01 -1.21132493e-01
5.78387966e-03 2.93792337e-01 -1.57544494e-01 -2.56999815e-03
6.89531744e-01 -1.27399528e+00 4.18969899e-01 1.26052365e-01
-2.39779770e-01 5.12635112e-01 8.95930111e-01 -5.67419171e-01
-1.98464537e+00 -4.46407378e-01 -1.27192698e-02 -9.10242438e-01
7.58501232e-01 -2.85020709e-01 -6.99246883e-01 7.70758212e-01
-2.80080020e-01 -1.29565388e-01 -3.86390835e-01 -3.73594761e-01
-4.05154303e-02 3.97891015e-01 -1.26194489e+00 7.63191521e-01
1.35393262e+00 -3.10435444e-01 -9.57650244e-01 4.62709785e-01
4.27774638e-01 -8.79461527e-01 -8.98822427e-01 3.22584152e-01
1.13973212e+00 -2.33524948e-01 1.15530622e+00 -1.10301161e+00
4.20718461e-01 -9.09304619e-02 -1.74074546e-02 -1.81139088e+00
-2.71316946e-01 -6.63298965e-01 -4.22668219e-01 6.58216178e-01
-1.21607527e-01 -5.61057150e-01 7.30456412e-01 3.53306681e-01
3.74933988e-01 -6.16214573e-01 -9.52138901e-01 -1.30624259e+00
2.67999619e-01 4.17618565e-02 5.34651995e-01 5.54642797e-01
-4.00369316e-02 7.76998475e-02 -4.56431359e-01 7.82794133e-02
8.19691360e-01 8.07260796e-02 1.25448883e+00 -1.03804636e+00
-7.92694092e-01 -4.47823524e-01 -1.93060160e-01 -1.19686604e+00
5.97256958e-01 -8.29717755e-01 5.85679531e-01 -1.24712312e+00
-8.22636634e-02 -4.91254240e-01 2.41162494e-01 4.30325538e-01
-1.26922801e-01 -3.14193368e-01 7.24844813e-01 4.04677570e-01
3.30523588e-02 6.78188860e-01 1.96230042e+00 -2.36115605e-01
-4.97729987e-01 2.73277164e-01 1.13261282e-01 4.69962537e-01
6.83931410e-01 -2.88062841e-01 -7.66301528e-02 -4.71685618e-01
-2.03504235e-01 3.93588960e-01 7.82738924e-01 -8.29427183e-01
-4.44271192e-02 -1.80068582e-01 3.67515504e-01 -7.08950013e-02
5.58058918e-01 -1.01532769e+00 1.69365987e-01 1.01478708e+00
-3.76369417e-01 -2.07657605e-01 1.54221043e-01 5.99169254e-01
3.11046153e-01 -1.67326361e-01 6.35069251e-01 -2.19081149e-01
-6.44577503e-01 2.43387461e-01 -1.49204195e-01 -7.67260715e-02
1.14459097e+00 1.75174162e-01 1.76643461e-01 -2.14505956e-01
-1.20445788e+00 2.71581620e-01 5.02433836e-01 7.64566898e-01
2.20914960e-01 -9.13995981e-01 -5.86630881e-01 2.51787215e-01
-2.38598421e-01 -6.22201599e-02 -7.01905265e-02 8.08123112e-01
-3.45942378e-01 2.57434428e-01 -6.24453664e-01 -7.06728458e-01
-1.01053965e+00 8.77045691e-01 4.05506164e-01 -9.04147476e-02
-9.97004509e-01 6.96272790e-01 -3.01026106e-01 -8.00004601e-01
6.41835213e-01 -2.75447637e-01 1.07628502e-01 -3.50877494e-01
-4.78024082e-03 7.77108550e-01 -3.72730762e-01 -1.48581862e-01
-1.44610122e-01 5.99347651e-01 3.28075260e-01 -2.66060233e-01
1.40547955e+00 3.45566779e-01 -1.26694879e-02 -2.14270689e-02
9.04676139e-01 -3.70396316e-01 -1.80192244e+00 -2.34703533e-02
-3.77093494e-01 -4.81249213e-01 -2.76861906e-01 -1.17900622e+00
-6.86071515e-01 8.84768367e-01 7.31237233e-01 -3.75584602e-01
5.65201938e-01 1.48458481e-01 4.92192864e-01 1.00438595e+00
7.99708068e-01 -1.11943233e+00 5.07336617e-01 6.60350859e-01
1.75408483e+00 -1.19342434e+00 -4.43349369e-02 -1.44032493e-01
-6.40071869e-01 1.16370976e+00 5.23776591e-01 -3.92722219e-01
5.14355600e-01 2.89069593e-01 9.85916778e-02 -2.95096468e-02
-1.84332520e-01 -8.14169049e-02 2.68390059e-01 8.10415745e-01
-3.83428484e-01 2.02701122e-01 2.48220731e-02 2.47238502e-01
-2.95606166e-01 3.25492680e-01 6.19861074e-02 1.25375235e+00
-1.02821924e-01 -1.09631443e+00 -4.23595548e-01 1.53737381e-01
-1.00371338e-01 4.87748384e-01 -3.36692065e-01 1.01867247e+00
-1.79972917e-01 2.63480574e-01 -2.45904535e-01 -3.45679492e-01
8.00608635e-01 7.11082071e-02 1.31915653e+00 -4.83624309e-01
-3.85088265e-01 -2.66206145e-01 -5.91102362e-01 -9.86213565e-01
-6.44050911e-02 -7.22060800e-01 -1.19889379e+00 4.03558239e-02
-2.26031065e-01 -2.65868038e-01 1.01001930e+00 9.50947344e-01
2.60269612e-01 9.12025273e-01 2.45164663e-01 -1.92252386e+00
-1.67974329e+00 -1.24226701e+00 -5.60960650e-01 3.85394275e-01
4.49669063e-01 -1.30090928e+00 -2.46428907e-01 -3.29959631e-01] | [4.743877410888672, 0.5872628688812256] |
4d19def0-0e2b-418b-9942-ae3952229b1b | learning-attraction-field-representation-for | 1812.02122 | null | http://arxiv.org/abs/1812.02122v2 | http://arxiv.org/pdf/1812.02122v2.pdf | Learning Attraction Field Representation for Robust Line Segment Detection | This paper presents a region-partition based attraction field dual
representation for line segment maps, and thus poses the problem of line
segment detection (LSD) as the region coloring problem. The latter is then
addressed by learning deep convolutional neural networks (ConvNets) for
accuracy, robustness and efficiency. For a 2D line segment map, our dual
representation consists of three components: (i) A region-partition map in
which every pixel is assigned to one and only one line segment; (ii) An
attraction field map in which every pixel in a partition region is encoded by
its 2D projection vector w.r.t. the associated line segment; and (iii) A
squeeze module which squashes the attraction field to a line segment map that
almost perfectly recovers the input one. By leveraging the duality, we learn
ConvNets to compute the attraction field maps for raw in-put images, followed
by the squeeze module for LSD, in an end-to-end manner. Our method rigorously
addresses several challenges in LSD such as local ambiguity and class
imbalance. Our method also harnesses the best practices developed in ConvNets
based semantic segmentation methods such as the encoder-decoder architecture
and the a-trous convolution. In experiments, our method is tested on the
WireFrame dataset and the YorkUrban dataset with state-of-the-art performance
obtained. Especially, we advance the performance by 4.5 percents on the
WireFrame dataset. Our method is also fast with 6.6~10.4 FPS, outperforming
most of existing line segment detectors. | ['Fu-Dong Wang', 'Gui-Song Xia', 'Song Bai', 'Nan Xue', 'Liangpei Zhang', 'Tianfu Wu'] | 2018-12-05 | learning-attraction-field-representation-for-1 | http://openaccess.thecvf.com/content_CVPR_2019/html/Xue_Learning_Attraction_Field_Representation_for_Robust_Line_Segment_Detection_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Xue_Learning_Attraction_Field_Representation_for_Robust_Line_Segment_Detection_CVPR_2019_paper.pdf | cvpr-2019-6 | ['line-segment-detection'] | ['computer-vision'] | [ 4.04738933e-01 8.10367987e-02 -5.23950875e-01 -3.15687537e-01
-7.69251585e-01 -7.20780373e-01 1.14818394e-01 1.81792572e-01
-4.50240999e-01 4.07859534e-01 -5.70558250e-01 -4.89968002e-01
2.54255146e-01 -9.77697790e-01 -1.13072658e+00 -4.20021981e-01
7.70287355e-03 2.19045311e-01 5.21765113e-01 -1.92317396e-01
4.09331799e-01 7.06485331e-01 -1.11152124e+00 3.47523481e-01
9.04937625e-01 1.61512315e+00 -4.33439240e-02 6.14980400e-01
-3.79565686e-01 5.42143881e-01 -4.09617633e-01 -2.75176436e-01
6.15097463e-01 -3.15877683e-02 -8.01747084e-01 1.87724769e-01
7.25885689e-01 -5.71940541e-01 -4.67877120e-01 9.59520519e-01
2.66752869e-01 -1.76199540e-01 3.09607387e-01 -1.24266148e+00
-2.92893916e-01 5.02678216e-01 -1.30267859e+00 2.44818434e-01
8.11159313e-02 -5.57703637e-02 8.20244908e-01 -8.70019853e-01
6.14611626e-01 8.76947522e-01 7.81761467e-01 1.75171196e-01
-1.03211462e+00 -6.48117483e-01 3.13385762e-02 -2.39289269e-01
-1.46049488e+00 -1.39086479e-02 7.64924407e-01 -3.46943736e-01
8.13285649e-01 1.04744896e-01 7.79833198e-01 5.05837321e-01
1.83898136e-01 1.19502389e+00 7.88838744e-01 -3.25241625e-01
3.77237760e-02 -1.55487567e-01 3.65207434e-01 9.07444537e-01
1.34409934e-01 -2.17487395e-01 -4.72421259e-01 4.35422003e-01
1.16038215e+00 -8.77826735e-02 -3.02264094e-01 -5.70229053e-01
-1.07162678e+00 5.68875968e-01 7.98444867e-01 -2.29968429e-01
-1.35028124e-01 3.65076900e-01 3.90314400e-01 -2.06946898e-02
3.80054921e-01 5.56644127e-02 -4.10034955e-01 8.85311961e-02
-1.37861335e+00 2.42522418e-01 6.81830287e-01 1.07742929e+00
1.08899522e+00 -1.85005665e-01 -2.68737316e-01 7.56133199e-01
2.07745284e-01 4.27653223e-01 -5.08877710e-02 -7.54129767e-01
9.38738346e-01 7.44004011e-01 -2.00357854e-01 -9.16931510e-01
-6.30878985e-01 -7.19393909e-01 -7.39801943e-01 3.99705768e-01
5.28817117e-01 -2.64870465e-01 -1.21807396e+00 1.42208636e+00
3.31119001e-01 1.50476769e-01 -1.01011328e-01 8.89681160e-01
6.50283873e-01 7.24150479e-01 -3.38612467e-01 3.04424405e-01
1.59036207e+00 -1.38976455e+00 -3.35782170e-01 -4.92364436e-01
6.01477146e-01 -8.03428590e-01 7.11360991e-01 4.61993963e-01
-1.21582305e+00 -5.60128570e-01 -1.42333555e+00 -4.55526590e-01
-2.89118677e-01 5.97510397e-01 7.38887489e-01 5.09308815e-01
-9.73441601e-01 5.32922208e-01 -8.26886892e-01 -5.60653023e-02
8.76177311e-01 5.32558441e-01 -2.59797722e-01 1.72074139e-03
-8.28285217e-01 2.25572959e-01 6.10384226e-01 3.76241744e-01
-2.80134916e-01 -8.82603109e-01 -1.04187286e+00 2.79272854e-01
4.58876461e-01 -4.04650956e-01 1.05454659e+00 -9.10804331e-01
-1.10875058e+00 9.90751743e-01 4.55807932e-02 -7.29418039e-01
7.06381440e-01 -3.43091637e-01 -3.20429504e-01 4.45448726e-01
4.24082339e-01 1.02025092e+00 5.72538912e-01 -1.09951007e+00
-1.15839219e+00 -2.77506649e-01 8.64522979e-02 9.48573425e-02
5.97088523e-02 -2.39164859e-01 -1.21249604e+00 -6.74967408e-01
4.60782111e-01 -8.15392435e-01 -7.66850263e-02 4.77641970e-01
-1.07219696e+00 -4.37766351e-02 1.07160103e+00 -4.99183238e-01
1.29773569e+00 -2.20340466e+00 -4.47240502e-01 5.77280760e-01
4.04624760e-01 3.87064606e-01 1.03286572e-01 1.36044830e-01
-2.75466055e-01 9.82752517e-02 -4.77802545e-01 -3.40440363e-01
-1.70193434e-01 -8.00969228e-02 -4.94953573e-01 6.67196631e-01
4.66055125e-01 9.76494312e-01 -5.71685672e-01 -4.90600139e-01
2.95884252e-01 2.37131953e-01 -4.18597013e-01 -1.18621968e-01
6.98005129e-03 -2.40290821e-01 -3.09649259e-01 7.45626926e-01
1.30316961e+00 -3.24775755e-01 -1.47063047e-01 -4.97759402e-01
-6.02530241e-01 4.59125359e-03 -1.22640276e+00 2.00032425e+00
1.47699177e-01 9.78302121e-01 -3.07333291e-01 -8.46390665e-01
1.37310946e+00 -3.05743188e-01 6.40719175e-01 -9.97778654e-01
8.61003548e-02 2.30604246e-01 -2.66502857e-01 -7.88657218e-02
7.24720776e-01 5.84620297e-01 -3.03111374e-01 2.89017409e-01
2.10305620e-02 1.26458451e-01 4.05745983e-01 2.84218401e-01
8.45654786e-01 4.45741475e-01 -1.76723242e-01 -4.23123777e-01
3.05814326e-01 2.38641515e-01 7.58323193e-01 7.87192523e-01
-1.87950060e-01 8.36857259e-01 9.74623024e-01 -5.23860216e-01
-1.11613262e+00 -1.24118018e+00 -3.54352653e-01 4.80995595e-01
7.53635943e-01 -3.46623510e-01 -1.15077722e+00 -6.20703638e-01
-4.17141467e-02 2.58755207e-01 -5.38110673e-01 -1.87582113e-02
-7.96151996e-01 -5.38635850e-01 8.18175912e-01 9.43667412e-01
1.21658313e+00 -5.46733558e-01 -8.01071882e-01 1.57639489e-01
5.96243516e-02 -1.28670776e+00 -7.00491607e-01 5.74291766e-01
-6.40453815e-01 -1.08856177e+00 -7.11842120e-01 -1.04410708e+00
7.07977593e-01 4.36988235e-01 8.64831209e-01 -1.30549073e-01
-2.98161626e-01 -2.48071805e-01 -9.47747007e-02 -1.83555752e-01
3.14109355e-01 2.68779516e-01 -4.84917194e-01 2.10530367e-02
2.81783998e-01 -1.11797586e-01 -9.76619124e-01 3.37196469e-01
-8.21825206e-01 6.10541821e-01 5.89998543e-01 6.90947890e-01
9.37491536e-01 -1.95963338e-01 4.12197635e-02 -9.98698413e-01
1.20799705e-01 -2.65528768e-01 -9.66062307e-01 2.49794707e-01
-3.51237506e-01 -3.72568518e-03 3.61826688e-01 1.13435499e-01
-7.31513202e-01 5.20883858e-01 -2.18757793e-01 -3.58073205e-01
-6.99390396e-02 2.11283490e-01 6.94179013e-02 -2.53575414e-01
4.66835529e-01 2.38380097e-02 -3.17205817e-01 -2.96189606e-01
4.24121022e-01 4.64352429e-01 1.13472533e+00 -3.89705062e-01
8.29277456e-01 6.93860292e-01 8.18309337e-02 -5.87556601e-01
-8.10092866e-01 -5.41241348e-01 -8.80119920e-01 -1.97060660e-01
1.06709898e+00 -9.42586601e-01 -6.33527875e-01 6.52966738e-01
-1.25124359e+00 -3.42794716e-01 -2.66649693e-01 9.19475704e-02
-5.19031405e-01 3.02000850e-01 -7.44829476e-01 -4.19350415e-01
-3.72810721e-01 -1.41040969e+00 1.29343486e+00 7.66796529e-01
1.80674151e-01 -5.77889085e-01 -2.35401958e-01 1.89949632e-01
-2.65035741e-02 6.20821834e-01 9.10723090e-01 -3.60217750e-01
-9.48761761e-01 -1.68629393e-01 -7.91983366e-01 2.21692488e-01
-4.48781967e-01 1.76629633e-01 -9.58877325e-01 -4.46063615e-02
-6.35606527e-01 -1.59134999e-01 1.08287227e+00 4.52587813e-01
1.33637261e+00 1.04494989e-01 -5.98803520e-01 1.26546872e+00
1.66438425e+00 3.59092057e-01 8.80709589e-01 5.41362405e-01
9.43878651e-01 2.99262822e-01 6.04408920e-01 3.33668798e-01
4.05116141e-01 6.91247404e-01 4.14730698e-01 -9.69932377e-01
-2.46057793e-01 -2.38532990e-01 -1.22704379e-01 1.27368227e-01
3.08405101e-01 -5.75370729e-01 -9.62950230e-01 3.94434154e-01
-1.88603640e+00 -6.65221930e-01 -5.95845401e-01 1.94668686e+00
4.67477798e-01 4.01401758e-01 4.86165099e-02 2.28222936e-01
7.40455866e-01 1.88158646e-01 -7.49733746e-01 -5.10930181e-01
-4.09095168e-01 2.88103878e-01 1.10285199e+00 1.35710418e-01
-1.61408949e+00 1.09089661e+00 5.23798800e+00 1.03226590e+00
-1.20169246e+00 -4.77509350e-01 1.18784380e+00 3.62089872e-01
-1.87200993e-01 2.30273381e-02 -9.54599679e-01 3.22210789e-01
1.44252136e-01 5.17310858e-01 7.96530247e-02 9.07772720e-01
-4.89254333e-02 -4.24963355e-01 -9.06845033e-01 1.19789326e+00
-7.19577214e-03 -1.62560284e+00 -2.79267341e-01 -1.63694622e-03
8.73116374e-01 1.16336308e-01 1.76984340e-01 1.80239864e-02
-1.21855728e-01 -9.26486313e-01 1.08964384e+00 1.87845141e-01
1.24612939e+00 -9.82438624e-01 4.21563864e-01 1.79305971e-02
-1.38747656e+00 1.45952165e-01 -1.33113325e-01 4.20166820e-01
1.41437128e-01 6.10335231e-01 -8.16641986e-01 5.72154224e-01
7.68338025e-01 6.70421958e-01 -4.02478516e-01 1.34056902e+00
-2.43503511e-01 4.91996199e-01 -5.40777445e-01 4.28331673e-01
6.75310373e-01 -2.79148668e-01 1.28368199e-01 1.41897786e+00
1.23592362e-01 -2.82171696e-01 8.53145123e-02 1.11984777e+00
-2.31572732e-01 -1.92549229e-01 -3.72084141e-01 3.27837497e-01
4.37666684e-01 1.33864951e+00 -1.37151623e+00 -1.68149382e-01
-4.70037043e-01 1.27357304e+00 3.44066262e-01 4.92512017e-01
-9.73526359e-01 -9.52056110e-01 6.32701695e-01 -1.06080137e-01
8.14174891e-01 -1.77430764e-01 -6.49765074e-01 -8.14200044e-01
2.33711988e-01 -2.91266799e-01 1.09754950e-01 -6.25228763e-01
-6.75693154e-01 6.48573935e-01 -3.51782829e-01 -1.25759161e+00
-5.04717082e-02 -6.46396875e-01 -5.31325340e-01 8.90796363e-01
-1.89669788e+00 -1.09834707e+00 -5.44241190e-01 4.68974352e-01
4.40065920e-01 2.12570190e-01 2.58202642e-01 4.89387006e-01
-9.15131569e-01 8.30124736e-01 2.27683395e-01 7.04542160e-01
3.13717961e-01 -1.27326369e+00 9.37491059e-01 1.11370313e+00
3.41320001e-02 2.46835291e-01 6.96034655e-02 -5.64250171e-01
-1.17591262e+00 -1.25838709e+00 4.51697230e-01 1.81378111e-01
2.56942213e-01 -6.81821346e-01 -7.75202751e-01 6.14261687e-01
-4.25602756e-02 4.08729494e-01 3.39179337e-01 -4.80607897e-01
-2.44279861e-01 -2.39909604e-01 -9.84402895e-01 4.52952445e-01
1.04101014e+00 -3.36514026e-01 1.05676136e-03 1.55544072e-01
7.89537609e-01 -1.06020617e+00 -5.67095459e-01 3.00758541e-01
5.36743104e-01 -1.16086972e+00 1.14070439e+00 -2.51401544e-01
6.87020838e-01 -6.43347204e-01 1.41380906e-01 -7.63293684e-01
-1.86026365e-01 -5.86458385e-01 8.27952102e-02 1.20454478e+00
4.85587507e-01 -2.79160738e-01 1.01156986e+00 4.42379504e-01
-5.60686409e-01 -1.25864041e+00 -8.38732958e-01 -3.96671087e-01
-2.41984904e-01 -6.04894400e-01 6.32414579e-01 6.61678493e-01
-5.13212979e-01 6.38322309e-02 -1.41609743e-01 3.19304556e-01
5.32800913e-01 3.71605784e-01 7.27698445e-01 -7.92025566e-01
-3.17894369e-02 -5.80802143e-01 -4.58213001e-01 -1.80518770e+00
-1.90978110e-01 -8.65094006e-01 1.26186505e-01 -1.44208944e+00
-8.86975005e-02 -6.65784478e-01 -1.84992120e-01 5.11466622e-01
6.17334023e-02 5.48345983e-01 2.98076063e-01 2.79695652e-02
-6.55312359e-01 9.80086625e-02 1.11044645e+00 -1.83624893e-01
-2.81967849e-01 -2.21201420e-01 -6.48963213e-01 8.37312937e-01
7.14644074e-01 -2.99724162e-01 -1.79165915e-01 -6.66419387e-01
1.34626195e-01 8.03035796e-02 4.96508062e-01 -1.20982099e+00
5.63624799e-01 2.75872976e-01 7.50565946e-01 -1.24443793e+00
2.47044846e-01 -6.25095606e-01 -3.82476181e-01 3.09607625e-01
-1.38645381e-01 8.39191824e-02 4.89565015e-01 4.11084861e-01
-2.13954464e-01 -1.74568102e-01 7.09075809e-01 3.17783177e-01
-1.38982368e+00 5.13370454e-01 8.21271837e-02 4.66426872e-02
1.31472003e+00 -7.55242944e-01 -6.25123739e-01 1.01758070e-01
-1.06083855e-01 5.46053886e-01 3.85056138e-01 1.85611576e-01
7.96857119e-01 -1.19627941e+00 -4.81189042e-01 6.42256141e-01
8.78235549e-02 5.78444839e-01 3.00765842e-01 7.52486587e-01
-1.31276429e+00 4.88988727e-01 -2.73407012e-01 -8.88053656e-01
-8.77409279e-01 2.19380036e-01 5.02942383e-01 -9.04151127e-02
-8.02536547e-01 1.12945747e+00 4.95707422e-01 -2.96897460e-02
2.08701193e-01 -6.72858775e-01 1.96292344e-02 -1.79681242e-01
4.00669008e-01 2.63754219e-01 1.10734381e-01 -5.34640670e-01
-3.83271873e-01 1.04213440e+00 -2.77016759e-01 6.66082650e-02
1.07450700e+00 9.28215906e-02 1.01353079e-01 -6.74919188e-02
1.49525297e+00 -2.41249934e-01 -1.76476336e+00 -1.04459696e-01
-1.26146004e-01 -4.20060128e-01 2.96589166e-01 -6.73469722e-01
-1.49387741e+00 7.92367339e-01 6.13092065e-01 -7.41422027e-02
1.26249254e+00 -1.87199995e-01 1.21291912e+00 3.22345197e-02
1.59405380e-01 -1.17540097e+00 -2.78168827e-01 4.18816924e-01
2.93998808e-01 -1.08411455e+00 -1.21774599e-01 -7.57374644e-01
-3.89343739e-01 1.66677213e+00 6.65949821e-01 -4.17292982e-01
3.81152034e-01 6.96237206e-01 -6.14297437e-03 -2.87517279e-01
-6.68327659e-02 -1.53137058e-01 3.90221983e-01 3.45626920e-01
6.40014634e-02 -3.39171104e-03 -5.53614907e-02 3.51246089e-01
-1.29905716e-01 -2.08001044e-02 2.76262432e-01 8.74931455e-01
-3.55049312e-01 -8.35658908e-01 -3.67836505e-01 4.76230651e-01
-2.03918338e-01 -6.91601485e-02 -2.58993395e-02 1.00778437e+00
4.08859074e-01 5.65761745e-01 7.63754964e-01 -4.57577586e-01
4.69512612e-01 -5.27594924e-01 6.48396760e-02 -1.82862714e-01
-4.67414260e-01 2.15991840e-01 -9.51591432e-02 -9.22291160e-01
-6.46612328e-03 -4.52359170e-01 -1.78637850e+00 -2.10837722e-01
-3.18963617e-01 -2.03861952e-01 6.99325681e-01 7.50478864e-01
2.90499896e-01 6.98428929e-01 5.53073227e-01 -5.96549213e-01
-5.79364970e-03 -2.97252536e-01 -5.81916630e-01 3.78762349e-03
4.55033749e-01 -3.48938555e-01 1.95783913e-01 -3.40857506e-02] | [8.313459396362305, -1.5467902421951294] |
dbef2cd9-ef65-420b-b096-77e7c95b81ae | over-the-air-membership-inference-attacks-as | 2006.14576 | null | https://arxiv.org/abs/2006.14576v1 | https://arxiv.org/pdf/2006.14576v1.pdf | Over-the-Air Membership Inference Attacks as Privacy Threats for Deep Learning-based Wireless Signal Classifiers | This paper presents how to leak private information from a wireless signal classifier by launching an over-the-air membership inference attack (MIA). As machine learning (ML) algorithms are used to process wireless signals to make decisions such as PHY-layer authentication, the training data characteristics (e.g., device-level information) and the environment conditions (e.g., channel information) under which the data is collected may leak to the ML model. As a privacy threat, the adversary can use this leaked information to exploit vulnerabilities of the ML model following an adversarial ML approach. In this paper, the MIA is launched against a deep learning-based classifier that uses waveform, device, and channel characteristics (power and phase shifts) in the received signals for RF fingerprinting. By observing the spectrum, the adversary builds first a surrogate classifier and then an inference model to determine whether a signal of interest has been used in the training data of the receiver (e.g., a service provider). The signal of interest can then be associated with particular device and channel characteristics to launch subsequent attacks. The probability of attack success is high (more than 88% depending on waveform and channel conditions) in identifying signals of interest (and potentially the device and channel information) used to build a target classifier. These results show that wireless signal classifiers are vulnerable to privacy threats due to the over-the-air information leakage of their ML models | ['Yalin E. Sagduyu', 'Kemal Davaslioglu', 'Yi Shi'] | 2020-06-25 | null | null | null | null | ['membership-inference-attack'] | ['computer-vision'] | [ 8.11921299e-01 1.41851991e-01 -1.59401134e-01 -6.99452311e-02
-1.00610387e+00 -1.14413989e+00 1.50343418e-01 2.31858805e-01
-1.25662088e-01 4.69121844e-01 -5.96117318e-01 -9.29447293e-01
-5.19478545e-02 -1.10463119e+00 -8.31147373e-01 -1.09838021e+00
-7.24886954e-01 -2.67335594e-01 -2.72408962e-01 2.25936219e-01
3.50994430e-02 9.06542480e-01 -9.26660538e-01 4.08171535e-01
-3.15044634e-02 1.63372159e+00 -3.79912823e-01 9.32760298e-01
2.86098659e-01 3.52093369e-01 -1.33373082e+00 -2.28809863e-01
3.78268689e-01 -1.68940231e-01 -3.30301762e-01 -9.29667890e-01
-1.03729360e-01 -2.34478369e-01 -4.07354921e-01 1.02093208e+00
6.92927122e-01 -5.70064127e-01 2.53461659e-01 -1.29681849e+00
5.71823716e-01 8.54368687e-01 2.06070125e-01 -9.33380723e-02
5.09617090e-01 1.71650015e-02 5.60309410e-01 -1.98015317e-01
1.69154704e-01 6.96759641e-01 8.63705158e-01 1.62963182e-01
-1.27168000e+00 -1.51482427e+00 -4.35805321e-01 -3.58161002e-01
-1.55460501e+00 -3.40055615e-01 9.79932547e-01 -1.49962813e-01
1.20693401e-01 7.50020504e-01 2.65059531e-01 1.58042741e+00
5.14591932e-01 3.24031323e-01 1.09248471e+00 -4.65968847e-01
6.89518094e-01 5.04655540e-01 -1.58654436e-01 4.57154840e-01
2.39734739e-01 9.15053964e-01 -3.31835598e-01 -1.13279510e+00
1.55441776e-01 -3.92072976e-01 -2.69917786e-01 -1.90077156e-01
-7.56233275e-01 5.72305620e-01 2.89428622e-01 1.96460575e-01
-3.64605993e-01 3.30996782e-01 2.67283976e-01 6.75326169e-01
-2.65870571e-01 7.73442745e-01 -5.06992877e-01 2.08530754e-01
-7.55453289e-01 4.94352095e-02 1.11538172e+00 6.99956834e-01
7.08334923e-01 1.54004350e-01 -1.30082324e-01 -7.81156048e-02
3.04259956e-01 1.01372707e+00 5.74998446e-02 -3.29437703e-01
5.19163370e-01 -2.77058721e-01 1.82099059e-01 -1.34171915e+00
-4.64575857e-01 -7.71618366e-01 -4.71018106e-01 2.45941192e-01
4.11158234e-01 -7.90026546e-01 -4.15139377e-01 1.87453878e+00
1.64199740e-01 4.29811001e-01 5.09358466e-01 2.72255033e-01
5.48409045e-01 6.31534398e-01 1.80622116e-01 4.85245697e-02
1.28439701e+00 2.23189130e-01 -3.76001805e-01 -3.84732634e-02
7.88362265e-01 -5.57917356e-01 1.92286164e-01 3.73862505e-01
-5.83682120e-01 -5.78748703e-01 -1.59467351e+00 1.07104039e+00
-5.52778721e-01 -1.55840948e-01 4.56609875e-01 1.58197582e+00
-3.36090595e-01 4.10581082e-01 -3.64522457e-01 9.67096016e-02
2.94285119e-01 8.37486625e-01 -9.29491743e-02 7.03661889e-02
-1.78072011e+00 2.63591766e-01 2.81789541e-01 1.20299160e-01
-1.08132064e+00 -9.41362083e-01 -6.85165644e-01 -2.23795068e-03
-2.79698428e-02 -2.64041692e-01 8.11981499e-01 -7.73557603e-01
-1.18152308e+00 4.37645555e-01 2.89903373e-01 -7.67957330e-01
3.47350985e-01 3.54924500e-01 -1.38676417e+00 2.88859040e-01
-3.33920568e-01 -3.54435176e-01 1.21417594e+00 -1.24666107e+00
-6.65185213e-01 -1.52414113e-01 2.50388592e-01 -6.29992247e-01
-4.49665897e-02 8.30603123e-04 3.24707329e-01 -6.26332045e-01
1.70918331e-01 -1.18788302e+00 9.74457264e-02 -2.82410771e-01
-7.56236851e-01 8.41203570e-01 1.10347664e+00 -6.07559443e-01
1.14134920e+00 -2.36048579e+00 -7.08640337e-01 1.31862366e+00
-2.06257030e-01 2.27992326e-01 2.27217376e-01 5.60135901e-01
-1.27973408e-01 4.00849581e-01 -1.15970172e-01 1.78240702e-01
-2.15544581e-01 -1.57891721e-01 -6.31516993e-01 7.89490879e-01
-6.49628118e-02 3.82919133e-01 -5.41588604e-01 2.20100328e-01
5.84423356e-02 3.72322381e-01 -1.03712536e-01 4.77863342e-01
1.53398454e-01 6.25138879e-01 -1.00460219e+00 4.38751072e-01
7.47283757e-01 4.15785879e-01 3.96886557e-01 -5.88879049e-01
1.74020961e-01 4.28364843e-01 -1.07039833e+00 9.90288138e-01
-1.02146292e+00 5.81869721e-01 3.60012293e-01 -9.32038605e-01
9.95555580e-01 7.41685808e-01 3.78807306e-01 -4.14083511e-01
3.39008957e-01 1.64286897e-01 2.77117312e-01 -4.87473637e-01
-4.23537105e-01 1.20913975e-01 -9.03384387e-01 7.62685180e-01
-8.32886770e-02 2.40250826e-01 -9.99003410e-01 -1.31968185e-01
1.54894400e+00 -3.69100004e-01 9.19957238e-04 -2.46286467e-01
8.08062792e-01 -4.50887948e-01 2.77272493e-01 1.18945515e+00
1.67185500e-01 -1.68220222e-01 3.21622759e-01 -2.01283485e-01
-4.75409031e-01 -1.11435401e+00 -3.60899895e-01 6.01868689e-01
1.48577645e-01 -2.36443095e-02 -4.07627970e-01 -7.19172657e-01
3.01637501e-01 7.98006415e-01 -5.78227520e-01 -7.25533068e-01
-3.50701213e-01 -3.55055898e-01 1.59060562e+00 3.06712855e-02
5.50074160e-01 -8.39798987e-01 -3.81019562e-01 3.61174703e-01
2.00077310e-01 -1.38655341e+00 7.50396773e-02 5.90306222e-01
-1.86321422e-01 -8.68685067e-01 2.66190410e-01 -3.87171865e-01
7.11962283e-01 -3.34553540e-01 3.32337141e-01 3.43936309e-02
-1.81584284e-01 4.82696593e-01 -1.75033629e-01 -5.49912691e-01
-1.03094912e+00 -6.23867773e-02 1.55075476e-01 6.20717943e-01
5.90352640e-02 -5.88096738e-01 -2.54225582e-01 4.70797479e-01
-5.92877090e-01 -7.99381614e-01 5.30447662e-01 3.81324381e-01
-7.78694078e-02 8.67691398e-01 5.36888957e-01 -9.82343674e-01
5.68641901e-01 -6.06760025e-01 -6.88092828e-01 2.70150274e-01
-1.62261054e-01 8.85131210e-02 8.64082336e-01 -6.81655228e-01
-6.83447123e-01 -2.57861405e-03 -2.54864514e-01 6.12884611e-02
-3.12258720e-01 5.90310454e-01 -9.87811089e-01 -8.11569870e-01
5.92342138e-01 1.54946059e-01 -2.97210425e-01 -2.32052356e-01
-2.45236054e-01 9.19123411e-01 5.17648101e-01 -6.40038908e-01
1.50262463e+00 2.93754995e-01 5.03535688e-01 -7.99307346e-01
-5.46975851e-01 4.76496704e-02 -1.58689350e-01 -1.82023734e-01
4.94271606e-01 -7.13113666e-01 -9.33786035e-01 5.33504188e-01
-8.50531757e-01 1.13438681e-01 4.48445976e-01 6.66131556e-01
5.96556487e-03 1.64367333e-02 -2.46511251e-01 -1.33396792e+00
-3.55447888e-01 -9.89948153e-01 4.48613852e-01 1.52606040e-03
-3.64760637e-01 -9.51617122e-01 -4.87236142e-01 2.14394599e-01
4.57146764e-01 8.02198410e-01 1.11039007e+00 -9.70936418e-01
-3.41186255e-01 -1.13679338e+00 3.96916091e-01 1.61965385e-01
-9.76663828e-02 -3.33879113e-01 -1.48346615e+00 -4.44048792e-01
4.47734743e-01 1.55236900e-01 -8.58881325e-02 2.69335747e-01
1.40965819e+00 -7.23727524e-01 -5.77835798e-01 1.04118311e+00
1.29290605e+00 5.47296107e-01 7.36462951e-01 -1.22663018e-03
2.91710705e-01 3.51365179e-01 5.42880237e-01 2.69329280e-01
-5.24980605e-01 6.37587905e-01 3.99618477e-01 -1.25284791e-01
6.86101198e-01 -4.36488956e-01 5.52756846e-01 -3.72065842e-01
5.24656713e-01 -4.11582500e-01 -4.07486051e-01 -2.81516016e-01
-9.10972834e-01 -1.05804086e+00 9.98433586e-03 2.62013125e+00
5.87661684e-01 6.46170437e-01 -2.22254414e-02 4.99164790e-01
6.78860486e-01 1.19944271e-02 -5.77818811e-01 -7.39763081e-01
1.43213468e-02 5.85934162e-01 1.33080733e+00 4.17710930e-01
-1.29502201e+00 2.09585592e-01 4.76124668e+00 6.04956210e-01
-1.39935064e+00 -2.45466053e-01 2.51757950e-01 4.53995526e-01
-2.30923623e-01 6.73348689e-03 -6.92112982e-01 5.48082590e-01
1.21730411e+00 2.20222443e-01 4.53719467e-01 4.92416233e-01
3.09999622e-02 1.77119330e-01 -1.19019747e+00 7.05158114e-01
-1.39512151e-01 -1.14184022e+00 -4.56395119e-01 4.59447205e-01
-8.23405236e-02 -5.34931839e-01 2.95805216e-01 2.98720062e-01
1.35831967e-01 -9.81718123e-01 5.90378463e-01 2.75409132e-01
9.42919195e-01 -1.08937168e+00 9.22712386e-01 4.68845397e-01
-1.11376083e+00 -2.87316442e-01 2.54154354e-01 2.01332211e-01
-2.35174969e-01 5.33488631e-01 -1.16298008e+00 7.01891422e-01
4.06643480e-01 -3.41414630e-01 -1.75847739e-01 6.76033258e-01
-3.83284390e-01 1.35022676e+00 -6.72567189e-01 -5.80433793e-02
6.83193430e-02 4.46866810e-01 7.29879260e-01 9.75848854e-01
4.32509452e-01 -1.97035577e-02 1.66492593e-02 7.72206426e-01
-1.38860613e-01 -3.33600193e-01 -5.48069775e-01 -6.62034051e-03
1.02897084e+00 1.02959764e+00 -6.83486909e-02 3.54147732e-01
-8.78365897e-03 5.35941124e-01 -7.95986712e-01 5.89523554e-01
-7.49078333e-01 -6.67605221e-01 7.71711826e-01 2.97288537e-01
2.19634354e-01 -1.67765990e-02 -1.64515659e-01 -2.49059558e-01
-5.82430959e-02 -9.04858708e-01 3.03636402e-01 -3.42807472e-01
-1.04646087e+00 3.43021035e-01 -2.49987006e-01 -1.45293319e+00
-4.06097710e-01 -3.26700240e-01 -7.68283725e-01 1.20007086e+00
-1.16720021e+00 -1.19997203e+00 1.00864112e-01 6.72330678e-01
-5.77176869e-01 -4.01913971e-01 1.37023056e+00 3.43727469e-02
-1.57358244e-01 1.17091048e+00 1.20048419e-01 7.82100320e-01
3.33527476e-02 -6.10527039e-01 3.69953096e-01 7.76921809e-01
-2.86784582e-02 7.09896088e-01 7.26833701e-01 -5.03157556e-01
-1.76265156e+00 -1.05467117e+00 3.93445879e-01 -1.31285405e-02
5.48162043e-01 -1.02697504e+00 -5.17197073e-01 3.38834941e-01
-4.86049920e-01 4.81349928e-03 1.34965646e+00 -1.43985137e-01
-3.96800131e-01 -4.88436401e-01 -1.91020334e+00 3.39569569e-01
1.31271288e-01 -8.42240810e-01 3.80011760e-02 -3.48734930e-02
4.67871249e-01 -1.03045464e-01 -1.00745904e+00 4.66729790e-01
1.09024286e+00 -3.89349848e-01 1.03221810e+00 -4.77344602e-01
-7.22931027e-01 -2.83543885e-01 -4.89476740e-01 -8.72256994e-01
1.73954859e-01 -1.20396686e+00 -1.97751403e-01 1.36481524e+00
8.08654964e-01 -8.87333691e-01 6.89521432e-01 5.73668003e-01
6.98621511e-01 -3.27185214e-01 -9.10165191e-01 -7.22387969e-01
-2.71684915e-01 -8.89082253e-01 1.06664014e+00 8.57981920e-01
-1.16012938e-01 -1.30761385e-01 -4.14349705e-01 1.38308001e+00
8.29275072e-01 -1.28884703e-01 9.35142756e-01 -1.11958170e+00
-5.76122046e-01 1.02119103e-01 -5.82381725e-01 -6.59404635e-01
7.86881894e-02 -9.20106232e-01 -1.97704881e-01 -4.02829170e-01
-9.26467419e-01 -1.16910839e+00 -6.13374889e-01 2.60980964e-01
4.93254304e-01 -7.69619122e-02 1.84443463e-02 -2.41460651e-01
4.29005742e-01 -3.75735849e-01 2.63631791e-01 -2.68388152e-01
-1.90708399e-01 1.25200033e+00 -4.87499565e-01 4.55597639e-01
1.11295652e+00 -9.10078228e-01 -1.32856876e-01 3.72746378e-01
1.68785200e-01 5.62978625e-01 6.44247949e-01 -1.52454090e+00
1.27482906e-01 2.26764664e-01 6.13152921e-01 -3.09207082e-01
1.74229413e-01 -1.58213055e+00 4.82167363e-01 7.59945691e-01
-5.25814533e-01 -5.53475976e-01 3.06062281e-01 5.55390120e-01
3.50136757e-01 -3.55365574e-01 5.25845170e-01 4.69413102e-01
-1.10204518e-01 8.47354829e-02 -5.61760247e-01 -4.67370659e-01
9.37756240e-01 -1.27840191e-01 -8.31194296e-02 -7.26335227e-01
-6.29507124e-01 -8.68474767e-02 6.98382780e-02 1.71821490e-01
3.87777895e-01 -9.19468582e-01 -3.58923107e-01 7.92630255e-01
1.27860457e-01 -5.89674652e-01 -1.67864319e-02 2.28713632e-01
-1.36319786e-01 2.80638814e-01 2.55983531e-01 -4.64175433e-01
-1.34897125e+00 3.80664349e-01 7.20215499e-01 -9.73988250e-02
-2.62254268e-01 6.02334857e-01 -1.93162933e-01 -2.47788757e-01
5.85119903e-01 -9.84816700e-02 1.99532539e-01 -2.72546142e-01
4.71445650e-01 -7.91940689e-02 3.70068014e-01 -2.85151005e-01
-5.91580033e-01 3.90719891e-01 5.38685024e-01 -1.30442053e-01
6.09599948e-01 2.62222767e-01 4.51999605e-01 1.33164793e-01
1.79205465e+00 6.90050423e-01 -7.00828135e-01 -2.33918771e-01
-1.00042492e-01 -3.14980686e-01 1.41885847e-01 -1.04491425e+00
-1.02943289e+00 6.80085659e-01 9.47231889e-01 6.44585729e-01
9.72098410e-01 -2.55694389e-01 8.91149223e-01 3.60571653e-01
8.52034509e-01 -6.82701528e-01 -6.78282082e-01 -2.63994068e-01
2.10695058e-01 -5.26605189e-01 -5.41446656e-02 -1.61081642e-01
-9.44517255e-02 1.06019127e+00 -3.25121105e-01 1.26610681e-01
1.26143181e+00 9.72164214e-01 2.48275265e-01 5.90267107e-02
-2.59065740e-02 5.11522293e-01 1.02056421e-01 1.08415973e+00
-3.23470145e-01 3.11130911e-01 3.08466047e-01 9.31143165e-01
-5.77521622e-01 -4.33251262e-01 2.77216315e-01 8.55411053e-01
-3.32495421e-01 -1.14084291e+00 -6.61571383e-01 4.85536754e-01
-8.23387742e-01 2.27497399e-01 -3.84798914e-01 5.28092146e-01
2.81046987e-01 1.27258217e+00 -2.49622017e-01 -8.54019225e-01
2.89533228e-01 -1.13554290e-02 5.40075973e-02 -1.70201212e-01
-7.92271614e-01 -4.17059004e-01 4.42210227e-01 -4.59341824e-01
3.51968631e-02 -6.17027342e-01 -1.28570783e+00 -1.68979913e-01
-3.75642329e-01 2.44965166e-01 1.08370352e+00 9.34049129e-01
4.81693670e-02 3.20722520e-01 1.48243713e+00 -2.83604294e-01
-6.82319283e-01 -4.53096360e-01 -7.23121524e-01 1.33557962e-02
5.73995411e-01 -4.62753236e-01 -8.80603790e-01 -5.90590656e-01] | [13.885719299316406, 5.821480751037598] |
bcdf643b-f97b-48b6-9695-cb69c3898f9e | a-method-for-automatically-animating-children | 2303.12741 | null | https://arxiv.org/abs/2303.12741v2 | https://arxiv.org/pdf/2303.12741v2.pdf | A Method for Animating Children's Drawings of the Human Figure | Children's drawings have a wonderful inventiveness, creativity, and variety to them. We present a system that automatically animates children's drawings of the human figure, is robust to the variance inherent in these depictions, and is simple and straightforward enough for anyone to use. We demonstrate the value and broad appeal of our approach by building and releasing the Animated Drawings Demo, a freely available public website that has been used by millions of people around the world. We present a set of experiments exploring the amount of training data needed for fine-tuning, as well as a perceptual study demonstrating the appeal of a novel twisted perspective retargeting technique. Finally, we introduce the Amateur Drawings Dataset, a first-of-its-kind annotated dataset, collected via the public demo, containing over 178,000 amateur drawings and corresponding user-accepted character bounding boxes, segmentation masks, and joint location annotations. | ['Jessica K. Hodgins', 'Somya Jain', 'Yifei Li', 'Qingyuan Zheng', 'Harrison Jesse Smith'] | 2023-03-07 | null | null | null | null | ['image-to-video'] | ['computer-vision'] | [ 8.34034681e-02 1.45358935e-01 3.03262860e-01 -4.35438544e-01
-2.09492579e-01 -1.18989182e+00 6.81527853e-01 -6.58829585e-02
-8.63293111e-02 3.26260448e-01 2.53799170e-01 4.85591888e-02
5.61462268e-02 -5.75359404e-01 -5.25509119e-01 -8.61617401e-02
-6.46388903e-02 6.88113093e-01 4.56862003e-01 -3.01651597e-01
6.09353423e-01 7.04859734e-01 -1.55766785e+00 1.54837385e-01
7.91571081e-01 5.19182086e-01 -6.46825805e-02 7.82254517e-01
-3.43778357e-02 4.25083667e-01 -9.01696026e-01 -1.19682193e+00
2.44699866e-01 -8.73598158e-02 -6.24303401e-01 1.16708696e-01
1.14661038e+00 -5.23598790e-01 -8.54562689e-03 7.18721449e-01
6.87146723e-01 -1.65186912e-01 7.18464255e-01 -1.22524190e+00
-1.13129187e+00 8.21390450e-01 -8.76604199e-01 2.05417275e-01
5.20507157e-01 2.46900886e-01 8.53107512e-01 -6.82796717e-01
1.06230009e+00 1.32782924e+00 7.38058329e-01 6.44004524e-01
-1.35378480e+00 -7.89938331e-01 2.04385176e-01 -4.26082075e-01
-1.38707662e+00 -3.33485246e-01 7.96982527e-01 -8.36501777e-01
7.59461164e-01 2.89243132e-01 1.30850053e+00 1.29823363e+00
-5.55266023e-01 9.27948952e-01 9.05856311e-01 -4.74169850e-01
2.28759646e-01 -1.51620060e-01 -4.68361109e-01 7.45669067e-01
9.71294865e-02 -3.96247327e-01 -4.04637665e-01 -1.60108104e-01
1.39679599e+00 -4.69731450e-01 6.41676188e-02 -6.65998876e-01
-1.09640217e+00 5.09951472e-01 4.23774183e-01 1.94701418e-01
1.60600513e-01 4.27584738e-01 3.03929359e-01 -8.79252106e-02
3.73743415e-01 6.70330465e-01 -2.61283994e-01 -5.55998087e-01
-8.68752956e-01 5.42362988e-01 3.08612823e-01 1.46501493e+00
-2.64404211e-02 -5.65925846e-03 3.23019087e-01 1.01645911e+00
-7.70961493e-02 2.45617673e-01 2.03567892e-01 -1.20538914e+00
3.45354915e-01 4.83912349e-01 8.68595466e-02 -1.13819993e+00
-4.56791550e-01 -1.04658499e-01 -3.73624533e-01 4.04382735e-01
5.77774346e-01 -3.17660384e-02 -8.00439596e-01 1.67044783e+00
4.45178211e-01 -1.73811272e-01 -5.68841159e-01 7.50717103e-01
1.10352230e+00 1.30983025e-01 8.56695622e-02 4.85596925e-01
1.31257975e+00 -8.33269477e-01 -1.94577694e-01 -1.49213318e-02
3.98475200e-01 -9.94670987e-01 1.18190527e+00 5.91559768e-01
-1.35272121e+00 -3.13660771e-01 -1.03327000e+00 -3.06952029e-01
-4.23442543e-01 3.88692655e-02 9.75555897e-01 9.77202952e-01
-9.77439046e-01 8.45990837e-01 -5.25841296e-01 -4.38261896e-01
8.65803242e-01 7.18423948e-02 -5.38566291e-01 1.79932684e-01
-3.92227292e-01 9.08293247e-01 1.40486568e-01 -3.69267672e-01
-3.07899415e-01 -7.04900086e-01 -5.94677329e-01 -1.47115439e-01
8.92756358e-02 -5.40136456e-01 1.19489408e+00 -1.10289240e+00
-1.37613928e+00 1.20408237e+00 5.22013724e-01 2.50326216e-01
9.07509029e-01 -3.12847674e-01 -3.04685086e-01 4.35570687e-01
2.98922777e-01 1.20304954e+00 7.69912958e-01 -1.05749547e+00
-5.32525897e-01 -3.07688147e-01 9.69319195e-02 -1.04530685e-01
-1.52121857e-01 2.17146754e-01 -5.21577299e-01 -1.31708717e+00
1.75579205e-01 -8.78046095e-01 -1.39177084e-01 4.10469979e-01
-3.60078990e-01 -1.54519632e-01 4.03496236e-01 -7.70796537e-01
1.06491673e+00 -2.40383625e+00 1.02746813e-02 1.58553094e-01
2.94529617e-01 1.19107395e-01 -2.52180755e-01 5.36245584e-01
-2.10086465e-01 6.30731106e-01 -1.89639211e-01 -2.37946078e-01
2.29049072e-01 -2.43201450e-01 -1.20228022e-01 4.34048206e-01
1.50074169e-01 9.53558922e-01 -9.33958590e-01 -6.15174115e-01
1.12371072e-01 4.99322683e-01 -4.99268919e-01 -3.07352580e-02
-6.45202249e-02 4.45933789e-01 -3.23456287e-01 7.39962935e-01
6.23616457e-01 9.96189425e-04 7.51905814e-02 1.63771093e-01
-3.04598987e-01 -1.54711101e-02 -9.91644144e-01 1.82906544e+00
-3.08408644e-02 9.43415701e-01 -2.81988889e-01 9.13250968e-02
9.36633408e-01 9.79947392e-03 -1.85545366e-02 -5.44782698e-01
8.18464011e-02 3.65029156e-01 -1.79738794e-02 -5.81805170e-01
6.69495761e-01 1.16500199e-01 -2.43976265e-01 5.79913616e-01
-8.64594579e-02 -8.15567315e-01 3.80413562e-01 4.43162143e-01
8.49009871e-01 6.37154520e-01 4.30465221e-01 -2.41319895e-01
-2.05883875e-01 -1.15949698e-01 1.35228679e-01 3.91908586e-01
1.31953834e-03 1.12196100e+00 5.83813131e-01 -9.37118530e-01
-1.67894638e+00 -9.94905889e-01 -1.69034377e-01 1.06277382e+00
-2.59274423e-01 -3.34708571e-01 -1.05955195e+00 -3.23183626e-01
-1.18026379e-02 7.24301994e-01 -9.53258514e-01 5.14001667e-01
-6.81896210e-01 -9.57588181e-02 8.45205784e-01 7.64173388e-01
2.98203170e-01 -9.83315527e-01 -1.00936711e+00 -5.75469136e-02
3.87841940e-01 -9.44389462e-01 -6.01912558e-01 -4.40497190e-01
-5.80823720e-01 -8.88819218e-01 -1.14577186e+00 -7.78598547e-01
8.44951391e-01 -5.33566847e-02 1.28324389e+00 2.08375201e-01
-5.26045978e-01 4.51771170e-01 -4.47229385e-01 -3.38668108e-01
-1.76332861e-01 1.55203342e-02 -2.38662422e-01 -6.39176130e-01
-1.78732961e-01 -1.12811744e+00 -5.22739947e-01 1.60294205e-01
-7.69518793e-01 5.01456261e-01 2.14989558e-01 3.10843319e-01
2.71635465e-02 -4.34753746e-01 3.45548272e-01 -9.41244602e-01
4.65945423e-01 -4.54499833e-02 -6.25331759e-01 2.37709537e-01
5.82784377e-02 -1.97670832e-01 3.55004519e-01 -5.88631094e-01
-7.08633363e-01 1.51409164e-01 -1.04807019e-01 -6.11272827e-02
-2.66398609e-01 -2.58087337e-01 -9.68239009e-02 -2.14184076e-01
6.43977880e-01 -3.33955705e-01 -6.09522998e-01 -5.91795802e-01
1.10132444e+00 5.06990314e-01 9.90650773e-01 -9.73959088e-01
7.44392991e-01 3.36000949e-01 -1.69551242e-02 -9.00302887e-01
-2.84280330e-01 1.54704913e-01 -1.30078816e+00 -2.76150614e-01
7.57438242e-01 -6.39441967e-01 -7.95590043e-01 4.76365149e-01
-1.40811288e+00 -2.62691408e-01 -3.62073779e-01 -1.09031215e-01
-5.56137383e-01 2.21543163e-01 -4.69734073e-01 -6.33049846e-01
-1.26096204e-01 -8.15442502e-01 9.65617597e-01 3.76857609e-01
-7.22031593e-01 -6.50659859e-01 9.29928720e-02 2.52934694e-01
8.17669928e-02 9.89217460e-01 1.17211902e+00 -4.30617243e-01
-4.06328350e-01 -7.69891813e-02 -4.10433948e-01 -1.29969478e-01
-1.27680615e-01 9.72165406e-01 -9.27254498e-01 -8.92797392e-03
-8.58807921e-01 -3.21113557e-01 4.65185672e-01 1.77786388e-02
9.83502269e-01 -3.38920861e-01 -1.32934153e-01 5.55896997e-01
1.08309579e+00 1.47921190e-01 6.00527883e-01 3.47442985e-01
1.00641775e+00 7.26381600e-01 1.12845600e-01 5.31597495e-01
6.59503043e-01 6.53117776e-01 2.25714937e-01 3.81179787e-02
-1.70667663e-01 -6.77666247e-01 -3.58002543e-01 6.85630977e-01
-5.16131163e-01 -2.19250277e-01 -9.76360798e-01 6.72809005e-01
-1.55189312e+00 -8.90161514e-01 -1.85999364e-01 2.05068660e+00
6.30368412e-01 1.12395532e-01 5.98682642e-01 -3.83946188e-02
8.40690792e-01 1.39778614e-01 -2.83102334e-01 -7.41667747e-01
-1.47286266e-01 5.06677270e-01 1.23093560e-01 -5.91026954e-02
-1.06457579e+00 1.12941086e+00 7.84579563e+00 8.40313315e-01
-8.58360469e-01 -1.55848324e-01 6.07101738e-01 -1.60914361e-01
-5.21060824e-01 -1.41025737e-01 -1.22218944e-01 4.83097494e-01
2.57793456e-01 4.29883376e-02 3.93030196e-01 9.80923295e-01
-2.28211492e-01 -9.43263248e-02 -1.22519159e+00 1.07267451e+00
1.22987457e-01 -1.41277266e+00 1.54574001e-02 -2.49802530e-01
8.63408387e-01 -4.98704702e-01 3.39496911e-01 -1.53803259e-01
5.03635705e-01 -1.28133547e+00 1.46624339e+00 3.75735015e-01
1.16244745e+00 -8.94803941e-01 -6.52616750e-03 -1.53801721e-02
-9.79379952e-01 1.87585846e-01 -5.21110147e-02 -3.41248453e-01
-3.14021036e-02 -3.98279727e-02 -6.92144871e-01 -5.34402356e-02
8.04298043e-01 5.70083320e-01 -1.24691784e+00 1.09555268e+00
-6.12763703e-01 6.88367486e-02 -4.02490199e-01 -3.70619029e-01
1.52482852e-01 -8.25796649e-02 3.21676582e-01 1.47581744e+00
2.17935011e-01 3.43397200e-01 -3.34976703e-01 8.63435388e-01
-1.19353086e-01 1.73949435e-01 -6.29573107e-01 -2.13517487e-01
8.05911899e-01 1.21819007e+00 -1.32616591e+00 -2.28109092e-01
-7.94734359e-02 1.14731669e+00 5.38452566e-01 9.73316357e-02
-8.00720215e-01 -4.53867823e-01 2.16404870e-01 2.12585062e-01
4.71726865e-01 -4.30836588e-01 -4.73465443e-01 -8.58715475e-01
4.45809886e-02 -8.54597926e-01 9.43425596e-02 -1.32994759e+00
-1.37467396e+00 5.42015374e-01 1.02613904e-01 -9.95454848e-01
-7.38299936e-02 -4.66608435e-01 -6.91690505e-01 3.73756945e-01
-3.55710536e-01 -1.61339164e+00 -1.71983793e-01 3.97763737e-02
4.84946668e-01 -2.85303555e-02 6.94652438e-01 2.05014288e-01
-6.10121727e-01 7.16852844e-01 -1.80385798e-01 3.76648545e-01
5.11824369e-01 -1.37281358e+00 1.46382535e+00 5.47333181e-01
2.68978298e-01 7.45406032e-01 7.31405973e-01 -6.11305177e-01
-8.23295772e-01 -4.37334627e-01 3.90612930e-01 -8.36016297e-01
7.45747328e-01 -6.56460881e-01 -5.54315567e-01 9.62663054e-01
2.35744864e-01 -2.71037757e-01 6.13188624e-01 -5.24081290e-02
-6.94168508e-01 5.45692861e-01 -1.39087725e+00 9.27277923e-01
1.46871257e+00 -2.41864607e-01 -8.26054633e-01 -3.01749120e-03
6.40932858e-01 -6.99037910e-01 -7.91371167e-01 5.82559928e-02
1.26612782e+00 -1.17948306e+00 9.25198197e-01 -4.65165854e-01
8.74303520e-01 -7.74567723e-02 1.65955292e-03 -1.27866387e+00
-5.09658992e-01 -8.94766331e-01 3.55827868e-01 1.73840415e+00
3.37140739e-01 -3.39569688e-01 6.82459056e-01 8.94662023e-01
3.81812528e-02 -7.79134750e-01 -8.74530315e-01 -4.89261538e-01
2.62470245e-01 -3.66149098e-01 1.08411109e+00 1.06449056e+00
-5.89406863e-02 2.00825319e-01 -1.48316160e-01 -1.73451573e-01
5.21142125e-01 -5.95362484e-03 1.06307065e+00 -1.20630848e+00
-2.28304639e-01 -8.68511856e-01 -6.30312622e-01 -8.72338653e-01
-3.67260963e-01 -5.00284970e-01 -4.62200463e-01 -1.44774854e+00
-4.65633795e-02 -5.94646215e-01 6.48306012e-01 3.70103449e-01
5.98764941e-02 6.88114643e-01 5.88371336e-01 8.37546065e-02
-5.67532480e-01 1.40793294e-01 1.50979805e+00 5.54697849e-02
-2.47503165e-02 -3.48456889e-01 -8.14312637e-01 1.27532649e+00
4.74035472e-01 -3.55870813e-01 -1.83896944e-01 -7.20677435e-01
5.93064964e-01 -3.80853325e-01 2.56269455e-01 -8.20480108e-01
-6.99044168e-02 -9.51104164e-02 9.53897655e-01 -7.62920082e-01
4.04239178e-01 -4.35367048e-01 3.59856486e-01 8.05441365e-02
-2.14871556e-01 3.21151376e-01 2.88761288e-01 -1.75308040e-03
5.63692927e-01 -3.03711712e-01 8.31757545e-01 -3.57407838e-01
-6.67443275e-01 -9.49918553e-02 -1.78388968e-01 2.48353347e-01
1.03549433e+00 -4.30058867e-01 -6.59711421e-01 -2.96070367e-01
-3.97525370e-01 1.00864775e-01 1.04301667e+00 4.12481725e-01
5.68117559e-01 -1.35683668e+00 -6.62469745e-01 1.59014016e-01
1.13010615e-01 1.04895197e-02 1.24493845e-01 3.29279006e-02
-1.14253724e+00 -2.11100802e-01 -6.49559796e-01 -2.80112654e-01
-1.15871465e+00 6.82033241e-01 -1.71294093e-01 5.14767766e-01
-9.86105442e-01 1.16078913e+00 1.76937237e-01 -3.81882280e-01
-1.75194114e-01 -3.56925994e-01 -1.34873152e-01 1.13493502e-01
6.11850500e-01 7.84550428e-01 -4.71999705e-01 -8.75636697e-01
-3.70230615e-01 9.61555898e-01 1.76984891e-01 -4.33218867e-01
1.49998510e+00 -5.08324243e-02 -2.72981953e-02 4.83782738e-01
7.68954575e-01 5.12733161e-01 -1.31079316e+00 3.45345467e-01
-1.02505095e-01 -9.07606483e-01 -6.95972264e-01 -9.30554986e-01
-8.07040334e-01 9.06464458e-01 2.79162019e-01 4.55187768e-01
6.46180987e-01 1.45805001e-01 7.13864923e-01 1.70970112e-02
6.36967182e-01 -9.17311966e-01 3.17071170e-01 7.42518082e-02
1.21522880e+00 -7.37590969e-01 4.45846200e-01 -5.68612456e-01
-7.30995297e-01 1.14302278e+00 5.97166359e-01 -3.70462716e-01
2.72521358e-02 2.14862809e-01 5.60312718e-02 -2.31354445e-01
-4.02277797e-01 1.60236090e-01 2.69690007e-01 1.05423164e+00
3.93430650e-01 3.12194526e-01 -1.22279756e-01 5.54305911e-01
-1.11107004e+00 -4.15936589e-01 6.83178604e-01 9.19192374e-01
-1.89101160e-01 -7.75750518e-01 -3.34617853e-01 3.05943508e-02
-2.72819310e-01 -6.17847331e-02 -9.79530454e-01 1.11410403e+00
3.73690158e-01 4.83629048e-01 2.26761371e-01 -2.05157146e-01
4.97045398e-01 -2.56376922e-01 1.11457121e+00 -6.08601093e-01
-6.73351645e-01 -7.98245966e-02 2.30316430e-01 -2.10930601e-01
-5.63850403e-02 -5.86351097e-01 -1.00803530e+00 -6.49357319e-01
-4.40139621e-02 -4.15125191e-01 6.85244441e-01 5.25435686e-01
1.55615896e-01 2.67486125e-01 3.46149087e-01 -1.35083497e+00
3.19237083e-01 -7.67085135e-01 -5.80600798e-01 4.22060192e-01
8.00107345e-02 -6.10185683e-01 -4.90340777e-02 7.49829486e-02] | [11.699957847595215, -0.2516021430492401] |
166c18ad-d9fc-4d16-86ea-69be490ef1d8 | dynamic-knowledge-distillation-with-a-single | 2106.09517 | null | https://arxiv.org/abs/2106.09517v3 | https://arxiv.org/pdf/2106.09517v3.pdf | Dynamic Knowledge Distillation With Noise Elimination for RGB-D Salient Object Detection | RGB-D salient object detection (SOD) demonstrates its superiority on detecting in complex environments due to the additional depth information introduced in the data. Inevitably, an independent stream is introduced to extract features from depth images, leading to extra computation and parameters. This methodology sacrifices the model size to improve the detection accuracy which may impede the practical application of SOD problems. To tackle this dilemma, we propose a dynamic distillation method along with a lightweight structure, which significantly reduces the computational burden while maintaining validity. This method considers the factors of both teacher and student performance within the training stage and dynamically assigns the distillation weight instead of applying a fixed weight on the student model. We also investigate the issue of RGB-D early fusion strategy in distillation and propose a simple noise elimination method to mitigate the impact of distorted training data caused by low quality depth maps. Extensive experiments are conducted on five public datasets to demonstrate that our method can achieve competitive performance with a fast inference speed (136FPS) compared to 10 prior methods. | ['Tania Stathaki', 'Hengyan Liu', 'Yinxiao Yu', 'Guangyu Ren'] | 2021-06-17 | null | null | null | null | ['rgb-d-salient-object-detection'] | ['computer-vision'] | [ 2.75421977e-01 -1.41298428e-01 1.13526091e-01 -4.38357472e-01
-4.94468540e-01 -7.32311904e-02 2.88910866e-01 3.22591782e-01
-8.39528739e-01 3.82911950e-01 -8.49850178e-02 -2.66074836e-01
-6.47219941e-02 -7.89619207e-01 -6.20595217e-01 -9.09905851e-01
2.74249613e-01 -1.11722626e-01 9.13694918e-01 7.71449432e-02
3.94037902e-01 4.41404402e-01 -2.08331013e+00 -2.81318091e-02
1.26797211e+00 1.01825917e+00 7.14344561e-01 5.65858185e-01
-1.21837176e-01 8.52478147e-01 -8.33331168e-01 -3.54084462e-01
4.12328184e-01 3.10758664e-03 -1.94538891e-01 1.14082247e-01
4.60479081e-01 -8.94965529e-01 -3.70882630e-01 1.08203709e+00
1.03621423e+00 2.01551035e-01 1.75634831e-01 -1.22461283e+00
-9.37992632e-02 1.34317249e-01 -1.22474813e+00 6.23736143e-01
2.85395443e-01 1.01372726e-01 3.04650426e-01 -8.88114929e-01
2.81276137e-01 1.10957527e+00 4.86318409e-01 3.37313771e-01
-8.57473791e-01 -8.83216739e-01 3.09045792e-01 4.28865314e-01
-1.29755652e+00 -2.88590968e-01 8.39524567e-01 -8.41353089e-02
5.06089568e-01 1.86487362e-01 8.68870437e-01 7.38457322e-01
-1.12258725e-01 1.15316343e+00 1.15940046e+00 -4.88632679e-01
3.09725821e-01 2.32977331e-01 -1.25677079e-01 8.44061792e-01
4.04165119e-01 1.64596997e-02 -7.39211619e-01 1.16582818e-01
8.78473341e-01 4.50444072e-02 -2.13946521e-01 -4.79276448e-01
-7.97689557e-01 5.29966474e-01 4.83816832e-01 -1.04996294e-01
-2.11615533e-01 -5.09893075e-02 2.36149773e-01 -9.14544612e-02
3.98391396e-01 -2.09144294e-01 -3.02410334e-01 -3.37308228e-01
-9.54441786e-01 1.64217263e-01 2.66654313e-01 8.00752759e-01
6.20272994e-01 -1.21333614e-01 -4.94427010e-02 5.99922538e-01
3.84819835e-01 4.70247149e-01 4.17813510e-01 -8.86575818e-01
5.58865190e-01 8.14988673e-01 -2.18036510e-02 -1.11553347e+00
-3.52630824e-01 -4.43841875e-01 -5.01908004e-01 3.17197204e-01
4.63744015e-01 1.03340549e-02 -9.08525467e-01 1.35389638e+00
1.08578622e+00 4.37035084e-01 -2.12551340e-01 1.11551404e+00
9.54598367e-01 4.22396034e-01 5.96839003e-03 -1.85001157e-02
1.31929386e+00 -8.25932384e-01 -4.93619412e-01 -2.85325974e-01
4.89993691e-01 -7.35093951e-01 9.12294090e-01 5.96997738e-01
-1.01534510e+00 -6.55198038e-01 -1.27456117e+00 -3.24728131e-01
-2.82634526e-01 2.77724326e-01 8.12289894e-01 7.87696779e-01
-6.95101976e-01 2.64115304e-01 -1.12580907e+00 -1.50146946e-01
4.48512614e-01 4.89159316e-01 -2.29208078e-02 -1.25494301e-01
-8.86267006e-01 7.40622282e-01 3.51983011e-01 1.65031821e-01
-7.10321128e-01 -8.29675257e-01 -6.77425742e-01 -5.05817533e-02
3.71030867e-01 -3.32395405e-01 1.15902686e+00 -5.51463842e-01
-1.50684786e+00 5.33786833e-01 -2.45292094e-02 -2.21968740e-01
7.21106589e-01 -3.75354767e-01 1.88612137e-02 3.76355708e-01
-1.12787001e-01 7.90450156e-01 8.76932740e-01 -9.44713295e-01
-1.28082716e+00 -5.17729640e-01 3.41942608e-01 4.82917994e-01
-6.54674590e-01 -1.47409379e-01 -7.60969698e-01 -5.15520334e-01
5.95004022e-01 -6.24476969e-01 -3.78868371e-01 3.65734369e-01
9.70397741e-02 -1.25779644e-01 1.02543890e+00 -4.49678212e-01
1.13059187e+00 -2.26953268e+00 -2.85727948e-01 -4.95297536e-02
2.88400203e-01 3.31508756e-01 2.84844458e-01 -1.59705967e-01
3.92526269e-01 -5.24363518e-01 -7.41834491e-02 -4.31322098e-01
-2.22175851e-01 3.11751366e-01 -1.19770989e-01 6.54137969e-01
1.39455691e-01 3.59862506e-01 -1.06124365e+00 -1.02783692e+00
5.43455660e-01 7.23826826e-01 -6.90952003e-01 2.73342818e-01
1.75382376e-01 1.56067476e-01 -5.86060405e-01 7.78898239e-01
1.12145913e+00 -4.73860428e-02 -2.62941957e-01 -5.10003045e-02
-4.08899009e-01 4.44852829e-01 -1.59516728e+00 1.81693971e+00
-2.00860277e-01 3.41165125e-01 2.78788775e-01 -9.58366990e-01
8.14490378e-01 6.38743341e-02 3.95824134e-01 -9.88671720e-01
3.54929924e-01 5.47875045e-03 -2.21268103e-01 -5.59250474e-01
6.60109460e-01 2.90684044e-01 1.49873123e-01 3.46677125e-01
-2.32775867e-01 -1.76000848e-01 7.69658238e-02 1.84347093e-01
9.02562141e-01 4.89755630e-01 1.45535786e-02 -1.38075486e-01
4.46562111e-01 -2.53201306e-01 8.39311898e-01 7.30297089e-01
-6.54044509e-01 4.30733383e-01 1.11076467e-01 -4.99087036e-01
-5.57411432e-01 -9.30628836e-01 -3.03453565e-01 1.12108839e+00
6.72651708e-01 -2.16072485e-01 -6.50114715e-01 -6.73948050e-01
7.71967918e-02 2.26750404e-01 -5.11152685e-01 -1.80433720e-01
-4.46726799e-01 -9.68758523e-01 4.16313291e-01 6.52748346e-01
8.40899110e-01 -4.66088355e-01 -1.52712369e+00 1.20642170e-01
-2.27592438e-01 -1.19802833e+00 -6.62918240e-02 2.35179171e-01
-1.09908700e+00 -8.13515604e-01 -5.80839038e-01 -5.81709325e-01
9.45996523e-01 7.88254380e-01 6.31109834e-01 1.36960119e-01
-4.19686258e-01 1.99083433e-01 -2.21720710e-01 -7.96873391e-01
1.58293813e-01 -2.02863403e-02 -7.42179304e-02 -4.02173787e-01
5.85460663e-01 -4.36998188e-01 -1.03119993e+00 1.53850615e-01
-8.95950079e-01 4.08982813e-01 6.80684805e-01 5.64522028e-01
4.15175438e-01 2.24959955e-01 3.00232083e-01 -3.18605602e-01
2.37927139e-01 -1.32605150e-01 -8.58980536e-01 1.42182022e-01
-6.60488844e-01 -9.83601343e-03 1.70062706e-01 -5.49653649e-01
-1.31174338e+00 2.31259018e-01 1.42023921e-01 -3.85259271e-01
9.20966938e-02 -6.93107620e-02 7.30812848e-02 -3.90718788e-01
2.88989037e-01 3.49079430e-01 -2.18246412e-03 -5.56216359e-01
4.94899005e-02 6.54167175e-01 6.17486715e-01 -6.82645500e-01
8.14432144e-01 6.78196073e-01 7.13199526e-02 -6.09322488e-01
-8.79420221e-01 -4.62229580e-01 -5.27307570e-01 -4.19757694e-01
4.20201749e-01 -1.26244056e+00 -8.45526755e-01 5.08123159e-01
-9.72748399e-01 1.83488995e-01 -2.53580529e-02 6.82954073e-01
-6.79560602e-02 4.65528369e-01 -6.62585795e-01 -1.01948476e+00
-3.17453235e-01 -1.06417954e+00 9.96547401e-01 6.81435525e-01
3.65866959e-01 -4.02204037e-01 -2.25267857e-01 5.00801623e-01
3.45574975e-01 2.15029866e-01 2.68868268e-01 1.08744558e-02
-8.59140694e-01 -9.15216133e-02 -5.80789864e-01 1.21132940e-01
-9.19391867e-03 6.78920671e-02 -1.24445748e+00 -3.79334599e-01
2.08815664e-01 -1.30597994e-01 7.47347891e-01 2.14234129e-01
1.38420236e+00 1.08402491e-01 -2.34569967e-01 6.00449800e-01
1.42339182e+00 1.33561745e-01 3.66756082e-01 5.76077998e-01
7.16375947e-01 6.27041280e-01 9.66181815e-01 7.49615729e-01
6.99158013e-01 4.34103996e-01 5.81746042e-01 -3.04184228e-01
-8.32209438e-02 -2.59247720e-01 2.57369787e-01 7.98305392e-01
-1.24364138e-01 8.56824741e-02 -7.24646032e-01 6.27268851e-01
-1.60889256e+00 -6.61255956e-01 -9.76030657e-04 2.25266361e+00
9.59678948e-01 4.96987134e-01 -4.11531292e-02 6.07510269e-01
5.57226717e-01 -4.51604975e-03 -4.92738664e-01 -1.57710925e-01
1.42125517e-01 2.63114005e-01 6.39534056e-01 1.74385533e-01
-9.96915102e-01 5.51687479e-01 5.69911289e+00 8.30901682e-01
-1.19855928e+00 6.18090257e-02 5.26571810e-01 -3.89513820e-01
7.64551014e-02 -1.21399149e-01 -1.04896343e+00 7.05474734e-01
4.92606848e-01 3.24601382e-02 -1.20495073e-01 9.47991133e-01
2.45472088e-01 -7.03422606e-01 -6.75809741e-01 9.59494472e-01
2.64492817e-02 -7.53423393e-01 -2.19673023e-01 -1.05462074e-01
6.61438406e-01 -8.18488598e-02 1.11484602e-01 2.30958670e-01
1.23239130e-01 -4.27631438e-01 7.89585352e-01 7.09525868e-02
2.78955311e-01 -9.34694707e-01 6.37816489e-01 4.45754498e-01
-1.19808137e+00 -1.77241072e-01 -5.83616376e-01 -3.60165238e-01
-3.00791562e-01 7.62013555e-01 -9.88097072e-01 4.59870607e-01
1.09243500e+00 3.93296659e-01 -7.16163218e-01 1.25109267e+00
-3.05213392e-01 3.58719826e-01 -7.58804083e-01 9.29727964e-03
2.55737484e-01 3.65607999e-02 3.34503025e-01 1.05113673e+00
3.10506821e-01 3.19927067e-01 1.61027029e-01 5.26871622e-01
2.60736823e-01 -4.19849604e-02 -1.48341700e-01 5.85065603e-01
7.17192292e-01 1.25606143e+00 -9.44883645e-01 -2.85557061e-01
-5.26246488e-01 7.17908263e-01 2.28759781e-01 6.34310171e-02
-1.07970440e+00 -4.41033632e-01 4.35659856e-01 2.49537677e-01
5.67060709e-01 -2.04917282e-01 -3.12748551e-01 -9.18757796e-01
1.25040606e-01 -5.22953391e-01 4.77801859e-01 -6.49826765e-01
-5.63720345e-01 1.20915778e-01 1.92687325e-02 -1.30040717e+00
1.78851515e-01 -2.61931300e-01 -4.45163786e-01 4.86542970e-01
-1.91450500e+00 -8.35509241e-01 -8.05777609e-01 6.14884496e-01
4.13859665e-01 5.24484694e-01 2.32401639e-01 7.56152213e-01
-8.63435745e-01 7.89749861e-01 -1.14501238e-01 -9.25737768e-02
6.98646963e-01 -1.16875076e+00 4.09140438e-02 1.18696141e+00
-2.80073702e-01 2.92578340e-01 7.16498733e-01 -3.42112601e-01
-1.35939038e+00 -7.38783300e-01 6.22120440e-01 -2.56674111e-01
2.15953752e-01 -2.55415142e-01 -8.17637205e-01 1.68206140e-01
-2.59551883e-01 2.46923283e-01 4.31369662e-01 -3.54421467e-01
1.20748626e-02 -4.19749379e-01 -1.40995419e+00 4.18404728e-01
9.88901138e-01 -1.56304285e-01 -5.52080512e-01 -2.33164161e-01
7.69328654e-01 -8.46735001e-01 -7.00642228e-01 6.59003198e-01
6.73508584e-01 -1.25798678e+00 1.07702708e+00 2.62657970e-01
3.80005836e-01 -6.50964499e-01 2.93377042e-03 -5.84952474e-01
5.34863546e-02 -3.99199635e-01 -5.17710924e-01 1.26095176e+00
-2.46257409e-01 -4.72203612e-01 1.05423379e+00 8.28826606e-01
8.63638818e-02 -9.85480964e-01 -1.11572766e+00 -5.19374728e-01
-6.04675174e-01 -2.94535100e-01 8.56547117e-01 6.95953131e-01
-2.55906880e-01 -1.34012252e-01 -6.92801401e-02 4.50548947e-01
9.29756463e-01 2.05666140e-01 8.79065633e-01 -1.12995398e+00
-1.40073314e-01 -1.55894578e-01 -5.83549261e-01 -1.48894167e+00
-5.28712153e-01 -1.67834178e-01 1.38144508e-01 -1.36583793e+00
1.77355945e-01 -7.44266093e-01 -5.53321600e-01 4.49301511e-01
-4.56320822e-01 2.66330242e-01 8.85480642e-02 -6.31576926e-02
-7.82135606e-01 6.18979216e-01 1.44507742e+00 1.69832364e-01
-1.81822762e-01 -1.35118037e-01 -5.45024455e-01 8.15919995e-01
4.67239380e-01 -6.46233201e-01 -6.21845543e-01 -6.93477869e-01
1.30287251e-02 -7.28378296e-02 3.66437405e-01 -1.32190919e+00
5.57824731e-01 -1.69528514e-01 7.62375891e-01 -9.27542329e-01
2.95431584e-01 -9.68242049e-01 -4.46234196e-01 6.94275916e-01
2.03693002e-01 2.67896429e-02 3.20005119e-01 5.02585113e-01
-9.80004892e-02 -1.78108707e-01 8.30743194e-01 9.14205983e-02
-8.43744874e-01 8.90023187e-02 -4.65264916e-02 -2.71733969e-01
1.24596024e+00 -6.93707347e-01 -1.81276530e-01 -5.01329973e-02
-1.83131881e-02 3.62069279e-01 4.65081364e-01 3.68220180e-01
5.99069655e-01 -1.09156215e+00 -3.02732736e-01 5.06955981e-01
-4.85658497e-02 5.69359183e-01 2.52023548e-01 8.51792336e-01
-6.30513251e-01 7.93418214e-02 -1.66898757e-01 -8.52514625e-01
-1.34236825e+00 3.74405205e-01 6.75525069e-02 -8.46405402e-02
-6.72926307e-01 1.25121069e+00 3.94620076e-02 -5.02692908e-02
7.91647851e-01 -5.51567733e-01 -9.90910679e-02 8.71582404e-02
6.04988158e-01 5.47400177e-01 1.33009315e-01 -2.82045603e-01
-4.60458606e-01 4.45360512e-01 -2.57219970e-01 1.53996915e-01
1.26662850e+00 -4.38408226e-01 3.28382015e-01 7.65971839e-02
7.84858346e-01 -8.79847109e-02 -1.74298966e+00 -3.36416990e-01
-1.65630326e-01 -9.63868499e-01 5.52622318e-01 -4.31814820e-01
-1.11914480e+00 8.05205405e-01 9.79547143e-01 -1.32288039e-01
1.47215629e+00 -4.10286248e-01 9.80647922e-01 7.32459575e-02
4.59165215e-01 -1.32267630e+00 1.66465819e-01 6.17636647e-03
1.14517123e-01 -1.47292507e+00 6.34371519e-01 -3.60271245e-01
-1.77825078e-01 8.74490857e-01 1.02317345e+00 1.25172529e-02
3.97104383e-01 5.48963785e-01 -1.27600245e-02 -1.04761317e-01
-5.71422637e-01 -1.41956836e-01 -4.01253365e-02 3.49897981e-01
4.53882255e-02 -2.22620189e-01 -3.00455362e-01 3.47990662e-01
-2.09676743e-01 2.64129620e-02 4.48051244e-01 1.47611213e+00
-6.97856247e-01 -8.79659891e-01 -4.95219111e-01 2.61524826e-01
-5.68731964e-01 1.76796705e-01 1.59264669e-01 7.17765093e-01
5.23602486e-01 9.14937675e-01 1.54653400e-01 -2.58521467e-01
2.60054231e-01 -4.05457646e-01 7.09390879e-01 -2.67013580e-01
-5.17192543e-01 3.28272618e-02 -4.79582161e-01 -5.79213262e-01
-7.64542699e-01 -5.73024035e-01 -1.35010445e+00 -5.34930229e-01
-6.79360747e-01 4.63951677e-02 8.99523675e-01 7.07266092e-01
3.86930525e-01 5.40050626e-01 6.90212667e-01 -7.71941781e-01
-6.13644004e-01 -6.81392670e-01 -3.00425529e-01 1.35220513e-01
4.20761555e-01 -9.36465263e-01 -4.29035187e-01 -1.49029344e-01] | [9.589129447937012, -0.8303123116493225] |
aeff4033-3358-4c7c-b32c-966b3aea1dd6 | a-case-study-of-empirical-bayes-in-user-movie | 1707.02294 | null | http://arxiv.org/abs/1707.02294v1 | http://arxiv.org/pdf/1707.02294v1.pdf | A case study of Empirical Bayes in User-Movie Recommendation system | In this article we provide a formulation of empirical bayes described by
Atchade (2011) to tune the hyperparameters of priors used in bayesian set up of
collaborative filter. We implement the same in MovieLens small dataset. We see
that it can be used to get a good initial choice for the parameters. It can
also be used to guess an initial choice for hyper-parameters in grid search
procedure even for the datasets where MCMC oscillates around the true value or
takes long time to converge. | ['Raghav Somani', 'Sreangsu Acharyya', 'Arabin Kumar Dey'] | 2017-07-07 | null | null | null | null | ['movie-recommendation'] | ['miscellaneous'] | [-4.37014073e-01 -1.92283839e-01 1.02370270e-01 -5.19830167e-01
-5.95743060e-01 -1.03469634e+00 8.01618934e-01 1.14316247e-01
-6.84667945e-01 8.06647956e-01 2.69090831e-01 -3.21710736e-01
-4.15832937e-01 -8.76100361e-01 -3.47822070e-01 -8.51459742e-01
1.31075069e-01 9.96652365e-01 8.43492866e-01 -2.56810516e-01
6.34066999e-01 3.82825881e-01 -1.20546949e+00 5.56637459e-02
1.49334639e-01 4.18255746e-01 2.96583265e-01 8.28027844e-01
1.26007959e-01 3.56098264e-01 -7.57029116e-01 -6.01109922e-01
5.09617627e-01 -1.25159681e-01 -6.85837090e-01 -1.15688341e-02
9.33472998e-04 -1.27207235e-01 2.96067842e-03 1.20993435e+00
5.35427272e-01 5.75036347e-01 1.04276562e+00 -8.46861303e-01
4.45138812e-02 9.62923527e-01 -5.82156479e-01 3.87880862e-01
2.00002104e-01 5.46522029e-02 7.23129511e-01 -7.25087941e-01
5.79214931e-01 1.44316792e+00 6.17243826e-01 1.94008294e-02
-1.21574008e+00 -6.32274270e-01 -2.61236936e-01 4.57983725e-02
-1.67648435e+00 -2.58786887e-01 1.62861392e-01 -6.12056017e-01
2.92678148e-01 7.64945000e-02 6.17968440e-01 1.11916387e+00
-3.37624960e-02 1.52969524e-01 1.23799598e+00 -6.35977149e-01
5.43276608e-01 3.91950041e-01 3.72199416e-01 3.63568872e-01
4.49675083e-01 1.48912907e-01 -1.71217054e-01 -6.37665153e-01
1.07977760e+00 -2.95701325e-01 -1.06613867e-01 2.60605037e-01
-9.27940726e-01 1.25701153e+00 -2.83220913e-02 2.13207439e-01
-4.72596049e-01 1.76531941e-01 9.86891761e-02 3.44196022e-01
1.67792559e-01 4.22070235e-01 -4.07727271e-01 -4.14617836e-01
-9.17692721e-01 6.23954058e-01 8.82176816e-01 7.26496994e-01
4.92941022e-01 -4.37497616e-01 -3.22250247e-01 9.77724433e-01
4.73110199e-01 3.62444043e-01 2.79062420e-01 -1.34215415e+00
-4.21274826e-02 -8.64609182e-02 8.46967041e-01 -5.59765160e-01
-2.57095277e-01 -4.15725172e-01 -4.62761015e-01 1.85559183e-01
1.01342332e+00 -5.11459351e-01 -7.17561483e-01 1.17415559e+00
5.22438943e-01 3.76644820e-01 -3.53462666e-01 8.27849150e-01
4.97726589e-01 6.42143726e-01 -1.31860137e-01 -2.00813562e-01
1.36472714e+00 -5.71122944e-01 -6.69284761e-01 2.77782828e-01
1.91070035e-01 -1.10543251e+00 9.30510581e-01 8.35983753e-01
-9.36343789e-01 -3.77051324e-01 -8.87096703e-01 6.56824887e-01
-3.39779764e-01 7.66733289e-02 7.16611624e-01 1.03936684e+00
-8.92853737e-01 8.33891034e-01 -6.99686229e-01 -4.35690820e-01
1.46571457e-01 3.44333649e-01 2.88275003e-01 2.60343820e-01
-1.10469770e+00 7.78730214e-01 7.61548460e-01 -1.69488862e-01
-9.40725446e-01 -2.46853769e-01 2.14955389e-01 -1.04815364e-01
5.77625036e-01 -6.20810330e-01 1.31059086e+00 -5.04148245e-01
-1.97173584e+00 4.00678605e-01 2.39803746e-01 -5.40898740e-01
4.90025401e-01 -1.70956522e-01 -1.18204772e-01 -2.82292426e-01
-4.46789980e-01 3.35720837e-01 8.21549952e-01 -8.36854219e-01
-3.81529450e-01 -3.86373661e-02 1.18560247e-01 1.26936898e-01
9.51928869e-02 4.05395657e-01 -7.64232039e-01 -6.76775455e-01
9.98226777e-02 -1.00547183e+00 -5.06140530e-01 -4.47029799e-01
-1.78539842e-01 -4.47374254e-01 1.49115697e-01 -2.63325632e-01
1.22532451e+00 -1.81563854e+00 3.80872823e-02 6.80752695e-01
-2.73101866e-01 9.58333313e-02 2.55556941e-01 7.58492887e-01
3.02913964e-01 1.77676499e-01 2.03107148e-01 7.76654482e-03
9.97088030e-02 3.27146709e-01 -7.27185830e-02 6.08333349e-01
-5.00819802e-01 2.70650983e-01 -6.34042203e-01 -5.34169495e-01
8.92526805e-02 4.49866802e-01 -5.60169399e-01 2.10720748e-01
-1.66509122e-01 2.78618693e-01 -3.65493000e-01 1.03883371e-01
6.56122267e-01 -3.84218603e-01 2.41732135e-01 -1.48998601e-02
-9.39299241e-02 2.21041739e-01 -2.01650929e+00 1.07231963e+00
1.14427455e-01 4.07830268e-01 6.48449138e-02 -6.52514040e-01
8.46482813e-01 3.31680745e-01 1.68572128e-01 3.09178054e-01
5.25171161e-01 -2.09234476e-01 9.43668932e-02 -2.62097776e-01
3.39449763e-01 -2.50229776e-01 3.57333094e-01 7.81398714e-01
1.37141198e-02 -3.32811534e-01 3.65923017e-01 2.56269485e-01
9.23790336e-01 -5.12064844e-02 3.55781645e-01 -5.29370248e-01
5.60869813e-01 -9.51939672e-02 2.59315044e-01 1.21103942e+00
3.31959337e-01 5.56924641e-01 4.25825000e-01 -3.23260128e-01
-1.09861577e+00 -7.27838099e-01 -6.93297148e-01 1.24676263e+00
-2.94113547e-01 -6.27561510e-01 -7.02777326e-01 -5.45374572e-01
-3.36544126e-01 4.09189850e-01 -5.87455451e-01 5.16596377e-01
-2.41641030e-01 -1.20528173e+00 1.81028292e-01 5.18077314e-02
1.31556585e-01 -8.52111220e-01 -3.06041449e-01 2.45424271e-01
-3.24946456e-02 -8.40203106e-01 -3.16093296e-01 2.33835056e-01
-8.41183364e-01 -1.01489222e+00 -5.99531412e-01 5.76513931e-02
4.71823722e-01 -6.97579831e-02 1.11837101e+00 2.06849679e-01
3.22966754e-01 2.05828976e-02 -6.83705568e-01 -6.47849739e-01
-5.51881433e-01 5.34245856e-02 2.55725626e-02 -3.44111919e-01
3.86476636e-01 -3.49231392e-01 -4.89848822e-01 7.85076380e-01
-8.16029847e-01 -3.54724646e-01 3.16075645e-02 6.72765553e-01
3.46405327e-01 7.81437516e-01 2.43879929e-01 -1.13240159e+00
1.15166056e+00 -7.22590268e-01 -1.19528377e+00 2.24756971e-01
-5.83870590e-01 1.77021295e-01 1.98837087e-01 -7.19237745e-01
-8.92173767e-01 -6.59122840e-02 -2.63635784e-01 -1.34803243e-02
-8.10904875e-02 5.10595858e-01 2.36292154e-01 -1.13743611e-01
8.31586480e-01 -4.84415889e-01 -4.30024981e-01 -9.72801387e-01
1.23300269e-01 9.33988988e-01 2.58778244e-01 -6.78931653e-01
5.71156442e-01 3.36845845e-01 -9.00625512e-02 -4.15037036e-01
-9.31419313e-01 -4.96904224e-01 -6.70503974e-01 -2.10528880e-01
4.58347797e-01 -8.00335169e-01 -7.36849546e-01 2.50227690e-01
-8.11610878e-01 -3.34614635e-01 1.10756420e-03 4.98884767e-01
-3.42080683e-01 1.70297623e-01 -2.74014980e-01 -9.54571903e-01
3.46049629e-02 -8.71535659e-01 5.41588068e-01 3.77224386e-01
-2.03093305e-01 -1.12880743e+00 3.47047895e-01 1.85434654e-01
3.18019420e-01 -1.91550516e-02 4.36057478e-01 -7.86682606e-01
-1.72012031e-01 -3.54234517e-01 1.63805977e-01 -8.37908387e-02
-2.08099321e-01 4.80454355e-01 -6.61152065e-01 -4.44220930e-01
-1.46768585e-01 7.86428452e-02 5.45052946e-01 7.40568042e-01
9.90822673e-01 -2.29698092e-01 -3.64463627e-01 4.33578968e-01
1.66267025e+00 1.61819965e-01 5.39484560e-01 4.02431667e-01
-1.69386819e-03 3.89325678e-01 8.04017603e-01 1.01804686e+00
5.20124659e-02 8.78909707e-01 6.44301772e-02 4.85134035e-01
2.30135411e-01 -1.52393326e-01 4.06827144e-02 4.82555568e-01
-1.60427779e-01 -4.08855438e-01 -9.59106445e-01 2.29282394e-01
-1.96405840e+00 -9.46235061e-01 -4.27565753e-01 2.58222389e+00
9.92615342e-01 3.54067028e-01 7.50625014e-01 1.74841229e-02
8.26406121e-01 -3.33874792e-01 -4.12454978e-02 -1.99991375e-01
8.84587467e-02 1.47652715e-01 6.64865732e-01 9.33784187e-01
-8.81522834e-01 1.01577866e+00 7.95358086e+00 9.95221317e-01
-4.94359344e-01 3.73051763e-01 2.89237648e-01 -8.44442919e-02
4.68590595e-02 4.70101833e-01 -1.24300742e+00 7.85599291e-01
1.23191416e+00 1.26615584e-01 4.83086705e-01 3.82575363e-01
3.92882138e-01 -7.64785588e-01 -7.12061405e-01 9.44549263e-01
-5.99053144e-01 -1.29248011e+00 -3.35480154e-01 1.41817346e-01
7.95020163e-01 8.66929591e-02 -5.86727977e-01 3.85289155e-02
1.08329237e+00 -8.96775782e-01 2.29263112e-01 5.73331475e-01
-7.47323856e-02 -8.81536424e-01 1.01874387e+00 6.75236106e-01
-4.91335690e-01 -1.08117647e-01 -1.03875697e+00 9.75404680e-03
2.73601502e-01 7.54727364e-01 -1.00824785e+00 -4.45053242e-02
7.39763021e-01 9.53118354e-02 -5.11593103e-01 1.67774725e+00
-1.49755374e-01 1.09070563e+00 -1.02974868e+00 -3.04509044e-01
2.45302275e-01 -6.09125197e-01 5.19786298e-01 1.10878015e+00
1.78615659e-01 2.23758399e-01 1.00549586e-01 3.06596071e-01
4.31299448e-01 2.99971610e-01 -1.99515134e-01 1.25493214e-01
9.32100952e-01 1.07487357e+00 -1.10287726e+00 -3.32162529e-01
1.43588213e-02 3.53180230e-01 2.53306657e-01 5.59688330e-01
-7.46957839e-01 -1.03967346e-01 8.65862444e-02 2.76535034e-01
7.07067549e-01 -3.47579956e-01 -3.29821594e-02 -8.79880846e-01
-6.49916708e-01 -9.64514911e-01 6.55348003e-01 -7.47869611e-01
-1.23990941e+00 5.77997386e-01 8.40792954e-01 -8.71526003e-01
-3.24129373e-01 -5.96687019e-01 -7.67595232e-01 8.23973238e-01
-6.12306833e-01 -3.67591619e-01 -2.42788885e-02 5.65586209e-01
2.35397339e-01 -3.39646220e-01 5.65039337e-01 -7.33278319e-02
-4.68126178e-01 1.76220506e-01 4.08320308e-01 -1.20744817e-01
8.86764050e-01 -1.24250519e+00 1.45065650e-01 6.41647041e-01
3.67395133e-01 8.81536126e-01 1.46325588e+00 -5.88937104e-01
-8.89089525e-01 -3.46215010e-01 3.28825176e-01 -5.91046929e-01
6.96495295e-01 -2.33981296e-01 -5.98214328e-01 6.09082043e-01
4.87069786e-01 -4.54754472e-01 7.69742131e-01 4.82774615e-01
5.65293431e-02 1.01483271e-01 -1.14474511e+00 4.34860468e-01
3.51386428e-01 4.26725186e-02 -2.93310881e-01 7.71917105e-01
7.20521435e-02 -2.93192118e-01 -1.27122867e+00 -2.72019710e-02
5.98127902e-01 -1.19060206e+00 9.57847774e-01 -5.85083663e-01
-2.68271387e-01 -3.03561270e-01 -2.19725892e-01 -1.30319917e+00
-6.01324320e-01 -1.07119930e+00 1.45894781e-01 1.38990879e+00
3.80561143e-01 -5.06013036e-01 6.05145454e-01 5.14611065e-01
7.68057883e-01 -5.07037938e-01 -6.93056285e-01 -4.23863918e-01
-3.45097408e-02 -5.49171686e-01 8.43050718e-01 6.13156736e-01
-2.62894899e-01 4.24093246e-01 -3.96178067e-01 2.44148627e-01
8.77413571e-01 -1.75286233e-01 9.90065932e-01 -1.61006010e+00
-7.91405797e-01 -2.95655340e-01 1.20033622e-01 -9.84353900e-01
-3.00835282e-01 -2.87769407e-01 -1.14633672e-01 -1.11119044e+00
2.28897512e-01 -7.23824322e-01 -2.99843699e-01 1.85640827e-01
-4.93265875e-02 2.33652234e-01 2.50345487e-02 1.15723476e-01
-3.44410419e-01 -2.93312524e-03 1.09797359e+00 5.23120642e-01
-2.75175929e-01 5.48866451e-01 -5.75597286e-01 8.00401807e-01
7.41798162e-01 -1.11441684e+00 -5.36170185e-01 1.13619260e-01
8.33673239e-01 1.34877190e-01 6.12500571e-02 -6.49034381e-01
3.24575335e-01 -5.42799830e-01 2.20988482e-01 -8.59307230e-01
1.44576967e-01 -8.60725105e-01 7.10920572e-01 -7.32422918e-02
-2.80872673e-01 8.89340416e-02 -9.59241837e-02 6.70734465e-01
2.29828194e-01 -1.36084783e+00 1.00492632e+00 -3.55799347e-01
5.72030209e-02 -2.72092875e-03 -6.89566970e-01 3.22698280e-02
4.48595762e-01 -1.71886817e-01 -3.27042229e-02 -7.68560946e-01
-8.86639714e-01 2.10107312e-01 4.89213318e-01 -3.04353256e-02
4.17791381e-02 -1.27580607e+00 -7.44113445e-01 3.93661410e-02
-4.35351849e-01 -4.41571712e-01 -3.56878996e-01 9.01223719e-01
-7.11171031e-01 -6.59264550e-02 -6.64599147e-03 -4.06693876e-01
-1.24597776e+00 2.52144486e-01 3.31326127e-01 -1.11268856e-01
-3.52050781e-01 8.75455201e-01 -4.96673971e-01 -1.20328136e-01
5.71976006e-01 6.56719878e-03 -4.62559342e-01 1.65314630e-01
8.42966437e-01 5.61563969e-01 -2.08093524e-01 -2.48308897e-01
-2.07669899e-01 4.45085704e-01 1.05759585e-02 -5.76055169e-01
1.56390953e+00 -3.65661651e-01 -1.92392796e-01 5.94131947e-01
6.99489772e-01 7.96096921e-02 -1.13719189e+00 -2.43372560e-01
1.04729652e-01 -7.10513473e-01 6.13054156e-01 -7.27541506e-01
-7.03267634e-01 4.13837194e-01 7.60152996e-01 5.75145483e-01
6.48913503e-01 -1.74439639e-01 1.07140847e-01 5.91347873e-01
5.52921593e-01 -1.35519874e+00 -3.24545622e-01 4.42019373e-01
6.84939981e-01 -1.31480348e+00 7.05280066e-01 -1.48881450e-01
-7.05269158e-01 1.29968750e+00 2.13841125e-01 -4.03119892e-01
1.41299808e+00 4.11372989e-01 -9.08045843e-02 -3.52171212e-01
-9.23154473e-01 -1.05733968e-01 4.00462806e-01 4.20952916e-01
5.13179481e-01 6.12976849e-02 -7.99830854e-01 2.31248021e-01
-3.32939297e-01 -5.11347279e-02 7.89814889e-01 4.97794539e-01
-7.21807003e-01 -1.45563698e+00 -7.22050786e-01 4.07498389e-01
-8.79189014e-01 -1.97805494e-01 -1.62998736e-01 6.00477993e-01
-3.14156115e-02 1.14816713e+00 -6.27122670e-02 -1.02860570e-01
-8.33043456e-02 1.66533619e-01 7.01097548e-01 -4.73111272e-01
-8.46498430e-01 7.23689497e-01 3.96791488e-01 -8.23960975e-02
-5.36308050e-01 -7.51426220e-01 -8.26929152e-01 -3.87364924e-01
-7.34369040e-01 5.71301997e-01 5.88620365e-01 8.57387006e-01
-1.76195562e-01 -2.64483273e-01 4.97303456e-01 -7.27428198e-01
-5.93923807e-01 -1.52155352e+00 -9.06274557e-01 1.67199180e-01
-1.69154599e-01 -9.61549580e-01 -5.54560244e-01 5.39274104e-02] | [6.703224182128906, 4.007948875427246] |
cf744abf-355f-40f2-a575-2b4820cda0a4 | thin-plate-spline-motion-model-for-image | 2203.14367 | null | https://arxiv.org/abs/2203.14367v2 | https://arxiv.org/pdf/2203.14367v2.pdf | Thin-Plate Spline Motion Model for Image Animation | Image animation brings life to the static object in the source image according to the driving video. Recent works attempt to perform motion transfer on arbitrary objects through unsupervised methods without using a priori knowledge. However, it remains a significant challenge for current unsupervised methods when there is a large pose gap between the objects in the source and driving images. In this paper, a new end-to-end unsupervised motion transfer framework is proposed to overcome such issue. Firstly, we propose thin-plate spline motion estimation to produce a more flexible optical flow, which warps the feature maps of the source image to the feature domain of the driving image. Secondly, in order to restore the missing regions more realistically, we leverage multi-resolution occlusion masks to achieve more effective feature fusion. Finally, additional auxiliary loss functions are designed to ensure that there is a clear division of labor in the network modules, encouraging the network to generate high-quality images. Our method can animate a variety of objects, including talking faces, human bodies, and pixel animations. Experiments demonstrate that our method performs better on most benchmarks than the state of the art with visible improvements in pose-related metrics. | ['HUI ZHANG', 'Jian Zhao'] | 2022-03-27 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Zhao_Thin-Plate_Spline_Motion_Model_for_Image_Animation_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Zhao_Thin-Plate_Spline_Motion_Model_for_Image_Animation_CVPR_2022_paper.pdf | cvpr-2022-1 | ['image-animation'] | ['computer-vision'] | [ 2.33340845e-01 -1.94365377e-04 2.47108918e-02 -2.94929206e-01
-4.83355373e-01 -2.99956977e-01 3.73874784e-01 -7.56404638e-01
-1.79119393e-01 7.15507627e-01 2.32444242e-01 3.43926698e-01
2.07499921e-01 -6.63430154e-01 -8.09583008e-01 -7.34561265e-01
2.39624754e-01 2.77877927e-01 3.99360567e-01 -2.41596267e-01
6.50350424e-03 4.67535913e-01 -1.42567432e+00 5.10662757e-02
1.02602136e+00 6.94276154e-01 3.64739478e-01 4.20792937e-01
4.25925516e-02 7.02616632e-01 -4.10176158e-01 -4.38053191e-01
5.08050382e-01 -5.86777985e-01 -8.76579583e-01 4.84924912e-01
7.28589237e-01 -5.84977329e-01 -4.47480112e-01 1.14396179e+00
3.00382763e-01 4.17189091e-01 6.50400341e-01 -1.35994637e+00
-3.55343968e-01 2.39729106e-01 -8.52014184e-01 -2.27594823e-01
2.88235635e-01 5.51927388e-01 7.36637771e-01 -9.45993125e-01
1.09306777e+00 1.56363451e+00 3.48191857e-01 6.40958190e-01
-1.42290521e+00 -8.06313694e-01 2.15682238e-02 1.26467615e-01
-1.13724327e+00 -5.68193257e-01 1.25720406e+00 -4.13466126e-01
2.48761952e-01 3.89305316e-02 6.09314084e-01 1.07055652e+00
1.96070999e-01 5.49450278e-01 8.01535010e-01 -1.79632664e-01
-3.58914584e-02 1.03248566e-01 -4.41762984e-01 7.66413271e-01
-4.38819714e-02 1.25728250e-01 -5.93870759e-01 1.63135052e-01
1.02036977e+00 -7.53161386e-02 -4.44253117e-01 -8.88669133e-01
-1.24304175e+00 7.64235198e-01 5.66311121e-01 2.09280580e-01
-3.74289393e-01 1.64981216e-01 4.41207327e-02 -1.05627544e-01
4.51407135e-01 3.58252138e-01 -1.50847688e-01 1.31507486e-01
-9.97438252e-01 1.86819285e-01 6.31639421e-01 7.60002017e-01
9.71024334e-01 2.65010834e-01 -9.71715972e-02 5.68122625e-01
3.85933995e-01 3.36104691e-01 6.35943562e-02 -1.52435875e+00
4.69264627e-01 5.41765690e-01 2.78949261e-01 -1.37206769e+00
-1.38774887e-01 -2.51896024e-01 -8.74313056e-01 4.13374156e-01
6.05547965e-01 -2.41417135e-03 -8.93860757e-01 1.87376547e+00
7.23006427e-01 3.72047305e-01 1.28677785e-02 1.31828082e+00
5.87375879e-01 7.66511977e-01 -7.82505721e-02 -1.76672488e-01
1.08464062e+00 -1.23702073e+00 -8.42434049e-01 -3.75470966e-01
3.73581916e-01 -7.92243123e-01 1.08937502e+00 2.04157829e-02
-1.28717101e+00 -7.05248535e-01 -8.66136432e-01 -1.83488548e-01
1.05820224e-01 1.09712062e-02 4.80103791e-01 3.13496202e-01
-8.82554471e-01 6.01597369e-01 -8.31234634e-01 -1.30634800e-01
6.15232825e-01 3.61083269e-01 -4.75840658e-01 -1.90533176e-01
-1.14622211e+00 7.16259658e-01 2.65173763e-01 1.60768822e-01
-8.66833866e-01 -7.20355511e-01 -1.06349289e+00 -5.54709509e-02
3.88494104e-01 -1.00435817e+00 7.72665203e-01 -1.38501728e+00
-1.71873772e+00 6.65301204e-01 -5.50283194e-02 -1.37936831e-01
8.12812626e-01 -2.20447570e-01 1.11616217e-02 3.91480476e-01
1.14164747e-01 1.31317616e+00 1.13984025e+00 -1.36427486e+00
-5.68547904e-01 -7.41548091e-02 6.12529628e-02 2.78733552e-01
-3.92867684e-01 -1.11032188e-01 -7.23149896e-01 -7.63697147e-01
-1.60362378e-01 -1.12314224e+00 -3.00014108e-01 4.84514505e-01
-3.54391217e-01 2.11197976e-02 1.22030282e+00 -6.68313682e-01
7.58038640e-01 -2.14424300e+00 6.47813439e-01 -1.25386342e-02
2.03267917e-01 3.00541073e-01 -3.02583486e-01 -2.94751357e-02
-4.42213304e-02 -3.17209125e-01 -5.45799434e-01 -5.35835922e-01
-3.63256246e-01 3.83752257e-01 -2.60427207e-01 5.73450029e-01
4.03080016e-01 9.88449633e-01 -9.33731139e-01 -6.76287353e-01
2.55242378e-01 7.17148006e-01 -8.47242236e-01 2.29813308e-01
-1.51537761e-01 1.15042305e+00 -3.72035086e-01 3.62735957e-01
8.13330173e-01 -2.71646470e-01 -2.15414613e-01 -2.85666347e-01
5.32348789e-02 5.14022112e-02 -1.09672105e+00 2.05263638e+00
-2.14049816e-01 7.02386558e-01 1.58152372e-01 -8.79025698e-01
7.57344663e-01 -1.33896526e-02 7.65768170e-01 -3.09664220e-01
2.05546543e-01 -2.16037937e-04 1.89172611e-01 -5.42585790e-01
3.43437046e-01 1.99211352e-02 1.17188081e-01 2.17723235e-01
-8.10881928e-02 -2.35175923e-01 2.13999391e-01 6.26556873e-02
5.91702223e-01 4.53387082e-01 -4.33800697e-01 -9.24777240e-02
6.04690790e-01 1.00649120e-02 8.83624732e-01 1.21094503e-01
-1.84374750e-01 9.22264814e-01 2.76139021e-01 -4.09710199e-01
-1.18026900e+00 -8.80411804e-01 1.86711311e-01 7.28421926e-01
3.85937959e-01 -6.51944280e-02 -8.65210712e-01 -7.07125306e-01
-2.10849583e-01 4.34440494e-01 -4.63419586e-01 -2.72518098e-01
-8.86538088e-01 -3.61140102e-01 1.40944064e-01 4.91384149e-01
7.51048923e-01 -1.06398964e+00 -7.43746579e-01 1.90677881e-01
-6.44374013e-01 -1.29992127e+00 -8.96163464e-01 -4.90416169e-01
-8.62707555e-01 -7.80783892e-01 -9.14251566e-01 -1.02597272e+00
1.02492130e+00 3.22827965e-01 7.77839124e-01 2.91192345e-02
-3.05893958e-01 9.71333832e-02 -9.98921320e-02 1.77856967e-01
-3.54048103e-01 -6.58290610e-02 -1.19709097e-01 3.70525181e-01
-3.31492573e-01 -6.26949012e-01 -8.92510414e-01 5.94990492e-01
-9.67403948e-01 2.97479063e-01 4.75574285e-01 8.40607166e-01
3.74276727e-01 -7.64022470e-02 1.83127925e-01 -4.81298804e-01
3.85576598e-02 -1.14225224e-01 -5.64574480e-01 -1.43227145e-01
-1.46277443e-01 9.38493460e-02 5.55508971e-01 -7.87260413e-01
-1.10163546e+00 6.59975350e-01 1.27236784e-01 -8.76645148e-01
1.13791898e-01 2.07801461e-02 -5.03596902e-01 -1.36447355e-01
4.99317288e-01 2.03227565e-01 4.03226912e-01 -2.26433158e-01
5.67917168e-01 1.79700404e-01 7.68897593e-01 -3.29510748e-01
1.25678301e+00 8.63268137e-01 6.38987646e-02 -6.48622692e-01
-6.91577673e-01 -1.73021734e-01 -8.42130601e-01 -3.44343394e-01
1.04347682e+00 -7.49010980e-01 -5.55309236e-01 4.77211267e-01
-1.31165493e+00 -2.86498368e-01 -1.67524442e-01 5.14752090e-01
-6.38581634e-01 3.24257016e-01 -4.39216703e-01 -4.67965961e-01
-2.41814345e-01 -1.44963670e+00 1.06361330e+00 3.17945659e-01
-2.27190405e-01 -8.42060804e-01 -1.00306667e-01 4.56776887e-01
2.95272261e-01 4.49133754e-01 5.75808227e-01 1.13358386e-01
-9.27903831e-01 1.44452468e-01 -9.98829529e-02 3.60120595e-01
3.32295597e-01 2.19558567e-01 -7.23464489e-01 -4.42308933e-01
-2.26494819e-02 -1.64767385e-01 7.85543323e-01 3.78180355e-01
1.01564801e+00 -4.08528805e-01 -3.79111350e-01 9.32920754e-01
1.01237941e+00 1.52560815e-01 9.16341066e-01 1.62868261e-01
1.10314155e+00 1.08271801e+00 7.26506710e-01 2.51713693e-02
3.98795545e-01 8.61786544e-01 4.55755591e-01 -3.17972034e-01
-3.59513640e-01 -2.62780875e-01 4.79626775e-01 6.25375390e-01
-3.06886509e-02 -3.09109990e-03 -6.28873646e-01 5.87976396e-01
-1.81039464e+00 -8.79370570e-01 -9.90876630e-02 2.18758941e+00
7.55399466e-01 2.46687196e-02 1.01146139e-01 -1.51751176e-01
8.38154018e-01 2.42916048e-01 -6.53275490e-01 1.93702742e-01
6.98333606e-02 -1.00275569e-01 2.16125488e-01 5.78228235e-01
-1.15196133e+00 1.17060530e+00 5.58188057e+00 5.41670263e-01
-1.34222698e+00 3.61317433e-02 6.59147441e-01 -4.17719670e-02
-1.93384260e-01 2.23755673e-01 -5.49063683e-01 5.86715102e-01
4.49012160e-01 -7.47795627e-02 2.70527244e-01 6.43685460e-01
3.17821532e-01 4.48578894e-02 -1.00522971e+00 9.73610163e-01
5.47135845e-02 -1.20901263e+00 1.08778477e-02 1.08712845e-01
9.69259739e-01 -4.17116314e-01 2.00822920e-01 -4.86401580e-02
1.34505957e-01 -8.23008478e-01 6.52583957e-01 6.33337259e-01
6.39002085e-01 -8.52331281e-01 2.48205170e-01 3.20621371e-01
-1.08619475e+00 2.47143134e-01 -4.02977616e-01 1.54033467e-01
4.27329391e-01 4.00703400e-01 -4.98952717e-01 4.81390268e-01
6.29007816e-01 7.07035780e-01 -2.83986419e-01 9.68807876e-01
-2.61958003e-01 1.82072088e-01 -2.18373835e-01 3.93142074e-01
2.10561901e-01 -3.17048818e-01 8.39841008e-01 7.60377407e-01
1.83575228e-01 -4.53942195e-02 2.04300553e-01 1.12247396e+00
-1.75712034e-01 7.95942172e-02 -6.45948350e-01 2.56785482e-01
1.67569324e-01 1.43764901e+00 -7.53546298e-01 -2.64099509e-01
-2.24654689e-01 1.20494032e+00 1.90712929e-01 4.19165462e-01
-1.02235079e+00 -2.26701289e-01 6.65892482e-01 3.56151313e-01
3.96401405e-01 -2.77172476e-01 1.07233323e-01 -1.29728806e+00
3.60351764e-02 -7.28582025e-01 7.76969688e-03 -7.98706770e-01
-1.03396964e+00 4.08760369e-01 -5.73637895e-02 -1.60353744e+00
-4.64438319e-01 -2.03395233e-01 -6.79467618e-01 6.93307996e-01
-1.36536276e+00 -1.09786022e+00 -5.55078506e-01 6.11225963e-01
6.25456452e-01 1.04358330e-01 2.83984959e-01 3.53824794e-01
-4.80645776e-01 4.48048651e-01 -2.18933463e-01 3.12930614e-01
1.00424850e+00 -6.94823682e-01 2.95787394e-01 9.70756471e-01
-2.40593851e-02 4.13664073e-01 6.15779102e-01 -5.87240040e-01
-1.40125144e+00 -1.36479950e+00 4.45907921e-01 -3.78487259e-01
2.21482158e-01 -3.69584501e-01 -1.21608019e+00 4.79463905e-01
2.32905880e-01 3.56075972e-01 1.09950967e-01 -6.54637456e-01
-8.88271406e-02 -2.07492515e-01 -1.09287155e+00 8.36647511e-01
1.12338221e+00 -1.84016213e-01 -3.75771374e-01 1.67836040e-01
6.94890797e-01 -5.34056723e-01 -6.10098243e-01 4.03086007e-01
5.56354463e-01 -7.60876179e-01 1.07358289e+00 -5.55312216e-01
9.18200672e-01 -5.93990684e-01 2.02157930e-01 -1.32416022e+00
-1.41293034e-01 -1.01936054e+00 -1.13681734e-01 1.32179701e+00
1.24949344e-01 -3.79942417e-01 8.69381964e-01 5.86231649e-01
8.95094797e-02 -5.17595828e-01 -7.65668690e-01 -6.63721442e-01
9.49014276e-02 3.55004109e-02 2.89409071e-01 1.08572161e+00
-3.80199164e-01 3.32289428e-01 -6.68964028e-01 1.17005818e-01
7.93415785e-01 2.20062301e-01 1.18205786e+00 -1.09168291e+00
-1.25670284e-01 -3.18882614e-01 -4.63349432e-01 -1.34840155e+00
4.50353801e-01 -8.63073766e-01 2.32439652e-01 -1.29679978e+00
1.07955031e-01 -2.14834288e-01 9.40064266e-02 4.02152598e-01
-3.04363877e-01 5.07462978e-01 3.88808757e-01 4.72472876e-01
-3.96416336e-01 9.41732526e-01 2.02424717e+00 -1.83913991e-01
-4.56720471e-01 -2.05442801e-01 -4.87135082e-01 9.32650924e-01
6.11105263e-01 -5.85822880e-01 -4.02940601e-01 -4.74584639e-01
-3.08664799e-01 2.04140902e-01 5.68569720e-01 -9.29650068e-01
2.09170014e-01 -2.99596965e-01 5.66954851e-01 -5.23055732e-01
5.26684880e-01 -8.50266993e-01 3.20860416e-01 4.97615099e-01
-1.67904735e-01 -1.75798073e-01 -3.91113050e-02 5.21661341e-01
-1.16855659e-01 5.45732789e-02 1.20870125e+00 1.14702508e-01
-5.27611196e-01 4.54053491e-01 8.23976845e-03 1.02579728e-01
1.14783490e+00 -1.61351770e-01 -8.02221335e-03 -5.85368693e-01
-5.71896613e-01 3.78111869e-01 7.62950361e-01 6.40535414e-01
7.29985476e-01 -1.38901901e+00 -7.51544595e-01 1.11304559e-01
-1.46214560e-01 4.12191540e-01 2.02763245e-01 1.01085627e+00
-6.05209291e-01 1.91979147e-02 -5.52066565e-01 -8.03500712e-01
-1.16755927e+00 4.82643753e-01 3.33730727e-01 -1.24136535e-02
-8.71849298e-01 5.38579643e-01 6.92673564e-01 -2.34097227e-01
1.22255661e-01 -1.51531279e-01 -9.23639722e-03 -2.41606124e-02
4.50790733e-01 3.18901747e-01 -4.35297072e-01 -1.15647018e+00
-2.26344094e-01 8.48912716e-01 -2.08604291e-01 -1.67181879e-01
1.30507278e+00 -2.40053385e-01 4.95939702e-02 -2.97348225e-03
1.37534821e+00 8.69804546e-02 -2.05641556e+00 -1.22067757e-01
-4.32942629e-01 -8.02479506e-01 -1.01558909e-01 -1.99467525e-01
-1.64765525e+00 8.04199100e-01 4.71929073e-01 -4.34437335e-01
1.09628284e+00 -4.06787060e-02 1.02167356e+00 5.79303280e-02
2.73647100e-01 -8.73979926e-01 4.32647824e-01 2.07257807e-01
9.51648057e-01 -1.20785880e+00 -3.47278155e-02 -7.32243359e-01
-6.93447113e-01 9.52386439e-01 8.26273978e-01 -3.20416421e-01
2.39474088e-01 5.99087253e-02 9.84116644e-02 5.14015481e-02
-4.52712625e-01 -1.13378003e-01 4.35351402e-01 6.15272880e-01
-5.93923628e-02 -4.00766015e-01 -2.16413394e-01 2.42529754e-02
-1.74105436e-01 -4.44719233e-02 4.62594420e-01 8.46423030e-01
-2.07116708e-01 -1.05497646e+00 -4.29524511e-01 -1.96991861e-02
-3.72016668e-01 1.62839547e-01 -2.81650811e-01 7.96919286e-01
1.04011949e-02 7.34176576e-01 1.43405095e-01 -1.53284326e-01
2.40201220e-01 -1.45548493e-01 5.26086986e-01 -4.17004317e-01
-1.47327751e-01 3.43531668e-01 -2.78332353e-01 -7.84943819e-01
-6.62617564e-01 -5.19394159e-01 -1.41564727e+00 -1.29566044e-01
-2.62459427e-01 -2.22939551e-01 4.47129071e-01 7.63590157e-01
4.02016968e-01 4.86993283e-01 8.77295494e-01 -1.35827160e+00
-2.84716100e-01 -7.01726198e-01 -1.79273725e-01 6.49020195e-01
3.50088984e-01 -9.08623576e-01 -3.51317644e-01 5.47686517e-01] | [10.908447265625, -0.8358219265937805] |
c6d37c04-60fa-4bf4-828c-9cfe4703d0c9 | apsnet-attention-based-point-cloud-sampling | 2210.05638 | null | https://arxiv.org/abs/2210.05638v1 | https://arxiv.org/pdf/2210.05638v1.pdf | APSNet: Attention Based Point Cloud Sampling | Processing large point clouds is a challenging task. Therefore, the data is often downsampled to a smaller size such that it can be stored, transmitted and processed more efficiently without incurring significant performance degradation. Traditional task-agnostic sampling methods, such as farthest point sampling (FPS), do not consider downstream tasks when sampling point clouds, and thus non-informative points to the tasks are often sampled. This paper explores a task-oriented sampling for 3D point clouds, and aims to sample a subset of points that are tailored specifically to a downstream task of interest. Similar to FPS, we assume that point to be sampled next should depend heavily on the points that have already been sampled. We thus formulate point cloud sampling as a sequential generation process, and develop an attention-based point cloud sampling network (APSNet) to tackle this problem. At each time step, APSNet attends to all the points in a cloud by utilizing the history of previously sampled points, and samples the most informative one. Both supervised learning and knowledge distillation-based self-supervised learning of APSNet are proposed. Moreover, joint training of APSNet over multiple sample sizes is investigated, leading to a single APSNet that can generate arbitrary length of samples with prominent performances. Extensive experiments demonstrate the superior performance of APSNet against state-of-the-arts in various downstream tasks, including 3D point cloud classification, reconstruction, and registration. | ['Shihao Ji', 'Xiulong Yang', 'Yang Ye'] | 2022-10-11 | null | null | null | null | ['3d-point-cloud-classification', 'point-cloud-classification'] | ['computer-vision', 'computer-vision'] | [ 1.77188084e-01 -1.51821047e-01 -1.86560750e-01 -3.74937713e-01
-1.11862266e+00 -1.34830445e-01 4.43574369e-01 2.10253835e-01
-1.75213024e-01 6.69521093e-01 -1.87396318e-01 8.67350698e-02
-4.33104672e-02 -1.15084124e+00 -1.04111302e+00 -7.20081449e-01
5.81238195e-02 1.01280451e+00 2.83945352e-01 1.73864424e-01
3.57383788e-01 1.03700924e+00 -1.72366822e+00 9.02377591e-02
1.03437936e+00 1.17498755e+00 7.84406245e-01 3.20831597e-01
-4.89901513e-01 1.11655675e-01 -6.32100582e-01 1.41439572e-01
2.66575456e-01 1.32107824e-01 -7.06179202e-01 4.65955734e-01
5.07806540e-01 -4.27727431e-01 -1.63836386e-02 9.11284029e-01
4.60968554e-01 2.75671482e-01 5.98416507e-01 -1.18517971e+00
-4.63039160e-01 3.39929044e-01 -5.97120225e-01 -2.99854614e-02
5.64335920e-02 4.18981701e-01 7.54808486e-01 -1.32675588e+00
4.26497728e-01 1.26237619e+00 5.05372107e-01 3.93093944e-01
-1.21456063e+00 -8.27581823e-01 5.47591567e-01 -1.78204089e-01
-1.43548453e+00 -1.68239698e-01 1.09116852e+00 -3.65277201e-01
6.05778575e-01 1.38271600e-01 1.08908522e+00 8.22044551e-01
-4.52633709e-01 1.02716827e+00 6.26084328e-01 2.76467646e-03
5.42226017e-01 -6.63706437e-02 -1.11385137e-01 2.41866007e-01
3.62343878e-01 -3.89946178e-02 -5.58413684e-01 -3.80682230e-01
9.82636333e-01 4.90191132e-01 -2.93614835e-01 -4.16878819e-01
-1.40565610e+00 6.65294766e-01 7.20301867e-01 -7.24427551e-02
-8.88487697e-01 1.22776397e-01 2.92312771e-01 7.57838115e-02
8.51531088e-01 3.06410909e-01 -5.63343525e-01 1.29295036e-01
-1.20389700e+00 4.72385794e-01 5.09211838e-01 1.55560768e+00
1.24542105e+00 -1.04651399e-01 -3.09368044e-01 6.79947555e-01
1.93360537e-01 6.62911475e-01 1.28728241e-01 -8.55275989e-01
6.84027433e-01 7.39977121e-01 3.20044875e-01 -7.15572774e-01
6.55237958e-02 -3.28056425e-01 -9.19864416e-01 2.16962159e-01
-5.01209535e-02 -5.56506179e-02 -9.27727878e-01 1.24833858e+00
8.99261236e-01 5.32148659e-01 -2.41750360e-01 9.46262240e-01
6.07373238e-01 9.13488030e-01 -1.61279663e-01 -1.55329317e-01
1.10303307e+00 -7.33955383e-01 -3.97507958e-02 -1.46976680e-01
1.01038292e-01 -5.78760386e-01 1.19140565e+00 4.91737098e-01
-1.04480278e+00 -8.48812580e-01 -7.79426813e-01 -1.23774126e-01
4.40293411e-03 -1.66068990e-02 4.51818436e-01 -3.39822657e-02
-7.36141145e-01 9.40351844e-01 -8.18376243e-01 -2.42168419e-02
1.02008665e+00 2.17854947e-01 1.51273608e-01 -3.98156583e-01
-5.52043140e-01 1.74580470e-01 2.49212697e-01 1.04137830e-01
-1.05928171e+00 -1.17068815e+00 -5.20217955e-01 2.42238417e-02
3.51446062e-01 -9.54004765e-01 1.21221924e+00 -6.31897748e-01
-1.23072433e+00 4.97712761e-01 -6.28898263e-01 -2.46318683e-01
5.59512019e-01 -3.12828481e-01 1.07889180e-03 2.46491686e-01
4.85013872e-01 9.14178252e-01 1.17419648e+00 -1.59479856e+00
-9.60497439e-01 -4.57625717e-01 -1.34853646e-01 3.67021829e-01
-1.96718797e-01 -5.08765996e-01 -7.68136024e-01 -3.38831633e-01
3.31373900e-01 -7.49893725e-01 -4.22814846e-01 3.97730112e-01
-4.61634666e-01 -6.07900679e-01 1.00177515e+00 -6.04774952e-02
5.72438240e-01 -2.12110043e+00 -1.46385819e-01 2.70641804e-01
2.91588455e-01 1.58005133e-01 -1.43404707e-01 3.65392029e-01
2.28884757e-01 -1.04458950e-01 -3.08116883e-01 -7.76253879e-01
-1.50432482e-01 3.15783978e-01 -6.48932159e-01 3.98372978e-01
4.48840648e-01 6.89848542e-01 -1.20244026e+00 -4.80673969e-01
4.06685442e-01 4.42122996e-01 -4.91524786e-01 2.87700325e-01
-8.00145447e-01 6.37601018e-01 -1.01193726e+00 9.37543750e-01
1.13785946e+00 -5.87990046e-01 -4.86451983e-01 -2.07090795e-01
-3.07242125e-02 1.78048313e-01 -1.04876661e+00 1.78224409e+00
-7.53010690e-01 1.41136244e-01 3.53926234e-02 -7.32913971e-01
1.11485100e+00 7.50308856e-02 6.70926511e-01 -8.07900950e-02
-2.66869605e-01 2.59078771e-01 -4.48299736e-01 -2.57673234e-01
7.15534627e-01 -9.32268873e-02 2.64956623e-01 2.02017024e-01
-3.53026271e-01 -6.70482337e-01 -1.39768794e-02 -2.77026501e-02
8.47719431e-01 1.94696099e-01 1.90048516e-02 3.15886773e-02
3.64172786e-01 4.31976646e-01 5.41263640e-01 7.04382837e-01
1.32646590e-01 9.11367297e-01 1.12983733e-01 -4.66965288e-01
-1.08454335e+00 -9.28079545e-01 -1.95294872e-01 5.61667681e-01
4.14104134e-01 -1.18529774e-01 -3.76167506e-01 -6.81666136e-01
4.88242328e-01 6.72037661e-01 -2.74713218e-01 1.40544280e-01
-5.34756362e-01 -2.55301803e-01 -5.33571988e-02 3.74853522e-01
3.42878312e-01 -1.33243918e+00 -5.56423843e-01 2.35084459e-01
1.23494025e-02 -9.36396122e-01 -4.62400228e-01 -3.17387842e-02
-1.38647294e+00 -1.07472992e+00 -7.73541093e-01 -6.48139536e-01
1.00267887e+00 7.04281688e-01 1.29771245e+00 1.16966091e-01
1.48454949e-01 3.12330037e-01 -4.45504397e-01 -6.70029819e-01
-3.66427191e-02 3.75938684e-01 -1.51691094e-01 1.76343307e-01
5.37270367e-01 -7.98762500e-01 -7.22489715e-01 1.95538044e-01
-8.46139729e-01 -9.68756974e-02 7.95143962e-01 6.94678128e-01
1.38388050e+00 -2.04254128e-02 5.78266740e-01 -1.08170688e+00
4.30071205e-01 -6.40133381e-01 -6.43196344e-01 -2.00901166e-01
-3.14063877e-01 -2.13141769e-01 8.61793280e-01 -3.84853870e-01
-7.30684996e-01 1.72041252e-01 -8.83157775e-02 -1.30000138e+00
-3.31225187e-01 3.00003737e-01 -2.68682361e-01 -4.99534085e-02
4.70409095e-01 5.31484663e-01 5.23052923e-02 -6.42679274e-01
1.50783613e-01 5.76009393e-01 1.99849203e-01 -8.00415754e-01
1.04897809e+00 8.36869657e-01 6.98898509e-02 -9.92445827e-01
-7.96079040e-01 -6.91514373e-01 -5.17398119e-01 -2.35080197e-01
4.08646196e-01 -1.12305999e+00 -7.45839953e-01 2.97938108e-01
-1.34566593e+00 -2.01673001e-01 -9.06043231e-01 2.48206273e-01
-5.88806808e-01 2.51446426e-01 -1.37690946e-01 -8.91178846e-01
-5.33475697e-01 -1.13801444e+00 1.73309124e+00 -1.92786083e-02
1.23374939e-01 -5.19133866e-01 -2.29053840e-01 7.47033060e-02
1.09663039e-01 3.22201282e-01 6.17818475e-01 -4.89169300e-01
-1.30400050e+00 -4.65927660e-01 -2.18304813e-01 3.42674881e-01
2.15587452e-01 -2.31277756e-02 -9.54700112e-01 -4.09575939e-01
1.29875243e-01 -3.87096524e-01 7.89853990e-01 4.64065969e-01
1.82984579e+00 -2.52071112e-01 -6.76066875e-01 7.80621469e-01
1.39988768e+00 -1.00083336e-01 4.28177208e-01 -4.69892360e-02
8.74502182e-01 4.06688988e-01 1.02328384e+00 6.86042309e-01
4.29373175e-01 3.37573230e-01 8.66762519e-01 1.00148365e-01
9.87736285e-02 -4.64744508e-01 -1.75231218e-01 7.46777296e-01
-4.44016457e-02 -1.02177627e-01 -7.45522559e-01 7.64977694e-01
-1.77642822e+00 -7.60219693e-01 -5.62455282e-02 2.37846899e+00
7.80120432e-01 4.93726879e-02 2.64367536e-02 -7.24324882e-02
7.28360951e-01 4.08684492e-01 -1.07479084e+00 5.34721255e-01
3.54664385e-01 3.05209190e-01 3.83163899e-01 3.06923576e-02
-9.40780520e-01 8.09701204e-01 4.92887974e+00 1.01872206e+00
-1.17668176e+00 -4.65409644e-02 6.30076170e-01 -4.42372501e-01
-4.05992031e-01 2.07688045e-02 -1.00369382e+00 8.26549351e-01
4.62589562e-01 -6.03055432e-02 3.29078704e-01 1.35071766e+00
3.01735520e-01 -2.75612250e-02 -1.23527658e+00 1.09541273e+00
-2.51375139e-01 -1.62572718e+00 3.93798321e-01 1.02128997e-01
7.58667529e-01 4.94588137e-01 -5.15019000e-02 2.50741065e-01
3.19391191e-01 -5.53539634e-01 7.87221968e-01 7.11293161e-01
8.84293735e-01 -8.45603704e-01 4.03808385e-01 6.57389641e-01
-1.30196917e+00 1.50963664e-01 -8.77332628e-01 1.88616261e-01
2.98290968e-01 1.27370453e+00 -9.26007271e-01 7.71315873e-01
7.35443830e-01 1.01880968e+00 -2.18947962e-01 1.34635997e+00
1.92010757e-02 4.33711708e-01 -6.15992248e-01 -2.13064864e-01
2.29819089e-01 -3.75115991e-01 7.31197715e-01 6.56577706e-01
6.98161364e-01 -8.66601360e-04 4.89765078e-01 1.07755399e+00
-1.95663020e-01 -8.24483708e-02 -6.62506640e-01 3.95598024e-01
1.03482056e+00 1.12226379e+00 -5.57281673e-01 -6.04834795e-01
-2.96726406e-01 7.80589938e-01 4.94245768e-01 2.41707265e-01
-3.79520029e-01 -3.14500123e-01 9.68775988e-01 4.15374786e-01
7.01188982e-01 -1.02181971e-01 -3.32940787e-01 -9.90771770e-01
3.53451222e-01 -4.90228325e-01 6.71123043e-02 -8.42700958e-01
-1.74653327e+00 5.74702203e-01 3.87288965e-02 -1.95812607e+00
7.57655725e-02 -1.36786684e-01 -7.97564805e-01 1.23186302e+00
-1.83617234e+00 -9.97531414e-01 -7.02055812e-01 4.79356706e-01
8.03356290e-01 1.29148647e-01 3.40454131e-01 1.81404501e-01
-2.04353318e-01 1.10327080e-01 -1.68319624e-02 -1.56949401e-01
2.81691462e-01 -1.10982919e+00 7.39864588e-01 6.73253238e-01
-1.02102429e-01 5.11476517e-01 2.17850789e-01 -9.00013685e-01
-1.50318301e+00 -1.83235645e+00 5.54048538e-01 -2.49609143e-01
1.70117423e-01 -3.19990069e-01 -1.07380879e+00 3.14695179e-01
-3.68074417e-01 4.80725408e-01 1.33115485e-01 -1.33563936e-01
1.02739513e-01 -4.64715898e-01 -1.25785220e+00 4.33707297e-01
1.30543363e+00 -3.10860455e-01 -2.67755508e-01 7.33212888e-01
1.11173618e+00 -5.44364154e-01 -7.28570223e-01 3.39752823e-01
-2.35032231e-01 -8.11937153e-01 1.14831960e+00 -3.36804241e-01
5.18598139e-01 -4.90497768e-01 1.47080636e-02 -1.37618887e+00
-3.34255904e-01 -5.20065606e-01 -2.74398416e-01 1.13683379e+00
1.18774727e-01 -5.85431099e-01 1.24664140e+00 4.22328174e-01
-5.67868590e-01 -9.07049119e-01 -9.75422621e-01 -7.43519723e-01
-1.70168787e-01 -6.08042002e-01 1.29849780e+00 8.04637194e-01
-6.59529567e-01 2.20893592e-01 -1.46176042e-02 5.28488815e-01
8.64725232e-01 7.57611871e-01 1.26861608e+00 -1.46947145e+00
-5.17912842e-02 -2.72236884e-01 -1.40609980e-01 -1.68668461e+00
-1.39802977e-01 -8.71890783e-01 2.32933044e-01 -1.66923201e+00
-3.01184356e-01 -1.08351040e+00 1.81730434e-01 9.07717422e-02
-3.56488049e-01 -1.54205039e-01 1.88963220e-01 7.08319902e-01
-3.71304244e-01 8.73366475e-01 1.62208593e+00 -1.07566923e-01
-3.28260541e-01 5.92237830e-01 -5.37246406e-01 5.02146125e-01
5.44542909e-01 -4.99526143e-01 -5.58924556e-01 -5.72063565e-01
-7.35981390e-02 2.07928464e-01 3.79485965e-01 -1.17631578e+00
2.68751532e-01 -2.77330250e-01 5.71248531e-01 -1.44306242e+00
6.80382252e-01 -1.03164661e+00 2.14026645e-01 1.54940560e-01
-1.94642991e-02 -1.70742601e-01 -1.16353370e-01 9.63368714e-01
-1.77713990e-01 -3.17409843e-01 5.73015392e-01 -4.76007760e-01
-4.29066092e-01 1.22751510e+00 2.96331137e-01 -6.63274303e-02
1.01362205e+00 -3.15859526e-01 1.37186289e-01 2.89880559e-02
-4.91028845e-01 5.31632960e-01 5.65292776e-01 5.80375679e-02
8.42273355e-01 -1.40683341e+00 -7.17484534e-01 3.27939749e-01
3.33263397e-01 1.48152363e+00 3.07797253e-01 5.31367540e-01
-4.47734654e-01 5.75864278e-02 3.25478852e-01 -1.08446467e+00
-6.44617736e-01 7.40660548e-01 -1.82978511e-01 -5.00091240e-02
-1.01773942e+00 8.35107327e-01 8.51879567e-02 -5.27000666e-01
1.36736423e-01 -7.58612454e-01 3.68918777e-02 5.65960184e-02
4.05397028e-01 2.81003058e-01 1.69653684e-01 -1.19232886e-01
-8.57486427e-02 5.78617811e-01 -7.95134827e-02 2.72512197e-01
1.59037697e+00 1.57364860e-01 -8.58939886e-02 5.58027804e-01
1.08505857e+00 -1.62293509e-01 -1.74664998e+00 -7.14462042e-01
-3.38635862e-01 -1.03941083e+00 -2.07556528e-03 -1.80313602e-01
-1.30568635e+00 7.27845192e-01 8.55777115e-02 3.66571039e-01
1.03283989e+00 1.45141780e-01 9.82649624e-01 3.54224175e-01
7.85661757e-01 -8.40632677e-01 -8.77153873e-02 3.49608809e-01
9.24073637e-01 -1.20065367e+00 1.28094926e-01 -4.92122978e-01
-5.49747765e-01 9.36351478e-01 7.73679852e-01 -4.20448393e-01
7.87335038e-01 -7.88662024e-03 -4.26276624e-01 -3.40206683e-01
-8.09238553e-01 -9.01594087e-02 5.95110804e-02 7.83734143e-01
-2.85087973e-01 8.64207596e-02 4.42211062e-01 2.15516835e-01
-1.27791211e-01 3.25508565e-01 1.55570716e-01 9.95426774e-01
-5.16799867e-01 -7.83440650e-01 -5.90446174e-01 1.09593248e+00
3.09859455e-01 2.55909748e-02 1.14466827e-02 4.98872310e-01
7.25576878e-02 3.10627997e-01 4.52215999e-01 -1.98962957e-01
4.84970570e-01 -2.46042982e-01 7.62021393e-02 -9.69054759e-01
-4.27049667e-01 8.87661800e-03 -3.27464640e-01 -4.60325360e-01
-4.03700709e-01 -9.59164202e-01 -1.15776050e+00 -4.70668137e-01
-3.68734479e-01 1.28423527e-01 4.88883823e-01 5.76660633e-01
8.41797471e-01 5.10790586e-01 1.00253344e+00 -1.69701588e+00
-6.72225535e-01 -8.58681321e-01 -5.32842100e-01 3.12760383e-01
5.45485616e-01 -5.50979733e-01 -2.16501459e-01 -2.08556086e-01] | [8.253817558288574, -3.519660472869873] |
5bcb77c2-4941-46f6-b9af-77855b81eb95 | fraunhofer-sit-at-checkthat-2023-mixing | 2307.00610 | null | https://arxiv.org/abs/2307.00610v1 | https://arxiv.org/pdf/2307.00610v1.pdf | Fraunhofer SIT at CheckThat! 2023: Mixing Single-Modal Classifiers to Estimate the Check-Worthiness of Multi-Modal Tweets | The option of sharing images, videos and audio files on social media opens up new possibilities for distinguishing between false information and fake news on the Internet. Due to the vast amount of data shared every second on social media, not all data can be verified by a computer or a human expert. Here, a check-worthiness analysis can be used as a first step in the fact-checking pipeline and as a filtering mechanism to improve efficiency. This paper proposes a novel way of detecting the check-worthiness in multi-modal tweets. It takes advantage of two classifiers, each trained on a single modality. For image data, extracting the embedded text with an OCR analysis has shown to perform best. By combining the two classifiers, the proposed solution was able to place first in the CheckThat! 2023 Task 1A with an F1 score of 0.7297 achieved on the private test set. | ['Inna Vogel', 'Raphael Frick'] | 2023-07-02 | null | null | null | null | ['optical-character-recognition'] | ['computer-vision'] | [ 7.62966573e-02 4.03800666e-01 -4.87681963e-02 -7.74215534e-02
-9.65310454e-01 -6.63455009e-01 8.90069366e-01 6.42698407e-01
-6.03069782e-01 6.01338983e-01 -2.26082996e-01 -2.82429338e-01
1.66183457e-01 -7.13897765e-01 -6.06782854e-01 -6.03205144e-01
3.57852690e-02 3.42381060e-01 6.62347496e-01 -1.07800633e-01
4.78235632e-01 3.08388829e-01 -1.91113985e+00 7.98927188e-01
5.46504080e-01 1.20602751e+00 1.20856605e-01 4.97736871e-01
-8.48294944e-02 8.08185875e-01 -6.59406364e-01 -9.77677166e-01
1.63602412e-01 -2.63495594e-01 -7.77536094e-01 -1.22780660e-02
5.55323958e-01 -4.48318362e-01 3.44070569e-02 1.27413392e+00
2.51327336e-01 -5.73197186e-01 2.76797324e-01 -1.24215841e+00
4.29619141e-02 8.13733876e-01 -3.98617446e-01 2.33607978e-01
7.48399436e-01 6.34121895e-02 8.18653882e-01 -7.36551583e-01
9.50024962e-01 7.56713212e-01 5.84781647e-01 1.09711610e-01
-1.05459356e+00 -5.03493249e-01 -6.54890954e-01 3.37730378e-01
-1.46445167e+00 -4.23247874e-01 4.08183694e-01 -5.87811053e-01
6.45345032e-01 3.99260819e-01 7.21582055e-01 1.05423760e+00
4.20425609e-02 7.91744649e-01 1.53908622e+00 -6.22686625e-01
4.40472662e-02 9.12971377e-01 -7.75400996e-02 6.98513508e-01
3.40361893e-01 -1.27069697e-01 -7.55790412e-01 -2.67282069e-01
-4.84910700e-03 -2.17556641e-01 -2.22696021e-01 4.81489208e-03
-1.07749426e+00 6.23018742e-01 2.58035243e-01 7.15351403e-01
-3.48235846e-01 -2.02568740e-01 5.72553694e-01 7.08188772e-01
7.60470510e-01 4.21717167e-01 -2.46649861e-01 -2.58714527e-01
-1.25902534e+00 9.79590192e-02 6.57795072e-01 3.42971146e-01
5.78287482e-01 -5.78883111e-01 8.84744804e-03 4.46703225e-01
1.93893731e-01 5.67121923e-01 6.65986359e-01 -1.84053466e-01
5.82165360e-01 5.88416040e-01 2.97388285e-01 -1.45242357e+00
-3.83344769e-01 -3.71658653e-01 -3.87307227e-01 8.90167356e-02
6.83129489e-01 1.11254096e-01 -4.49908346e-01 1.07943010e+00
4.85974342e-01 5.36235012e-02 -2.88879007e-01 7.89335132e-01
4.74180549e-01 3.98803681e-01 -2.92973608e-01 -1.29310057e-01
1.66880119e+00 -5.07448494e-01 -7.23073602e-01 2.34509990e-01
7.76248097e-01 -1.04055667e+00 7.07912028e-01 8.34658742e-01
-9.06601012e-01 -2.81591326e-01 -1.18258762e+00 7.05424324e-02
-7.68920481e-01 2.28468850e-01 1.76307470e-01 1.07389247e+00
-7.73139000e-01 7.34758377e-01 -5.71561038e-01 -2.82519072e-01
5.28379977e-01 2.22296529e-02 -6.48031175e-01 -6.32578656e-02
-1.10954213e+00 8.87030005e-01 6.07710183e-01 -9.20509398e-02
-3.50877196e-01 -9.61961970e-02 -3.42189401e-01 3.57947014e-02
3.15390259e-01 -2.06456166e-02 7.85727143e-01 -1.21077454e+00
-1.27116573e+00 1.48642850e+00 1.90227002e-01 -7.63740063e-01
1.26135159e+00 -9.91509631e-02 -6.71442688e-01 6.30535305e-01
5.42778596e-02 2.84681469e-01 1.40316772e+00 -9.43006158e-01
-9.00441468e-01 -2.52690554e-01 -1.06508888e-01 -4.36288923e-01
-4.11457211e-01 2.22008049e-01 -2.62185931e-01 -5.24155200e-01
1.38352752e-01 -1.00871599e+00 6.67530477e-01 -3.00715744e-01
-6.05585754e-01 8.88699219e-02 9.35184002e-01 -1.04611945e+00
1.16310692e+00 -2.17835116e+00 -2.88126707e-01 5.07399499e-01
2.26665139e-01 6.47723734e-01 2.98438281e-01 4.39390838e-01
-1.25036053e-02 2.13230118e-01 1.46647096e-01 -2.64127433e-01
-1.29162878e-01 -4.18611944e-01 -2.07434520e-01 8.30268919e-01
1.92267075e-01 6.48906171e-01 -8.78332913e-01 -4.98858958e-01
4.38304543e-02 4.50918764e-01 -2.87059963e-01 -1.64656639e-01
-2.60624494e-02 3.28466058e-01 -2.56229103e-01 4.47013736e-01
9.05456960e-01 -2.24867404e-01 1.81966975e-01 6.14034431e-03
-3.09292406e-01 4.33524042e-01 -1.29385638e+00 1.26994944e+00
-2.91717589e-01 9.95303154e-01 1.55397728e-02 -9.01024163e-01
7.48604178e-01 5.80202460e-01 2.48859361e-01 -7.51944780e-01
5.07340610e-01 6.50511801e-01 -2.53628910e-01 -7.48001933e-01
6.49568379e-01 6.00264408e-02 9.53524932e-02 5.65557182e-01
2.01892685e-02 3.21589932e-02 3.34562182e-01 2.49338657e-01
1.15189552e+00 -2.62356043e-01 1.07206665e-01 -1.68800175e-01
8.38915884e-01 -3.29137072e-02 -9.40579772e-02 6.75959110e-01
-2.64737815e-01 4.20930862e-01 6.18270218e-01 -4.16560769e-01
-9.52900887e-01 -4.57446575e-01 -2.10402563e-01 5.92593491e-01
-7.88210630e-02 -3.83674920e-01 -7.41693974e-01 -8.45143497e-01
5.73522709e-02 4.36765313e-01 -4.37561125e-01 3.17647378e-03
-9.28353369e-02 -3.56958538e-01 6.74816012e-01 -4.54565674e-01
6.55762851e-01 -6.04023933e-01 -8.17014456e-01 -3.73674231e-03
-5.47877550e-01 -1.33119965e+00 2.63960660e-01 -4.15202565e-02
-5.51319599e-01 -1.22484803e+00 -5.04814923e-01 -3.67131889e-01
3.38356972e-01 3.03525835e-01 9.00156796e-01 6.45383716e-01
-2.45031998e-01 7.22967535e-02 -7.33251035e-01 -3.47882032e-01
-8.18145096e-01 1.11079276e-01 -2.47490346e-01 5.53199589e-01
1.87984899e-01 -1.57863006e-01 -3.39131713e-01 4.33235824e-01
-1.07110882e+00 -1.12348236e-01 4.24941272e-01 5.80119967e-01
1.04419708e-01 2.90196121e-01 2.94048488e-01 -9.34603155e-01
3.81168008e-01 -5.57722926e-01 -8.15595925e-01 7.48879611e-02
-4.81301516e-01 -1.73251748e-01 4.42176104e-01 -2.52328575e-01
-6.61475658e-01 1.65130943e-01 -2.99278229e-01 3.28895114e-02
-1.91967100e-01 4.61477548e-01 9.03604925e-02 -2.10517734e-01
6.03512645e-01 6.01462796e-02 8.09652060e-02 -4.74513113e-01
7.86065459e-02 9.94263172e-01 1.59014806e-01 2.64022738e-01
9.55996215e-01 6.73588157e-01 -1.25931665e-01 -1.05855846e+00
-7.23032773e-01 -7.40066469e-01 -5.95220864e-01 -3.64310592e-01
7.01494634e-01 -9.05126512e-01 -8.44264150e-01 8.15243661e-01
-9.98115718e-01 2.83527702e-01 1.40778854e-01 4.26409692e-01
-1.33815542e-01 5.26808918e-01 -4.28720862e-01 -1.00982165e+00
7.33825669e-04 -1.00599849e+00 7.65840709e-01 -2.26666480e-01
-8.58570188e-02 -5.04687309e-01 -2.18170494e-01 8.35662603e-01
4.03058708e-01 2.21185148e-01 1.81444705e-01 -9.95551109e-01
-5.72665930e-01 -1.03152645e+00 -3.73886526e-01 4.03081298e-01
-3.18947405e-01 8.46671239e-02 -1.32145452e+00 -1.01685636e-01
5.47604188e-02 -4.48770523e-01 9.39564049e-01 -1.91664755e-01
8.06709349e-01 -4.44136322e-01 -1.83203563e-01 -5.15188053e-02
1.31767941e+00 -3.39829743e-01 9.52819467e-01 5.25338888e-01
2.67714858e-01 5.89803874e-01 7.22984612e-01 5.98147511e-01
8.20606798e-02 9.13916171e-01 7.78406024e-01 3.08727711e-01
1.44816518e-01 -1.53376400e-01 3.34696889e-01 2.88676918e-01
8.99860710e-02 -3.43165219e-01 -8.64834666e-01 5.46002805e-01
-1.55348682e+00 -1.46257365e+00 -7.31699228e-01 2.49387312e+00
3.88533920e-01 4.97571588e-01 4.04344946e-01 7.96320677e-01
8.54159474e-01 -9.07288417e-02 4.32192892e-01 -2.04605609e-01
-1.92487672e-01 -1.00645922e-01 7.39755571e-01 3.52817446e-01
-1.33070576e+00 4.92095023e-01 4.87160540e+00 1.01563013e+00
-1.41637337e+00 3.62680435e-01 4.87658560e-01 4.49033231e-02
6.60775006e-02 -1.07353777e-01 -6.96871638e-01 9.44200456e-01
1.06751776e+00 4.54853922e-01 1.96226895e-01 6.19170785e-01
2.65672773e-01 -6.72187984e-01 -7.02440083e-01 1.04475081e+00
2.20940411e-01 -1.41555691e+00 -3.64023060e-01 2.01502383e-01
4.11332250e-01 1.20717034e-01 -3.90269645e-02 -6.34289086e-02
-4.50861454e-01 -6.72963381e-01 1.18114245e+00 4.40432191e-01
4.28743422e-01 -5.32257438e-01 1.11304843e+00 7.07292557e-01
-6.61054254e-01 -7.87845552e-02 1.00448452e-01 -8.48404132e-03
2.33653083e-01 1.05202627e+00 -1.06614816e+00 5.19579291e-01
6.46378994e-01 4.16607291e-01 -7.33712375e-01 1.27661967e+00
-2.28604510e-01 4.91544962e-01 -4.94156063e-01 -5.95236309e-02
2.30932478e-02 2.35566333e-01 6.19757593e-01 1.34389830e+00
4.91268724e-01 -4.36968207e-01 -1.53076246e-01 4.46322709e-01
-2.24334523e-02 3.02113652e-01 -5.15934587e-01 -2.93389201e-01
1.79705679e-01 1.29512095e+00 -1.12586665e+00 -3.63988310e-01
-2.18322322e-01 1.19080961e+00 7.94315264e-02 -3.79365504e-01
-9.13590074e-01 -2.86643654e-01 -2.67612226e-02 6.65071607e-01
4.14507747e-01 -8.23575184e-02 1.42464295e-01 -1.29172790e+00
2.42045581e-01 -9.57327187e-01 3.49937886e-01 -6.58588707e-01
-8.44169438e-01 6.06961012e-01 -2.27740049e-01 -1.42167640e+00
-6.20166808e-02 -6.43278122e-01 -2.74641272e-02 5.07480025e-01
-1.51900697e+00 -1.03922129e+00 -2.23886669e-01 5.55301368e-01
-1.83952544e-02 -1.35928899e-01 5.16435146e-01 7.40993142e-01
-3.22188348e-01 4.78944659e-01 2.04463117e-02 1.22596733e-01
6.67668521e-01 -9.46766853e-01 -8.05879012e-02 8.65224361e-01
4.64467525e-01 2.30995715e-01 9.61747646e-01 -5.69202781e-01
-1.18535697e+00 -4.79498863e-01 1.44062662e+00 -3.08683395e-01
9.61318374e-01 -2.09457815e-01 -9.14049149e-01 1.63121045e-01
1.21789150e-01 -1.15392134e-02 5.75746477e-01 -7.57068694e-02
-5.87215841e-01 -2.03768417e-04 -1.48400283e+00 -1.24845453e-01
4.23711896e-01 -7.99547195e-01 -4.13587779e-01 5.82378268e-01
7.96701536e-02 -3.89238209e-01 -5.86773217e-01 -1.31887421e-01
6.60282910e-01 -1.43636405e+00 6.68105483e-01 -1.70419842e-01
5.76330721e-01 -2.20774129e-01 1.37695014e-01 -9.99017119e-01
3.56026530e-01 -5.46369433e-01 1.79184526e-01 1.41240287e+00
5.14605701e-01 -7.66601741e-01 5.69603205e-01 2.02411845e-01
3.23752403e-01 9.67618264e-03 -1.24777484e+00 -6.79563403e-01
-6.32608891e-01 -5.65760434e-01 3.97755861e-01 1.04041326e+00
1.79022908e-01 -3.61262299e-02 -5.04746079e-01 1.38339311e-01
2.46626496e-01 -1.32723093e-01 8.57866883e-01 -1.36546206e+00
-4.37494427e-01 -4.07616377e-01 -8.11327219e-01 -4.73928511e-01
-1.47188380e-01 -8.09582591e-01 -3.19433838e-01 -8.08868408e-01
6.56477809e-02 -4.36382800e-01 -3.52141671e-02 2.89392024e-01
2.55505741e-01 8.16447258e-01 2.96685785e-01 3.45514178e-01
-4.97103751e-01 -1.76406458e-01 7.28249133e-01 -1.14848092e-01
1.13674432e-01 5.78430668e-02 -4.37876992e-02 7.49137521e-01
7.16518700e-01 -8.00675035e-01 4.40592207e-02 -1.48431346e-01
9.44854796e-01 -1.16607789e-02 7.15481758e-01 -1.10957134e+00
-6.32936582e-02 3.51470232e-01 -5.82221523e-02 -6.74563825e-01
2.34396219e-01 -1.18507385e+00 1.86302811e-01 7.55765676e-01
-9.16980654e-02 -9.68155712e-02 -4.12626751e-03 6.47700965e-01
-3.54579240e-01 -5.53181827e-01 7.86053181e-01 -1.62398785e-01
-4.01485294e-01 -5.24356604e-01 -5.47471344e-01 -2.92840093e-01
1.00198185e+00 -2.05732167e-01 -5.00161946e-01 -3.33507806e-01
-7.24158108e-01 -2.53528565e-01 5.19810736e-01 1.87169790e-01
2.27569163e-01 -8.28386426e-01 -6.21426046e-01 2.06519753e-01
3.27847958e-01 -8.06486726e-01 2.95397013e-01 1.13892794e+00
-8.32670450e-01 2.61395603e-01 -2.88808435e-01 -6.81672573e-01
-1.69806421e+00 5.60537934e-01 3.36334184e-02 -4.30410594e-01
-2.55742550e-01 8.21852446e-01 -1.01257133e+00 1.26109466e-01
-3.53726596e-02 -1.17455542e-01 -2.78576493e-01 8.06039751e-01
8.57143044e-01 4.42343444e-01 7.65589237e-01 -1.01900077e+00
-3.91775161e-01 7.90018812e-02 5.93976304e-02 -2.29097292e-01
1.20142329e+00 -3.25863957e-01 -2.16579378e-01 2.97411382e-01
1.29171181e+00 5.49290180e-01 -5.37666678e-01 8.24289992e-02
2.78479517e-01 -9.69672382e-01 3.13416243e-01 -7.08656013e-01
-1.05143261e+00 7.56929398e-01 6.41781151e-01 1.07395971e+00
8.56820583e-01 -7.98094198e-02 6.01295710e-01 1.07640542e-01
4.67912614e-01 -1.30675614e+00 9.26856771e-02 2.24529833e-01
6.50550663e-01 -1.67946637e+00 1.48413375e-01 -4.56423998e-01
-4.93003458e-01 1.11900699e+00 -2.71315843e-01 1.13401026e-01
6.64237380e-01 -6.03297129e-02 -3.76841500e-02 -3.35601062e-01
-3.69393766e-01 -2.49554574e-01 2.88414091e-01 2.26972908e-01
3.68912160e-01 6.60074130e-02 -7.86018729e-01 9.17910933e-02
-1.95656434e-01 7.89344385e-02 7.28653193e-01 8.35244238e-01
-5.22482395e-01 -1.03827655e+00 -6.89610362e-01 3.86975646e-01
-1.01836228e+00 1.61856353e-01 -4.92056936e-01 7.20398128e-01
4.97705758e-01 1.16160095e+00 6.02556684e-04 -4.45889503e-01
-5.94711816e-03 3.23815972e-01 2.73417383e-01 -2.62796700e-01
-9.86061811e-01 -2.40991876e-01 4.03548747e-01 -6.65565848e-01
-6.50079548e-01 -8.43177974e-01 -6.28679395e-01 -4.58768159e-01
-6.78331912e-01 6.30846992e-02 1.14086449e+00 1.08611751e+00
3.58309150e-01 -2.32109968e-02 6.08326256e-01 -7.38796473e-01
-2.98854113e-01 -7.02416241e-01 -5.20593703e-01 6.88504398e-01
5.10415614e-01 -6.14643216e-01 -5.44676483e-01 1.64005294e-01] | [8.148065567016602, 10.224099159240723] |
4d624737-535b-45c8-80a5-60ec24b28c49 | a-survey-of-deep-learning-techniques-for-the | 2202.06372 | null | https://arxiv.org/abs/2202.06372v2 | https://arxiv.org/pdf/2202.06372v2.pdf | A Survey of Deep Learning Techniques for the Analysis of COVID-19 and their usability for Detecting Omicron | The Coronavirus (COVID-19) outbreak in December 2019 has become an ongoing threat to humans worldwide, creating a health crisis that infected millions of lives, as well as devastating the global economy. Deep learning (DL) techniques have proved helpful in analysis and delineation of infectious regions in radiological images in a timely manner. This paper makes an in-depth survey of DL techniques and draws a taxonomy based on diagnostic strategies and learning approaches. DL techniques are systematically categorized into classification, segmentation, and multi-stage approaches for COVID-19 diagnosis at image and region level analysis. Each category includes pre-trained and custom-made Convolutional Neural Network architectures for detecting COVID-19 infection in radiographic imaging modalities; X-Ray, and Computer Tomography (CT). Furthermore, a discussion is made on challenges in developing diagnostic techniques such as cross-platform interoperability and examining imaging modality. Similarly, a review of the various methodologies and performance measures used in these techniques is also presented. This survey provides an insight into the promising areas of research in DL for analyzing radiographic images, and further accelerates the research in designing customized DL based diagnostic tools for effectively dealing with new variants of COVID-19 and emerging challenges. | ['Muhammad Waleed Khan', 'Anabia Sohail', 'Asiya Batool', 'Mahrukh Saif', 'Saddam Hussain Khan', 'Asifullah Khan'] | 2022-02-13 | null | null | null | null | ['covid-19-detection'] | ['medical'] | [ 3.76249999e-01 -3.79833549e-01 -2.49322236e-01 -2.38315195e-01
-9.24441278e-01 -6.73903525e-01 1.57360941e-01 5.24212241e-01
-4.63739604e-01 3.63135755e-01 3.12232375e-02 -5.86183369e-01
-3.22788805e-01 -4.93402779e-01 -2.97169030e-01 -9.14918959e-01
-5.17363131e-01 9.34338808e-01 -4.71192002e-02 3.55057448e-01
-1.00640059e-01 1.08928204e+00 -9.18422818e-01 5.17995059e-01
3.89984399e-01 9.30092931e-01 5.61118662e-01 1.14646232e+00
3.46050672e-02 7.91757822e-01 -6.74445927e-01 -8.15729871e-02
-1.11168705e-01 -3.64627302e-01 -9.18915272e-01 -5.15778810e-02
3.48903239e-01 -4.98329848e-01 -1.12224825e-01 4.32685405e-01
6.86623573e-01 -3.60361516e-01 1.02180684e+00 -8.59839201e-01
-2.88846821e-01 1.47639379e-01 -5.90711057e-01 1.10137939e+00
3.10592297e-02 1.75767958e-01 2.20844701e-01 -5.13189375e-01
9.18149769e-01 8.05606663e-01 1.17634153e+00 5.01459658e-01
-5.53662717e-01 -4.57814097e-01 -2.79907852e-01 3.23811680e-01
-8.94563079e-01 5.60933888e-01 2.00651422e-01 -9.52248275e-01
1.20269632e+00 4.00169015e-01 9.61488366e-01 9.78225589e-01
1.13929939e+00 8.20154130e-01 8.44357550e-01 -1.19115368e-01
-6.12900034e-03 -4.66585040e-01 3.02283436e-01 6.76233292e-01
3.44080299e-01 -6.65455237e-02 2.42718369e-01 -5.79416037e-01
9.44642186e-01 3.69334519e-01 -1.45249382e-01 -2.88710117e-01
-1.34743047e+00 1.15522099e+00 5.01357496e-01 4.95589346e-01
-4.65598822e-01 3.70426811e-02 1.08824885e+00 -4.12534960e-02
2.12201491e-01 3.04134399e-01 -3.72160017e-01 2.24052772e-01
-7.59121299e-01 9.58362743e-02 1.30623788e-01 3.05327624e-01
-2.39266977e-01 5.22402301e-03 -2.01525435e-01 8.60507309e-01
2.91717857e-01 6.81776226e-01 5.59748113e-01 -7.48401284e-01
-8.78726467e-02 2.13266522e-01 -4.46182966e-01 -6.78993404e-01
-1.17567790e+00 -4.54936951e-01 -1.12837029e+00 1.21376760e-01
2.74559688e-02 -3.20827305e-01 -1.31968701e+00 1.17373908e+00
2.46777028e-01 -3.89871858e-02 -3.34017754e-01 8.23955655e-01
1.08371758e+00 3.12741607e-01 3.27980071e-01 -1.61907211e-01
1.98204851e+00 -8.10414374e-01 -6.10615432e-01 3.28447483e-02
6.63726628e-01 -7.86232591e-01 4.67717618e-01 2.61298180e-01
-8.60977232e-01 -6.66622669e-02 -1.04171586e+00 1.50410995e-01
-4.55877662e-01 -3.10651667e-04 7.74806678e-01 6.44388795e-01
-9.03861880e-01 1.38111591e-01 -1.21897507e+00 -6.52583063e-01
7.83856750e-01 3.25304955e-01 -1.45705566e-01 -2.70455986e-01
-9.92345154e-01 1.30386746e+00 -5.89999929e-03 2.46874336e-02
-1.26529956e+00 -8.65240932e-01 -6.16824925e-01 -4.83252943e-01
-1.99100450e-02 -1.14220870e+00 1.40046561e+00 -2.27209598e-01
-7.04095125e-01 1.28989565e+00 1.79830939e-01 -3.94053310e-01
3.19481820e-01 9.13544074e-02 -4.25512850e-01 5.97642720e-01
2.60122865e-01 5.68109810e-01 5.99235177e-01 -1.05092919e+00
-7.32484758e-01 -4.23698723e-01 -1.83341444e-01 -2.02883407e-02
5.29959261e-01 5.77150941e-01 -2.09824786e-01 -8.93437147e-01
-3.84902149e-01 -1.02003109e+00 -5.24878323e-01 2.27217063e-01
-7.93375671e-02 -3.12624246e-01 1.06244469e+00 -7.10029066e-01
7.30400026e-01 -1.76919973e+00 -3.39492202e-01 1.11840427e-01
8.01327407e-01 4.25373226e-01 1.34336561e-01 8.31805915e-02
-1.80588692e-01 -3.01266965e-02 -3.21744680e-01 2.20318343e-02
-5.67594409e-01 3.01595747e-01 1.70604303e-01 9.02409732e-01
4.05510254e-02 1.03279781e+00 -7.37707436e-01 -6.38777256e-01
5.29653549e-01 8.09864163e-01 -2.25451246e-01 1.78267002e-01
4.54962533e-03 7.50813127e-01 -4.31434333e-01 9.45658863e-01
6.11787260e-01 -8.27807724e-01 1.01442769e-01 -4.63597253e-02
6.02116771e-02 -1.93297654e-01 -3.66953015e-01 1.34314370e+00
-5.71308970e-01 7.64780223e-01 2.70746440e-01 -9.23409998e-01
2.39478484e-01 6.63801789e-01 9.54711020e-01 -4.46906298e-01
6.31999314e-01 2.34474197e-01 5.45178019e-02 -8.76635313e-01
-1.19699590e-01 -2.69932985e-01 4.67698015e-02 7.03071058e-01
-1.55738562e-01 -3.41362357e-02 1.38018578e-01 3.90853807e-02
1.11796606e+00 -5.80232680e-01 3.81255746e-01 -1.18223883e-01
5.55433095e-01 4.55063552e-01 2.77030587e-01 1.01105583e+00
-5.33387363e-01 6.53610945e-01 1.04883857e-01 -8.27688277e-01
-1.03280020e+00 -1.41928887e+00 -5.92865825e-01 6.01064444e-01
-1.33469477e-01 2.42600948e-01 -6.14445031e-01 -6.73851013e-01
-2.16775388e-01 1.31360024e-01 -8.43219340e-01 2.46441573e-01
-9.96147096e-01 -1.05458879e+00 7.36889243e-01 9.06836033e-01
5.09744026e-02 -1.24304366e+00 -1.29019046e+00 1.80224746e-01
-2.82836974e-01 -1.06358147e+00 -6.70126155e-02 1.12910420e-01
-9.29066777e-01 -1.49702334e+00 -1.26907206e+00 -1.35357094e+00
4.47759151e-01 2.69847691e-01 1.01703167e+00 2.84627974e-01
-1.29526114e+00 5.98112226e-01 -1.43579632e-01 -7.05539167e-01
-5.18166304e-01 -1.54253334e-01 -1.32407084e-01 -8.93849611e-01
4.46146309e-01 -3.02661546e-02 -8.39783967e-01 -1.02406368e-01
-8.79580557e-01 -2.35884130e-01 6.53055727e-01 8.00891638e-01
7.92872727e-01 -2.88040698e-01 3.71707737e-01 -8.16676795e-01
7.70674646e-01 -7.03933775e-01 -1.41915485e-01 3.12676430e-01
-5.63779652e-01 -6.51347816e-01 1.64274231e-01 -5.51524237e-02
-8.63460362e-01 -4.26623046e-01 -2.86651880e-01 -4.70090479e-01
-3.28842968e-01 3.46251518e-01 9.54079926e-01 -1.85093135e-01
3.94578695e-01 3.60946241e-03 2.74700969e-01 -2.69183099e-01
1.28708392e-01 7.13811696e-01 6.88459158e-01 -2.87073970e-01
1.77680373e-01 9.18092191e-01 2.51352768e-02 -9.64284360e-01
-6.10564113e-01 -7.43289173e-01 -5.62332511e-01 -5.07406592e-01
1.51958203e+00 -5.91830373e-01 -4.61658806e-01 7.55354226e-01
-1.20247734e+00 5.29016554e-02 -1.26026228e-01 6.64250433e-01
-4.22039568e-01 3.51875007e-01 -1.28749776e+00 -1.18436851e-01
-1.03877044e+00 -1.88424098e+00 1.05238652e+00 1.68286800e-01
-3.53096187e-01 -1.24846935e+00 4.53562617e-01 4.63073552e-01
5.64744294e-01 7.21611977e-01 1.43523276e+00 -6.18103147e-01
-3.09129864e-01 -3.23394328e-01 -5.43073535e-01 1.61540121e-01
2.05419600e-01 -1.08417451e-01 -6.31460547e-01 -3.74620140e-01
1.19069397e-01 -1.88696429e-01 5.87245762e-01 1.09809935e+00
1.29891455e+00 4.08888936e-01 -6.64336801e-01 6.18489385e-01
1.32232189e+00 8.72621834e-01 1.74612716e-01 4.88457173e-01
5.68756223e-01 2.80122787e-01 3.04237098e-01 2.73199737e-01
1.77318364e-01 1.61489502e-01 3.23670059e-01 -6.01051211e-01
-2.97656447e-01 7.34958529e-01 -5.17436087e-01 6.10389650e-01
-2.84394562e-01 -2.57806242e-01 -1.33645582e+00 5.68273425e-01
-1.12872982e+00 -6.53526366e-01 -3.67314398e-01 1.36233771e+00
3.93157333e-01 -2.50392258e-01 4.27823812e-02 -3.48209530e-01
7.58691311e-01 -1.18815161e-01 -6.86231077e-01 -5.80430567e-01
-3.32847685e-02 5.07803917e-01 4.22820747e-01 1.49832502e-01
-1.35221732e+00 2.00228333e-01 8.05477810e+00 4.79670435e-01
-1.45881176e+00 5.24152219e-01 5.69934487e-01 -3.56641077e-02
6.36109114e-02 -7.48244107e-01 -3.61247003e-01 2.75770724e-01
5.45477629e-01 1.14073016e-01 -2.33045205e-01 7.06360340e-01
1.64716542e-01 -8.22926834e-02 -7.44439483e-01 8.26671481e-01
2.83447921e-01 -1.90089631e+00 -1.07034527e-01 7.47531280e-02
6.16059065e-01 8.57548296e-01 9.69894230e-02 1.54491793e-02
9.76093113e-02 -9.46524501e-01 7.65906572e-02 2.95065582e-01
1.07068753e+00 -6.11986995e-01 1.20790279e+00 -2.94666886e-01
-1.05628955e+00 2.55218357e-01 1.06998226e-02 6.14815414e-01
4.42791700e-01 3.40149283e-01 -1.15037322e+00 4.23408628e-01
1.12329078e+00 3.69181991e-01 -2.38079935e-01 1.28619504e+00
1.04307003e-01 4.04175460e-01 -2.37029776e-01 2.22032085e-01
4.96518970e-01 1.42070875e-02 6.35850370e-01 1.77448738e+00
-1.01454973e-01 1.59367502e-01 1.92436263e-01 3.89865965e-01
2.68789172e-01 6.99556991e-02 -4.96554762e-01 7.84336776e-02
3.02083399e-02 1.17431295e+00 -1.29030681e+00 -5.55510819e-01
-5.58288336e-01 5.40424883e-01 -2.86179155e-01 2.38150001e-01
-1.01030290e+00 -8.39804336e-02 5.64097822e-01 8.16538036e-02
1.92433923e-01 -1.37600154e-01 -3.69028211e-01 -5.50987065e-01
-5.91007590e-01 -8.23750436e-01 1.02669275e+00 -7.53544390e-01
-1.49412322e+00 6.35870337e-01 2.58090496e-01 -1.08818781e+00
-3.57420832e-01 -9.05069888e-01 -7.02400148e-01 4.95972961e-01
-1.19207597e+00 -1.13178229e+00 -3.03169936e-01 3.80946994e-01
6.29746556e-01 -1.84751421e-01 1.09861577e+00 2.92572469e-01
-3.52547854e-01 4.17502820e-01 4.09484506e-01 1.68553054e-01
4.96463358e-01 -1.02364767e+00 9.33143646e-02 3.86876553e-01
-6.72247529e-01 7.24737346e-01 1.81001276e-01 -7.07850873e-01
-9.09744203e-01 -9.47846830e-01 4.93985742e-01 -3.20333064e-01
5.93470633e-01 5.66014089e-02 -6.53428376e-01 8.26585710e-01
4.59122241e-01 -1.47936106e-01 1.19810426e+00 -4.95764673e-01
-3.67163047e-02 4.93262261e-01 -1.58952892e+00 3.03815931e-01
4.62491661e-01 -3.33918631e-01 -6.32324278e-01 7.85972416e-01
6.16582632e-01 -5.71431875e-01 -1.21080899e+00 8.01041543e-01
6.95035756e-01 -7.02970624e-01 1.17468905e+00 -6.40213013e-01
3.62788677e-01 -3.01575456e-02 2.23119512e-01 -7.53306210e-01
-3.24359804e-01 2.22880319e-02 1.02489240e-01 1.59915790e-01
2.72314865e-02 -6.54926717e-01 7.66201138e-01 -5.16983457e-02
-3.77766460e-01 -1.27972043e+00 -8.94870043e-01 -3.65234554e-01
3.39064687e-01 -5.38209617e-01 2.21307382e-01 1.09486008e+00
-4.52689052e-01 -2.27249131e-01 3.93308818e-01 5.17620444e-02
7.47645259e-01 1.20309792e-01 1.34928450e-01 -9.67101157e-01
1.44727733e-02 -7.45103359e-01 -5.40367424e-01 -4.20236319e-01
-4.15957630e-01 -9.15452659e-01 -3.29204440e-01 -2.05550361e+00
5.16248584e-01 -5.27138412e-01 -4.21111315e-01 2.87589997e-01
2.03827053e-01 5.89050889e-01 -2.27974840e-02 4.69426960e-01
-1.42327145e-01 -3.15669984e-01 1.47926044e+00 -4.10617679e-01
2.90633470e-01 8.22572336e-02 -1.76371709e-01 8.87593627e-01
8.33143175e-01 -4.99820054e-01 -2.86619335e-01 -3.45432132e-01
-9.92740393e-02 7.62548819e-02 3.57577592e-01 -8.39429080e-01
-3.08018140e-02 1.16651542e-01 5.88130295e-01 -1.33279991e+00
2.94767208e-02 -6.32853448e-01 4.76645082e-02 1.07065785e+00
5.35915047e-02 5.64934731e-01 3.76531959e-01 2.49194294e-01
-9.22283381e-02 -6.12087213e-02 1.24074793e+00 -3.79405081e-01
-7.48763561e-01 3.50983769e-01 -1.27194548e+00 6.70380443e-02
1.32890701e+00 -2.94070482e-01 -3.51270467e-01 7.74172251e-04
-9.81970906e-01 3.93045187e-01 1.18310839e-01 4.41833317e-01
6.53041959e-01 -8.28462183e-01 -7.53707945e-01 1.26639619e-01
1.88049629e-01 2.07462057e-01 8.17675650e-01 1.44170415e+00
-1.49207902e+00 7.13847339e-01 -2.42566273e-01 -1.18563175e+00
-1.54713523e+00 5.82675576e-01 7.50500798e-01 -6.45128489e-01
-9.06403482e-01 1.01790035e+00 3.95116866e-01 -2.94393390e-01
2.86010861e-01 -3.05583864e-01 -3.29604805e-01 -4.19578552e-02
5.55171072e-01 2.48327494e-01 4.17538643e-01 -6.19592786e-01
-7.24960804e-01 7.03357339e-01 -5.10933995e-01 4.74309504e-01
1.25048327e+00 -1.49338767e-01 -2.53592730e-01 2.37724587e-01
1.32399607e+00 -5.85870862e-01 -4.37930107e-01 7.15827942e-02
-4.13388580e-01 2.17486918e-01 1.03507668e-01 -9.58254039e-01
-1.19289839e+00 8.53651226e-01 1.24564457e+00 -2.10986882e-02
9.99200106e-01 5.80606461e-01 1.11235166e+00 -1.60694033e-01
2.85412014e-01 -6.58549428e-01 4.77376729e-02 2.56568223e-01
5.98999500e-01 -1.03936660e+00 1.40314892e-01 -2.76595980e-01
-4.79273587e-01 1.18299949e+00 3.51249546e-01 -2.44600654e-01
9.71770406e-01 7.00134814e-01 5.82045555e-01 -9.07695353e-01
-2.18915805e-01 2.94951856e-01 4.32073064e-02 9.46174145e-01
6.13216758e-01 3.40045094e-01 -2.65374154e-01 3.34586240e-02
1.45526454e-01 -1.93077430e-01 3.47562701e-01 1.37432790e+00
-2.85653323e-01 -8.56673181e-01 -6.40121162e-01 8.24133396e-01
-7.57295668e-01 5.67510054e-02 -1.23166062e-01 1.07155156e+00
4.08365339e-01 4.57141906e-01 2.53412217e-01 -9.61857736e-02
6.67762384e-02 -2.47055337e-01 6.41862988e-01 -4.59529847e-01
-6.54282510e-01 1.96267813e-01 -1.04817085e-01 -4.46835101e-01
-6.34318769e-01 -7.44352996e-01 -1.54648733e+00 -7.02101141e-02
-4.56267484e-02 -1.89288035e-01 8.79734576e-01 1.11666620e+00
-1.34722628e-02 9.51267838e-01 1.45638332e-01 -6.17953897e-01
6.07701503e-02 -7.07824588e-01 -4.23704326e-01 -1.95449938e-05
4.94803637e-01 -7.31038511e-01 -1.19174600e-01 -7.67632946e-02] | [15.536113739013672, -1.7459293603897095] |
bbb8fd15-872f-4484-bc6a-152f55ee4680 | video-diffusion-models-with-local-global | 2306.02562 | null | https://arxiv.org/abs/2306.02562v1 | https://arxiv.org/pdf/2306.02562v1.pdf | Video Diffusion Models with Local-Global Context Guidance | Diffusion models have emerged as a powerful paradigm in video synthesis tasks including prediction, generation, and interpolation. Due to the limitation of the computational budget, existing methods usually implement conditional diffusion models with an autoregressive inference pipeline, in which the future fragment is predicted based on the distribution of adjacent past frames. However, only the conditions from a few previous frames can't capture the global temporal coherence, leading to inconsistent or even outrageous results in long-term video prediction. In this paper, we propose a Local-Global Context guided Video Diffusion model (LGC-VD) to capture multi-perception conditions for producing high-quality videos in both conditional/unconditional settings. In LGC-VD, the UNet is implemented with stacked residual blocks with self-attention units, avoiding the undesirable computational cost in 3D Conv. We construct a local-global context guidance strategy to capture the multi-perceptual embedding of the past fragment to boost the consistency of future prediction. Furthermore, we propose a two-stage training strategy to alleviate the effect of noisy frames for more stable predictions. Our experiments demonstrate that the proposed method achieves favorable performance on video prediction, interpolation, and unconditional video generation. We release code at https://github.com/exisas/LGC-VD. | ['You He', 'Zhizhuo Jiang', 'Yu Liu', 'Lu Zhang', 'Siyuan Yang'] | 2023-06-05 | null | null | null | null | ['video-generation', 'video-prediction', 'unconditional-video-generation'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [-9.59778130e-02 -2.78766543e-01 -2.15039670e-01 -3.35696697e-01
-4.72317576e-01 8.47486034e-03 6.70681834e-01 -3.62005860e-01
8.16481337e-02 5.83748102e-01 7.64645934e-01 -9.16245133e-02
2.56138146e-01 -6.52496696e-01 -7.78961003e-01 -7.44286001e-01
2.36822769e-01 -3.63383979e-01 3.83743376e-01 -5.45413606e-02
1.67704567e-01 1.29742026e-01 -1.34298027e+00 6.83052480e-01
7.83977270e-01 1.16482210e+00 7.49129415e-01 7.10164845e-01
1.15139388e-01 1.11783576e+00 -3.58854622e-01 -2.73368418e-01
8.69702473e-02 -7.67610610e-01 -3.11945081e-01 1.36925772e-01
1.75122708e-01 -8.88645113e-01 -6.86394632e-01 9.09726560e-01
6.14597678e-01 2.27439940e-01 4.30558056e-01 -1.09454942e+00
-1.03848517e+00 4.48168099e-01 -5.14894247e-01 7.96529576e-02
5.31899869e-01 5.92147529e-01 7.91581094e-01 -1.09032643e+00
7.28668034e-01 1.35303342e+00 4.04499441e-01 7.25287557e-01
-1.03072357e+00 -8.04078996e-01 5.06435394e-01 4.92473155e-01
-1.27362251e+00 -6.30931199e-01 8.91712546e-01 -5.09737313e-01
8.67745101e-01 2.16250923e-02 7.83933222e-01 1.43712795e+00
4.27785039e-01 8.88775408e-01 6.40372336e-01 -6.06550388e-02
2.05079421e-01 -2.14435488e-01 -4.50862408e-01 5.47178149e-01
-4.47838843e-01 3.27103913e-01 -7.59045780e-01 1.97636798e-01
1.11051929e+00 1.92464292e-02 -4.44022208e-01 4.16800752e-02
-1.36054862e+00 5.10244966e-01 3.12383413e-01 1.01003803e-01
-7.21182883e-01 3.09924513e-01 2.72822678e-01 1.41971171e-01
6.16860271e-01 -1.21969640e-01 -1.84927344e-01 -2.16211975e-01
-1.04090500e+00 2.81000406e-01 2.51770794e-01 1.21749973e+00
7.04596043e-01 4.11423653e-01 -5.53565919e-01 6.69167340e-01
4.84664202e-01 2.76859283e-01 4.08476144e-01 -1.29948175e+00
5.04018068e-01 5.83649874e-02 3.07575494e-01 -1.19138598e+00
1.86088532e-01 -2.58831918e-01 -1.12859023e+00 -4.42904755e-02
-1.67193294e-01 -2.53100514e-01 -6.89911008e-01 1.73381615e+00
2.30154634e-01 6.82054222e-01 -1.88497350e-01 9.76569355e-01
6.09654248e-01 1.25304615e+00 1.17065065e-01 -5.14005244e-01
7.23218679e-01 -1.24029303e+00 -8.63237262e-01 1.86591577e-02
3.76268804e-01 -8.92235100e-01 9.18445885e-01 3.33408654e-01
-1.22004545e+00 -1.02617526e+00 -8.16678345e-01 -2.53255069e-01
3.45335126e-01 2.45659336e-01 3.82859796e-01 4.65122461e-02
-1.17530787e+00 6.18389785e-01 -9.06880438e-01 1.85350284e-01
3.53865474e-01 -1.73865817e-02 -1.33379723e-03 -2.71703213e-01
-1.17521501e+00 4.33906615e-01 1.87839851e-01 3.89805794e-01
-1.15486741e+00 -6.48435056e-01 -8.19937348e-01 7.91134462e-02
7.78775960e-02 -6.58732951e-01 1.01207137e+00 -1.16608906e+00
-1.62092650e+00 9.99276042e-02 -5.99101186e-01 -3.43663156e-01
5.84619105e-01 -2.54949123e-01 -3.90517145e-01 -4.82162088e-02
1.03582464e-01 1.09646356e+00 1.10206664e+00 -1.13442338e+00
-5.96984267e-01 8.68635401e-02 -9.81831104e-02 2.47253075e-01
-1.82407588e-01 -1.19099878e-01 -8.40432882e-01 -1.17081678e+00
-2.65995637e-02 -7.36939311e-01 -3.49149853e-01 2.52142847e-01
-1.75209478e-01 -1.39207229e-01 9.79055166e-01 -9.57210898e-01
1.53928375e+00 -2.42636633e+00 3.24321479e-01 -2.56667852e-01
-6.88410910e-06 1.08009405e-01 -2.26403549e-01 2.24729732e-01
4.06772047e-02 -1.64994150e-02 -1.79306176e-02 -6.30980134e-01
-1.38852149e-01 2.56068885e-01 -5.65778077e-01 5.40332273e-02
3.50337535e-01 7.83911765e-01 -9.56099629e-01 -4.72864509e-01
3.16588104e-01 7.65354753e-01 -8.82123172e-01 4.92684156e-01
-5.18318832e-01 8.81671846e-01 -4.28619921e-01 4.04446721e-01
6.40009165e-01 -2.78858066e-01 5.59263751e-02 -3.62552196e-01
-2.21042052e-01 2.39705548e-01 -1.08391809e+00 1.96621406e+00
-4.98744130e-01 7.11813748e-01 -9.65576470e-02 -6.14699244e-01
8.70529950e-01 4.15987104e-01 2.75351763e-01 -7.12306380e-01
-1.40786096e-01 7.50710443e-02 -2.67002165e-01 -5.22237182e-01
7.15494752e-01 6.71826825e-02 3.66244376e-01 7.61225894e-02
-1.16677426e-01 9.50173065e-02 5.02298772e-02 2.56710291e-01
6.90958917e-01 6.56655192e-01 -1.22252502e-01 6.72943592e-02
4.54579175e-01 -4.36803579e-01 1.12518072e+00 2.42603883e-01
-2.65426993e-01 9.12946641e-01 5.05487919e-01 -3.24572414e-01
-1.17499232e+00 -9.34754848e-01 3.38819534e-01 8.42194796e-01
3.02516729e-01 -6.73717141e-01 -5.10999978e-01 -3.02119195e-01
-4.26296234e-01 8.43396008e-01 -4.20268923e-01 -1.85763553e-01
-7.23839462e-01 -3.80739540e-01 8.41432884e-02 5.79626858e-01
6.56661034e-01 -1.00564528e+00 -1.67711511e-01 4.74225044e-01
-4.15271074e-01 -9.85884368e-01 -8.78675818e-01 -5.85667491e-01
-7.49180377e-01 -4.93736207e-01 -9.41385329e-01 -8.02382827e-01
3.96164417e-01 3.17117989e-01 8.56831491e-01 1.33622244e-01
3.64055395e-01 -6.31020889e-02 -3.76989365e-01 2.07039222e-01
-4.13156122e-01 -4.78304148e-01 -5.35353497e-02 2.20869020e-01
-2.87442319e-02 -6.37349129e-01 -1.11997163e+00 4.37939525e-01
-7.34352887e-01 7.09922552e-01 3.40547472e-01 9.73076820e-01
8.20983231e-01 -2.23357212e-02 6.05682611e-01 -4.01209772e-01
4.89065558e-01 -6.84199870e-01 -3.99585009e-01 8.76354501e-02
-3.89371961e-01 -2.34405369e-01 7.08168566e-01 -5.73388994e-01
-1.46781158e+00 -1.20464772e-01 -4.31210607e-01 -9.31026876e-01
7.46731386e-02 4.11570400e-01 -2.95047700e-01 3.62908036e-01
2.01549888e-01 5.18276870e-01 -2.26956382e-01 -3.33169848e-01
2.38726407e-01 5.19373655e-01 3.27417821e-01 -4.47805405e-01
1.90762132e-01 2.75171638e-01 -2.35977203e-01 -6.54008746e-01
-5.24278343e-01 4.92909104e-02 -2.83627450e-01 -4.43749458e-01
1.02893138e+00 -1.39022827e+00 -2.69527942e-01 6.38087869e-01
-1.30716383e+00 -7.74410903e-01 -8.35001990e-02 5.12969673e-01
-5.96025646e-01 5.69094718e-01 -9.41733539e-01 -6.31734252e-01
-1.99960813e-01 -1.53742719e+00 1.00453544e+00 1.55625686e-01
-1.80512905e-01 -8.31651688e-01 -1.95681453e-01 1.54789835e-01
4.58977729e-01 -7.21698580e-03 5.92319250e-01 2.95306593e-01
-9.67828631e-01 1.99230894e-01 -2.98126280e-01 5.91552138e-01
8.40813816e-02 3.13844919e-01 -6.78945422e-01 -1.71131894e-01
-4.05413797e-03 -2.15179529e-02 8.15624356e-01 5.88765621e-01
1.44464529e+00 -3.86027336e-01 -2.07474306e-01 7.36570477e-01
1.20383394e+00 3.46111685e-01 9.06189263e-01 -1.41156569e-01
8.94326389e-01 3.32973093e-01 8.39492202e-01 7.29037702e-01
5.76738358e-01 6.09696507e-01 1.99863076e-01 2.68119454e-01
-6.06113136e-01 -6.46675646e-01 6.44309402e-01 1.12851357e+00
-1.71322450e-01 -4.94127274e-01 -5.57392955e-01 6.13030612e-01
-1.96026242e+00 -1.11890924e+00 2.68135928e-02 1.94435990e+00
8.09682608e-01 1.14170536e-01 -3.00181299e-01 -2.20944807e-01
7.25533068e-01 4.51526254e-01 -4.35066640e-01 -1.62462398e-01
-1.87355980e-01 -2.63042331e-01 -8.56814384e-02 6.52633905e-01
-7.47756958e-01 9.21851397e-01 5.46245003e+00 1.11377072e+00
-1.20598888e+00 2.29993507e-01 1.09549820e+00 -1.46049872e-01
-6.12525880e-01 3.39396484e-03 -7.48060882e-01 1.00353146e+00
8.06435287e-01 1.16308182e-01 2.96497732e-01 6.86573982e-01
8.23910773e-01 -7.18264207e-02 -9.11652625e-01 9.16390359e-01
-8.71026069e-02 -1.63113689e+00 2.94618666e-01 -1.71533659e-01
1.09723282e+00 -1.97434202e-01 3.17192376e-01 2.38133281e-01
-1.42300650e-02 -6.79964483e-01 1.02032030e+00 7.68731594e-01
9.10396338e-01 -5.66211402e-01 2.74814904e-01 4.03324842e-01
-1.33603239e+00 -1.85712054e-01 -3.52339953e-01 -6.67965859e-02
5.83116114e-01 5.20038843e-01 -2.71249771e-01 4.40103710e-01
7.04920769e-01 1.26532960e+00 -1.63729101e-01 6.81535065e-01
-3.21945429e-01 7.02166617e-01 2.49923375e-02 2.43081480e-01
3.07244748e-01 -3.18530738e-01 5.04773319e-01 1.13503242e+00
7.68795907e-01 2.61235535e-01 7.03669116e-02 8.49131823e-01
5.68694919e-02 -1.71424165e-01 -3.06647360e-01 1.96020871e-01
4.67802614e-01 7.07332373e-01 -1.77801937e-01 -5.33531010e-01
-5.53967535e-01 1.25774324e+00 1.54807001e-01 5.79899669e-01
-1.19000757e+00 1.65868238e-01 6.50969625e-01 8.11622888e-02
5.69726706e-01 -3.97463828e-01 -1.49313688e-01 -1.38046515e+00
1.08118065e-01 -8.33496571e-01 3.31324749e-02 -1.18883574e+00
-1.07605970e+00 5.75612545e-01 -2.27926299e-01 -1.41729391e+00
-3.02085578e-01 -1.14565074e-01 -7.17360377e-01 9.25564349e-01
-1.38444066e+00 -1.18327200e+00 -2.78167665e-01 7.00022042e-01
1.07466161e+00 6.18348159e-02 4.06044424e-01 5.18081665e-01
-6.97675943e-01 3.36517036e-01 -3.27950306e-02 -2.09716246e-01
8.66128683e-01 -7.81509459e-01 3.92425120e-01 1.12247193e+00
-3.03790420e-01 4.78330791e-01 6.76076770e-01 -8.99775147e-01
-1.19276488e+00 -1.31736624e+00 8.22771132e-01 6.90091401e-02
4.40412283e-01 -3.42463814e-02 -1.03054738e+00 6.73133910e-01
3.20800245e-01 4.26101796e-02 3.28925908e-01 -4.09383059e-01
-1.89775392e-01 -2.19346046e-01 -6.63248181e-01 9.36597228e-01
1.11005187e+00 -5.18795907e-01 2.86264382e-02 1.82300299e-01
1.03201807e+00 -3.96389633e-01 -9.32237327e-01 3.14891577e-01
4.61182684e-01 -1.23730028e+00 7.84693837e-01 -1.89259294e-02
1.03701973e+00 -4.03440326e-01 -2.89878964e-01 -1.09236038e+00
-5.23270607e-01 -8.15280259e-01 -4.67828721e-01 1.37745488e+00
2.17701346e-01 -1.49403155e-01 6.77854717e-01 5.81226349e-01
-3.78173828e-01 -9.57078397e-01 -7.85630286e-01 -3.99223685e-01
-1.24613233e-01 -5.73773324e-01 5.85243106e-01 7.80505359e-01
-3.28732371e-01 2.11257115e-01 -1.02194452e+00 1.62604526e-01
3.26030642e-01 1.58540815e-01 6.94134355e-01 -2.98711002e-01
-5.51605344e-01 -3.50744277e-01 -1.67797759e-01 -1.85168362e+00
1.14981225e-03 -3.83873284e-01 1.41865507e-01 -1.50017381e+00
3.04593574e-02 -4.03898597e-01 -2.27671608e-01 8.56670458e-03
-3.37713927e-01 4.60760780e-02 2.44730949e-01 3.53478163e-01
-5.55375040e-01 1.00628912e+00 1.62761569e+00 1.66705400e-02
-1.15100980e-01 -1.34257972e-01 -2.58014232e-01 5.02784312e-01
6.41728342e-01 -5.26544973e-02 -6.30525291e-01 -8.74979377e-01
1.57722592e-01 4.51320440e-01 3.62047106e-01 -9.52858686e-01
2.84571588e-01 -3.65630150e-01 5.65501809e-01 -6.56970561e-01
6.97785556e-01 -5.00104129e-01 4.95522887e-01 3.06240976e-01
-3.56245965e-01 1.17091998e-01 -1.23937726e-01 8.83471310e-01
-4.61030275e-01 1.82100654e-01 7.06259668e-01 -2.01407005e-03
-1.07155466e+00 7.22013533e-01 -3.79455030e-01 -3.93892944e-01
9.27522898e-01 -1.40101612e-01 -9.53670144e-02 -6.82785988e-01
-6.46571159e-01 2.42476046e-01 6.08936310e-01 4.77659732e-01
1.05253398e+00 -1.50873268e+00 -7.25800157e-01 2.99382776e-01
-2.81628013e-01 1.49204880e-01 8.70479107e-01 8.09921026e-01
-5.37660658e-01 6.90897405e-02 -1.11327447e-01 -5.62313259e-01
-1.01869142e+00 7.53910780e-01 4.17244509e-02 -1.46953821e-01
-5.70872545e-01 1.04507029e+00 5.31347990e-01 2.50466019e-01
7.87084624e-02 -3.06494564e-01 2.91769523e-02 -2.02639550e-01
5.93028784e-01 2.67823964e-01 -4.90716487e-01 -7.68707633e-01
1.19853564e-01 3.30498725e-01 -8.05899203e-02 -5.42828962e-02
1.10422456e+00 -6.36458397e-01 -2.18987726e-02 3.20771188e-01
1.03030646e+00 -4.84802201e-02 -2.02805138e+00 -4.72367220e-02
-5.96974671e-01 -7.29971170e-01 1.92383289e-01 -4.01852220e-01
-1.23011708e+00 9.96194065e-01 4.10447329e-01 -1.77682638e-01
1.41506231e+00 -3.42667133e-01 1.09115386e+00 -2.51142114e-01
2.25041524e-01 -8.94384384e-01 2.24722132e-01 4.66049641e-01
1.09449315e+00 -1.09993458e+00 -9.83406678e-02 -4.09524888e-01
-8.56046617e-01 9.26323950e-01 7.91896224e-01 -7.68884122e-02
7.32376873e-01 1.79283675e-02 -1.66555718e-01 3.58398974e-01
-1.28238857e+00 2.55987138e-01 1.69506505e-01 5.16890228e-01
4.83057678e-01 -5.00473939e-02 -1.44535691e-01 5.18838882e-01
1.78554788e-01 1.50954753e-01 3.89131010e-01 6.79674327e-01
-2.21890077e-01 -9.55839992e-01 -1.45206023e-02 1.21322073e-01
-2.80376643e-01 -2.81554013e-01 2.37172216e-01 2.47653812e-01
3.76057059e-01 9.41106796e-01 1.27704501e-01 -6.09559834e-01
-2.36741938e-02 -2.50773102e-01 2.68003285e-01 -2.00554714e-01
-2.02493861e-01 6.37663662e-01 1.30881974e-03 -8.40220809e-01
-4.88838792e-01 -6.44206285e-01 -9.60423768e-01 -4.62402225e-01
-1.68422669e-01 -2.19718561e-01 2.24386021e-01 6.39732838e-01
7.53574789e-01 6.00564063e-01 7.07088172e-01 -9.79066432e-01
-1.57019988e-01 -9.69175100e-01 -3.18646550e-01 3.17375302e-01
3.62205237e-01 -4.56279576e-01 -6.93780482e-02 4.31883037e-01] | [10.718304634094238, -0.6861099600791931] |
7c98b1e5-288c-4674-9045-dd4d78eec392 | answering-complex-logical-queries-on | 2212.09567 | null | https://arxiv.org/abs/2212.09567v3 | https://arxiv.org/pdf/2212.09567v3.pdf | Answering Complex Logical Queries on Knowledge Graphs via Query Computation Tree Optimization | Answering complex logical queries on incomplete knowledge graphs is a challenging task, and has been widely studied. Embedding-based methods require training on complex queries, and cannot generalize well to out-of-distribution query structures. Recent work frames this task as an end-to-end optimization problem, and it only requires a pretrained link predictor. However, due to the exponentially large combinatorial search space, the optimal solution can only be approximated, limiting the final accuracy. In this work, we propose QTO (Query Computation Tree Optimization) that can efficiently find the exact optimal solution. QTO finds the optimal solution by a forward-backward propagation on the tree-like computation graph, i.e., query computation tree. In particular, QTO utilizes the independence encoded in the query computation tree to reduce the search space, where only local computations are involved during the optimization procedure. Experiments on 3 datasets show that QTO obtains state-of-the-art performance on complex query answering, outperforming previous best results by an average of 22%. Moreover, QTO can interpret the intermediate solutions for each of the one-hop atoms in the query with over 90% accuracy. The code of our paper is at https://github.com/bys0318/QTO. | ['Lei Hou', 'Juanzi Li', 'Xin Lv', 'Yushi Bai'] | 2022-12-19 | null | null | null | null | ['complex-query-answering'] | ['knowledge-base'] | [-1.90685123e-01 1.49711862e-01 -6.32607639e-01 -2.92874575e-01
-1.09372592e+00 -5.75629473e-01 4.28380147e-02 4.90929276e-01
-3.63180041e-01 4.81607676e-01 6.57476112e-02 -3.82809967e-01
-3.05017412e-01 -1.12273538e+00 -1.02715874e+00 -1.44336894e-01
5.96616082e-02 9.80021656e-01 5.91323316e-01 -1.69487000e-01
-1.79892164e-02 1.22558080e-01 -1.14130509e+00 2.22876415e-01
8.29256177e-01 1.16157341e+00 -6.46080822e-02 6.68571591e-01
-3.60628158e-01 8.37717295e-01 -3.42607707e-01 -7.41685033e-01
1.33247022e-02 -1.26856431e-01 -1.14614809e+00 -7.70937979e-01
4.30366367e-01 -2.55765319e-01 -7.25297689e-01 1.00904071e+00
5.58242977e-01 -7.77256042e-02 1.25950783e-01 -1.20066559e+00
-6.42701268e-01 9.06364024e-01 -1.35563910e-01 9.81004238e-02
5.58473408e-01 -1.51953682e-01 1.67793047e+00 -1.13983762e+00
5.69817901e-01 1.19795191e+00 2.88785100e-01 1.27429709e-01
-1.35448980e+00 -5.82123756e-01 2.93933973e-02 5.76739490e-01
-1.96675193e+00 -3.66668612e-01 6.50251806e-01 5.69444411e-02
1.06724262e+00 3.46452534e-01 4.91404444e-01 2.75743335e-01
-1.26679838e-01 9.85606670e-01 5.20483851e-01 -1.40175223e-01
2.52904147e-01 -8.66125673e-02 3.54375541e-01 1.17900133e+00
2.62162298e-01 -1.91217825e-01 -7.85641789e-01 -5.69648147e-01
2.41708800e-01 3.21814902e-02 -2.76602805e-01 -4.31190491e-01
-1.04631031e+00 9.29270267e-01 8.69610429e-01 6.68744277e-03
-2.53014982e-01 6.91813886e-01 3.08008432e-01 4.51783895e-01
2.27235883e-01 2.80762434e-01 -5.75796068e-01 -1.89866200e-01
-6.44841969e-01 3.39498371e-01 1.31874156e+00 1.17627656e+00
1.19528902e+00 -5.94942749e-01 -3.56233448e-01 6.76513135e-01
2.74701655e-01 5.63737035e-01 -1.42504767e-01 -8.44598532e-01
7.92518020e-01 8.18000078e-01 -3.14590745e-02 -9.58308518e-01
-2.01877490e-01 -5.54446578e-01 -4.37983781e-01 -5.46452105e-01
3.74348313e-01 6.94486201e-02 -5.00639796e-01 1.60613883e+00
5.74922919e-01 -1.28686726e-02 7.10434187e-03 8.87759745e-01
6.52213395e-01 6.39935017e-01 -3.01057249e-01 9.34700221e-02
1.42005146e+00 -1.31366384e+00 -5.29352307e-01 -1.15806967e-01
1.14639199e+00 -7.20637441e-01 1.15357053e+00 2.42973968e-01
-1.06333601e+00 -5.98080680e-02 -7.80313909e-01 -5.48587561e-01
-2.95100063e-01 -8.11885521e-02 7.62397826e-01 3.62992764e-01
-9.63465869e-01 9.49363708e-02 -9.05707777e-01 -8.82514194e-02
4.36907738e-01 4.47311878e-01 -1.12202466e-01 -6.54180884e-01
-1.23668075e+00 5.46543658e-01 5.87053955e-01 4.97919880e-02
-7.03299820e-01 -8.74790370e-01 -7.61286259e-01 3.09994280e-01
1.08414805e+00 -1.01068914e+00 1.24973047e+00 -8.84234812e-03
-1.22660708e+00 3.42308968e-01 -8.59485626e-01 -3.29029173e-01
2.22692162e-01 -3.69411469e-01 -1.37818933e-01 2.18208075e-01
3.51262884e-03 3.72548372e-01 4.93598551e-01 -1.02602851e+00
-4.33452040e-01 -5.27404070e-01 5.21533728e-01 7.47063905e-02
-4.15046185e-01 -3.06989998e-01 -1.35097218e+00 -2.61640728e-01
3.59212875e-01 -9.47671294e-01 -5.89928962e-03 8.96528140e-02
-5.98054171e-01 -6.55222237e-01 5.81701219e-01 -5.66407084e-01
1.80049455e+00 -2.07092190e+00 2.17274934e-01 4.98940647e-01
5.29809535e-01 -1.74962476e-01 -1.77472103e-02 7.65676320e-01
5.40767252e-01 1.23958386e-01 -2.69430876e-01 -2.33966991e-01
4.54273939e-01 4.11737323e-01 -4.15742904e-01 2.23201126e-01
-5.41951992e-02 1.44993091e+00 -9.79682088e-01 -7.80231893e-01
-2.49163866e-01 2.32433498e-01 -9.57688928e-01 1.65522203e-01
-8.24987590e-01 -1.33364648e-01 -8.12290609e-01 9.12957311e-01
5.41471124e-01 -6.16279602e-01 3.29126865e-01 -4.26299691e-01
3.80706638e-01 4.60224062e-01 -1.01975369e+00 2.02346373e+00
-6.83092535e-01 3.25822413e-01 6.70144558e-02 -8.37023377e-01
6.93854213e-01 -1.25808856e-02 3.95406485e-01 -9.02033329e-01
-3.45275313e-01 6.26683176e-01 -2.91646600e-01 -5.22521436e-01
4.97530252e-01 3.47471982e-01 -1.79300562e-01 5.13212681e-01
-3.23750377e-01 -2.50863045e-01 2.45736539e-01 5.90607285e-01
1.38844478e+00 -2.28785798e-01 -2.12141126e-01 1.07550040e-01
5.01911700e-01 -1.16966469e-02 4.45459008e-01 8.07511568e-01
3.78539562e-01 2.05732137e-01 7.87849069e-01 -2.36573771e-01
-5.34668326e-01 -9.81470823e-01 1.82769060e-01 1.19413793e+00
4.09110993e-01 -8.36699367e-01 -5.18757343e-01 -6.85494840e-01
3.16843182e-01 7.24164724e-01 -2.23606363e-01 -3.09165120e-01
-7.73670554e-01 -4.15683955e-01 6.02380633e-01 3.79991978e-01
3.37503880e-01 -6.22237802e-01 -3.07345569e-01 3.03057194e-01
-3.22156310e-01 -1.25152564e+00 -6.94286466e-01 -1.18749492e-01
-9.40535545e-01 -1.18012059e+00 -3.55081141e-01 -6.94709003e-01
6.99462056e-01 1.54483140e-01 1.28685474e+00 5.70061982e-01
-1.71536937e-01 4.29226965e-01 -2.21647948e-01 -1.78274557e-01
2.11315319e-01 5.35559118e-01 -5.12481987e-01 -4.82314639e-02
4.84771013e-01 -4.16606456e-01 -7.22715020e-01 3.26946795e-01
-1.07463932e+00 -3.33840668e-01 5.73140681e-01 9.59905684e-01
1.02464950e+00 2.50759095e-01 2.50954539e-01 -1.05456734e+00
6.82671010e-01 -5.79347789e-01 -9.12381709e-01 6.53326571e-01
-8.51996303e-01 6.40478015e-01 5.04569650e-01 -1.16081508e-02
-6.33210182e-01 2.04931237e-02 -6.43301010e-02 -4.63160455e-01
5.12000084e-01 1.08149636e+00 -7.34921470e-02 -1.04947053e-01
4.31890965e-01 3.02803040e-01 -2.13728517e-01 -5.45113862e-01
4.94090468e-01 2.13130906e-01 3.63140225e-01 -9.03411567e-01
9.41049516e-01 4.71388310e-01 1.07889637e-01 -5.03818691e-01
-1.15837550e+00 -5.99304795e-01 -1.94349587e-01 2.20668405e-01
4.95323747e-01 -7.73171127e-01 -1.07422853e+00 5.55980951e-02
-1.29117107e+00 -3.56364816e-01 -1.86896980e-01 4.48294252e-01
-1.60457015e-01 3.59268785e-01 -4.85582769e-01 -5.05773425e-01
-4.87069428e-01 -1.15166497e+00 1.18341506e+00 -9.50067341e-02
-1.86586604e-02 -1.05246890e+00 -7.21091628e-02 7.06804335e-01
4.15046483e-01 -2.01311097e-01 1.35609269e+00 -4.99301136e-01
-1.46813810e+00 -3.52899015e-01 -4.61332828e-01 -5.91263771e-02
-4.14280236e-01 -4.19070929e-01 -6.78303361e-01 -3.18626136e-01
-2.75943369e-01 -5.56395769e-01 1.08508348e+00 -2.44216710e-01
1.42531800e+00 -4.03647006e-01 -3.91308248e-01 7.93436229e-01
1.52231431e+00 -4.85133141e-01 4.69057828e-01 -1.60372987e-01
7.64899135e-01 3.36882323e-01 6.68330967e-01 3.96628678e-01
1.05319452e+00 8.23348820e-01 6.52433336e-01 2.92015612e-01
-1.48234114e-01 -7.39958823e-01 2.40739837e-01 8.90216470e-01
4.04672354e-01 -3.75604987e-01 -1.02496171e+00 5.65486729e-01
-1.98702872e+00 -5.43185055e-01 -2.72918314e-01 2.28305149e+00
9.03834343e-01 1.57908410e-01 -2.31985748e-01 -7.33760670e-02
2.63302982e-01 2.30931401e-01 -7.99184263e-01 -1.24017522e-01
1.21664628e-01 4.65493649e-01 7.41682589e-01 9.52026129e-01
-5.12254357e-01 1.12583113e+00 5.19153357e+00 1.01683855e+00
-9.03858304e-01 2.13900194e-01 8.92597362e-02 -1.90141365e-01
-8.52906466e-01 4.96535540e-01 -9.01701570e-01 1.89407587e-01
6.95703506e-01 -3.34410846e-01 7.65905082e-01 6.64225578e-01
-3.04949909e-01 -1.96668372e-01 -1.28800046e+00 9.94821131e-01
3.44531536e-02 -1.34670544e+00 -2.38912809e-03 1.66147854e-02
4.82540309e-01 1.32617936e-01 -1.02878734e-01 5.60834765e-01
1.72257826e-01 -9.33940113e-01 4.75170255e-01 6.08959258e-01
6.60408676e-01 -7.22501457e-01 7.31658936e-01 3.44427973e-01
-1.53680634e+00 -4.89292294e-02 -3.19429278e-01 2.48784065e-01
1.52808294e-01 8.39359879e-01 -7.53481805e-01 7.31047392e-01
5.58475792e-01 3.65390927e-01 -6.05714262e-01 8.72807324e-01
-4.70962912e-01 6.42943501e-01 -8.65900755e-01 -3.44049811e-01
5.55643849e-02 7.91264251e-02 3.45193744e-01 9.15482581e-01
2.76283443e-01 8.70846882e-02 4.27146018e-01 9.36102629e-01
-6.37206078e-01 1.63249433e-01 -2.56823599e-01 -3.35908979e-01
8.69669974e-01 8.71282160e-01 -2.24984646e-01 -2.84793884e-01
-4.85871881e-01 9.73694444e-01 7.28222132e-01 5.48597157e-01
-8.22814107e-01 -7.34629810e-01 3.84170771e-01 -4.41020401e-03
6.50311470e-01 -2.47153088e-01 -1.74177006e-01 -1.23235190e+00
7.86853671e-01 -6.83366537e-01 7.80720115e-01 -4.01663899e-01
-1.02481532e+00 3.12828958e-01 2.45647095e-02 -5.41365445e-01
-1.09374166e-01 -1.68414533e-01 -2.70087868e-01 9.00686443e-01
-1.81416476e+00 -9.60470915e-01 -3.96878004e-01 6.73806369e-01
-2.25733593e-02 1.67760983e-01 8.84007335e-01 5.10729551e-01
-5.29510379e-01 1.06114423e+00 -8.95287562e-03 4.54921648e-02
4.33220685e-01 -1.11474264e+00 1.81876838e-01 4.94960397e-01
3.47816557e-01 6.80687904e-01 2.97662199e-01 -4.05751944e-01
-2.27910972e+00 -9.11173820e-01 1.42286003e+00 -4.15472925e-01
8.25431287e-01 -3.65900576e-01 -1.07188606e+00 5.42201817e-01
-1.71878770e-01 3.52694303e-01 6.12063706e-01 3.07728082e-01
-8.10610652e-01 -4.24829394e-01 -7.68869460e-01 4.44929242e-01
1.13393581e+00 -9.80822325e-01 -3.52804542e-01 5.96800804e-01
1.18413770e+00 -6.67874515e-01 -1.20293081e+00 2.52998978e-01
4.42693502e-01 -7.12354958e-01 1.01277220e+00 -3.72317106e-01
1.21674955e-01 -4.35199380e-01 -2.95126975e-01 -6.76059067e-01
8.91671926e-02 -7.93145418e-01 -9.49811578e-01 7.09248185e-01
8.11767936e-01 -8.40581238e-01 1.01045895e+00 6.42229319e-01
3.75506394e-02 -1.43868661e+00 -9.11182404e-01 -7.17405260e-01
-2.78617680e-01 -7.73221135e-01 8.75016034e-01 5.14235616e-01
-2.34894589e-01 5.99080920e-01 4.71376069e-02 5.38313031e-01
7.02080846e-01 5.85763991e-01 9.94202852e-01 -1.06075740e+00
-5.87880909e-01 -2.40776107e-01 -1.67699963e-01 -1.73611641e+00
2.78717995e-01 -1.26319599e+00 -3.18326890e-01 -1.81174016e+00
3.94319408e-02 -7.78053463e-01 -1.19131461e-01 5.67911923e-01
-1.41018480e-01 -2.12603867e-01 2.01076254e-01 2.88883597e-01
-9.23937619e-01 7.57163703e-01 1.25131476e+00 -3.84852469e-01
-3.41440327e-02 6.56582490e-02 -7.32633829e-01 2.83129275e-01
6.41542494e-01 -7.74405181e-01 -6.85005963e-01 -9.47251022e-01
8.49539340e-01 2.62456030e-01 4.91210699e-01 -6.20377064e-01
8.28958333e-01 1.28962189e-01 -3.44254911e-01 -7.38546312e-01
6.42090559e-01 -9.31648672e-01 -3.46849710e-02 6.62369549e-01
-3.38909626e-01 1.61038488e-01 -1.18534930e-01 8.31896067e-01
-5.48509955e-01 -3.27900410e-01 1.88306808e-01 1.40786350e-01
-5.81003964e-01 7.13975072e-01 5.33194005e-01 5.53923130e-01
7.37456143e-01 3.46580297e-01 -3.77530307e-01 -5.14121413e-01
-4.84327257e-01 8.51305723e-01 2.65776187e-01 1.32322684e-01
7.54921317e-01 -1.26160860e+00 -5.90515375e-01 -5.58967218e-02
3.75123650e-01 5.49172819e-01 1.20936342e-01 1.00575626e+00
-6.18451834e-01 7.71307111e-01 8.94406140e-01 -4.72696632e-01
-9.64423239e-01 5.78803003e-01 2.15452135e-01 -6.41993165e-01
-4.25495148e-01 1.00974059e+00 -1.96708694e-01 -4.65561628e-01
5.15272081e-01 -3.02113235e-01 3.72802377e-01 -1.70526028e-01
2.75755256e-01 4.25899655e-01 1.68089021e-03 -9.44323242e-02
-4.09414172e-01 6.15410566e-01 -1.91330716e-01 -1.83497500e-02
1.05731952e+00 5.00575081e-02 -6.36499286e-01 8.97046998e-02
1.68864298e+00 1.93292290e-01 -5.08444130e-01 -8.97191286e-01
2.35177547e-01 -5.61874688e-01 1.39562875e-01 -5.83109677e-01
-1.14874327e+00 8.18690717e-01 8.70401412e-02 8.12396854e-02
1.02113163e+00 3.56648237e-01 1.29857683e+00 1.21530569e+00
5.71951568e-01 -9.16241944e-01 1.11746617e-01 7.37411022e-01
8.09832573e-01 -1.22582746e+00 1.45134181e-01 -6.69196665e-01
-1.41582653e-01 9.88972723e-01 5.10125399e-01 2.07975149e-01
8.10356379e-01 -1.54978588e-01 -4.10246253e-01 -5.91555655e-01
-9.99129832e-01 -1.15396887e-01 2.75794297e-01 4.44399379e-02
1.91723868e-01 6.46484941e-02 -2.62619078e-01 3.85084659e-01
-2.57160395e-01 7.78347487e-03 -7.03655630e-02 1.02833104e+00
-2.46306732e-01 -1.37728822e+00 -4.33246829e-02 5.37816405e-01
-2.86834091e-01 -2.63479531e-01 -3.37292552e-01 5.34092486e-01
-3.04993749e-01 9.95969594e-01 -3.03176492e-01 -4.23839688e-01
5.64136982e-01 1.80197358e-01 3.04904103e-01 -6.26482248e-01
-3.00593436e-01 -3.96857709e-01 2.91079402e-01 -9.27006304e-01
1.15160523e-02 -1.35609150e-01 -1.60139954e+00 -4.32680905e-01
-3.94214332e-01 4.25819576e-01 4.57381427e-01 5.92247427e-01
8.09513211e-01 3.03092927e-01 4.50903833e-01 1.48514882e-01
-8.52603734e-01 -5.85165679e-01 -1.07401267e-01 1.41267374e-01
2.50856578e-01 -3.26806933e-01 -1.07503310e-01 -3.90781671e-01] | [9.324214935302734, 7.716960430145264] |
9fa80e15-1e7c-4479-baa8-0b26b1335c09 | cic-lt-edi-acl2022-are-transformers-the-only | null | null | https://aclanthology.org/2022.ltedi-1.28 | https://aclanthology.org/2022.ltedi-1.28.pdf | CIC@LT-EDI-ACL2022: Are transformers the only hope? Hope speech detection for Spanish and English comments | Hope is an inherent part of human life and essential for improving the quality of life. Hope increases happiness and reduces stress and feelings of helplessness. Hope speech is the desired outcome for better and can be studied using text from various online sources where people express their desires and outcomes. In this paper, we address a deep-learning approach with a combination of linguistic and psycho-linguistic features for hope-speech detection. We report our best results submitted to LT-EDI-2022 which ranked 2nd and 3rd in English and Spanish respectively. | ['Alexander Gelbukh', 'Grigori Sidorov', 'Sabur Butt', 'Fazlourrahman Balouchzahi'] | null | null | null | null | ltedi-acl-2022-5 | ['hope-speech-detection'] | ['natural-language-processing'] | [-1.05028617e+00 3.49820137e-01 -8.35201323e-01 -3.28049809e-01
-7.36450195e-01 1.97356850e-01 7.63235152e-01 6.58360541e-01
-3.14175457e-01 9.53120053e-01 1.49173141e+00 2.64883220e-01
1.12303108e-01 -6.49422646e-01 3.84658635e-01 -1.02866471e-01
-9.79642272e-02 1.22115813e-01 -7.29314029e-01 -6.42864347e-01
1.86761260e-01 5.79384387e-01 -8.75818551e-01 3.29239786e-01
2.82295614e-01 9.39519167e-01 -4.80215043e-01 3.74229878e-01
-2.95578927e-01 1.46240056e+00 -2.27010846e-01 -6.94991469e-01
-3.21134686e-01 -2.90414423e-01 -9.49110150e-01 -1.41030580e-01
-3.01681697e-01 -2.54357457e-01 -5.30037820e-01 8.44897807e-01
7.50120997e-01 1.75588727e-01 2.69622117e-01 -1.33626008e+00
-1.03937197e+00 7.40421116e-01 -9.12889913e-02 1.22697599e-01
8.79946351e-01 1.48825720e-01 8.41189325e-01 -1.13178051e+00
5.45365334e-01 1.53098965e+00 6.35151327e-01 8.29607725e-01
-5.36454320e-01 -4.38429475e-01 -8.26500714e-01 2.48811603e-01
-1.06750643e+00 -1.29765224e+00 1.24566185e+00 2.49036085e-02
1.44004369e+00 2.08185688e-01 1.10262871e+00 1.18806183e+00
3.21178883e-01 1.19551802e+00 8.43306780e-01 -3.05565923e-01
2.11756214e-01 4.73736167e-01 -8.07426572e-02 8.85767460e-01
-6.02877796e-01 1.63661107e-01 -1.20448339e+00 -1.93789288e-01
3.30714881e-02 1.51370749e-01 1.12964533e-01 3.47200125e-01
-1.14496195e+00 1.27798736e+00 4.71008360e-01 5.30335963e-01
-7.74597108e-01 -1.44324347e-01 5.86288869e-01 3.32154900e-01
6.11940086e-01 4.97461446e-02 -1.09193526e-01 -7.88685143e-01
-1.07230759e+00 -1.30261630e-01 9.14113760e-01 2.99781948e-01
4.57064770e-02 3.28481287e-01 -3.15415412e-01 1.17494488e+00
6.59948409e-01 4.18222100e-01 9.96870697e-01 -1.00580633e+00
-1.70884170e-02 6.29054189e-01 -1.19886845e-01 -1.10290837e+00
-8.10264230e-01 -4.93783981e-01 -9.46842372e-01 2.05940474e-02
-3.28596026e-01 -3.46377283e-01 -2.86701232e-01 1.76243091e+00
-2.11673211e-02 -4.29454297e-01 7.66257882e-01 5.67539275e-01
1.57266867e+00 7.27276802e-01 2.25216210e-01 -6.44387901e-01
1.07099676e+00 -1.14928412e+00 -1.26490092e+00 -6.08658135e-01
5.30669808e-01 -1.04389119e+00 1.10006905e+00 4.08605665e-01
-1.40195417e+00 1.49880826e-01 -7.71843433e-01 -5.10911465e-01
-4.21816081e-01 1.01224415e-01 1.10526514e+00 6.98510230e-01
-1.30865037e+00 5.57306707e-01 -4.08140451e-01 -9.38652098e-01
8.41270387e-01 6.50581047e-02 -5.54776073e-01 2.95657396e-01
-1.30003071e+00 1.10409713e+00 -8.09796751e-02 -2.48665661e-01
-6.28773987e-01 -1.37048811e-01 -8.20506036e-01 1.81724548e-01
-3.15233976e-01 -4.63602424e-01 1.03828132e+00 -5.57097673e-01
-1.62739277e+00 1.23414516e+00 -3.03711683e-01 -7.09648609e-01
5.34524433e-02 -8.24398473e-02 -8.73665631e-01 -1.62940636e-01
-8.03941581e-03 1.03317845e+00 5.27510762e-01 -4.13253844e-01
-1.48778886e-01 -5.37978828e-01 -2.16309413e-01 2.58403152e-01
-9.29219007e-01 4.69610363e-01 3.80382985e-01 -3.08827430e-01
4.23878804e-03 -4.83678043e-01 -1.17833447e-02 8.41417983e-02
-3.67458522e-01 -6.84432149e-01 7.36688673e-01 -7.93970346e-01
1.17140567e+00 -2.18965673e+00 -2.85967618e-01 -7.32231915e-01
4.93015349e-01 -1.91675015e-02 8.47012401e-02 7.20268607e-01
1.22269981e-01 2.31979892e-01 4.63759035e-01 -8.10153127e-01
8.37107971e-02 2.05759630e-01 5.67480363e-02 5.56062758e-01
2.11883888e-01 8.96805525e-01 -9.77896452e-01 -6.77858114e-01
2.46837407e-01 8.31658185e-01 -4.62771237e-01 2.80896604e-01
4.17090058e-01 -1.93019599e-01 -3.70199025e-01 9.40515220e-01
1.26971006e-01 -1.31190956e-01 -3.25211048e-01 1.65275037e-02
-3.21285069e-01 7.26778805e-01 -1.44174278e-01 1.65247297e+00
-6.15849257e-01 9.48734462e-01 -8.35437607e-03 -7.92776406e-01
1.34670579e+00 6.27973676e-01 3.82448733e-01 -7.36903131e-01
3.99939865e-01 3.44535476e-03 -4.52757686e-01 -6.09842539e-01
4.99344945e-01 -8.14865649e-01 -2.22125560e-01 5.91966584e-02
1.13770217e-01 -4.00206685e-01 -3.59687001e-01 4.98201400e-01
7.76997745e-01 -8.16166520e-01 9.18250024e-01 -1.95845142e-01
5.43127060e-01 -3.01550657e-01 5.76907873e-01 1.65989131e-01
-1.11840284e+00 2.54242331e-01 7.21873641e-01 -6.02985084e-01
-7.44081020e-01 -9.94102418e-01 7.88569599e-02 7.03701079e-01
-6.87778354e-01 -2.59224206e-01 -1.14209250e-01 -1.96209535e-01
-4.36240524e-01 1.35496783e+00 -4.94371712e-01 -4.42269772e-01
1.48634598e-01 -4.09870178e-01 3.25659811e-01 1.82430863e-01
4.95117188e-01 -1.64802969e+00 -4.47289169e-01 7.91359544e-02
-5.44642806e-01 -8.64208341e-01 -2.86765069e-01 1.26086205e-01
-8.54149640e-01 -1.94491461e-01 -5.59376419e-01 -7.13728905e-01
-2.26499230e-01 -1.34766400e-01 1.24326468e+00 -8.37450475e-02
-1.21558763e-01 1.69269904e-01 -1.88975409e-01 -4.36551958e-01
-4.44128811e-01 -2.34737247e-01 2.96215564e-01 -4.24363792e-01
6.82806909e-01 -7.64835954e-01 -2.90922910e-01 -8.40573013e-01
-1.74496159e-01 7.19839036e-02 2.50072449e-01 7.92640090e-01
-1.68305695e-01 -1.90867856e-01 8.99690092e-01 -4.22770865e-02
1.22075367e+00 -8.84722114e-01 3.68016988e-01 -7.06390366e-02
-4.98472124e-01 -2.99494974e-02 1.81430593e-01 3.12074590e-02
-8.29139233e-01 -3.27418670e-02 -6.19818568e-01 -9.38198343e-02
-1.11069247e-01 6.39566183e-01 7.26372674e-02 3.60952020e-01
2.32080266e-01 -2.44279560e-02 -9.33906808e-02 -4.19750452e-01
3.77747744e-01 1.17707396e+00 3.98808390e-01 1.16065398e-01
-1.65408671e-01 4.31799740e-01 -9.76566747e-02 -1.29296851e+00
-1.07950914e+00 -4.04424876e-01 -1.68568715e-01 -3.76156896e-01
7.79255331e-01 -9.53162968e-01 -1.01945353e+00 2.77624428e-01
-1.31137931e+00 2.72311687e-01 -4.80070174e-01 4.44208086e-01
-5.15616596e-01 7.46241584e-02 -7.46494830e-01 -1.52005386e+00
-1.14549637e+00 -7.02027738e-01 7.15149879e-01 5.72131097e-01
-4.87026393e-01 -1.19200814e+00 1.97194025e-01 6.12586081e-01
6.03510559e-01 1.88802391e-01 5.55800974e-01 -6.61644697e-01
5.77527046e-01 -2.94725925e-01 -1.00771770e-01 1.54569447e-01
2.13800684e-01 -4.26488668e-01 -1.03408098e+00 9.39305276e-02
5.27625680e-01 -1.34480214e+00 7.40097404e-01 3.64044875e-01
3.85674030e-01 -7.55493164e-01 1.40905365e-01 4.33067679e-02
1.28448522e+00 -1.10774063e-01 6.58844054e-01 1.58718973e-01
4.05341852e-03 6.19729340e-01 2.01861754e-01 1.16092336e+00
6.67942762e-01 2.05878630e-01 3.49411368e-01 6.23391867e-02
-3.71874273e-01 -4.76918191e-01 7.01433778e-01 1.09416008e+00
4.01885718e-01 -3.44524145e-01 -1.07577789e+00 9.52988863e-01
-1.37377715e+00 -1.27759361e+00 -5.43763787e-02 1.84166133e+00
8.50652874e-01 1.63146764e-01 3.02401304e-01 1.09735526e-01
7.11107776e-02 5.46150327e-01 -3.01943570e-01 -9.69770133e-01
-3.55236977e-01 6.02501631e-02 -8.19083899e-02 7.62748897e-01
-9.08515036e-01 1.00715911e+00 7.12090588e+00 6.56740129e-01
-1.15851045e+00 4.48101044e-01 1.15405798e+00 -5.73473871e-01
-2.67655402e-01 -2.65723944e-01 -3.63121390e-01 -9.91730392e-02
1.30762112e+00 -5.38424253e-01 6.18505180e-01 8.82961094e-01
8.51068318e-01 -1.75983727e-01 -7.63517439e-01 1.23980606e+00
3.33524436e-01 -1.31824362e+00 -5.90873241e-01 -2.50747830e-01
2.70478994e-01 4.86030787e-01 3.12789083e-01 3.93615305e-01
1.15590364e-01 -1.30693781e+00 4.96295542e-01 8.29928577e-01
4.82847273e-01 -9.35572386e-01 6.50778651e-01 4.95437980e-01
-5.39713264e-01 -1.85939416e-01 -2.26325467e-01 -4.86035526e-01
4.14116651e-01 7.67028451e-01 -7.82870650e-01 -2.20208913e-01
4.21756774e-01 1.04475927e+00 -2.16727450e-01 6.18600249e-01
-6.29379898e-02 4.54272091e-01 -3.36457133e-01 -7.51320481e-01
3.79698783e-01 3.51256907e-01 7.12087750e-01 1.15443242e+00
5.28461576e-01 7.62627721e-02 -3.26165743e-03 9.70768511e-01
-4.22647774e-01 3.99565578e-01 -7.78324842e-01 -7.82573104e-01
1.56058609e-01 1.39438927e+00 -2.66388595e-01 -1.27095178e-01
-3.75010163e-01 7.47465730e-01 5.43801665e-01 -3.35597694e-01
-2.59542406e-01 2.04828039e-01 5.82096517e-01 -1.04167216e-01
-5.90825260e-01 7.26398686e-03 -4.11354095e-01 -1.15154409e+00
-4.57757384e-01 -5.65115511e-01 3.54261249e-01 -9.82460797e-01
-1.39513612e+00 7.21923947e-01 -6.87021255e-01 -6.98990941e-01
-2.87907451e-01 -1.72656983e-01 -7.88463533e-01 6.64318383e-01
-1.42781389e+00 -1.26369369e+00 2.00857833e-01 4.18974519e-01
9.35954630e-01 -5.19454837e-01 1.24025738e+00 1.65506631e-01
-3.53573710e-01 4.55361456e-01 -9.90988910e-02 -1.17839880e-01
4.83024329e-01 -8.50723386e-01 -1.03672847e-01 3.10721487e-01
9.83214900e-02 1.70484915e-01 9.43579257e-01 -5.70326269e-01
-1.42134464e+00 -3.94287407e-01 2.04943419e+00 -1.31049022e-01
9.51084554e-01 -1.12337247e-03 -3.00593734e-01 1.30191490e-01
9.00519967e-01 -2.61317372e-01 9.12822008e-01 3.94628912e-01
1.71699196e-01 -1.03784874e-02 -1.74111116e+00 4.44469392e-01
5.61650872e-01 -6.92504525e-01 -6.93829775e-01 8.40764225e-01
7.33922720e-01 2.07734004e-01 -9.18117881e-01 -1.67864040e-01
4.66348469e-01 -1.30854356e+00 9.98761773e-01 -7.57341743e-01
9.90564823e-01 8.66011858e-01 -3.53801310e-01 -1.22663152e+00
-1.54285163e-01 -7.66889155e-01 -1.28141180e-01 1.44079196e+00
3.74967843e-01 -1.57844007e-01 9.16995704e-01 9.27502692e-01
1.35361075e-01 -8.60308111e-01 -1.21760106e+00 -2.39531234e-01
3.19123566e-01 -5.03174603e-01 2.93172479e-01 1.07626820e+00
1.01738405e+00 8.40615094e-01 -9.01280046e-01 -6.21205509e-01
4.54002082e-01 -3.76071185e-01 1.54496487e-02 -1.02613521e+00
2.80393898e-01 -6.93886817e-01 -2.39405081e-01 -3.20613921e-01
2.09037706e-01 -7.90279627e-01 -4.05130953e-01 -1.70997226e+00
6.08033717e-01 1.98111106e-02 -2.66816944e-01 8.03987682e-01
4.20270205e-01 -9.95289087e-02 2.35289633e-01 -3.68341841e-02
-5.12689173e-01 1.09262228e+00 9.08075035e-01 -1.43453047e-01
-2.49722511e-01 -1.15619056e-01 -9.34795022e-01 8.21103215e-01
1.26689386e+00 -4.13690388e-01 -2.11320728e-01 5.44302501e-02
4.43585843e-01 8.73770654e-01 1.06687061e-01 -8.46993268e-01
1.60068333e-01 -3.15500289e-01 2.20406473e-01 -8.00095677e-01
9.51475382e-01 -2.91686803e-01 -2.80063689e-01 5.92640877e-01
-6.91087902e-01 -6.21628687e-02 7.77885914e-02 1.37092397e-01
-2.33541310e-01 -2.86944449e-01 9.66415405e-01 -1.44660607e-01
-1.59032077e-01 2.44306438e-02 -4.85342890e-01 -1.43144326e-02
4.62165475e-01 4.92945939e-01 -2.55630583e-01 -1.23340392e+00
-9.11454499e-01 2.50173640e-02 4.25664373e-02 7.05849767e-01
1.17325783e+00 -1.69605398e+00 -9.25529361e-01 -1.22813523e-01
-1.08841226e-01 -1.21585286e+00 3.82561907e-02 9.81478333e-01
-1.32891284e-02 6.62402868e-01 -3.18686545e-01 2.56784290e-01
-1.24124849e+00 6.55243993e-01 1.93748727e-01 -3.92762542e-01
-4.92080420e-01 1.07282460e+00 -6.99900746e-01 -4.10859585e-01
3.32095385e-01 1.30488932e-01 -8.35998178e-01 5.82013428e-02
7.46101022e-01 3.88331175e-01 -3.83536816e-01 -1.05616593e+00
-6.19434059e-01 -4.55598772e-01 2.82201886e-01 -2.98582256e-01
1.88869989e+00 -3.93403202e-01 -4.61552739e-01 6.43593550e-01
1.54273927e+00 -3.26304398e-02 -2.17195883e-01 6.92914426e-02
7.18903765e-02 -3.29604030e-01 9.99318600e-01 -1.10606754e+00
-1.13388336e+00 1.18600571e+00 1.04895747e+00 3.18685383e-01
1.03291976e+00 3.10854793e-01 1.32456517e+00 3.34291399e-01
2.43721940e-02 -1.11898029e+00 3.64860803e-01 4.27239150e-01
1.27346742e+00 -1.70768404e+00 -1.08451657e-01 3.22314918e-01
-8.76262367e-01 1.17398572e+00 9.49641913e-02 2.17613354e-01
9.53338325e-01 2.05477744e-01 8.05916861e-02 -2.42520005e-01
-1.00901747e+00 6.54749423e-02 6.70063794e-02 4.89291072e-01
1.16716647e+00 4.16035503e-01 -8.20126116e-01 9.86900806e-01
-3.59810293e-01 4.17082250e-01 3.69871318e-01 3.73275012e-01
-8.72563422e-01 -1.03606761e+00 -3.92838977e-02 4.67667371e-01
-6.77521288e-01 -4.56742227e-01 -6.75215006e-01 3.41977216e-02
-2.12190300e-01 1.47657382e+00 -3.75991672e-01 -3.43350440e-01
-2.16657549e-01 2.30032876e-01 -9.07433107e-02 -2.24160209e-01
-5.85607231e-01 1.73865631e-01 7.55122602e-01 -5.28812468e-01
-6.09443545e-01 -8.69358838e-01 -1.32071674e+00 -8.17862570e-01
2.28470996e-01 1.84575710e-02 7.37549901e-01 1.04901779e+00
2.60967791e-01 4.00444306e-02 5.65313756e-01 -3.47519189e-01
-4.93389428e-01 -9.49289143e-01 -6.28654957e-01 -9.77255851e-02
5.46313822e-01 -2.55827427e-01 -3.53946298e-01 -3.86964649e-01] | [9.034977912902832, 10.748452186584473] |
4e6a9aed-010f-4914-951f-3df3c4dd2bf1 | a-systematic-review-of-transfer-learning | 2105.13793 | null | https://arxiv.org/abs/2105.13793v1 | https://arxiv.org/pdf/2105.13793v1.pdf | A systematic review of transfer learning based approaches for diabetic retinopathy detection | Cases of diabetes and related diabetic retinopathy (DR) have been increasing at an alarming rate in modern times. Early detection of DR is an important problem since it may cause permanent blindness in the late stages. In the last two decades, many different approaches have been applied in DR detection. Reviewing academic literature shows that deep neural networks (DNNs) have become the most preferred approach for DR detection. Among these DNN approaches, Convolutional Neural Network (CNN) models are the most used ones in the field of medical image classification. Designing a new CNN architecture is a tedious and time-consuming approach. Additionally, training an enormous number of parameters is also a difficult task. Due to this reason, instead of training CNNs from scratch, using pre-trained models has been suggested in recent years as transfer learning approach. Accordingly, the present study as a review focuses on DNN and Transfer Learning based applications of DR detection considering 38 publications between 2015 and 2020. The published papers are summarized using 9 figures and 10 tables, giving information about 22 pre-trained CNN models, 12 DR data sets and standard performance metrics. | ['Atilla Özgür', 'Hamit Erdem', 'Büşra Kübra Karaca', 'Burcu Oltu'] | 2021-05-28 | null | null | null | null | ['diabetic-retinopathy-detection'] | ['medical'] | [-2.75187641e-02 -1.08830400e-01 -2.08900243e-01 -4.69072104e-01
-1.37496784e-01 -1.07603677e-01 3.35487098e-01 -8.96664783e-02
-5.83378375e-01 8.74845326e-01 1.53285742e-01 -3.56546760e-01
-8.38758945e-02 -8.19864452e-01 -7.28334039e-02 -6.85684621e-01
8.65062997e-02 3.14404845e-01 1.04635231e-01 -8.79383311e-02
4.31633383e-01 9.36382234e-01 -1.73727429e+00 3.47848922e-01
1.20140791e+00 9.38941121e-01 1.12622239e-01 6.45246923e-01
-3.65158617e-01 1.10724640e+00 -6.31038725e-01 -4.68333930e-01
2.42710814e-01 -5.04284859e-01 -6.04514837e-01 -1.16058819e-01
4.88361150e-01 -6.57777071e-01 -4.28707987e-01 8.93279195e-01
1.11831069e+00 -1.19319037e-01 5.77868938e-01 -4.38553452e-01
-7.74104953e-01 2.91466683e-01 -5.61013818e-01 5.00420868e-01
-1.12307385e-01 7.59570897e-02 3.47243518e-01 -7.26375937e-01
2.36691713e-01 1.14191592e+00 5.32261252e-01 8.81099582e-01
-9.97499466e-01 -6.73895478e-01 -4.57627147e-01 6.20254993e-01
-1.10706520e+00 -2.68897176e-01 5.19979656e-01 -7.61741281e-01
7.00040162e-01 -1.74768344e-02 9.07396793e-01 9.87827957e-01
1.82819828e-01 6.79839790e-01 1.19152820e+00 -3.28247726e-01
2.10220844e-01 9.73324478e-02 5.78437559e-02 2.34597504e-01
7.50853956e-01 1.42519683e-01 1.13662392e-01 2.18464687e-01
1.00400603e+00 -6.30187765e-02 -1.81266248e-01 -7.31663629e-02
-7.69179106e-01 9.17721987e-01 6.12691700e-01 3.60282362e-01
-4.25226420e-01 -1.50291979e-01 5.43029666e-01 1.94011033e-01
2.29843214e-01 3.90169144e-01 -2.40579188e-01 -1.32469565e-01
-7.33993113e-01 6.05488010e-02 3.40840250e-01 4.31944579e-01
3.01298529e-01 1.82428032e-01 -9.32002440e-02 1.15723741e+00
1.60610583e-02 1.41456693e-01 3.87390405e-01 -6.92835689e-01
8.22983310e-02 8.51742566e-01 -4.34054770e-02 -9.09464180e-01
-3.95364642e-01 -7.06121087e-01 -1.39279902e+00 6.59173429e-01
6.48249984e-01 -3.13134819e-01 -1.30921364e+00 8.22706699e-01
-9.35952663e-02 -2.61029005e-01 2.23651320e-01 8.64320874e-01
1.06425416e+00 3.69882166e-01 2.75667846e-01 6.27872953e-03
1.26264668e+00 -6.91700101e-01 -6.72256827e-01 -1.22216403e-01
4.48176980e-01 -7.13710785e-01 5.16074002e-01 7.28245974e-01
-8.23926330e-01 -5.06339610e-01 -6.71833515e-01 -3.75245690e-01
-3.32085133e-01 7.27002919e-01 6.34306908e-01 6.64235413e-01
-1.02890086e+00 2.52776921e-01 -5.57504416e-01 -9.39969540e-01
8.90804172e-01 3.51858795e-01 -1.94445476e-01 -5.22363663e-01
-7.97646403e-01 1.30854928e+00 2.57116437e-01 5.09317040e-01
-4.83440220e-01 -3.86146069e-01 -2.89119899e-01 -2.95703799e-01
-1.12615287e-01 -7.67093897e-01 1.01711905e+00 -9.33597505e-01
-1.36064601e+00 1.08394074e+00 8.25014040e-02 -5.89473724e-01
6.70687377e-01 -3.77940029e-01 -7.75503635e-01 1.32021293e-01
-4.10177916e-01 5.59930325e-01 4.93653178e-01 -9.58470583e-01
-9.46144223e-01 -3.15765172e-01 9.64844376e-02 -1.50646254e-01
-8.28348771e-02 3.18546265e-01 9.75358561e-02 -4.00292665e-01
-5.16171791e-02 -4.74654585e-01 -2.97382206e-01 2.98765063e-01
-2.91340590e-01 -4.72664803e-01 5.51773548e-01 -7.43817449e-01
1.17471111e+00 -1.81923437e+00 -1.67116433e-01 -7.65207931e-02
5.07662892e-01 8.60101223e-01 1.90413594e-01 1.14416100e-01
-1.79065794e-01 1.28898621e-01 4.68385145e-02 2.28856683e-01
-7.41638720e-01 2.48842716e-01 2.74326682e-01 3.11428159e-01
2.24571362e-01 5.41487694e-01 -6.79718137e-01 -4.21795785e-01
5.33387065e-01 9.46980596e-01 -2.52746224e-01 2.30527908e-01
2.42416784e-01 8.63042533e-01 -4.17573899e-01 8.12584579e-01
7.35959589e-01 -1.68327853e-01 -5.33392429e-02 -4.68223155e-01
-4.01213527e-01 -1.81463778e-01 -7.23462880e-01 9.73498821e-01
-1.87437370e-01 9.51148033e-01 -4.12385702e-01 -1.29859555e+00
1.23210204e+00 3.61728638e-01 2.59154856e-01 -9.56486702e-01
4.82190490e-01 4.86448616e-01 4.50527281e-01 -1.08197427e+00
-4.76710089e-02 1.03342645e-01 7.93150902e-01 -3.23804080e-01
-2.92019665e-01 5.14262259e-01 2.51526862e-01 -5.74873507e-01
9.00393426e-01 -2.06641957e-01 6.11200035e-01 7.17714950e-02
8.01337779e-01 2.17042416e-02 7.42707372e-01 3.66134107e-01
-2.81060696e-01 6.70084894e-01 6.25650585e-01 -1.11772096e+00
-1.09989679e+00 -5.72655320e-01 -4.38059449e-01 7.58146346e-02
-4.25909996e-01 4.87600356e-01 -3.53961051e-01 -3.35173666e-01
-4.26378548e-02 5.03510654e-01 -8.27165544e-01 1.86882406e-01
-6.76004887e-01 -8.41621578e-01 4.29196358e-01 6.85673356e-01
1.10781884e+00 -1.01624548e+00 -6.01689935e-01 9.61820334e-02
1.91738456e-01 -8.80859733e-01 3.45672339e-01 -2.20137611e-01
-1.33220422e+00 -1.29941571e+00 -1.50029027e+00 -1.01456249e+00
9.36022937e-01 2.15925157e-01 9.37984049e-01 3.56950201e-02
-8.37438941e-01 -2.47632310e-01 -1.99517533e-01 -4.71619368e-01
-2.97788352e-01 3.24431248e-02 -3.34492087e-01 9.71988440e-02
9.05507863e-01 -4.45452452e-01 -1.06275797e+00 -2.10153759e-02
-4.86782253e-01 -9.31622311e-02 1.34431732e+00 6.51903868e-01
3.66103381e-01 -1.20662786e-01 7.25115716e-01 -7.47878432e-01
3.97024125e-01 -2.54258335e-01 -8.26624811e-01 7.83203840e-02
-7.30883181e-01 -3.00409883e-01 3.28112185e-01 -2.26325303e-01
-9.07153606e-01 -1.50536960e-02 -9.10143703e-02 -1.77870125e-01
-5.50171077e-01 4.01482284e-01 8.53263512e-02 -2.61494964e-01
8.36022615e-01 4.37691696e-02 3.80144149e-01 -9.14575934e-01
4.93197218e-02 9.18911636e-01 3.34369242e-01 -1.98997948e-02
4.42853600e-01 2.30890900e-01 1.34592742e-01 -9.56806839e-01
-8.48535419e-01 -3.82667959e-01 -5.34888268e-01 -3.98980945e-01
1.34473550e+00 -9.13904309e-01 -6.42478049e-01 7.78791666e-01
-1.12301970e+00 -1.13384150e-01 6.53974712e-02 8.38480353e-01
-3.15175951e-02 1.41400024e-01 -4.30071890e-01 -6.96537793e-01
-5.14122427e-01 -1.15392900e+00 1.93256497e-01 6.63249195e-01
9.46740434e-02 -8.52283597e-01 1.00653745e-01 4.23735887e-01
8.74525070e-01 6.07001126e-01 1.14147460e+00 -3.56357306e-01
-3.87088209e-01 -4.24343675e-01 -8.23928952e-01 8.05156112e-01
4.71120209e-01 2.61257023e-01 -9.11877573e-01 -7.66808391e-02
-3.89338583e-01 5.88195361e-02 8.89655471e-01 7.93998122e-01
1.28589714e+00 -2.25681923e-02 -2.89170444e-01 3.39752018e-01
1.78075874e+00 8.05535018e-01 1.30028272e+00 6.28968656e-01
5.26327431e-01 4.67045456e-01 2.10201070e-01 1.95471451e-01
2.70837069e-01 2.44680747e-01 4.69134510e-01 -5.34864366e-01
-3.41149151e-01 3.79382730e-01 -1.49738058e-01 3.95978391e-01
-9.34818685e-01 -2.27354392e-01 -1.02896011e+00 6.81066871e-01
-1.32185137e+00 -1.00916553e+00 -4.86017525e-01 2.26555490e+00
6.00138962e-01 -4.27983589e-02 2.03487962e-01 1.07372291e-01
9.17792559e-01 -4.15287435e-01 -4.88138020e-01 -3.78188908e-01
-1.13948971e-01 3.61063898e-01 5.13624728e-01 2.32117362e-02
-9.74349499e-01 4.58861887e-01 6.44892597e+00 9.50673670e-02
-1.41861045e+00 -2.25497976e-01 6.49484992e-01 1.12089112e-01
3.08455050e-01 -3.42179716e-01 -6.85606956e-01 5.17416477e-01
7.38568246e-01 2.94049263e-01 4.98597659e-02 7.89653718e-01
5.68735003e-01 -3.74523193e-01 -7.77878582e-01 1.15224314e+00
-9.29435194e-02 -1.19843495e+00 2.99408555e-01 1.91198662e-01
6.41551077e-01 1.03460401e-02 -1.78582277e-02 1.62857696e-01
-1.07819624e-01 -1.29569113e+00 -2.99731851e-01 9.81951118e-01
8.71040583e-01 -1.01339960e+00 1.37769210e+00 -2.82736987e-01
-6.07778668e-01 -2.88476765e-01 -6.95240498e-01 9.05494690e-02
-1.07952982e-01 9.74457204e-01 -4.88920093e-01 2.80957311e-01
9.79740560e-01 9.63771284e-01 -6.80414081e-01 1.83600771e+00
-2.82910705e-01 7.46664882e-01 1.63438752e-01 4.80197147e-02
-2.71755587e-02 -3.09410006e-01 3.04853588e-01 9.75172937e-01
5.27920604e-01 1.44950792e-01 -3.89299124e-01 7.04458117e-01
1.20475642e-01 1.04870968e-01 -4.57868844e-01 -7.97734931e-02
-6.90120310e-02 1.24914265e+00 -3.46909165e-01 -1.23078547e-01
-7.64733493e-01 5.01959562e-01 -1.56551987e-01 3.96278381e-01
-3.86031359e-01 -6.31178916e-01 7.31219292e-01 3.63085449e-01
2.52440035e-01 -4.05236222e-02 -3.42314243e-01 -8.49370539e-01
-2.65581250e-01 -7.53984928e-01 2.86826521e-01 -6.99731290e-01
-1.25213456e+00 6.52984738e-01 -4.15806413e-01 -1.51191390e+00
2.34052569e-01 -8.49775255e-01 -4.24618065e-01 1.07251656e+00
-1.97055769e+00 -7.42622554e-01 -8.68470192e-01 2.54886210e-01
4.07826930e-01 -4.33598846e-01 8.20218146e-01 7.97892392e-01
-9.50583041e-01 2.71111190e-01 8.99697468e-02 5.26030004e-01
8.59795988e-01 -1.19419730e+00 -1.44490972e-01 6.03647470e-01
-1.01284182e+00 5.02237439e-01 5.16974568e-01 -4.55602288e-01
-7.70387530e-01 -1.10275662e+00 9.63965416e-01 1.23236477e-01
2.02850983e-01 3.09553504e-01 -8.78574610e-01 5.08750975e-01
3.11192930e-01 1.82009459e-01 6.50681555e-01 -2.25995183e-01
-4.78196852e-02 -6.06124282e-01 -1.25666034e+00 5.50171316e-01
6.16681337e-01 -2.12724730e-02 -3.92173201e-01 4.62929271e-02
-5.46658337e-02 -3.51595521e-01 -9.88254070e-01 5.42926967e-01
6.68284774e-01 -1.16892445e+00 7.60718524e-01 -4.34045941e-01
7.07700729e-01 -3.76081884e-01 1.79920793e-01 -9.51275349e-01
-1.76381722e-01 -1.24471501e-01 6.93135038e-02 9.40992892e-01
1.92705125e-01 -9.05041397e-01 8.15247953e-01 4.05053109e-01
-1.13266587e-01 -7.48209655e-01 -6.26254618e-01 -5.27082503e-01
5.99057525e-02 2.36200154e-01 9.90051031e-02 7.45383024e-01
-7.44784057e-01 1.70019075e-01 -2.73308218e-01 2.35323384e-02
6.64628506e-01 -3.85194153e-01 5.56836903e-01 -1.73455656e+00
3.44544470e-01 -6.07463121e-01 -9.58742261e-01 -7.42382824e-01
-6.05400801e-01 -5.21245003e-01 -3.25008094e-01 -2.27479768e+00
9.20088589e-02 -2.25132033e-01 -3.61743033e-01 5.88436961e-01
1.84209183e-01 1.89050466e-01 -4.04061079e-01 1.94961578e-01
2.01359391e-01 7.94196427e-02 1.51097357e+00 -1.20984398e-01
-2.55050927e-01 3.88224870e-01 -6.70094550e-01 5.72576880e-01
1.07993436e+00 -1.98557928e-01 -2.98953593e-01 -5.10940433e-01
3.09285462e-01 -3.45280468e-01 4.31118220e-01 -1.27821469e+00
2.02511534e-01 1.24399364e-02 7.21177995e-01 -8.15266430e-01
8.12383462e-03 -7.04510212e-01 -1.64985452e-02 5.83803713e-01
-2.26156399e-01 -1.34900659e-01 2.10833877e-01 3.46818238e-01
-3.32993507e-01 -2.16862764e-02 1.23830676e+00 -2.75134146e-01
-9.08974230e-01 1.97313994e-01 -6.96165621e-01 -3.29399616e-01
9.63966906e-01 -5.79903007e-01 -3.12883735e-01 -1.66581094e-01
-8.44033241e-01 1.12927094e-01 3.87204885e-02 3.52499753e-01
6.61222756e-01 -9.41299140e-01 -9.29850519e-01 -1.69740394e-01
2.32085943e-01 1.27930954e-01 4.56308663e-01 1.10827458e+00
-1.00246930e+00 5.17760038e-01 -7.07044065e-01 -5.27529776e-01
-1.31166530e+00 3.36613297e-01 7.21683204e-01 1.07453056e-01
-1.01441896e+00 5.29046953e-01 -1.30484045e-01 1.68825343e-01
4.24848020e-01 -4.47074801e-01 -1.08320296e+00 2.16221720e-01
7.97275543e-01 7.86726236e-01 1.92463920e-01 -1.13449045e-01
-1.02525569e-01 8.91910195e-01 -4.06566978e-01 6.30242646e-01
1.42214334e+00 -4.30215038e-02 -4.07976717e-01 1.71568766e-01
8.43281448e-01 -4.31554973e-01 -7.71726251e-01 -8.45628679e-02
1.05787866e-01 -3.27689141e-01 3.78878862e-01 -1.13575518e+00
-1.47194445e+00 1.07103229e+00 1.16290259e+00 -1.26446038e-03
1.28372526e+00 -3.09382766e-01 6.13593102e-01 2.78176546e-01
1.91494644e-01 -9.04741406e-01 -1.86157942e-01 2.04017416e-01
9.30816948e-01 -1.26271737e+00 -9.52885766e-03 -2.86310554e-01
-4.29048121e-01 1.78040123e+00 6.12049997e-01 -1.91206008e-01
6.28566265e-01 -3.07545364e-01 5.10350645e-01 -1.14203192e-01
-1.42497048e-01 -3.06700587e-01 2.25518614e-01 7.41231501e-01
7.44615197e-01 -1.33166760e-01 -8.15097809e-01 1.10460892e-01
1.30897880e-01 5.21495759e-01 6.98160589e-01 7.49800324e-01
-5.37708044e-01 -1.20428383e+00 2.55239336e-03 8.83855641e-01
-6.63113415e-01 9.26964059e-02 -2.41378665e-01 8.67882490e-01
3.17634016e-01 8.52266729e-01 4.52029221e-02 -1.35744691e-01
4.02350366e-01 -2.65822142e-01 3.15676540e-01 -5.43427706e-01
-2.65837550e-01 -2.87918091e-01 2.87437439e-01 -3.35242301e-01
-8.57676983e-01 -3.94450516e-01 -9.82445419e-01 -3.78546268e-01
-7.93639645e-02 -3.07642758e-01 7.23713994e-01 8.01653266e-01
3.91266882e-01 6.20956242e-01 3.70052546e-01 -2.47105137e-01
-3.22952747e-01 -1.12100911e+00 -4.97459173e-01 -3.64801198e-01
4.29539084e-01 -5.74225664e-01 -3.08119714e-01 1.31107140e-02] | [15.83752727508545, -3.98468017578125] |
b70d536c-13ef-4b46-864e-9a10769eaddf | unsupervised-text-summarization-of-long | null | null | https://aclanthology.org/2022.rocling-1.3 | https://aclanthology.org/2022.rocling-1.3.pdf | Unsupervised Text Summarization of Long Documents using Dependency-based Noun Phrases and Contextual Order Arrangement | Unsupervised extractive summarization has recently gained importance since it does not require labeled data. Among unsupervised methods, graph-based approaches have achieved outstanding results. These methods represent each document by a graph, with sentences as nodes and word-level similarity among sentences as edges. Common words can easily lead to a strong connection between sentence nodes. Thus, sentences with many common words can be misinterpreted as salient sentences for a summary. This work addresses the common word issue with a phrase-level graph that (1) focuses on the noun phrases of a document based on grammar dependencies and (2) initializes edge weights by term-frequency within the target document and inverse document frequency over the entire corpus. The importance scores of noun phrases extracted from the graph are then used to select the most salient sentences. To preserve summary coherence, the order of the selected sentences is re-arranged by a flow-aware orderBERT. The results reveal that our unsupervised framework outperformed other extractive methods on ROUGE as well as two human evaluations for semantic similarity and summary coherence. | ['Yi-Shin Chen', 'Hsiao-Yen Lan', 'Yen-Hao Huang'] | null | null | null | null | rocling-2022-11 | ['unsupervised-extractive-summarization', 'extractive-summarization'] | ['natural-language-processing', 'natural-language-processing'] | [ 2.97995389e-01 3.14968348e-01 -4.18932855e-01 -2.42936343e-01
-4.78088200e-01 -5.65330446e-01 4.41356540e-01 9.59447324e-01
-3.69688660e-01 6.61228836e-01 1.02658582e+00 7.89179206e-02
-1.91654146e-01 -8.38556707e-01 -7.86161795e-02 -5.76667249e-01
-1.11880720e-01 1.82998374e-01 3.58291566e-01 -3.35622370e-01
9.68935132e-01 1.70594320e-01 -1.34198225e+00 2.39739701e-01
1.24537206e+00 2.72303045e-01 3.88605922e-01 5.54163873e-01
-7.31861532e-01 6.06951237e-01 -9.31023598e-01 -4.00899738e-01
-2.28899136e-01 -8.86430025e-01 -8.65256011e-01 3.50624204e-01
4.23382074e-01 1.32671863e-01 -2.36350179e-01 1.34879231e+00
3.42921197e-01 4.28038538e-01 5.08107245e-01 -7.80806124e-01
-4.34589863e-01 1.10362256e+00 -6.20727658e-01 4.64485198e-01
6.04500473e-01 -3.77545506e-01 1.66033423e+00 -8.52428436e-01
8.57000291e-01 1.33948541e+00 2.49101952e-01 1.56536669e-01
-9.88186538e-01 -5.40149771e-02 3.82731766e-01 1.70543939e-01
-1.07189846e+00 -2.01357916e-01 1.09738064e+00 -6.05404861e-02
1.20710659e+00 4.85747546e-01 7.62724876e-01 5.32699287e-01
5.18765867e-01 6.80935919e-01 3.92391473e-01 -6.72187924e-01
3.93350869e-01 -2.12266549e-01 7.74784446e-01 6.87518835e-01
6.28572106e-01 -7.72580504e-01 -7.99264908e-01 -1.89359963e-01
-3.19394469e-02 -1.26155883e-01 -3.10181230e-01 1.79023314e-02
-1.21536207e+00 9.78322625e-01 3.51757884e-01 7.39526212e-01
-7.41318941e-01 -7.96912462e-02 6.44175649e-01 1.62275255e-01
6.23066306e-01 6.93308234e-01 6.29249215e-02 7.41453618e-02
-1.11763823e+00 1.23108588e-01 8.78681064e-01 8.87923777e-01
8.10973704e-01 -1.50413334e-01 -5.18596828e-01 7.05978334e-01
2.44899422e-01 2.41939098e-01 4.88397151e-01 -6.68563366e-01
5.02701759e-01 1.11740136e+00 -3.71858478e-01 -1.79149044e+00
-2.39848897e-01 -5.62226534e-01 -7.99884558e-01 -4.59962159e-01
-2.82208949e-01 -1.39818678e-03 -6.17138386e-01 1.40929353e+00
1.10372134e-01 -1.77756906e-01 9.47592035e-02 6.14393055e-01
1.18217015e+00 9.00027394e-01 2.12137103e-01 -7.70571411e-01
1.44924164e+00 -1.00882292e+00 -1.08450902e+00 -3.45287085e-01
5.78290701e-01 -7.50889719e-01 9.67005134e-01 -2.65846830e-02
-1.11943328e+00 -3.19514900e-01 -1.04824042e+00 -4.43722727e-03
-2.43819714e-01 -1.94398552e-01 3.15582275e-01 1.92259759e-01
-1.22729981e+00 6.74763143e-01 -3.70595366e-01 -6.95920110e-01
1.22332022e-01 3.88022438e-02 -1.17433742e-01 1.57552153e-01
-1.21824467e+00 7.84185350e-01 9.64697182e-01 -1.75375849e-01
-9.64126214e-02 -3.00441265e-01 -1.04697537e+00 4.36518997e-01
5.32897711e-01 -8.22482407e-01 8.64753425e-01 -8.16601992e-01
-1.03143620e+00 7.13871121e-01 -5.44385970e-01 -3.96220386e-01
-2.82263935e-01 -1.09055676e-01 -2.95210689e-01 7.60711253e-01
3.70223999e-01 3.96055996e-01 7.90189087e-01 -1.18184841e+00
-8.07735801e-01 -2.48055890e-01 -1.53542444e-01 7.34627604e-01
-7.36783147e-01 1.65294737e-01 -5.21221280e-01 -9.10481334e-01
4.96318966e-01 -4.38485503e-01 -1.87115133e-01 -8.13373029e-01
-7.84403682e-01 -6.64494932e-01 7.17446923e-01 -6.42740190e-01
1.99303532e+00 -1.99239969e+00 3.82471353e-01 2.63130367e-01
5.76257646e-01 -9.06125903e-02 -1.23857170e-01 8.58460069e-01
1.91388488e-01 3.88573289e-01 -3.65687191e-01 -5.82368299e-02
-1.25970274e-01 6.54258877e-02 -2.34970018e-01 -3.66010591e-02
-4.45045047e-02 7.84698606e-01 -1.42207360e+00 -9.87404525e-01
-3.65690999e-02 -6.09827675e-02 -3.35900605e-01 7.35839680e-02
1.90495234e-02 -1.08457416e-01 -6.27764702e-01 2.89948970e-01
1.94028750e-01 -1.97563767e-01 3.77634674e-01 -4.14351404e-01
-1.63335204e-01 5.93630791e-01 -7.99199343e-01 1.45610845e+00
-6.23915493e-02 6.37448192e-01 -3.61430287e-01 -9.78701949e-01
1.08534551e+00 5.72323762e-02 4.20171469e-01 -4.36816394e-01
-8.88894573e-02 2.73415428e-02 -8.41183066e-02 -6.50687397e-01
1.17100573e+00 1.44997612e-01 -4.51579630e-01 4.26625341e-01
-6.11239411e-02 -3.70751202e-01 9.45541441e-01 1.08942580e+00
1.07974625e+00 -4.20923114e-01 6.63634539e-01 -5.81239998e-01
7.11793125e-01 1.83600560e-01 4.96359169e-01 7.73850262e-01
1.23925999e-01 4.91549045e-01 7.49643326e-01 -1.40799567e-01
-8.46513689e-01 -9.79366541e-01 4.01597351e-01 1.03929746e+00
4.36992854e-01 -1.13349581e+00 -8.23194027e-01 -8.54606450e-01
-1.48941144e-01 1.16598773e+00 -4.68200386e-01 -3.24073970e-01
-5.65274537e-01 -3.88723165e-01 7.66700804e-02 1.57211915e-01
3.42395395e-01 -1.27760577e+00 -5.23448229e-01 3.70725751e-01
-4.94561642e-01 -8.42940807e-01 -8.21670771e-01 -2.62076948e-02
-9.12247181e-01 -8.55344296e-01 -4.23825681e-01 -1.04607558e+00
1.15957260e+00 5.52716374e-01 1.29193020e+00 3.34570378e-01
8.13036188e-02 3.71472627e-01 -8.17702889e-01 -1.82208851e-01
-4.98478681e-01 2.90912211e-01 -3.22810970e-02 -2.22595066e-01
4.43029135e-01 -3.70528877e-01 -5.38483381e-01 -1.77919909e-01
-9.66922939e-01 3.34265046e-02 4.70815897e-01 6.14185750e-01
6.01761520e-01 3.25591922e-01 8.01419675e-01 -9.48945642e-01
1.43701947e+00 -2.69349307e-01 9.89272539e-03 5.32014072e-01
-6.96315110e-01 1.96616873e-01 5.83434403e-01 3.95006500e-02
-1.04733038e+00 -4.06917661e-01 2.91939110e-01 2.69867599e-01
2.01706454e-01 9.53216076e-01 6.68134466e-02 6.06880248e-01
4.89690602e-01 4.65806693e-01 -3.32418442e-01 -1.56704277e-01
3.91933411e-01 6.55138612e-01 5.10091424e-01 -2.64470339e-01
5.34374774e-01 1.89030901e-01 -8.15817863e-02 -1.21580994e+00
-9.81057346e-01 -7.61765480e-01 -8.25741470e-01 -4.99637663e-01
8.59864652e-01 -4.56689477e-01 -1.72247455e-01 1.29236598e-02
-1.37923038e+00 4.62964088e-01 -4.84925449e-01 4.19634312e-01
-1.70806097e-03 9.90953028e-01 -3.79661500e-01 -6.89139903e-01
-7.98076510e-01 -7.15171695e-01 9.09888983e-01 5.50338984e-01
-7.79122651e-01 -1.04454947e+00 2.50967238e-02 1.98551901e-02
1.40571088e-01 1.04198009e-01 1.09902787e+00 -9.18260157e-01
-9.87021774e-02 -2.40723565e-01 -3.18355113e-02 1.29721239e-01
6.33543193e-01 2.62194753e-01 -2.65211076e-01 -2.29009748e-01
-4.01617847e-02 1.28172442e-01 1.12629175e+00 5.92621148e-01
6.59875154e-01 -5.99568605e-01 -4.96804595e-01 -1.70114115e-01
1.12008727e+00 1.15383193e-02 4.07023400e-01 2.23033428e-01
7.48409808e-01 8.65014315e-01 5.38755476e-01 3.74282122e-01
3.22845429e-01 1.87196314e-01 1.90216988e-01 7.05347285e-02
-1.21743210e-01 -4.14682776e-01 3.31628263e-01 1.61519969e+00
1.59135133e-01 -6.20035708e-01 -6.25782549e-01 7.19611943e-01
-1.90796399e+00 -1.11879361e+00 -4.35099632e-01 2.03065348e+00
7.67759800e-01 5.16580999e-01 4.66046482e-02 2.30309889e-01
1.07713473e+00 6.99554503e-01 -1.42093495e-01 -5.14827609e-01
-3.04697931e-01 -1.19585477e-01 4.90582958e-02 5.61322629e-01
-7.86039829e-01 1.12872660e+00 6.06624746e+00 7.42173195e-01
-5.90642929e-01 -3.79729092e-01 4.25443769e-01 6.19237013e-02
-7.42069900e-01 2.68406749e-01 -7.12527931e-01 4.81019288e-01
5.62408447e-01 -9.34924245e-01 -1.32409483e-01 5.50769567e-01
4.47199792e-01 -5.38925290e-01 -7.02957273e-01 6.36625469e-01
4.92501497e-01 -1.30162084e+00 5.99088013e-01 -2.86585271e-01
7.56248355e-01 -4.66784477e-01 -4.05285120e-01 3.40659171e-02
-4.49241884e-02 -5.14099360e-01 3.11991900e-01 4.53412324e-01
1.80173591e-01 -9.34544027e-01 9.05234337e-01 2.25597814e-01
-1.37775183e+00 3.61169726e-01 -5.14553428e-01 1.72484461e-02
3.86855721e-01 9.07011509e-01 -7.01095819e-01 8.66279840e-01
3.47496718e-01 9.48890805e-01 -6.39345467e-01 8.19137573e-01
-6.25135541e-01 7.17942476e-01 -1.72073208e-02 -6.10366166e-01
3.86858910e-01 -5.40706933e-01 1.10968804e+00 1.56730342e+00
2.38799900e-01 9.73303691e-02 3.04155380e-01 4.67154741e-01
-2.37434030e-01 6.97004735e-01 -5.99734247e-01 -2.50928938e-01
7.02858269e-01 1.43253624e+00 -1.32444429e+00 -6.81696713e-01
-2.89277196e-01 1.01121461e+00 1.50967360e-01 2.36709595e-01
-2.79681116e-01 -7.58264720e-01 7.83228874e-02 -7.68949911e-02
1.65127456e-01 -1.60687223e-01 -2.64593452e-01 -9.78969872e-01
1.30533073e-02 -4.75813925e-01 5.97149968e-01 -6.38614058e-01
-1.17879701e+00 8.07871640e-01 3.40832286e-02 -1.08597767e+00
-1.84799984e-01 1.42034456e-01 -9.86350775e-01 5.23585260e-01
-1.23447311e+00 -4.39298779e-01 -1.90326452e-01 1.43736362e-01
9.86392081e-01 -1.18210159e-01 6.68905199e-01 -2.91732073e-01
-6.60389662e-01 1.59795985e-01 -1.52479559e-01 -9.61126462e-02
4.73608375e-01 -1.46312356e+00 4.75656778e-01 1.22789633e+00
3.46623391e-01 8.15580010e-01 9.53156114e-01 -1.09494174e+00
-9.95514035e-01 -8.96789134e-01 1.77761626e+00 1.25308454e-01
6.51423573e-01 3.30713876e-02 -1.07197356e+00 1.81183070e-01
8.22914660e-01 -6.60376728e-01 6.50267005e-01 -6.92666620e-02
8.18708166e-02 2.92820502e-02 -7.53848433e-01 8.80006254e-01
1.09924722e+00 -1.39884040e-01 -1.24139190e+00 4.36722100e-01
8.76643777e-01 8.08368400e-02 -5.11467993e-01 7.64550120e-02
5.66796176e-02 -8.54114532e-01 5.12554348e-01 -5.25461733e-01
6.66280091e-01 -2.13548183e-01 2.06870511e-01 -1.74082196e+00
-6.02922738e-01 -7.77325153e-01 -1.18881211e-01 1.45647490e+00
5.38590252e-01 -3.07272553e-01 5.34977138e-01 1.43152907e-01
-2.59127349e-01 -4.53520447e-01 -5.02504110e-01 -5.48593462e-01
-4.83634830e-01 2.38464430e-01 2.30532005e-01 7.56603539e-01
7.09342837e-01 9.23183322e-01 -6.30192310e-02 -1.48643747e-01
6.41784966e-01 3.24906647e-01 2.52168924e-01 -1.39780855e+00
2.81934977e-01 -7.36568034e-01 -2.57837534e-01 -8.74694169e-01
3.11773956e-01 -1.11592495e+00 1.79391250e-01 -2.22531009e+00
5.04344463e-01 3.89137655e-01 -1.14729531e-01 1.58311695e-01
-7.09370792e-01 -2.02658549e-01 1.65242907e-02 3.31500053e-01
-9.86173511e-01 5.44907749e-01 1.18073082e+00 -2.03582108e-01
-4.84959126e-01 -2.76944667e-01 -8.27019393e-01 8.46241415e-01
9.25866187e-01 -5.60006201e-01 -5.50817549e-01 -2.17305660e-01
3.24238926e-01 -1.82694376e-01 -2.80453712e-01 -6.58108115e-01
5.74963987e-01 -1.17540568e-01 -9.84157473e-02 -8.58096063e-01
-3.94255489e-01 -4.70649898e-01 -2.89122045e-01 5.78711092e-01
-6.84827626e-01 3.27367276e-01 -1.46848246e-01 6.54321671e-01
-4.67999488e-01 -6.08565331e-01 4.09613878e-01 -1.35005042e-01
-8.68554831e-01 -1.62589639e-01 -5.33485532e-01 2.13744372e-01
7.83724189e-01 -4.27298397e-01 -2.23798290e-01 -5.46273589e-01
-4.49289203e-01 3.35359097e-01 2.77627170e-01 4.74947959e-01
9.24791098e-01 -1.11620796e+00 -1.08811510e+00 -3.29317808e-01
2.37875625e-01 2.46314649e-02 6.14886358e-02 4.85460013e-01
-3.91186625e-01 3.31126660e-01 1.04017839e-01 -4.76629704e-01
-1.62204909e+00 3.77898514e-01 -3.27098638e-01 -4.31183070e-01
-9.04729009e-01 6.83308423e-01 -1.13307200e-02 2.24807084e-01
1.08557336e-01 -1.58762798e-01 -7.04069495e-01 5.29139340e-01
5.28956354e-01 4.34970379e-01 1.60322804e-02 -7.39623904e-01
-4.48707908e-01 5.64104557e-01 -3.28212410e-01 -1.38796493e-03
1.33617437e+00 -4.08904284e-01 -6.18544936e-01 3.55274856e-01
1.05185485e+00 3.45392734e-01 -5.76047421e-01 -2.55050540e-01
5.48463821e-01 -2.44601429e-01 1.24135599e-01 -3.17046970e-01
-8.07139099e-01 1.96446165e-01 -4.03023601e-01 8.55916917e-01
1.11921799e+00 1.97670281e-01 7.37439811e-01 5.17182946e-01
-4.45568748e-02 -1.32995176e+00 3.32041889e-01 6.28813922e-01
1.01330662e+00 -8.37404609e-01 3.72034758e-01 -7.45471358e-01
-7.83011436e-01 1.33701277e+00 3.36476684e-01 -1.81368068e-01
2.28512779e-01 2.29611918e-02 -2.92463928e-01 -5.26091933e-01
-5.73603928e-01 -2.80724972e-01 5.55682003e-01 2.34190211e-01
6.67288482e-01 -1.44694194e-01 -1.29555082e+00 2.74651080e-01
-3.76306027e-01 -5.10371864e-01 6.62563682e-01 9.22040224e-01
-9.81403530e-01 -8.67095649e-01 -1.22964852e-01 6.69018447e-01
-2.62130708e-01 -3.25393111e-01 -8.62925768e-01 4.05585855e-01
-4.77230728e-01 1.26684356e+00 2.48472348e-01 -2.46105924e-01
4.36807781e-01 4.99162152e-02 1.94562674e-01 -9.90984797e-01
-7.19372690e-01 1.02343999e-01 3.15124184e-01 -1.76032737e-01
-5.56414485e-01 -5.69608033e-01 -1.64695942e+00 -1.28439024e-01
-3.71406019e-01 6.64018273e-01 4.14300859e-01 8.54516089e-01
4.54307020e-01 8.96560073e-01 7.81300128e-01 -5.41519642e-01
-3.07416379e-01 -1.04965580e+00 -5.76823413e-01 6.53509200e-01
-6.83848262e-02 -1.50860116e-01 -6.02904677e-01 1.50927141e-01] | [12.493996620178223, 9.540916442871094] |
9993fc6b-0ac9-4af2-81a4-baecb4f49e08 | hybrid-multimodal-fusion-for-humor-detection | 2209.11949 | null | https://arxiv.org/abs/2209.11949v1 | https://arxiv.org/pdf/2209.11949v1.pdf | Hybrid Multimodal Fusion for Humor Detection | In this paper, we present our solution to the MuSe-Humor sub-challenge of the Multimodal Emotional Challenge (MuSe) 2022. The goal of the MuSe-Humor sub-challenge is to detect humor and calculate AUC from audiovisual recordings of German football Bundesliga press conferences. It is annotated for humor displayed by the coaches. For this sub-challenge, we first build a discriminant model using the transformer module and BiLSTM module, and then propose a hybrid fusion strategy to use the prediction results of each modality to improve the performance of the model. Our experiments demonstrate the effectiveness of our proposed model and hybrid fusion strategy on multimodal fusion, and the AUC of our proposed model on the test set is 0.8972. | ['Meng Wang', 'Xiao Sun', 'Yunwei Shi', 'Yasi Peng', 'Yu Feng', 'Mingzheng Li', 'Jingwei Liu', 'Weifeng Liu', 'Haojie Xu'] | 2022-09-24 | null | null | null | null | ['humor-detection'] | ['natural-language-processing'] | [-4.25275683e-01 -2.85359234e-01 2.00324684e-01 1.52051389e-01
-1.03730834e+00 -3.28560561e-01 3.96061093e-01 -1.00258484e-01
-3.74799639e-01 4.53580737e-01 6.92889929e-01 3.21683347e-01
2.42376477e-01 -1.80080190e-01 -3.18168104e-01 -2.99760073e-01
2.53943324e-01 -8.78280774e-02 1.34861737e-01 -4.33966190e-01
2.66344488e-01 -2.28708908e-01 -1.44692063e+00 8.42744231e-01
6.71430707e-01 1.05122066e+00 -2.54995465e-01 9.48716283e-01
3.94144416e-01 1.42277312e+00 -6.41884565e-01 -9.49575663e-01
-1.44737422e-01 -8.13983738e-01 -6.81341708e-01 -2.84621298e-01
4.37655866e-01 -1.93524465e-01 -6.15464747e-01 1.01380479e+00
9.94298756e-01 5.00277936e-01 5.19742072e-01 -1.30957508e+00
-1.16059020e-01 1.06439245e+00 -4.24860626e-01 4.29087654e-02
7.20157743e-01 2.21412063e-01 9.57792103e-01 -1.20452690e+00
5.73324800e-01 1.13013315e+00 9.16049540e-01 5.45779586e-01
-7.79710293e-01 -6.37827814e-01 -5.71575701e-01 8.88088644e-01
-1.27599204e+00 -5.14887869e-01 1.09461093e+00 -5.84068477e-01
6.83513224e-01 2.25625724e-01 5.16712546e-01 1.31205714e+00
1.15761548e-01 1.11651671e+00 1.07034993e+00 -2.50053078e-01
-6.81556389e-02 1.22140117e-01 1.98461711e-01 6.09849930e-01
-5.36887646e-01 -2.23212138e-01 -1.16274726e+00 -9.86111984e-02
1.94189757e-01 -3.79133016e-01 -3.15527052e-01 3.77373695e-01
-1.16948056e+00 7.27396607e-01 2.65121311e-01 3.44848335e-01
-3.44200343e-01 1.28992200e-01 7.60912538e-01 3.29230189e-01
2.42667019e-01 3.35619628e-01 3.82612526e-01 -5.40726125e-01
-1.16074002e+00 5.01303196e-01 7.49198735e-01 3.62480462e-01
6.09966367e-02 -3.41274478e-02 -4.85643417e-01 1.13465941e+00
1.71434253e-01 3.42616379e-01 6.61471903e-01 -1.17954981e+00
4.16824847e-01 4.51111078e-01 -1.57992184e-01 -1.17501271e+00
-7.19741881e-01 -3.94009531e-01 -5.96083522e-01 -1.07675292e-01
1.20438673e-01 -3.31084073e-01 -3.07052106e-01 1.79299629e+00
1.29343392e-02 2.21588403e-01 -7.86711574e-02 1.18936181e+00
1.42816854e+00 5.57805598e-01 6.11828268e-02 -2.67164707e-01
1.09926665e+00 -1.13730788e+00 -1.01745546e+00 1.96906880e-01
6.36955738e-01 -8.12895656e-01 8.92262638e-01 6.92062557e-01
-1.42088938e+00 -7.12029696e-01 -1.25845373e+00 -1.66941166e-01
5.30141480e-02 2.44798318e-01 -1.84469998e-01 2.52266884e-01
-5.33648610e-01 4.34673160e-01 -2.26309940e-01 -9.89087000e-02
-1.44153044e-01 -2.58777142e-02 -3.69203091e-01 2.73302108e-01
-1.69739461e+00 1.08577347e+00 5.42939305e-01 -1.70844961e-02
-8.10721278e-01 -4.43313390e-01 -7.33919919e-01 2.95776147e-02
-3.93200889e-02 -5.81952751e-01 1.31817579e+00 -8.61444652e-01
-1.18953216e+00 7.08035052e-01 8.51960033e-02 -6.90028131e-01
4.56563592e-01 -4.61134166e-01 -4.42517519e-01 2.54771441e-01
-2.16453329e-01 4.05971438e-01 6.18124545e-01 -1.23809886e+00
-6.84131742e-01 -1.00486994e-01 -2.51276284e-01 3.47256958e-01
-6.80691361e-01 2.58007377e-01 -3.27209175e-01 -4.32239830e-01
-2.51681030e-01 -7.97046125e-01 3.57184500e-01 -1.06145453e+00
-2.07028359e-01 -2.24638090e-01 6.49361730e-01 -1.37431335e+00
1.98067331e+00 -2.45785522e+00 4.94123667e-01 -7.04465434e-02
4.86336797e-01 -1.45619106e-03 -1.20604873e-01 4.83465850e-01
1.36669993e-01 -2.69328922e-01 4.04871479e-02 -4.99606103e-01
2.46764988e-01 -4.12444651e-01 -3.01813453e-01 1.59095943e-01
-3.63137662e-01 8.73034298e-01 -7.39432156e-01 -6.89406574e-01
1.87502816e-01 4.14952219e-01 -3.70003134e-01 3.93211454e-01
9.76234600e-02 8.24014992e-02 3.38868737e-01 5.52512705e-01
2.97323287e-01 2.23079279e-01 2.15482086e-01 -3.85216981e-01
1.46872163e-01 2.90086955e-01 -9.01455641e-01 1.50402093e+00
-1.96405903e-01 8.35701466e-01 -2.01657563e-01 -2.13292435e-01
1.03687394e+00 5.21252275e-01 5.15596867e-01 -7.61585534e-01
4.39614624e-01 2.09452242e-01 4.20877896e-02 -7.90402889e-01
7.32527792e-01 -4.56305057e-01 -4.32170630e-01 5.12811430e-02
5.76557040e-01 8.22855607e-02 1.56110272e-01 3.41505915e-01
1.08166742e+00 3.21454974e-03 2.25394070e-01 4.08616543e-01
6.88174784e-01 -2.43436843e-01 4.14782405e-01 4.98500884e-01
-4.54031467e-01 7.75289059e-01 6.08764827e-01 -8.02274197e-02
-8.87535989e-01 -6.92500055e-01 2.18604743e-01 1.23182309e+00
-1.34755403e-01 -7.58193552e-01 -7.35961020e-01 -5.60860574e-01
-2.18088031e-01 9.20486033e-01 -5.97139537e-01 -3.99696320e-01
-3.02942216e-01 -5.39016306e-01 9.55043077e-01 3.77772450e-01
6.10915422e-01 -1.06284106e+00 -6.65213823e-01 1.45955414e-01
-1.16268921e+00 -1.31364477e+00 -3.08252960e-01 -4.46543321e-02
-2.82658130e-01 -9.45880055e-01 -5.62031031e-01 -4.67276692e-01
-4.58395183e-01 3.79874185e-02 8.81301880e-01 -1.15381911e-01
2.20948726e-01 5.32532096e-01 -5.60079396e-01 -1.51611075e-01
-3.53291392e-01 -5.12397699e-02 3.05566099e-02 1.22510970e-01
3.74787688e-01 -4.15857971e-01 -4.27983582e-01 2.89404094e-01
-5.02332032e-01 1.09686516e-01 1.93970129e-01 1.03097713e+00
1.18983403e-01 -1.88757733e-01 6.64184928e-01 -2.57534802e-01
1.07572997e+00 -6.69711947e-01 1.44069299e-01 3.14684361e-02
-1.88662499e-01 -6.18543327e-01 2.32050359e-01 -4.92475033e-01
-9.11061883e-01 -2.88161002e-02 -4.36307527e-02 -8.43900323e-01
2.43675426e-01 8.82329166e-01 -9.08192098e-02 1.05592564e-01
7.11786151e-01 4.12884429e-02 -2.83944935e-01 -3.68369699e-01
2.21530214e-01 1.01336277e+00 1.33797503e+00 -4.89243537e-01
2.82010913e-01 -2.54675210e-01 -1.33532332e-02 -5.38559020e-01
-9.76483047e-01 -5.68675756e-01 -3.95132691e-01 -1.12879860e+00
8.90293241e-01 -1.08401251e+00 -1.15759587e+00 5.32922268e-01
-1.41542804e+00 -1.95290055e-03 8.38653445e-02 5.09671688e-01
-5.94106376e-01 6.36144698e-01 -7.67464876e-01 -1.26801527e+00
-6.38232648e-01 -7.02948630e-01 5.51424623e-01 2.67760634e-01
-5.48023343e-01 -7.11413622e-01 5.87443829e-01 1.00916672e+00
-5.47701865e-02 5.12205422e-01 5.77458560e-01 -9.75613892e-01
4.42009181e-01 -4.11430895e-01 -3.72806303e-02 6.58292115e-01
-6.25195563e-01 -2.03778669e-01 -1.15783370e+00 9.48556215e-02
1.00886263e-01 -8.34722340e-01 1.27584100e+00 1.26149254e-02
7.47596443e-01 -1.43304572e-01 3.59155834e-01 1.06183857e-01
9.82941091e-01 -1.15898371e-01 7.22598195e-01 3.34419012e-01
5.61795115e-01 6.15489364e-01 7.03820169e-01 8.07801008e-01
6.27759457e-01 8.20626318e-01 4.07739490e-01 3.74023646e-01
-2.80003935e-01 -4.90435451e-01 9.18318093e-01 1.22673535e+00
-2.99383879e-01 -1.04942583e-01 -1.01961970e+00 5.18261969e-01
-2.25474834e+00 -1.31317854e+00 -5.38501978e-01 2.15201068e+00
7.81314194e-01 -1.10707842e-02 6.49319112e-01 3.75992060e-01
5.94877243e-01 3.01940963e-02 1.96685463e-01 -5.92174947e-01
-4.69308227e-01 -7.99925178e-02 -2.51548439e-01 5.53179383e-01
-1.04706109e+00 7.96623707e-01 6.15862942e+00 1.04610372e+00
-8.71469796e-01 4.46185827e-01 -1.44840525e-02 -6.35764539e-01
-7.54456744e-02 -3.26213628e-01 -3.92260224e-01 5.99853814e-01
1.27302229e+00 -1.18388385e-01 5.27683198e-01 4.91625398e-01
1.28020689e-01 -1.98107883e-01 -8.55974793e-01 1.31962967e+00
7.78255343e-01 -8.36349010e-01 -1.53979734e-01 -2.56061286e-01
5.74621856e-01 -1.08131632e-01 1.27459556e-01 8.34174216e-01
-1.93015203e-01 -1.07181811e+00 9.00907218e-01 8.39595437e-01
2.02260643e-01 -9.93130565e-01 1.24474490e+00 3.56891930e-01
-6.96449459e-01 -2.69184917e-01 -6.70842379e-02 -2.59541012e-02
2.70495534e-01 3.10738415e-01 -6.89493179e-01 4.56172407e-01
5.36901593e-01 4.05586183e-01 -7.99220979e-01 1.32693565e+00
-1.56926975e-01 8.17757964e-01 7.26964101e-02 3.65606397e-02
4.64862473e-02 3.75811517e-01 8.17081034e-01 1.61291051e+00
4.09876645e-01 1.50110181e-02 -4.58948268e-03 6.24130428e-01
-3.14847589e-01 3.81111503e-01 -3.57211202e-01 -1.52863950e-01
3.57372522e-01 1.48414135e+00 1.17019452e-01 -1.73914030e-01
-1.19568244e-01 8.80154490e-01 3.36522728e-01 6.29077852e-02
-1.13292587e+00 -3.46539706e-01 -1.90265104e-01 -6.46277592e-02
-9.21672508e-02 -2.54162345e-02 -4.86239702e-01 -1.12542832e+00
-3.04332554e-01 -1.18327725e+00 7.57164955e-01 -1.13659644e+00
-1.06964886e+00 5.53607166e-01 -1.20783024e-01 -1.25490081e+00
-4.08067495e-01 -3.20595056e-01 -6.88128889e-01 5.57751834e-01
-6.83545470e-01 -1.40288377e+00 -4.18534309e-01 6.77600741e-01
3.67431760e-01 -4.04034764e-01 4.59635943e-01 3.68092567e-01
-3.76504093e-01 7.89283037e-01 -3.17749083e-01 1.62902892e-01
1.05951548e+00 -1.04667044e+00 -5.42944193e-01 6.52740598e-01
2.43396729e-01 1.28994673e-01 1.15233469e+00 -5.39274633e-01
-1.12205589e+00 -5.44229150e-01 1.08234334e+00 -6.59898520e-01
8.45200062e-01 1.14332967e-01 -8.23584437e-01 3.13701123e-01
6.32803857e-01 -7.64006257e-01 1.00534749e+00 3.72180402e-01
-6.00052595e-01 1.39221013e-01 -8.10468674e-01 3.12788814e-01
2.96502203e-01 -5.58337510e-01 -8.80092382e-01 -8.32582414e-02
3.88349295e-01 -4.28958684e-01 -1.01568127e+00 7.31208920e-01
9.39360261e-01 -1.03093398e+00 4.67984587e-01 -6.34543061e-01
1.20408535e+00 -1.32646531e-01 -7.17482209e-01 -1.22541630e+00
-2.51494050e-01 -3.94626588e-01 -4.56920415e-01 1.41854990e+00
1.74723089e-01 3.18704665e-01 1.05046943e-01 2.05627397e-01
-1.15402021e-01 -4.34069008e-01 -1.17811775e+00 -3.31163615e-01
-1.93525553e-02 -7.41891980e-01 -7.23776296e-02 9.77443755e-01
8.78062069e-01 7.70564973e-01 -1.15844703e+00 -3.27476144e-01
7.01814234e-01 -1.87935621e-01 9.53859806e-01 -9.25238669e-01
-4.24361348e-01 -5.45310855e-01 -5.96440136e-01 -3.33605766e-01
2.73023605e-01 -8.63417208e-01 1.98183302e-02 -1.14506853e+00
9.63987887e-01 8.86664867e-01 -8.07018876e-01 5.57956338e-01
-1.31794751e-01 5.93764782e-01 8.52579713e-01 8.36804360e-02
-1.15540457e+00 8.32802415e-01 7.92623341e-01 -2.49282986e-01
-2.38437653e-01 -3.19896996e-01 -4.64816093e-01 6.49368763e-01
5.70058703e-01 -2.57781148e-01 8.01464692e-02 8.95614401e-02
4.90950435e-01 5.19726515e-01 8.08622777e-01 -1.32381344e+00
1.71321496e-01 9.07115489e-02 3.02048385e-01 -9.10571218e-01
6.63180351e-01 -3.25877637e-01 1.88284367e-01 2.46297866e-01
-4.01823550e-01 -2.84869850e-01 2.86302447e-01 1.37554616e-01
-5.85092604e-01 -2.94623345e-01 7.44356215e-01 3.09218317e-01
-2.97465175e-01 -5.00053823e-01 -3.31746101e-01 -6.72083870e-02
6.73194230e-01 2.11639956e-01 -5.18316567e-01 -8.91960025e-01
-1.03802288e+00 3.92033905e-01 6.67710304e-02 5.35535753e-01
8.68387997e-01 -1.78969705e+00 -1.10707009e+00 -5.32941639e-01
1.32637292e-01 -1.07581937e+00 7.04843223e-01 1.40041792e+00
-1.78052504e-02 5.35283014e-02 -4.98651296e-01 -2.50377148e-01
-1.69905841e+00 3.96825135e-01 4.55592364e-01 -2.21264139e-01
-2.70831615e-01 4.86937791e-01 -1.70922607e-01 -2.12419778e-01
2.39124700e-01 6.40052319e-01 -4.73383605e-01 3.78421336e-01
5.83441436e-01 8.47488523e-01 -7.32470080e-02 -9.57764387e-01
-2.74577588e-01 1.79929677e-02 2.32530519e-01 -7.63345122e-01
1.07497990e+00 -2.54940726e-02 -3.10359418e-01 9.36353683e-01
9.07022834e-01 1.02097400e-01 -3.97106767e-01 -1.25031874e-01
-1.78550631e-01 -1.53676674e-01 2.68545151e-01 -1.11876333e+00
-6.51980042e-01 1.09142983e+00 7.15517461e-01 -3.54964398e-02
1.20416605e+00 -2.02173069e-01 1.03371739e+00 1.92489207e-01
-6.99694902e-02 -1.35865271e+00 3.52887809e-01 8.53755236e-01
1.24096203e+00 -9.33887243e-01 -3.33799809e-01 2.37923339e-01
-1.27960646e+00 1.28124321e+00 7.97640026e-01 6.35639504e-02
2.80667961e-01 -7.26690814e-02 -6.32705912e-02 -8.17763060e-02
-1.21925950e+00 -1.90153554e-01 8.37478817e-01 1.22627236e-01
5.87304711e-01 8.73355046e-02 -9.72680807e-01 1.67595601e+00
-3.34943801e-01 1.45177513e-01 5.58829725e-01 2.13756084e-01
-8.09463143e-01 -4.29921329e-01 -5.81750274e-01 -2.98767090e-02
-4.56321508e-01 6.86369464e-03 -1.04842174e+00 4.10742313e-01
2.57575303e-01 1.13108563e+00 -5.32626271e-01 -1.58095992e+00
3.32544744e-01 6.31840169e-01 5.23167133e-01 1.18773930e-01
-1.09020436e+00 4.17081177e-01 4.28932250e-01 -4.14339602e-01
-3.29847813e-01 -5.89388728e-01 -1.05484486e+00 -5.56772530e-01
-2.27522440e-02 2.32572675e-01 5.39240181e-01 8.81819010e-01
-1.67400558e-02 3.95072520e-01 9.13312912e-01 -7.62167275e-01
-4.89062577e-01 -1.44591010e+00 -4.99934614e-01 5.56735456e-01
4.46611270e-02 -7.07118869e-01 -4.48465228e-01 4.77459095e-02] | [13.223134994506836, 5.124037742614746] |
6a4a8d54-ab4b-4f17-969a-e20bfb9e43d3 | freehand-ultrasound-image-simulation-with | 1707.05392 | null | http://arxiv.org/abs/1707.05392v1 | http://arxiv.org/pdf/1707.05392v1.pdf | Freehand Ultrasound Image Simulation with Spatially-Conditioned Generative Adversarial Networks | Sonography synthesis has a wide range of applications, including medical
procedure simulation, clinical training and multimodality image registration.
In this paper, we propose a machine learning approach to simulate ultrasound
images at given 3D spatial locations (relative to the patient anatomy), based
on conditional generative adversarial networks (GANs). In particular, we
introduce a novel neural network architecture that can sample anatomically
accurate images conditionally on spatial position of the (real or mock)
freehand ultrasound probe. To ensure an effective and efficient spatial
information assimilation, the proposed spatially-conditioned GANs take
calibrated pixel coordinates in global physical space as conditioning input,
and utilise residual network units and shortcuts of conditioning data in the
GANs' discriminator and generator, respectively. Using optically tracked B-mode
ultrasound images, acquired by an experienced sonographer on a fetus phantom,
we demonstrate the feasibility of the proposed method by two sets of
quantitative results: distances were calculated between corresponding
anatomical landmarks identified in the held-out ultrasound images and the
simulated data at the same locations unseen to the networks; a usability study
was carried out to distinguish the simulated data from the real images. In
summary, we present what we believe are state-of-the-art visually realistic
ultrasound images, simulated by the proposed GAN architecture that is stable to
train and capable of generating plausibly diverse image samples. | ['Li-Lin Lee', 'Yipeng Hu', 'Tom Vercauteren', 'Weidi Xie', 'Eli Gibson', 'Dean C. Barratt', 'J. Alison Noble'] | 2017-07-17 | null | null | null | null | ['medical-procedure'] | ['medical'] | [ 5.44344246e-01 5.95964134e-01 4.45155948e-01 -1.06225386e-01
-8.95765305e-01 -6.47600174e-01 3.10662925e-01 -4.49495614e-01
-3.19577217e-01 8.04979742e-01 2.36808900e-02 -3.83616954e-01
-8.68156701e-02 -6.93337679e-01 -1.01748884e+00 -9.76284504e-01
-3.65979224e-01 3.91478151e-01 -8.35165158e-02 2.61865016e-02
-1.64514288e-01 6.45587742e-01 -1.04227448e+00 -8.06304812e-02
9.00772095e-01 8.90552342e-01 1.68288991e-01 1.07184458e+00
5.06223977e-01 7.16430664e-01 -8.09650302e-01 -3.83669466e-01
4.24341232e-01 -1.04197693e+00 -5.13019323e-01 -2.57175475e-01
4.60788608e-01 -4.69604105e-01 -4.28596824e-01 8.35446775e-01
9.58460510e-01 1.48176834e-01 8.99298310e-01 -7.90395200e-01
-7.38024950e-01 4.29953486e-01 -3.64079714e-01 3.39392237e-02
3.57518464e-01 2.98427224e-01 4.14296575e-02 -4.00086939e-01
8.20126176e-01 5.92777848e-01 8.67915213e-01 9.01592433e-01
-1.08690572e+00 -6.57371104e-01 -6.72448695e-01 -3.80306721e-01
-1.27205026e+00 -2.50676066e-01 7.45403588e-01 -6.51912868e-01
2.18351454e-01 6.33113205e-01 7.23201454e-01 1.20098543e+00
7.81634331e-01 2.35353947e-01 1.26717806e+00 -5.05121708e-01
3.11615944e-01 1.44378349e-01 -8.33211482e-01 9.76681590e-01
5.20688817e-02 5.53193092e-01 1.06357727e-02 -1.25697285e-01
1.52416348e+00 -2.89055735e-01 -8.76553833e-01 -5.57604373e-01
-1.09720075e+00 7.84033060e-01 6.41461432e-01 5.33525050e-01
-7.56909490e-01 3.92709553e-01 -3.30386609e-02 -8.69469419e-02
-2.89452765e-02 7.40040779e-01 1.83158815e-01 -7.75286481e-02
-7.84029305e-01 -1.75705045e-01 8.38686466e-01 8.03780615e-01
1.46774352e-01 3.87624472e-01 -3.36566746e-01 3.73338521e-01
2.16750517e-01 6.53363228e-01 7.56677270e-01 -8.20569158e-01
1.38170749e-01 -7.69365802e-02 1.28374860e-01 -1.14622629e+00
-1.96454912e-01 -5.67476273e-01 -1.11431503e+00 2.71345496e-01
1.67321891e-01 -5.37108064e-01 -1.43184626e+00 1.69978905e+00
4.40340966e-01 9.38396931e-01 4.55250382e-01 1.13357389e+00
1.18612015e+00 6.09394372e-01 7.57932588e-02 -3.00984383e-01
9.42234874e-01 -5.22723019e-01 -8.04271638e-01 5.41728437e-02
3.86793435e-01 -6.85431182e-01 7.97668397e-01 1.93227261e-01
-1.59381640e+00 -6.71537101e-01 -1.13656843e+00 5.20824492e-01
6.26980886e-02 -3.42948660e-02 4.15498912e-01 9.97707546e-01
-1.16038871e+00 5.47306120e-01 -1.12718296e+00 3.66632082e-02
2.82235712e-01 1.57435566e-01 -5.53001702e-01 6.88012689e-02
-1.29675722e+00 8.79298031e-01 1.97589651e-01 4.92596179e-01
-1.21528566e+00 -9.65057135e-01 -1.18925011e+00 -6.43795654e-02
-1.96147617e-02 -8.75251651e-01 1.05573511e+00 -9.94237721e-01
-1.76729071e+00 8.68403256e-01 4.12423104e-01 -3.45213056e-01
7.43201554e-01 1.73440009e-01 -5.01886070e-01 4.42277819e-01
-1.20284483e-01 4.65321362e-01 7.89649308e-01 -1.77603328e+00
1.75808787e-01 2.27801185e-02 -2.17045635e-01 1.56218812e-01
3.67192835e-01 -2.72379279e-01 -3.14909786e-01 -8.72128010e-01
2.43675366e-01 -8.92330348e-01 -3.74515951e-01 -7.82884806e-02
-3.65297854e-01 8.64135861e-01 2.65477270e-01 -1.04795146e+00
9.08209860e-01 -2.09030294e+00 1.77143887e-01 6.81360543e-01
-4.36437724e-04 3.85056376e-01 -6.80911988e-02 1.84257761e-01
-3.36245924e-01 5.58025539e-02 -5.31412244e-01 1.04633227e-01
-3.60492319e-01 1.96551710e-01 -1.11816125e-02 6.93834960e-01
-2.88967788e-01 1.29535449e+00 -1.13887119e+00 -4.42385137e-01
4.75372821e-01 7.08101153e-01 -2.95254409e-01 7.91702926e-01
3.84852707e-01 1.25203991e+00 -3.98645252e-01 2.15255082e-01
7.50261962e-01 2.28340387e-01 1.27907082e-01 -3.93079787e-01
2.03236073e-01 -3.35665762e-01 -8.86176527e-01 1.83008528e+00
-7.67976820e-01 2.69464552e-01 2.36990556e-01 -7.90564477e-01
7.08910108e-01 7.68060148e-01 3.50438535e-01 -8.20847750e-01
3.51637870e-01 1.83348611e-01 2.48328120e-01 -8.40662837e-01
3.65099008e-03 -5.17874658e-01 1.36409458e-02 4.55340296e-01
1.82596982e-01 -6.96954250e-01 -3.42419237e-01 -8.05139989e-02
8.20396543e-01 1.35641649e-01 9.81075615e-02 -1.22126095e-01
4.90251452e-01 -4.01080668e-01 1.52016684e-01 1.01865566e+00
1.09625436e-01 1.23797834e+00 3.84385139e-01 -2.27868721e-01
-9.85405862e-01 -1.21673250e+00 -9.65933055e-02 1.10638648e-01
2.25298896e-01 4.00859624e-01 -1.02468121e+00 -6.19217634e-01
-3.65796089e-01 8.08992684e-01 -1.04149902e+00 -2.67214507e-01
-7.70781457e-01 -4.29695398e-01 7.50136614e-01 4.63611633e-01
3.36600095e-01 -1.34415877e+00 -8.38848472e-01 2.82277822e-01
4.40766215e-02 -8.55666697e-01 -2.29142666e-01 -4.42167297e-02
-5.78485787e-01 -9.77437317e-01 -1.23899567e+00 -8.57995868e-01
1.00997543e+00 -5.50757229e-01 1.12117362e+00 -5.43301329e-02
-5.00669599e-01 5.83569825e-01 -3.53579879e-01 -2.88371772e-01
-1.13344288e+00 -3.88114572e-01 -2.47526616e-01 -7.20714107e-02
-7.53469169e-01 -7.22379208e-01 -1.18469322e+00 2.44323954e-01
-1.16898131e+00 8.49637985e-02 7.16359377e-01 1.07932186e+00
4.38349724e-01 -2.82158166e-01 3.65400881e-01 -1.06238019e+00
6.41779423e-01 -4.27834332e-01 -4.88051921e-01 3.31798792e-01
1.38745243e-02 -1.58520386e-01 5.62775433e-01 -3.44441861e-01
-1.21167743e+00 -9.21166018e-02 -5.43697417e-01 -5.63060403e-01
-2.33890146e-01 4.94505495e-01 1.20724387e-01 -5.11307955e-01
9.36412990e-01 5.34317136e-01 2.66445190e-01 9.87445042e-02
3.39960605e-01 4.28913802e-01 9.02474642e-01 -5.12830198e-01
6.67494833e-01 3.26487541e-01 1.82401612e-01 -5.33208013e-01
-6.17569908e-02 2.37676919e-01 -3.66202474e-01 -3.96964431e-01
9.25176620e-01 -5.11677742e-01 -6.64213359e-01 5.01979172e-01
-9.24327910e-01 -5.89676380e-01 -5.09198308e-01 7.74466276e-01
-7.57539749e-01 3.71987343e-01 -7.41286218e-01 -5.19103765e-01
-5.04808426e-01 -1.38815129e+00 8.80460918e-01 3.46684486e-01
3.51991393e-02 -1.21574020e+00 1.36753008e-01 4.07745540e-02
6.82724535e-01 1.04730797e+00 8.91468883e-01 -2.11178288e-01
-4.38498974e-01 -3.70203525e-01 2.28179067e-01 5.62933445e-01
3.44897479e-01 -2.36934662e-01 -1.00588238e+00 -5.21627784e-01
2.52918124e-01 -1.39684483e-01 2.70780921e-01 1.01853573e+00
1.12201583e+00 -2.55137622e-01 -2.75789708e-01 1.02642894e+00
1.44026196e+00 6.50768459e-01 1.01685870e+00 -4.28372830e-01
6.34218156e-01 2.48773322e-01 2.87820578e-01 8.28469843e-02
-3.29205334e-01 2.94906527e-01 5.44829309e-01 -7.24701703e-01
-4.03992265e-01 -4.08706158e-01 -3.61815274e-01 6.24071002e-01
-3.14529628e-01 -5.15701532e-01 -6.59732461e-01 3.21878850e-01
-1.17588115e+00 -5.62981784e-01 3.60752672e-01 2.34493756e+00
6.21276975e-01 -2.41699889e-01 -4.54102665e-01 -2.51217335e-01
6.65348887e-01 -1.27313972e-01 -2.37031356e-01 -3.72270405e-01
2.32570559e-01 8.69455814e-01 5.24245977e-01 6.31262183e-01
-8.82939339e-01 3.73342752e-01 6.38420582e+00 5.79612315e-01
-1.62032259e+00 3.15435916e-01 8.36761057e-01 2.77183771e-01
-4.68489558e-01 -4.95139569e-01 3.56149703e-01 4.91836876e-01
7.26398945e-01 3.06664973e-01 1.29302531e-01 4.66658324e-01
7.70941600e-02 -1.85252950e-01 -9.54044938e-01 8.30566227e-01
1.32434517e-01 -1.51933229e+00 7.34539144e-03 -2.32431278e-01
1.04918420e+00 -5.20714343e-01 3.63326520e-01 -6.78654537e-02
4.92302105e-02 -1.53548801e+00 4.28208292e-01 8.15690160e-01
1.52966225e+00 -7.23462284e-01 9.85243857e-01 2.68244326e-01
-5.19431710e-01 4.99051511e-01 1.59646884e-01 5.63548028e-01
3.53727520e-01 2.60762665e-02 -1.17358327e+00 7.19372213e-01
3.01864058e-01 -6.82457443e-03 -1.38309002e-01 1.00764382e+00
-2.78956205e-01 6.29339993e-01 -1.20171949e-01 2.19944671e-01
3.70319247e-01 -1.48278385e-01 4.69205499e-01 1.07170463e+00
6.56148911e-01 4.59116071e-01 -4.44624126e-01 9.66002285e-01
6.82452768e-02 5.09100221e-03 -6.59822166e-01 3.44320297e-01
1.55745104e-01 1.18547654e+00 -8.21813047e-01 -1.26147255e-01
-8.24567396e-03 9.76581931e-01 -3.17934990e-01 6.82696164e-01
-1.00596154e+00 -3.77943337e-01 -6.43565059e-02 1.43999219e-01
1.26189515e-01 1.61632851e-01 -7.68621713e-02 -7.34970868e-01
-9.74128917e-02 -5.92559934e-01 1.41578525e-01 -1.15698087e+00
-9.36351836e-01 1.05021596e+00 1.32101551e-02 -1.30947709e+00
-6.76490128e-01 -3.41810107e-01 -7.45018899e-01 1.30530620e+00
-1.08642030e+00 -1.08899617e+00 -5.91574073e-01 3.89710784e-01
-1.89798489e-01 -2.79906951e-02 1.07814300e+00 2.92695522e-01
8.15566257e-02 7.66321599e-01 1.40704751e-01 3.02501976e-01
3.03909242e-01 -1.16888452e+00 3.05012375e-01 8.15845132e-01
-2.43687555e-01 5.90914428e-01 7.23501623e-01 -6.31648302e-01
-1.21701133e+00 -9.17395413e-01 -1.41713813e-01 -1.03247128e-01
3.13216038e-02 -1.27430901e-01 -7.98997939e-01 7.56630361e-01
4.35872495e-01 3.19181889e-01 4.50444490e-01 -8.69232953e-01
5.82849681e-01 1.43922910e-01 -1.81906247e+00 4.74063814e-01
6.43374264e-01 -9.61373821e-02 -2.88364589e-01 3.00370485e-01
3.36666673e-01 -1.47906029e+00 -1.08225393e+00 6.58591092e-01
5.86903453e-01 -9.37186837e-01 9.25891757e-01 -2.59898394e-01
6.47632241e-01 -1.30935907e-01 4.59142864e-01 -1.68395317e+00
1.14083163e-01 -9.22471046e-01 3.20205867e-01 7.42021143e-01
2.37846479e-01 -8.59549403e-01 9.96968567e-01 4.64282930e-01
-2.82223374e-01 -7.97036529e-01 -1.10558808e+00 -2.72241831e-01
1.17207982e-01 -1.18781850e-01 5.12356937e-01 8.85802329e-01
-3.79338771e-01 -1.82210729e-01 -3.65963161e-01 4.52675313e-01
4.47862744e-01 -1.44988477e-01 4.78548586e-01 -5.16654670e-01
-4.77943629e-01 -1.92495435e-01 -6.11209989e-01 -9.28401768e-01
-1.21476710e-01 -6.29311383e-01 2.91747093e-01 -1.36779809e+00
-2.48418644e-01 -9.07617569e-01 -7.91637599e-02 4.93434928e-02
-7.88149685e-02 5.09897947e-01 -1.25913724e-01 -2.72713155e-01
3.12716156e-01 2.85892159e-01 2.01602983e+00 1.39726087e-01
-5.49105518e-02 2.65587687e-01 -3.55033070e-01 3.43366951e-01
4.18127954e-01 -3.36055905e-01 -5.75481832e-01 -2.04468578e-01
-1.28607973e-01 8.78928959e-01 4.42909747e-01 -1.09354496e+00
9.96841639e-02 2.00990096e-01 6.99176013e-01 -2.15617910e-01
3.72795612e-01 -1.04256916e+00 8.11960101e-01 7.19047129e-01
-2.78081745e-01 -2.44179711e-01 3.49449068e-01 1.96254447e-01
-2.82123059e-01 -4.64505613e-01 9.95968819e-01 -3.44668120e-01
-2.01131195e-01 1.73491687e-01 -2.09108174e-01 -4.78113890e-02
1.17768884e+00 -3.55203807e-01 2.03519344e-01 -7.91774273e-01
-1.02918458e+00 -2.83812165e-01 2.69906461e-01 -8.92188028e-02
9.93318856e-01 -1.10961008e+00 -8.08592856e-01 7.79622376e-01
-2.56767154e-01 2.89517432e-01 8.93490851e-01 8.33775699e-01
-1.39118922e+00 8.40387866e-02 -2.87986368e-01 -7.19058156e-01
-6.55135095e-01 4.81523871e-01 9.92666960e-01 -7.48612210e-02
-4.78914112e-01 9.87279892e-01 3.08247626e-01 -5.97618520e-01
-1.05361916e-01 -4.08201158e-01 -9.29240510e-02 -7.36239433e-01
4.56578918e-02 -7.75745511e-02 1.75674230e-01 -5.49333215e-01
1.06378617e-02 6.77455723e-01 3.97235811e-01 -2.17060417e-01
9.08666968e-01 2.00911433e-01 2.64087945e-01 -1.20858274e-01
1.00430167e+00 2.03538120e-01 -1.32925653e+00 1.38780266e-01
-8.81143630e-01 -6.32023156e-01 -9.67615098e-02 -1.07881498e+00
-1.55747283e+00 7.04194188e-01 1.02187884e+00 2.12001324e-01
1.20695162e+00 -1.13716736e-01 6.45389199e-01 -4.05361623e-01
2.61738807e-01 -1.73746109e-01 -1.17557347e-01 -2.38923252e-01
1.10416341e+00 -9.28743243e-01 -2.52734244e-01 -6.40787005e-01
-7.23130524e-01 8.79036844e-01 5.13210356e-01 -2.76517600e-01
4.41382676e-01 5.11422873e-01 5.64157903e-01 -6.95087165e-02
1.94773301e-01 4.82754260e-01 3.72999281e-01 8.66595387e-01
3.66902918e-01 9.05827209e-02 -1.19151071e-01 2.61634290e-01
-1.54346362e-01 -1.30614359e-02 5.90077579e-01 8.10873747e-01
2.56376892e-01 -6.24821544e-01 -4.32438850e-01 1.51678249e-01
-6.23293042e-01 -9.70141292e-02 3.91021401e-01 1.04881287e+00
2.11894453e-01 3.16765219e-01 5.45391627e-02 5.13090342e-02
4.56026256e-01 -3.64557177e-01 9.80536222e-01 -3.94146591e-01
-6.46706760e-01 -4.88537699e-02 -3.09588581e-01 -3.30298632e-01
-1.96750864e-01 -3.21489960e-01 -1.15750718e+00 2.50249624e-01
-1.05470926e-01 3.69427860e-01 8.82719874e-01 5.86979866e-01
1.78306624e-01 8.92778456e-01 7.12885976e-01 -1.03671038e+00
-2.89827853e-01 -1.09955013e+00 -7.10979402e-01 4.86274987e-01
4.06393737e-01 -3.97797108e-01 -4.01236057e-01 6.39399588e-02] | [14.13900089263916, -2.0197365283966064] |
e93a208b-c497-42cc-80c1-ecf282109af0 | re-embedding-words | null | null | https://aclanthology.org/P13-2087 | https://aclanthology.org/P13-2087.pdf | Re-embedding words | null | ['Igor Labutov', 'Hod Lipson'] | 2013-08-01 | null | null | null | acl-2013-8 | ['subjectivity-analysis'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.445281505584717, 3.647340774536133] |
7866dc14-7415-4785-aafc-1136af25b80f | the-surprising-effectiveness-of-diffusion | 2306.01923 | null | https://arxiv.org/abs/2306.01923v1 | https://arxiv.org/pdf/2306.01923v1.pdf | The Surprising Effectiveness of Diffusion Models for Optical Flow and Monocular Depth Estimation | Denoising diffusion probabilistic models have transformed image generation with their impressive fidelity and diversity. We show that they also excel in estimating optical flow and monocular depth, surprisingly, without task-specific architectures and loss functions that are predominant for these tasks. Compared to the point estimates of conventional regression-based methods, diffusion models also enable Monte Carlo inference, e.g., capturing uncertainty and ambiguity in flow and depth. With self-supervised pre-training, the combined use of synthetic and real data for supervised training, and technical innovations (infilling and step-unrolled denoising diffusion training) to handle noisy-incomplete training data, and a simple form of coarse-to-fine refinement, one can train state-of-the-art diffusion models for depth and optical flow estimation. Extensive experiments focus on quantitative performance against benchmarks, ablations, and the model's ability to capture uncertainty and multimodality, and impute missing values. Our model, DDVM (Denoising Diffusion Vision Model), obtains a state-of-the-art relative depth error of 0.074 on the indoor NYU benchmark and an Fl-all outlier rate of 3.26\% on the KITTI optical flow benchmark, about 25\% better than the best published method. For an overview see https://diffusion-vision.github.io. | ['David J. Fleet', 'Deqing Sun', 'Mohammad Norouzi', 'Abhishek Kar', 'Junhwa Hur', 'Charles Herrmann', 'Saurabh Saxena'] | 2023-06-02 | null | null | null | null | ['monocular-depth-estimation'] | ['computer-vision'] | [-6.21068291e-02 -1.07890576e-01 1.68073587e-02 -3.42033319e-02
-9.73068774e-01 -5.46769023e-01 7.26704180e-01 -4.94063497e-01
-4.95363772e-01 1.05461884e+00 3.50992262e-01 -1.26290426e-01
1.34115843e-02 -5.93594491e-01 -6.42137945e-01 -9.11791921e-01
-8.45621452e-02 4.35226381e-01 1.22098990e-01 2.94622600e-01
3.47386658e-01 2.83237398e-01 -1.32176042e+00 1.10681178e-02
1.22337961e+00 1.05956614e+00 1.25473410e-01 1.28020573e+00
2.39680195e-03 1.31231260e+00 -4.70438927e-01 -3.15421432e-01
3.47144037e-01 -3.72078955e-01 -5.65718353e-01 4.63514477e-02
9.62576330e-01 -9.19714510e-01 -9.77767169e-01 9.81601119e-01
6.05760813e-01 1.91433623e-01 9.10210729e-01 -1.03337741e+00
-1.01228881e+00 2.38151774e-01 -7.63131261e-01 5.70185304e-01
3.14484149e-01 8.81820738e-01 5.92591286e-01 -9.00256276e-01
8.55605066e-01 1.21814847e+00 7.21039951e-01 7.76915729e-01
-1.36115646e+00 -3.78543645e-01 -6.76115602e-03 1.36051714e-01
-1.19472730e+00 -6.89148486e-01 4.32739049e-01 -8.45764577e-01
9.58904624e-01 -2.03810513e-01 5.02744794e-01 1.54462743e+00
3.28422189e-01 7.35988975e-01 1.29057765e+00 -4.14192602e-02
3.01087648e-01 -8.16673189e-02 -1.43895984e-01 9.20294523e-01
2.78181821e-01 6.08080149e-01 -6.79791987e-01 -4.50435467e-02
1.15574145e+00 -4.18669820e-01 -7.05385208e-01 -2.40751952e-01
-1.31112778e+00 7.11307228e-01 4.90255654e-01 -2.13704184e-01
-3.08241725e-01 8.50566089e-01 4.99672890e-02 3.91676351e-02
6.66206598e-01 2.03825876e-01 -2.92822838e-01 -5.49303472e-01
-1.23881114e+00 2.66336858e-01 8.68595898e-01 9.02908683e-01
7.91967571e-01 4.27642107e-01 -4.07242835e-01 4.79806274e-01
4.74121869e-01 8.82420957e-01 2.70334661e-01 -1.91348934e+00
3.29813033e-01 -2.28546456e-01 2.47968912e-01 -6.34552419e-01
-2.10483849e-01 -5.67412198e-01 -1.11367750e+00 6.79234326e-01
9.47583079e-01 -4.54982042e-01 -1.33119488e+00 1.62719810e+00
1.09329186e-01 7.72820354e-01 4.48224694e-02 9.62810099e-01
7.69591630e-01 5.25498092e-01 -2.52111822e-01 -1.32463574e-01
7.88508832e-01 -1.00244093e+00 -4.70041662e-01 -3.67704123e-01
3.90284956e-01 -8.45240891e-01 7.95967281e-01 6.29966736e-01
-1.31224787e+00 -4.32204545e-01 -7.95452118e-01 -4.02717799e-01
1.62991524e-01 -1.76128715e-01 8.05399716e-01 7.34815836e-01
-1.40393686e+00 7.52137244e-01 -9.57153440e-01 -5.64715900e-02
8.73645961e-01 -4.12435122e-02 -2.49249980e-01 -5.41942894e-01
-7.65232265e-01 7.18464196e-01 -2.48561546e-01 2.07348496e-01
-1.40477729e+00 -1.12662876e+00 -9.17887568e-01 -3.39898199e-01
1.17261499e-01 -1.37776303e+00 1.02430487e+00 -5.04743218e-01
-1.72486579e+00 5.96294403e-01 -3.37013185e-01 -6.13794863e-01
1.13908112e+00 -1.91948891e-01 1.41395509e-01 3.39200586e-01
1.56272784e-01 1.14997542e+00 1.13467729e+00 -1.31526875e+00
-5.22655785e-01 -2.11908028e-01 3.51694040e-03 2.39345562e-02
2.02578425e-01 -5.71818948e-01 -5.22541761e-01 -6.30577028e-01
7.00945593e-03 -8.45893443e-01 -3.20590764e-01 4.56351131e-01
-3.52018863e-01 2.68667132e-01 4.94861901e-01 -6.29430175e-01
7.80153394e-01 -1.78444648e+00 2.46266887e-01 -5.79418689e-02
6.88607514e-01 -1.00960247e-01 -2.22129643e-01 -1.73935816e-01
4.91061568e-01 1.15383111e-01 -7.08540618e-01 -8.31838787e-01
-4.43557501e-02 3.35263461e-01 -1.45413533e-01 7.50042200e-01
1.84455514e-01 9.50116575e-01 -1.12193930e+00 -4.55638349e-01
6.09091699e-01 8.72071207e-01 -8.85417640e-01 -4.80814502e-02
-5.54349124e-02 9.36052322e-01 -9.21819657e-02 7.15348482e-01
8.44117701e-01 -3.15904200e-01 -4.76250321e-01 -1.32470340e-01
4.98627722e-02 -6.13951720e-02 -1.22791028e+00 2.05698037e+00
-6.15818977e-01 1.14164233e+00 9.20842290e-02 -2.76733011e-01
4.99640316e-01 4.80401404e-02 5.29241145e-01 -4.02250350e-01
-1.03790633e-01 3.13229740e-01 -8.16883817e-02 -4.04276758e-01
3.77226889e-01 8.51396993e-02 5.17779291e-01 2.73187220e-01
3.28491330e-01 -5.68203092e-01 2.16697246e-01 3.39731246e-01
1.31931186e+00 4.33634013e-01 -3.22031140e-01 -1.43556073e-01
2.74283141e-01 -1.39745623e-01 5.48981130e-01 1.15833378e+00
-5.64805150e-01 1.09643757e+00 3.91494691e-01 -1.41820222e-01
-1.01052809e+00 -1.27561784e+00 -2.85829365e-01 3.78003538e-01
2.12476358e-01 -1.38360441e-01 -6.59595490e-01 -6.26065552e-01
1.84245452e-01 6.20222747e-01 -8.71783197e-01 3.95356901e-02
-3.23741257e-01 -9.50620115e-01 6.72254980e-01 4.62755680e-01
7.45369911e-01 -7.00235963e-01 -4.36734438e-01 1.92448497e-01
-3.40790182e-01 -1.37917161e+00 -4.70365047e-01 -2.02805534e-01
-8.88350904e-01 -1.07893062e+00 -1.12541628e+00 -1.82412699e-01
3.80346298e-01 6.65907562e-02 1.32544672e+00 -2.06419200e-01
-2.55891353e-01 6.82949722e-01 1.21698558e-01 -5.87250106e-02
-2.79252648e-01 -1.91781461e-01 -1.07757486e-01 -1.82153821e-01
-5.11329547e-02 -7.46000946e-01 -1.21859407e+00 1.72835648e-01
-6.80774748e-01 -2.95602292e-01 4.06523436e-01 1.00974917e+00
4.04496163e-01 -3.60544533e-01 9.70884413e-02 -6.27664626e-01
4.34801340e-01 -4.55057085e-01 -6.27901971e-01 -2.07560465e-01
-7.54548430e-01 1.08172610e-01 1.84649006e-01 -4.63497758e-01
-1.26682568e+00 -2.47864500e-01 -3.85095365e-02 -9.46531296e-01
-1.25060394e-01 9.59898233e-02 4.53131527e-01 -4.39709574e-01
1.00651693e+00 2.54572004e-01 -9.93336178e-03 -1.19104303e-01
6.15227520e-01 2.59332508e-01 7.92862833e-01 -6.59912825e-01
6.03211045e-01 1.04052603e+00 1.96300223e-01 -8.64624679e-01
-7.11659908e-01 -1.62087768e-01 -4.52666491e-01 -2.69014716e-01
9.47458088e-01 -1.14451146e+00 -6.63744986e-01 9.51047122e-01
-1.20367932e+00 -7.49473274e-01 -5.46989501e-01 6.97043359e-01
-8.19878697e-01 6.19783282e-01 -8.37509274e-01 -8.59915495e-01
-7.19424337e-02 -1.28110898e+00 1.02733397e+00 2.03316480e-01
5.06779812e-02 -1.34002769e+00 1.10849701e-01 4.99253541e-01
6.39187217e-01 2.91989058e-01 1.51903361e-01 3.63972723e-01
-1.01577199e+00 2.76346117e-01 -4.90294397e-01 7.37527490e-01
-4.83732224e-02 3.89484704e-01 -1.37813652e+00 -1.46877676e-01
5.69742918e-02 -3.85894835e-01 1.49678528e+00 1.09234130e+00
8.67182434e-01 3.30105126e-02 -1.37250409e-01 1.28358376e+00
1.30654573e+00 -1.86377421e-01 8.79651248e-01 1.60882846e-01
9.75204587e-01 4.10148680e-01 8.41365159e-02 5.93890190e-01
6.13454700e-01 1.43036202e-01 5.22932708e-01 2.11373538e-01
-7.44403541e-01 -6.50854483e-02 2.07102343e-01 3.30657750e-01
-2.58344382e-01 -3.68109643e-01 -9.18691158e-01 5.30453742e-01
-1.68972075e+00 -9.36085582e-01 -3.53963286e-01 1.96646965e+00
7.08442748e-01 3.07745039e-01 -2.47122556e-01 -1.68835402e-01
2.50459790e-01 3.43111932e-01 -7.63758302e-01 3.95011753e-02
-4.36340541e-01 6.15260080e-02 7.99897134e-01 1.05707216e+00
-9.94440138e-01 1.07415700e+00 6.64161682e+00 7.32655525e-01
-9.23027933e-01 1.47926778e-01 9.93727088e-01 -3.01782757e-01
-4.76289392e-01 -2.06809901e-02 -6.93136036e-01 5.53651452e-01
6.68356657e-01 1.97938606e-01 7.04470515e-01 2.92103022e-01
4.03698444e-01 -6.10514462e-01 -9.61526453e-01 1.31145132e+00
2.24267486e-02 -1.62879670e+00 -9.59726945e-02 2.41098449e-01
1.31227148e+00 5.20843387e-01 3.26899558e-01 1.42222340e-03
5.35744190e-01 -1.11087525e+00 7.09728658e-01 9.88838911e-01
8.91201138e-01 -2.80731529e-01 6.75704002e-01 1.69922635e-01
-6.82097256e-01 -1.47438291e-02 -3.49025249e-01 -6.69222400e-02
4.09211040e-01 1.00888646e+00 -1.80339143e-01 2.92164356e-01
8.86307299e-01 1.05655718e+00 -3.00877810e-01 1.24493885e+00
-3.87956202e-01 6.20461464e-01 -5.93537927e-01 2.69192904e-01
3.91158849e-01 -3.12728763e-01 8.00981760e-01 1.16384888e+00
3.92886102e-01 -1.48747981e-01 -1.79958284e-01 1.18399107e+00
-7.77240992e-02 -6.39661431e-01 -5.70752203e-01 3.78841788e-01
3.08205217e-01 9.59528804e-01 -2.17185020e-01 -3.95400375e-01
-3.20957929e-01 1.11194408e+00 2.18061015e-01 8.36194754e-01
-8.13182950e-01 -5.18540740e-02 1.15275121e+00 4.56866343e-04
3.06674451e-01 -5.49783528e-01 -6.70062900e-01 -1.39358497e+00
-8.34241807e-02 -4.22614336e-01 7.37923235e-02 -9.30123150e-01
-1.46144652e+00 4.98629034e-01 -1.99206084e-01 -9.43979740e-01
-5.31540632e-01 -6.24617875e-01 -4.66826826e-01 1.12673044e+00
-1.97435725e+00 -8.76570344e-01 -6.02565885e-01 6.74188197e-01
4.36709493e-01 1.10458560e-01 3.71357173e-01 2.26192906e-01
-3.71994495e-01 3.15508604e-01 1.61170453e-01 -5.36217541e-03
8.43555629e-01 -1.47372007e+00 4.35240060e-01 1.02729034e+00
1.73279032e-01 2.69978642e-01 7.92961836e-01 -5.29614389e-01
-1.23023224e+00 -8.77825320e-01 4.94576126e-01 -9.61272776e-01
8.38819742e-01 -2.28656810e-02 -7.21082687e-01 5.80056071e-01
1.65066406e-01 6.07139289e-01 1.03387125e-01 -3.41654748e-01
-3.61198306e-01 2.55815443e-02 -1.30764449e+00 5.77776968e-01
1.34616780e+00 -6.12006068e-01 -1.35236919e-01 2.62285054e-01
4.69903529e-01 -7.47213721e-01 -7.90046990e-01 2.37416565e-01
5.37131727e-01 -1.55059528e+00 1.11525559e+00 -8.39180201e-02
6.78773642e-01 -2.91858643e-01 -1.02732137e-01 -1.34821486e+00
-2.28596747e-01 -1.05494142e+00 -6.10850871e-01 9.85465586e-01
2.73930401e-01 -6.15083814e-01 1.03817892e+00 5.53946078e-01
-1.08524285e-01 -5.72544873e-01 -9.90318596e-01 -6.57580197e-01
2.97873139e-01 -8.38270664e-01 -3.79958213e-03 7.92776823e-01
-7.36495614e-01 7.92447254e-02 -3.85042846e-01 2.06443995e-01
1.32776499e+00 -4.10501957e-01 7.80943215e-01 -9.36466634e-01
-4.40307885e-01 -6.81399822e-01 -4.05175656e-01 -1.56047153e+00
8.49283189e-02 -4.84252274e-01 6.23629764e-02 -1.82319403e+00
-2.52742887e-01 -3.06835324e-01 5.41434856e-03 -1.91631094e-02
-2.11962059e-01 4.12487239e-01 5.84072396e-02 3.33525538e-01
-2.45386839e-01 7.73194671e-01 1.72785008e+00 -3.44252586e-01
-2.94406921e-01 -1.11128263e-01 -5.53010285e-01 7.62595654e-01
6.23470068e-01 -3.69147807e-01 -5.89840829e-01 -9.36708450e-01
-7.94919394e-03 7.92534426e-02 6.90513849e-01 -1.23594785e+00
4.51099873e-01 -5.02960384e-02 4.98927534e-01 -2.46268004e-01
6.60836756e-01 -4.06460673e-01 -2.96736389e-01 3.41553688e-01
-3.54648568e-02 -1.66163474e-01 1.71442568e-01 8.62986326e-01
-1.02475382e-01 2.50213761e-02 8.33005726e-01 -2.62330592e-01
-8.32478881e-01 5.85229814e-01 -3.13857138e-01 5.92994213e-01
6.86566234e-01 -3.99072409e-01 -6.99700713e-01 -6.46739423e-01
-6.89411044e-01 1.26897827e-01 4.92417157e-01 9.57510173e-02
7.27688968e-01 -9.20391679e-01 -1.07631707e+00 1.01570226e-01
-8.85258168e-02 3.16244543e-01 4.61876124e-01 1.00736940e+00
-8.56315196e-01 -3.92510444e-02 -5.08121625e-02 -9.62277293e-01
-5.13184369e-01 1.11121237e-01 6.94650352e-01 -2.54405029e-02
-6.41350389e-01 1.34053409e+00 2.00975329e-01 -1.69231445e-01
2.46176556e-01 -6.88234746e-01 3.57480824e-01 -1.89009055e-01
4.36267436e-01 7.25348711e-01 -2.90290594e-01 -2.49128744e-01
-3.22844088e-01 7.66981781e-01 3.45935315e-01 -5.24502814e-01
9.63925660e-01 -4.50929344e-01 7.67280757e-02 2.72612959e-01
1.00295019e+00 2.06701141e-02 -2.30656052e+00 -1.12171814e-01
-5.21744013e-01 -7.02164829e-01 5.34563601e-01 -8.29211712e-01
-1.22538662e+00 1.05911458e+00 6.28003716e-01 -1.60445690e-01
7.68395245e-01 -1.84312925e-01 6.85632944e-01 1.74874306e-01
1.41400129e-01 -8.30756187e-01 -1.10240122e-02 6.01036251e-01
7.22302914e-01 -1.59755433e+00 -6.94812536e-02 -3.89092058e-01
-5.67389488e-01 9.30966794e-01 4.91748512e-01 -3.12736839e-01
9.45992887e-01 5.32224238e-01 2.33063549e-01 1.57750040e-01
-8.87424469e-01 -2.81475097e-01 2.57483572e-01 1.03392470e+00
1.65464148e-01 -2.74866223e-01 3.21960598e-01 -1.78619668e-01
-1.96783151e-02 2.91889012e-01 7.33174205e-01 6.42474353e-01
-2.49022156e-01 -4.99791682e-01 -1.42748475e-01 4.28874224e-01
-2.49582827e-01 -4.43663836e-01 5.46101704e-02 6.51778936e-01
-1.75294429e-02 1.16456139e+00 2.70925879e-01 -1.38032809e-01
7.14786127e-02 -4.19223011e-01 9.03473496e-01 -1.52379215e-01
-2.59208977e-01 -1.17071113e-02 3.25370505e-02 -9.38408315e-01
-4.89212662e-01 -5.92600942e-01 -9.79640782e-01 -7.26423085e-01
9.47220922e-02 -3.86073470e-01 5.60282469e-01 8.16605866e-01
4.32830185e-01 5.05838454e-01 3.33940476e-01 -1.12990987e+00
-3.35396707e-01 -1.00957584e+00 -5.18304050e-01 1.77431807e-01
8.09236169e-01 -6.56104207e-01 -9.01181936e-01 -2.53079645e-02] | [8.786431312561035, -2.2106196880340576] |
53a59cbc-3cb7-4b78-8a10-fcabc61e95f6 | association-metrics-in-neural-transition | null | null | https://aclanthology.org/W19-7722 | https://aclanthology.org/W19-7722.pdf | Association Metrics in Neural Transition-Based Dependency Parsing | null | ['Dani{\\"e}l de Kok', 'Sebastian P{\\"u}tz', 'Patricia Fischer'] | 2019-08-01 | null | null | null | ws-2019-8 | ['transition-based-dependency-parsing'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.21635627746582, 3.826930046081543] |
d77dad25-2102-4c05-bc37-c21a2aa1563d | decoupling-features-in-hierarchical | 2210.09782 | null | https://arxiv.org/abs/2210.09782v3 | https://arxiv.org/pdf/2210.09782v3.pdf | Decoupling Features in Hierarchical Propagation for Video Object Segmentation | This paper focuses on developing a more effective method of hierarchical propagation for semi-supervised Video Object Segmentation (VOS). Based on vision transformers, the recently-developed Associating Objects with Transformers (AOT) approach introduces hierarchical propagation into VOS and has shown promising results. The hierarchical propagation can gradually propagate information from past frames to the current frame and transfer the current frame feature from object-agnostic to object-specific. However, the increase of object-specific information will inevitably lead to the loss of object-agnostic visual information in deep propagation layers. To solve such a problem and further facilitate the learning of visual embeddings, this paper proposes a Decoupling Features in Hierarchical Propagation (DeAOT) approach. Firstly, DeAOT decouples the hierarchical propagation of object-agnostic and object-specific embeddings by handling them in two independent branches. Secondly, to compensate for the additional computation from dual-branch propagation, we propose an efficient module for constructing hierarchical propagation, i.e., Gated Propagation Module, which is carefully designed with single-head attention. Extensive experiments show that DeAOT significantly outperforms AOT in both accuracy and efficiency. On YouTube-VOS, DeAOT can achieve 86.0% at 22.4fps and 82.0% at 53.4fps. Without test-time augmentations, we achieve new state-of-the-art performance on four benchmarks, i.e., YouTube-VOS (86.2%), DAVIS 2017 (86.2%), DAVIS 2016 (92.9%), and VOT 2020 (0.622). Project page: https://github.com/z-x-yang/AOT. | ['Yi Yang', 'Zongxin Yang'] | 2022-10-18 | null | null | null | null | ['semi-supervised-video-object-segmentation', 'video-object-segmentation'] | ['computer-vision', 'computer-vision'] | [-3.37479323e-01 -2.00871885e-01 -2.69610196e-01 -2.19857305e-01
-4.74346191e-01 -2.65505314e-01 2.30726823e-01 -6.24057464e-02
-5.32998979e-01 3.54414344e-01 3.64993438e-02 -1.95242912e-01
3.44831407e-01 -7.56250620e-01 -8.68487000e-01 -5.47714889e-01
3.12256329e-02 9.72872525e-02 7.81856179e-01 2.29383871e-01
-5.20692468e-02 1.94504574e-01 -1.29058468e+00 2.21200928e-01
7.80896306e-01 1.24129581e+00 3.13448310e-01 7.24712253e-01
-2.40325570e-01 8.20767522e-01 -4.83317435e-01 -5.08162141e-01
2.04035550e-01 -1.38197377e-01 -7.94180095e-01 1.50139838e-01
4.83299851e-01 -6.66587234e-01 -5.05385399e-01 9.94549453e-01
3.05761606e-01 -2.15939164e-01 4.92105544e-01 -1.66363692e+00
-8.91928673e-01 5.37644923e-01 -9.24166501e-01 2.83497125e-01
-2.75423080e-01 5.13829529e-01 9.13353384e-01 -1.10937726e+00
4.73568678e-01 1.24709070e+00 5.58391929e-01 5.85573018e-01
-9.38831270e-01 -7.50855029e-01 3.85833263e-01 3.08111221e-01
-1.34780622e+00 -1.92688867e-01 5.96619189e-01 -6.41438663e-01
6.30916059e-01 1.24687077e-02 8.56661022e-01 7.39179134e-01
-7.37709627e-02 1.29521537e+00 8.31417203e-01 3.79009657e-02
-8.68890062e-02 1.87731132e-01 3.46193314e-01 8.91888916e-01
3.20288241e-01 -2.06328869e-01 -4.56885517e-01 3.18156242e-01
8.55836391e-01 1.91216215e-01 -3.11329633e-01 -2.20157370e-01
-1.10055053e+00 6.55912519e-01 7.88244963e-01 8.63403529e-02
-2.62256175e-01 4.46182251e-01 6.07371449e-01 -1.66981861e-01
4.12120789e-01 -4.46483977e-02 -5.28912306e-01 -1.81176797e-01
-9.80920672e-01 -2.60403398e-02 4.11776155e-01 1.28659499e+00
7.22875595e-01 3.24893832e-01 -4.29836392e-01 6.82944417e-01
6.68750882e-01 3.29694837e-01 4.69585210e-01 -1.06151628e+00
4.80717957e-01 5.45853019e-01 2.02970550e-05 -8.56208563e-01
-1.16271526e-01 -4.60140437e-01 -7.33048260e-01 7.43341073e-02
2.46268600e-01 -1.63237065e-01 -1.16291273e+00 1.61214221e+00
4.24161166e-01 2.52613842e-01 -6.67514056e-02 1.07197297e+00
1.18860054e+00 1.12042463e+00 3.49332571e-01 -8.18891823e-02
1.25850332e+00 -1.68281329e+00 -5.67258477e-01 -1.01433434e-01
3.57013106e-01 -5.89042664e-01 1.19590855e+00 1.32788196e-01
-1.28485417e+00 -8.25519025e-01 -9.75422025e-01 -2.60035694e-01
-1.55825645e-01 2.50632405e-01 5.71180046e-01 3.97354692e-01
-1.01189375e+00 4.18120086e-01 -9.96580660e-01 -1.87292248e-01
9.83945370e-01 3.51759762e-01 -4.43340205e-02 -4.13133837e-02
-1.02462566e+00 3.50031048e-01 3.52308303e-01 8.96774884e-03
-1.14579999e+00 -9.48815346e-01 -7.88169324e-01 1.89493760e-01
2.95111895e-01 -6.59701467e-01 1.27088654e+00 -8.61012340e-01
-1.34814930e+00 5.07161438e-01 -2.03524813e-01 -4.80295897e-01
5.78877985e-01 -4.36728477e-01 -7.15064108e-02 3.57237577e-01
2.72192746e-01 1.19132280e+00 9.33637321e-01 -1.26737106e+00
-8.48980486e-01 -1.16049133e-01 1.41390026e-01 1.13472417e-01
-8.62732947e-01 -7.44217858e-02 -1.30420578e+00 -7.00195670e-01
-1.90504432e-01 -6.98994875e-01 7.48712718e-02 4.40996915e-01
-4.31316167e-01 -4.35078740e-01 1.15590847e+00 -6.46810591e-01
1.34265053e+00 -2.22806501e+00 1.02583572e-01 -4.56838459e-01
5.17226160e-01 6.54595733e-01 -5.86128682e-02 -8.40074793e-02
2.35806689e-01 4.26326275e-01 -1.95331588e-01 -5.37900329e-01
-8.19846243e-02 1.39841482e-01 1.77931227e-02 2.67853528e-01
3.99042010e-01 1.08817983e+00 -9.06182170e-01 -7.12848365e-01
2.90082216e-01 6.63989961e-01 -7.89173961e-01 2.05551550e-01
-3.11088730e-02 1.47349268e-01 -3.17324698e-01 8.43574524e-01
8.67177188e-01 -3.95110667e-01 -3.39883387e-01 -5.22444665e-01
-2.66518593e-01 5.48510775e-02 -7.79541194e-01 1.43074489e+00
-2.23294854e-01 8.45526695e-01 -1.55722603e-01 -6.95581436e-01
6.01848304e-01 1.04114488e-01 3.82601291e-01 -6.71012044e-01
2.38139749e-01 1.38295246e-02 -1.67511970e-01 -5.24116516e-01
4.33981389e-01 1.27628848e-01 1.99848264e-01 -3.53332460e-02
4.50502038e-01 2.81094015e-01 3.15878332e-01 4.10129756e-01
8.76577973e-01 2.81327546e-01 -2.29618326e-01 -1.43163726e-01
3.80110592e-01 -1.64370283e-01 9.19665575e-01 4.09917951e-01
-6.26258969e-01 6.03297770e-01 4.84436512e-01 -1.97845325e-01
-8.80861759e-01 -1.17179978e+00 8.53364356e-03 8.94282460e-01
5.07161379e-01 -6.17234111e-01 -8.38210821e-01 -9.69236910e-01
1.23920599e-02 4.82236356e-01 -5.31434774e-01 -1.85614944e-01
-5.28234243e-01 -6.13928556e-01 4.22375083e-01 9.33506310e-01
9.03837025e-01 -1.01178062e+00 -5.45926213e-01 9.51027945e-02
-1.13373846e-01 -1.39471090e+00 -8.21142554e-01 -2.19553143e-01
-9.43128407e-01 -7.98325419e-01 -1.17595327e+00 -9.18665349e-01
6.67616069e-01 4.48068857e-01 8.22500527e-01 -4.38061133e-02
-1.96094260e-01 4.01879638e-01 -3.91427845e-01 -2.42081732e-01
6.57175332e-02 2.43479796e-02 -1.23829156e-01 1.80405691e-01
2.10549831e-01 -4.15063053e-01 -8.27469230e-01 2.87599921e-01
-7.41102040e-01 3.24188083e-01 6.06839299e-01 7.66263306e-01
6.28822803e-01 1.27002327e-02 3.63949865e-01 -6.15856826e-01
1.29081547e-01 -4.79989827e-01 -5.49165905e-01 -5.64057417e-02
-5.04448891e-01 -2.55982339e-01 6.77124560e-01 -5.50683200e-01
-9.36657071e-01 -1.26441360e-01 -1.64974928e-01 -9.45152044e-01
4.57532555e-02 2.80588865e-01 -2.79537231e-01 -5.78160770e-02
9.17009488e-02 3.33015472e-01 -3.06677014e-01 -3.88631403e-01
3.61187249e-01 6.62424862e-01 5.26359797e-01 -2.92825937e-01
9.30976093e-01 3.94234985e-01 -4.57540959e-01 -7.54690707e-01
-7.28635609e-01 -4.92765158e-01 -2.98973352e-01 -3.23142052e-01
1.07995987e+00 -1.05556500e+00 -6.85456932e-01 8.13570678e-01
-1.22225332e+00 -5.47822058e-01 -2.93102920e-01 4.82470363e-01
-2.64449567e-01 3.98164093e-01 -8.96687686e-01 -5.52041233e-01
-4.56704140e-01 -1.34547591e+00 8.95950854e-01 6.64364696e-01
1.99799672e-01 -7.94972897e-01 -4.25438881e-01 4.72687483e-01
3.32086235e-01 -7.29647651e-02 5.74029624e-01 -2.19391599e-01
-9.49834824e-01 7.85621181e-02 -8.56718779e-01 7.62921453e-01
-5.47500420e-03 2.49514252e-01 -1.00038767e+00 -3.39382350e-01
-2.50265300e-01 -2.35472426e-01 1.08599770e+00 3.91062051e-01
1.28433144e+00 -2.64976948e-01 -2.13932559e-01 9.16144133e-01
1.43794537e+00 2.79314667e-01 5.65145493e-01 3.41739565e-01
1.22714150e+00 2.03915715e-01 6.31036341e-01 3.25044990e-01
7.79393554e-01 6.56362057e-01 5.55603445e-01 -1.41963348e-01
-5.48357487e-01 -3.19281161e-01 5.08157194e-01 1.00776958e+00
-2.13016905e-02 -2.64883161e-01 -7.01316476e-01 8.67895901e-01
-1.77791429e+00 -5.74801087e-01 -2.70171553e-01 1.94143534e+00
9.68814313e-01 3.88730079e-01 2.18827412e-01 5.95164038e-02
7.63459980e-01 1.90743417e-01 -6.53200984e-01 -2.17387334e-01
1.50125504e-01 -1.34477794e-01 7.07745314e-01 3.72493535e-01
-1.29863787e+00 1.13134015e+00 4.48525429e+00 8.77439320e-01
-1.19528019e+00 4.17211890e-01 6.14067018e-01 -1.61318425e-02
-4.39756066e-02 -1.40175536e-01 -1.07567346e+00 8.72853398e-01
7.16248214e-01 -1.60381734e-01 6.85303807e-02 8.81837666e-01
3.21337744e-03 8.91632363e-02 -8.53402138e-01 1.11339462e+00
-1.34055251e-02 -1.37982523e+00 2.64520526e-01 -1.37748823e-01
6.60060585e-01 1.17515884e-01 1.26733080e-01 4.98295218e-01
-9.66632068e-02 -5.83891213e-01 1.04907322e+00 1.90413997e-01
8.40674937e-01 -6.25721812e-01 6.62229061e-01 3.58328000e-02
-1.51691055e+00 1.26714977e-02 -2.67992526e-01 2.30688497e-01
3.71400088e-01 5.07150292e-01 -3.18988085e-01 4.16674495e-01
1.09393144e+00 1.02970111e+00 -5.34356236e-01 1.36204946e+00
-3.15637410e-01 8.86985540e-01 -2.61077285e-01 1.22262090e-01
5.52894771e-01 1.60764620e-01 4.50472116e-01 1.37619984e+00
1.02999508e-01 -1.76064409e-02 6.03910089e-02 8.32580924e-01
-3.74372423e-01 -8.32751617e-02 -8.47251192e-02 -4.70764600e-02
3.80183101e-01 1.27160299e+00 -7.13126898e-01 -5.72558463e-01
-5.17343521e-01 1.16862094e+00 2.58103937e-01 4.79063898e-01
-1.35422134e+00 -5.95930219e-01 6.98543429e-01 2.14551091e-01
7.79360712e-01 -2.11646944e-01 -9.55549330e-02 -1.12513018e+00
1.52425423e-01 -3.55278581e-01 2.04753146e-01 -7.59953201e-01
-1.10138297e+00 5.66538036e-01 -9.98120904e-02 -1.17869067e+00
5.30946195e-01 -5.77922761e-01 -5.88824153e-01 6.38603747e-01
-1.73587549e+00 -1.21021008e+00 -5.64331770e-01 4.59079772e-01
7.96971262e-01 1.17096737e-01 2.78696984e-01 7.77644575e-01
-8.84842098e-01 8.71766508e-01 -2.09858462e-01 4.46206927e-01
5.18156230e-01 -1.22872770e+00 4.72010791e-01 8.63673151e-01
-5.72758727e-03 3.08161974e-01 2.12751031e-01 -4.41550702e-01
-1.20944679e+00 -1.68327916e+00 6.95564508e-01 -2.85469383e-01
6.78363323e-01 -2.53693163e-01 -1.06763959e+00 6.87893450e-01
1.58599034e-01 4.24733371e-01 3.79193723e-01 -4.39611286e-01
-4.47054476e-01 -2.99646050e-01 -1.00892699e+00 6.64178133e-01
9.37461257e-01 -3.57614398e-01 -2.80662328e-01 1.27800167e-01
1.22737718e+00 -4.41673100e-01 -8.64023030e-01 4.31495845e-01
5.17644405e-01 -8.63594592e-01 8.71749341e-01 -3.30201924e-01
5.29509664e-01 -6.02924287e-01 -5.25713302e-02 -8.44967425e-01
-4.09051567e-01 -5.60355306e-01 -5.30522823e-01 1.46459138e+00
2.00292483e-01 -5.51135957e-01 7.57678330e-01 3.14579666e-01
-3.88823867e-01 -1.20523453e+00 -6.45788014e-01 -9.33495343e-01
4.35419306e-02 -4.97280538e-01 2.88597345e-01 6.58489048e-01
-4.90565002e-01 3.62896949e-01 -3.58811527e-01 2.66532779e-01
7.06862867e-01 -4.61624563e-02 7.11753309e-01 -8.80582333e-01
-2.02400804e-01 -4.32994157e-01 -6.68419659e-01 -1.63214660e+00
-1.76249415e-01 -7.54909873e-01 -4.80635129e-02 -1.85723245e+00
3.94187242e-01 -4.10651743e-01 -4.41447556e-01 6.10118210e-01
-3.27814519e-01 5.23027539e-01 6.28290832e-01 2.06108645e-01
-9.14693952e-01 7.78894126e-01 1.45684266e+00 -2.81787157e-01
-1.29546270e-01 -2.10808396e-01 -6.35053635e-01 9.03749287e-01
9.40975428e-01 -4.73769933e-01 -3.94926190e-01 -8.82221341e-01
-3.49673331e-01 -3.22703153e-01 5.23032486e-01 -9.95586753e-01
2.78413028e-01 1.17391050e-01 3.33327651e-01 -7.60141969e-01
3.01913053e-01 -6.67341769e-01 -4.08511639e-01 5.17099857e-01
1.46014476e-02 -1.27625301e-01 4.40853149e-01 5.40924132e-01
-2.98397869e-01 -7.07756653e-02 8.36729407e-01 1.27590314e-01
-1.03896928e+00 8.25800002e-01 -2.00580075e-01 2.96432406e-01
1.17001581e+00 -2.92072266e-01 -3.96080822e-01 -1.27517506e-01
-6.10162497e-01 6.45902753e-01 2.52996773e-01 6.09403908e-01
8.49685848e-01 -1.42736900e+00 -5.24097562e-01 1.88091416e-02
1.82207692e-02 2.53246993e-01 4.60326642e-01 9.79174733e-01
-5.64734757e-01 2.69598991e-01 -1.17619760e-01 -8.06465685e-01
-1.25682271e+00 4.66140389e-01 2.61286031e-02 -2.00813413e-02
-5.63130260e-01 1.26820993e+00 5.42063951e-01 2.54782774e-02
4.53943580e-01 -4.86310333e-01 -1.18058883e-01 1.22753896e-01
5.25431931e-01 3.67631108e-01 -3.90111566e-01 -7.21075654e-01
-4.60079312e-01 7.39061952e-01 -2.54395396e-01 1.72353610e-01
1.12167501e+00 -4.01447490e-02 1.00048698e-01 4.77286398e-01
1.50479567e+00 -2.15966001e-01 -1.65527141e+00 -1.47520542e-01
-3.57172340e-01 -5.65704405e-01 2.09788248e-01 -5.95505059e-01
-1.62523556e+00 1.28449488e+00 6.54641628e-01 2.10153945e-02
1.15544748e+00 3.96231525e-02 1.24842632e+00 -1.52670190e-01
2.59051085e-01 -7.10534990e-01 2.94732004e-01 4.11589533e-01
6.13366723e-01 -1.25526226e+00 -1.31984949e-01 -4.52648640e-01
-7.03001499e-01 8.54602337e-01 9.71107304e-01 -1.25916777e-02
6.79358602e-01 2.01909140e-01 -1.64350227e-01 5.65625774e-03
-6.21442378e-01 -2.92697340e-01 2.87094384e-01 4.94752079e-01
2.26581514e-01 4.63803597e-02 -1.37722716e-01 7.16063678e-01
3.18441838e-01 1.43552631e-01 3.94765019e-01 8.14922810e-01
-4.03802007e-01 -6.63577020e-01 1.58278865e-03 2.80967832e-01
-4.66374248e-01 -2.96126544e-01 -8.90232250e-03 7.80631483e-01
3.23302627e-01 6.71992242e-01 9.75691155e-02 -4.69271898e-01
2.36549124e-01 -2.78636783e-01 3.03592324e-01 -5.53952336e-01
-2.55171865e-01 2.22873777e-01 -1.93735719e-01 -5.76143026e-01
-3.52009982e-01 -4.67332274e-01 -1.57158875e+00 -2.58879513e-01
-2.89379418e-01 3.33860554e-02 4.24610108e-01 5.85098147e-01
4.28661555e-01 9.78154838e-01 4.77346659e-01 -9.50541973e-01
-5.56796901e-02 -6.70327425e-01 -2.43875831e-01 1.97100267e-01
3.43115032e-01 -6.34655774e-01 -2.65631914e-01 1.96712345e-01] | [9.317607879638672, -0.08782043308019638] |
81c680c9-fd79-4777-8c86-6f28598f56a9 | sentu-sentiment-analysis-of-tweets-by | null | null | https://aclanthology.org/S15-2108 | https://aclanthology.org/S15-2108.pdf | SeNTU: Sentiment Analysis of Tweets by Combining a Rule-based Classifier with Supervised Learning | null | ['Soujanya Poria', 'Prerna Chikersal', 'Erik Cambria'] | 2015-06-01 | null | null | null | semeval-2015-6 | ['twitter-sentiment-analysis'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.295368671417236, 3.6487231254577637] |
9d6f42cd-e9f0-4056-8c9c-81af77b09cd2 | asc-net-unsupervised-medical-anomaly | 2112.09135 | null | https://arxiv.org/abs/2112.09135v1 | https://arxiv.org/pdf/2112.09135v1.pdf | ASC-Net: Unsupervised Medical Anomaly Segmentation Using an Adversarial-based Selective Cutting Network | In this paper we consider the problem of unsupervised anomaly segmentation in medical images, which has attracted increasing attention in recent years due to the expensive pixel-level annotations from experts and the existence of a large amount of unannotated normal and abnormal image scans. We introduce a segmentation network that utilizes adversarial learning to partition an image into two cuts, with one of them falling into a reference distribution provided by the user. This Adversarial-based Selective Cutting network (ASC-Net) bridges the two domains of cluster-based deep segmentation and adversarial-based anomaly/novelty detection algorithms. Our ASC-Net learns from normal and abnormal medical scans to segment anomalies in medical scans without any masks for supervision. We evaluate this unsupervised anomly segmentation model on three public datasets, i.e., BraTS 2019 for brain tumor segmentation, LiTS for liver lesion segmentation, and MS-SEG 2015 for brain lesion segmentation, and also on a private dataset for brain tumor segmentation. Compared to existing methods, our model demonstrates tremendous performance gains in unsupervised anomaly segmentation tasks. Although there is still room to further improve performance compared to supervised learning algorithms, the promising experimental results and interesting observations shed light on building an unsupervised learning algorithm for medical anomaly identification using user-defined knowledge. | ['Yi Hong', 'Haibo Xu', 'Wenbo Sun', 'Raunak Dey'] | 2021-12-16 | null | null | null | null | ['brain-tumor-segmentation'] | ['medical'] | [ 5.19059539e-01 6.71162784e-01 -3.98341753e-02 -5.92807174e-01
-7.98116326e-01 -4.85739827e-01 2.75513351e-01 3.68684709e-01
-6.23196840e-01 2.18692958e-01 -2.10282028e-01 -4.82923210e-01
3.07576507e-01 -7.00546503e-01 -6.93430960e-01 -8.95963609e-01
-1.80948302e-01 8.10278118e-01 4.84323204e-01 5.51540330e-02
-6.34449422e-02 6.34977996e-01 -9.87245381e-01 2.36545086e-01
1.26339936e+00 1.21105325e+00 -3.60675782e-01 5.72646201e-01
-3.52822185e-01 6.68399096e-01 -3.87506634e-01 -4.43495810e-01
5.25434375e-01 -5.33573210e-01 -1.15256703e+00 3.63303006e-01
5.14329195e-01 -4.87759486e-02 -3.17954004e-01 1.51003754e+00
2.86400199e-01 -1.10647447e-01 7.64147222e-01 -1.38647914e+00
-3.37109715e-01 6.75255656e-01 -7.70252705e-01 5.62451005e-01
-2.16594040e-01 4.71178025e-01 7.78985739e-01 -4.63214606e-01
6.84006453e-01 6.56262755e-01 6.77455485e-01 6.63017809e-01
-1.12486851e+00 -5.44696331e-01 1.59785897e-01 2.06931144e-01
-1.10974729e+00 -5.95513880e-02 6.73898578e-01 -6.16992354e-01
4.66647506e-01 2.93381989e-01 6.53220117e-01 1.21249890e+00
2.89420694e-01 1.16289687e+00 7.08889365e-01 -2.86610693e-01
3.86662841e-01 -2.00022787e-01 2.96475857e-01 8.53097260e-01
2.03515515e-01 -3.42811905e-02 1.21460810e-01 -2.13933706e-01
5.84968626e-01 2.97696978e-01 -1.39777571e-01 -5.28368354e-01
-1.22648108e+00 8.74209166e-01 5.17235577e-01 3.44170392e-01
-4.40817118e-01 -1.20521918e-01 7.85966396e-01 4.11912739e-01
6.29961848e-01 5.43992519e-01 -3.10626447e-01 2.32858256e-01
-1.10591412e+00 -7.36930296e-02 6.27063692e-01 6.62106335e-01
3.41710329e-01 1.65049776e-01 -2.75873363e-01 7.13841379e-01
1.33166805e-01 5.27323633e-02 8.05414915e-01 -6.93844557e-01
1.01631440e-01 7.49532759e-01 -4.50783968e-01 -7.52548456e-01
-5.64648390e-01 -3.65228027e-01 -1.29559422e+00 2.69343615e-01
8.50272000e-01 -8.90495256e-02 -1.66939199e+00 1.37095010e+00
2.84248173e-01 4.56938297e-01 -9.16191787e-02 7.80038893e-01
6.21684134e-01 9.87913460e-02 -6.18944934e-04 1.06273979e-01
1.08405960e+00 -1.03752458e+00 -6.72198653e-01 -1.35732278e-01
8.23437333e-01 -3.41428876e-01 9.55456257e-01 5.78121781e-01
-9.12138939e-01 4.53610905e-02 -7.66612351e-01 2.34113216e-01
-5.30745864e-01 -4.06739771e-01 7.35523283e-01 8.68949115e-01
-9.84270215e-01 3.98178399e-01 -1.54846752e+00 -2.58895308e-01
1.30175447e+00 4.71313536e-01 -2.33950540e-01 -2.56585516e-02
-8.16934764e-01 3.92647207e-01 5.29125333e-01 -9.06604677e-02
-9.03977334e-01 -7.41941988e-01 -9.69883859e-01 -3.12261373e-01
5.84577382e-01 -5.02747715e-01 1.07203948e+00 -1.50079727e+00
-1.12729239e+00 1.34907401e+00 2.80554801e-01 -1.08958292e+00
8.42481256e-01 9.09199566e-02 -5.56978881e-01 4.38025475e-01
2.63857186e-01 7.86114573e-01 8.61595571e-01 -9.60794985e-01
-5.58380008e-01 -5.20423949e-01 -2.39241585e-01 -2.97827333e-01
-1.46331675e-02 -2.53663689e-01 -3.59687388e-01 -1.16176558e+00
5.40858448e-01 -9.38737571e-01 -8.96107376e-01 5.22275232e-02
-9.95386839e-01 -3.11838798e-02 9.68144178e-01 -6.16047204e-01
7.87505567e-01 -2.09424591e+00 7.87133053e-02 7.18387306e-01
4.25791562e-01 6.78723752e-02 -7.66982883e-02 -4.63674307e-01
-4.32672530e-01 2.81873375e-01 -1.13865471e+00 -1.23052485e-01
-4.19268966e-01 6.02879703e-01 -1.20942160e-01 7.28726208e-01
2.43242636e-01 9.76885200e-01 -9.50004876e-01 -6.20956779e-01
1.33377671e-01 -2.37425745e-01 -8.52160513e-01 1.43819198e-01
-3.87376845e-01 9.96686041e-01 -4.75653887e-01 1.23021936e+00
6.61377192e-01 -1.19375750e-01 -4.14686054e-01 1.98019102e-01
4.01061922e-01 -3.16716343e-01 -8.45523953e-01 2.00197339e+00
1.61057323e-01 2.94739634e-01 1.43037632e-01 -1.54995406e+00
3.92722219e-01 2.19531655e-01 1.14090157e+00 -5.24708748e-01
2.07866460e-01 2.26891041e-01 3.26846331e-01 -5.97340941e-01
1.05757594e-01 -8.20520893e-02 -5.81904761e-02 4.08883244e-01
2.24647269e-01 -3.27927947e-01 1.83558762e-01 4.59255666e-01
1.45953095e+00 -4.27599996e-01 9.48744081e-03 -2.92412221e-01
5.75406134e-01 2.40814939e-01 6.87736928e-01 9.69678044e-01
-6.31924033e-01 9.86473680e-01 7.00130105e-01 -6.22979403e-01
-7.49216199e-01 -1.34204078e+00 -4.77377653e-01 9.43534672e-01
7.79307187e-02 2.01366376e-02 -1.09511149e+00 -1.53359914e+00
-5.73499426e-02 6.27166092e-01 -8.74328375e-01 -2.39329264e-01
-5.27796566e-01 -1.05379152e+00 1.01236784e+00 4.98284280e-01
4.60179776e-01 -1.30831766e+00 -4.50197995e-01 -3.26887867e-03
-1.32485822e-01 -1.06942475e+00 -5.77729583e-01 3.11789215e-01
-8.79180849e-01 -1.49662697e+00 -6.73946798e-01 -6.70676887e-01
1.25536311e+00 -3.93820941e-01 1.16788805e+00 1.06733255e-01
-8.69476438e-01 7.10851610e-01 -4.26583052e-01 -7.37216830e-01
-5.92174888e-01 1.80151224e-01 -4.73869517e-02 1.71349987e-01
4.05607074e-01 -5.42142332e-01 -5.95050037e-01 3.83149207e-01
-1.34838438e+00 -2.73547798e-01 6.27332687e-01 9.39475894e-01
1.11602855e+00 -4.84798104e-02 3.20806146e-01 -1.58763027e+00
-1.21094575e-02 -8.35390449e-01 -3.89503539e-01 -1.68686435e-01
-5.50110459e-01 -1.80359423e-01 4.97996300e-01 -1.42190024e-01
-6.72634959e-01 2.62324423e-01 -5.07916272e-01 -5.15455544e-01
-7.08762765e-01 3.16896141e-01 -9.54472721e-02 -2.93840766e-02
8.09735119e-01 8.93289447e-02 2.37076566e-01 -2.74535656e-01
2.16714367e-01 2.13794202e-01 1.07169151e+00 -3.85855913e-01
8.69583547e-01 7.89245188e-01 -6.18143119e-02 -6.40886784e-01
-9.38407779e-01 -6.83690846e-01 -7.87340999e-01 1.27408057e-01
1.24783301e+00 -4.76622820e-01 -1.01113714e-01 8.50340128e-01
-6.25359952e-01 -4.36584920e-01 -1.05200398e+00 1.62469774e-01
-6.02999926e-01 5.46764672e-01 -6.80626750e-01 -1.38483867e-01
-3.65907997e-01 -1.41318667e+00 8.25640023e-01 2.58021951e-02
-1.20484367e-01 -9.84366715e-01 -9.94958505e-02 3.55145007e-01
3.14827263e-01 8.49565506e-01 9.13331509e-01 -1.44038570e+00
-5.17509878e-01 -4.00696933e-01 -2.03297082e-02 5.88449538e-01
1.81651637e-01 -1.83334008e-01 -7.98094153e-01 -2.46156201e-01
2.38499455e-02 -2.94665396e-01 1.16194546e+00 6.60787523e-01
2.15670896e+00 -1.48447916e-01 -4.05551612e-01 1.13695419e+00
9.60685611e-01 1.75271407e-01 7.62982965e-01 3.97268981e-01
1.02767324e+00 3.01286280e-01 3.13565135e-01 1.66885361e-01
-4.12568823e-02 6.56926110e-02 9.37875152e-01 -5.31293571e-01
2.24609882e-01 2.74046183e-01 -7.46771991e-02 4.22043562e-01
6.64518308e-03 -1.89984262e-01 -1.34494448e+00 7.39282668e-01
-1.62545455e+00 -7.53015101e-01 -2.09677175e-01 1.92415833e+00
6.91780090e-01 3.41401815e-01 1.32348776e-01 1.69120252e-01
6.26416326e-01 6.74305484e-02 -1.04722965e+00 -3.46625090e-01
-4.64319810e-02 4.71256703e-01 6.59209788e-01 1.14117963e-02
-1.59183609e+00 8.04921269e-01 6.10882521e+00 8.60641360e-01
-1.05481851e+00 1.99971423e-01 1.26142037e+00 5.61428741e-02
-1.30901940e-03 -5.01052797e-01 -1.60647873e-02 6.48179233e-01
6.90158963e-01 1.15736790e-01 -1.27988219e-01 9.75061834e-01
-1.55429363e-01 -1.17197581e-01 -1.15936089e+00 7.19465315e-01
3.53119858e-02 -1.16083646e+00 2.24971980e-01 6.65795654e-02
9.60199177e-01 3.90043557e-01 1.67919010e-01 1.89999044e-01
4.04142618e-01 -1.25814188e+00 2.69562960e-01 4.41956758e-01
6.91312015e-01 -9.28710580e-01 1.05827785e+00 2.32960686e-01
-5.64004719e-01 9.02178362e-02 -9.60909054e-02 5.97260475e-01
-4.31653708e-02 8.55299592e-01 -7.42918432e-01 5.31236887e-01
7.76260853e-01 7.36901820e-01 -6.92750990e-01 1.21808529e+00
-9.59639102e-02 1.10990095e+00 -4.67579931e-01 8.95161152e-01
5.56634843e-01 -2.33104676e-01 8.48474383e-01 1.27642155e+00
-4.00855765e-03 9.54023562e-03 4.84904617e-01 9.11366343e-01
-3.17855239e-01 4.99309860e-02 -5.52275419e-01 2.54558697e-02
-2.59003669e-01 1.28788507e+00 -1.33488369e+00 -3.00484240e-01
-3.18157017e-01 1.24412000e+00 -2.49366701e-01 2.28071406e-01
-8.38571250e-01 -3.19373608e-01 5.65851867e-01 2.27419689e-01
6.70373291e-02 2.35789418e-01 -6.62703216e-01 -1.09593236e+00
-2.42323175e-01 -9.50814009e-01 9.52972293e-01 -1.31656462e-02
-1.54920733e+00 5.01012802e-01 -1.40666291e-01 -1.10372663e+00
-2.86653936e-01 -5.56599081e-01 -1.00415421e+00 3.46636266e-01
-1.32463825e+00 -1.00503206e+00 -3.23225498e-01 8.16477895e-01
6.22635961e-01 -4.26887691e-01 7.40845919e-01 1.95969895e-01
-6.85523331e-01 9.28749681e-01 -2.14087721e-02 7.62516081e-01
6.30310357e-01 -1.77752566e+00 5.00279188e-01 1.14426136e+00
2.54760176e-01 -1.08337566e-01 4.68135566e-01 -5.47410131e-01
-6.52389169e-01 -1.41161847e+00 -3.96586098e-02 -5.59876084e-01
6.23617470e-01 -7.55423158e-02 -1.37426877e+00 1.00762606e+00
3.05441450e-02 7.74253607e-01 9.21680808e-01 -4.02618110e-01
-1.82116330e-01 2.88951129e-01 -1.74060953e+00 3.99199188e-01
1.04480493e+00 1.74753217e-03 -5.76999128e-01 6.27933145e-01
6.70021951e-01 -8.90847623e-01 -6.89391613e-01 6.13794982e-01
-5.00972942e-02 -9.34462488e-01 9.63511407e-01 -7.62976646e-01
5.54830134e-01 -3.11051030e-02 2.24324241e-01 -1.41296959e+00
1.50825113e-01 -4.34114397e-01 8.34319517e-02 7.45458782e-01
3.37571859e-01 -8.40923429e-01 1.12903821e+00 7.09994137e-01
-6.58410966e-01 -9.20706749e-01 -9.59401667e-01 -5.72614372e-01
4.11589593e-01 -6.82220638e-01 4.19136286e-01 1.32222414e+00
-3.91721070e-01 -4.66150850e-01 3.54672581e-01 2.53303587e-01
6.67408407e-01 -3.22742105e-01 5.32866597e-01 -1.16994596e+00
-7.45026618e-02 -7.44125009e-01 -8.82656991e-01 -4.01501447e-01
2.89261162e-01 -1.25227845e+00 -1.45727554e-02 -1.12498415e+00
-2.04920396e-02 -4.62846786e-01 -7.17465401e-01 7.09828198e-01
-1.11693949e-01 5.21764219e-01 -4.17578578e-01 1.22162066e-01
-7.37764835e-01 1.87147290e-01 1.18435192e+00 -5.44949412e-01
-2.46481404e-01 3.64708602e-01 -5.15613616e-01 1.36752486e+00
7.83218503e-01 -6.68340921e-01 -7.31355250e-02 -9.17558596e-02
-3.33313227e-01 -3.83267254e-01 4.78140652e-01 -9.94374692e-01
3.30651194e-01 -7.09660770e-03 3.50236952e-01 -6.22811913e-01
-5.09333968e-01 -9.55482185e-01 -4.75526392e-01 5.10774374e-01
-3.47931445e-01 -1.26410738e-01 3.59851748e-01 7.48758972e-01
-3.31632942e-01 -2.10732222e-01 1.01755047e+00 -3.55304211e-01
-8.08737755e-01 7.60337055e-01 -6.25400066e-01 4.98065859e-01
1.28965688e+00 -3.92335027e-01 3.81960608e-02 -6.63569346e-02
-1.28516102e+00 4.85928178e-01 3.39990795e-01 2.63375729e-01
4.71772641e-01 -1.02063835e+00 -6.75894618e-01 3.45767945e-01
2.71738291e-01 8.14842999e-01 3.06790084e-01 1.10336542e+00
-7.60450065e-01 -1.40677691e-01 -3.00835639e-01 -1.03574598e+00
-8.79016042e-01 3.85833204e-01 6.28317416e-01 -6.00596726e-01
-1.04574430e+00 8.90676677e-01 3.01490724e-01 -4.90636021e-01
3.87385875e-01 -5.81744432e-01 9.20497254e-02 -1.60814092e-01
2.33592659e-01 2.11427316e-01 2.53157616e-01 -5.13400078e-01
-2.61692375e-01 4.97436747e-02 -5.86190283e-01 2.60599434e-01
1.12351727e+00 2.43733093e-01 -2.73313552e-01 2.14203492e-01
9.66506839e-01 -1.16081357e-01 -9.25667882e-01 -2.91650444e-01
3.38113368e-01 -2.59645760e-01 3.29119042e-02 -9.14147139e-01
-1.85593045e+00 5.52219331e-01 8.52976739e-01 3.36223394e-01
1.18903518e+00 1.29948482e-01 1.11091483e+00 5.22106588e-01
5.67353144e-03 -1.14600837e+00 2.31464088e-01 4.20565218e-01
5.03018498e-01 -1.63546467e+00 -3.27528536e-01 -5.08050919e-01
-4.98368293e-01 8.36136341e-01 8.13685596e-01 -2.51681089e-01
6.35814369e-01 4.04426426e-01 2.65098929e-01 -5.07216752e-01
7.64659718e-02 -1.84779391e-01 3.28658462e-01 7.08426893e-01
-1.48860607e-02 1.66599989e-01 2.00566158e-01 7.01561630e-01
-1.21686012e-01 -3.78073692e-01 5.35522103e-01 8.95449162e-01
-2.08553076e-01 -9.86399293e-01 -3.21249902e-01 1.12713766e+00
-1.04970860e+00 3.79828662e-02 -3.61985862e-01 8.85566950e-01
3.37445825e-01 4.39446747e-01 4.73805219e-01 1.57000661e-01
3.24707687e-01 3.50210309e-01 1.24285229e-01 -7.47957170e-01
-6.50395155e-01 -2.04451948e-01 -4.49434429e-01 -7.94336677e-01
-1.59881026e-01 -6.12212121e-01 -1.53365850e+00 2.63831615e-01
-1.09863251e-01 2.66097859e-03 2.83562660e-01 1.01794565e+00
8.23950917e-02 7.92531133e-01 4.94368434e-01 -5.68275809e-01
-1.48602784e-01 -7.45474756e-01 -6.75355077e-01 9.00090456e-01
3.78336459e-01 -3.52819234e-01 -4.54504788e-01 8.42707008e-02] | [14.513944625854492, -2.0344181060791016] |
e866a317-15d4-4f11-ab75-c1533656d923 | osan-a-one-stage-alignment-network-to-unify | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Liu_OSAN_A_One-Stage_Alignment_Network_To_Unify_Multimodal_Alignment_and_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Liu_OSAN_A_One-Stage_Alignment_Network_To_Unify_Multimodal_Alignment_and_CVPR_2023_paper.pdf | OSAN: A One-Stage Alignment Network To Unify Multimodal Alignment and Unsupervised Domain Adaptation | Extending from unimodal to multimodal is a critical challenge for unsupervised domain adaptation (UDA). Two major problems emerge in unsupervised multimodal domain adaptation: domain adaptation and modality alignment. An intuitive way to handle these two problems is to fulfill these tasks in two separate stages: aligning modalities followed by domain adaptation, or vice versa. However, domains and modalities are not associated in most existing two-stage studies, and the relationship between them is not leveraged which can provide complementary information to each other. In this paper, we unify these two stages into one to align domains and modalities simultaneously. In our model, a tensor-based alignment module (TAL) is presented to explore the relationship between domains and modalities. By this means, domains and modalities can interact sufficiently and guide them to utilize complementary information for better results. Furthermore, to establish a bridge between domains, a dynamic domain generator (DDG) module is proposed to build transitional samples by mixing the shared information of two domains in a self-supervised manner, which helps our model learn a domain-invariant common representation space. Extensive experiments prove that our method can achieve superior performance in two real-world applications. The code will be publicly available. | ['Bo Ren', 'Haoyuan Peng', 'Chen Lin', 'Di Yin', 'Changchong Lu', 'Lingfeng Qiao', 'Ye Liu'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['unsupervised-domain-adaptation'] | ['methodology'] | [ 3.98643196e-01 -2.29484200e-01 -6.08698130e-01 -5.04748702e-01
-5.72614849e-01 -8.24836433e-01 7.88948417e-01 -9.43172947e-02
-1.59609437e-01 6.17711246e-01 4.47383165e-01 8.22103322e-02
1.30929232e-01 -5.56122124e-01 -3.35585415e-01 -7.16038644e-01
3.49671602e-01 3.60847712e-01 1.26566738e-01 -4.28109616e-01
1.02255113e-01 -6.10611355e-03 -1.29301798e+00 2.92991698e-01
1.30088449e+00 7.52845466e-01 2.06856191e-01 1.20396182e-01
-4.98040646e-01 3.69845897e-01 -7.76852071e-02 -3.88731152e-01
1.55988917e-01 -8.07163358e-01 -9.65356827e-01 3.69862229e-01
-6.41843975e-02 -2.77800113e-01 -2.36859396e-01 1.15354872e+00
5.11057556e-01 2.45810643e-01 8.99430335e-01 -1.36404872e+00
-5.44657052e-01 6.08799577e-01 -7.23756313e-01 -2.53693342e-01
5.85666299e-01 9.74802598e-02 9.76198792e-01 -7.57338166e-01
6.32401824e-01 1.25258780e+00 1.79292038e-01 7.52822042e-01
-1.42839026e+00 -7.67187476e-01 3.61102164e-01 1.95622042e-01
-1.12631643e+00 -5.06183147e-01 1.35465074e+00 -5.20144582e-01
2.22350255e-01 -1.84826404e-02 5.25774717e-01 1.41948736e+00
-3.84088486e-01 1.01705706e+00 1.21481287e+00 -5.19471169e-01
2.67271996e-01 4.41046953e-01 4.18969952e-02 1.71486750e-01
-1.32034803e-02 -8.06391686e-02 -8.00477445e-01 -1.74928568e-02
6.82618439e-01 1.29185230e-01 -1.83439285e-01 -6.78759515e-01
-1.57722402e+00 6.55532897e-01 2.93295771e-01 4.41845715e-01
-2.59215027e-01 -7.37302065e-01 5.32031834e-01 4.19095755e-01
1.35679066e-01 3.98595363e-01 -1.64683893e-01 -1.15039714e-01
-4.47707713e-01 -2.89951414e-02 5.54971635e-01 1.09638870e+00
9.14964139e-01 -2.82366484e-01 -1.32257506e-01 1.23610091e+00
5.45444846e-01 4.95355010e-01 7.89606869e-01 -5.11404991e-01
7.28774428e-01 1.03615975e+00 9.76332370e-03 -9.94672060e-01
-1.98384076e-01 3.16389017e-02 -9.96237874e-01 -3.00561160e-01
5.16079545e-01 -1.00856595e-01 -8.28502417e-01 2.09797883e+00
4.79427487e-01 7.89031684e-02 2.83367634e-01 1.07552731e+00
8.57161045e-01 4.20049220e-01 2.68764764e-01 -5.31776622e-02
1.51708543e+00 -7.84874022e-01 -7.68172622e-01 -4.54551011e-01
3.79514784e-01 -8.72721195e-01 1.19223797e+00 2.37343069e-02
-7.59033024e-01 -6.57950759e-01 -1.08640051e+00 -8.68386403e-02
-3.05028707e-01 8.13879147e-02 5.01952171e-01 3.31098288e-01
-3.99268091e-01 1.77732825e-01 -6.09283268e-01 -6.54187202e-01
1.75011083e-01 1.75607085e-01 -5.28518438e-01 -3.25337023e-01
-1.55322611e+00 6.25568211e-01 6.36119187e-01 -1.06362179e-01
-5.54756522e-01 -1.67939514e-01 -1.07077634e+00 -3.74413699e-01
2.56862462e-01 -7.53288150e-01 1.14677155e+00 -1.26887274e+00
-1.64370894e+00 9.66088653e-01 -3.41494799e-01 5.70348985e-02
2.81309128e-01 9.08475742e-02 -6.51820004e-01 1.24586277e-01
2.19281495e-01 6.67029202e-01 8.45123649e-01 -1.44850552e+00
-7.11693227e-01 -4.49200273e-01 7.89572150e-02 6.27033234e-01
-9.12954688e-01 -1.52531415e-01 -7.88703382e-01 -6.03720725e-01
3.77104402e-01 -9.02304113e-01 -4.60596867e-02 -2.92324871e-01
-4.06065643e-01 -1.33504450e-01 6.25241697e-01 -6.16081774e-01
1.15138626e+00 -2.28240013e+00 7.24478602e-01 3.68267328e-01
1.70355767e-01 5.46207204e-02 -3.80772471e-01 4.68939573e-01
-1.16899289e-01 -1.90483138e-01 -4.58918303e-01 -4.48996395e-01
1.35209188e-01 3.70831192e-01 -3.53062868e-01 1.77906215e-01
2.77514189e-01 6.91972435e-01 -1.09173131e+00 -6.53146982e-01
-1.47917373e-02 1.66703820e-01 -3.33379626e-01 5.65460145e-01
-7.34818429e-02 1.13911664e+00 -6.06307924e-01 9.23785090e-01
9.17711735e-01 -2.54921377e-01 6.33082867e-01 -5.00737727e-01
2.09840864e-01 1.30570695e-01 -1.33347309e+00 2.06383204e+00
-2.85698026e-01 3.23370397e-01 -6.36650994e-02 -1.32234347e+00
1.14342499e+00 2.80554324e-01 5.45701742e-01 -8.53391886e-01
1.65068343e-01 2.56639510e-01 -2.60357410e-02 -6.28956258e-01
4.13845360e-01 -4.35256213e-01 -3.16371709e-01 5.35868704e-01
3.35426301e-01 9.59419757e-02 2.73092240e-01 2.14354292e-01
6.33313298e-01 2.42198601e-01 3.10070723e-01 1.62570342e-01
7.42600799e-01 6.19483329e-02 8.06301832e-01 2.96487957e-01
-3.16138625e-01 6.87192559e-01 4.23284948e-01 -1.51390024e-02
-9.51529920e-01 -1.04815209e+00 -5.41546047e-02 1.23354530e+00
6.12711012e-01 -2.17835903e-01 -5.26106238e-01 -8.02933574e-01
-1.99134797e-01 2.97585338e-01 -6.49783015e-01 -3.94040704e-01
-3.64238471e-01 -5.17850935e-01 3.61849636e-01 5.57852626e-01
6.79339290e-01 -7.63007581e-01 5.54812662e-02 7.28350282e-02
-6.35683477e-01 -1.29207718e+00 -5.30245423e-01 9.68681462e-03
-9.10562396e-01 -9.86862540e-01 -1.04879951e+00 -9.75790620e-01
9.11175609e-01 5.24461448e-01 9.95712459e-01 -2.73107201e-01
5.54687619e-01 5.02320111e-01 -5.95564187e-01 -3.99847552e-02
-3.97956878e-01 2.78712064e-01 2.40878627e-01 3.73770833e-01
5.74210286e-01 -7.79522002e-01 -4.52909291e-01 6.02821827e-01
-1.16793072e+00 3.36315483e-01 6.96599960e-01 1.12798595e+00
5.48725069e-01 -1.89163730e-01 5.22574484e-01 -8.62372160e-01
7.27399051e-01 -6.63246393e-01 -1.64159566e-01 3.82675380e-01
-2.99871236e-01 1.39385432e-01 7.92636812e-01 -8.65318716e-01
-1.26153111e+00 3.85617703e-01 1.88834921e-01 -5.16022921e-01
-4.14179891e-01 8.26721668e-01 -7.48510063e-01 1.91444635e-01
4.16097105e-01 6.39118552e-01 2.85358101e-01 -5.27873099e-01
6.28389120e-01 7.78736055e-01 5.47631145e-01 -8.74022424e-01
1.13146925e+00 3.68728399e-01 -3.27085465e-01 -4.85407412e-01
-5.55712938e-01 -6.10523760e-01 -1.02298343e+00 -1.63802221e-01
6.45361900e-01 -1.00838792e+00 -3.42060864e-01 4.44868326e-01
-9.15963173e-01 -1.47200804e-02 -9.74038988e-02 5.68560719e-01
-3.54406029e-01 6.80565059e-01 -2.79398441e-01 -4.71303850e-01
1.11875646e-01 -1.14338028e+00 8.34227026e-01 6.86193585e-01
-1.16603866e-01 -1.21364689e+00 3.75351012e-01 4.31730181e-01
2.66725242e-01 3.43055427e-02 7.83189774e-01 -9.04609263e-01
-1.18541114e-01 -5.26047312e-02 -3.15897346e-01 3.48494470e-01
5.36198497e-01 -2.11155921e-01 -7.99733877e-01 -3.50914598e-01
-1.92178458e-01 -4.14921582e-01 5.31346977e-01 -1.53783739e-01
6.98179007e-01 -7.83164650e-02 -3.93762797e-01 5.22679925e-01
9.88747239e-01 1.88384175e-01 4.66262311e-01 4.83125210e-01
8.36545825e-01 8.10168505e-01 8.48190010e-01 3.86762440e-01
7.30464041e-01 7.38924623e-01 6.73455074e-02 -1.59165204e-01
-3.61051448e-02 -4.92932975e-01 4.60933954e-01 1.11416852e+00
-1.16989408e-02 1.78901792e-01 -1.06551170e+00 5.26690960e-01
-2.01122737e+00 -7.38597572e-01 1.86644241e-01 2.20277476e+00
1.15783370e+00 -9.25188437e-02 4.77759719e-01 -1.25209093e-01
8.64736319e-01 -8.68333504e-02 -7.26687074e-01 1.80366352e-01
-1.48541242e-01 -3.10592830e-01 -9.98762771e-02 1.13084704e-01
-1.15535426e+00 7.82856822e-01 5.65452528e+00 6.72864199e-01
-1.24630725e+00 -6.25906140e-02 2.74458975e-01 3.17406863e-01
-6.34530544e-01 1.72873154e-01 -5.11779904e-01 7.47466385e-01
4.72934157e-01 -2.55098403e-01 4.47657466e-01 7.70738602e-01
-4.43785787e-02 1.79986387e-01 -1.24382317e+00 1.03569257e+00
1.15627060e-02 -8.12771142e-01 2.05005467e-01 -4.23986651e-02
6.83646083e-01 -4.76661742e-01 -2.89001931e-02 5.60912311e-01
2.30721563e-01 -5.98268449e-01 3.19175571e-01 5.63215733e-01
7.96140075e-01 -5.98585725e-01 6.59891963e-01 3.31706852e-01
-1.37119591e+00 -1.59622505e-02 -2.41664797e-01 2.09484652e-01
7.05075040e-02 3.26091409e-01 -6.83677495e-01 1.03627360e+00
4.68563139e-01 9.77536857e-01 -5.61159074e-01 8.44118118e-01
-1.64309204e-01 4.80545104e-01 -1.00864314e-01 2.37186506e-01
-1.12432770e-01 -4.48752940e-01 6.28893793e-01 1.03047562e+00
1.57362759e-01 5.53099290e-02 3.36940795e-01 7.65879750e-01
-5.58324866e-02 2.33422086e-01 -6.82443440e-01 -3.15471828e-01
7.82705545e-01 1.21834528e+00 -4.67954189e-01 -3.88961852e-01
-6.29272938e-01 1.02180398e+00 2.59318411e-01 5.19059181e-01
-7.81780720e-01 -4.43101138e-01 6.45525515e-01 -1.03151828e-01
-5.72335571e-02 -1.96668759e-01 -3.21208835e-01 -1.76165080e+00
-1.87644511e-02 -1.28471935e+00 6.73516273e-01 -4.45586652e-01
-1.84747517e+00 3.95363390e-01 8.30581561e-02 -2.06044745e+00
-2.74152964e-01 -3.89669389e-01 -4.32324857e-01 1.06698215e+00
-1.66126025e+00 -1.51607978e+00 -6.03375375e-01 1.02996027e+00
3.20667952e-01 -4.19751346e-01 7.62860715e-01 4.10208791e-01
-8.05650413e-01 8.49544704e-01 2.41171077e-01 3.90753597e-01
1.19996309e+00 -1.14474475e+00 -5.23723662e-02 9.64659989e-01
-9.11112130e-02 9.69404936e-01 3.94753128e-01 -4.94940907e-01
-1.57052910e+00 -7.66226172e-01 4.39939499e-01 -2.30485365e-01
6.01060629e-01 -2.29000866e-01 -1.22446465e+00 3.98681641e-01
2.67236859e-01 -2.68284798e-01 1.09862053e+00 4.03624922e-01
-6.73945665e-01 -2.77845114e-01 -8.33923638e-01 6.98360085e-01
9.16515887e-01 -6.37574375e-01 -8.06026399e-01 1.07633229e-03
4.83314276e-01 -3.99709791e-01 -1.00288355e+00 4.54472363e-01
4.54152226e-01 -8.21183085e-01 8.68299961e-01 -6.38286650e-01
7.13407278e-01 -5.05578279e-01 -2.52411813e-01 -1.44457006e+00
-1.25398442e-01 -4.72241342e-01 -1.69534549e-01 1.78933179e+00
2.00429186e-01 -7.38511741e-01 4.86405522e-01 6.02703989e-01
1.20637007e-01 -3.71283442e-01 -6.67627871e-01 -6.88662052e-01
1.06412269e-01 -2.26495251e-01 8.45724225e-01 1.56846130e+00
4.58314478e-01 6.54646635e-01 -4.15764779e-01 8.08781981e-02
4.59493548e-01 3.88337076e-01 1.09368598e+00 -1.12248361e+00
-1.90262288e-01 -5.72419345e-01 -1.28076613e-01 -1.40492415e+00
1.56554729e-01 -9.74079549e-01 2.08268110e-02 -1.31257987e+00
6.05404198e-01 -5.34044981e-01 -5.01504302e-01 6.45333052e-01
-4.19318497e-01 -1.21663168e-01 1.13062240e-01 6.75306857e-01
-5.63815415e-01 9.00162339e-01 1.39605594e+00 -3.26508820e-01
-4.50497001e-01 -1.57996073e-01 -1.11397851e+00 5.75899005e-01
6.80800974e-01 -1.74225479e-01 -6.55485213e-01 -3.45961452e-01
-1.58743501e-01 1.39372930e-01 3.37748863e-02 -9.14978266e-01
2.91557670e-01 -5.46045184e-01 3.32702875e-01 -3.48903298e-01
1.70968413e-01 -8.96884263e-01 -1.46147043e-01 -6.94461763e-02
-3.43278915e-01 -1.65477291e-01 1.41892329e-01 5.38357079e-01
-6.79882944e-01 -3.63547541e-02 7.39912331e-01 1.09012946e-01
-1.09258342e+00 2.27738693e-01 2.40949653e-02 -3.75863235e-03
8.72361362e-01 -3.42790961e-01 -2.13011995e-01 -4.54755157e-01
-6.59373701e-01 5.37009060e-01 5.78833759e-01 6.16273999e-01
5.12311041e-01 -1.79179537e+00 -6.01175129e-01 2.69217908e-01
6.15758121e-01 -3.09815556e-02 4.63745266e-01 9.60483134e-01
2.13250205e-01 -5.04455110e-03 -5.04106522e-01 -7.97489166e-01
-1.11753106e+00 4.76626277e-01 1.97433561e-01 -2.86638469e-01
-4.67654504e-02 4.90471393e-01 2.11308092e-01 -8.10456395e-01
1.96784750e-01 2.14296490e-01 -5.99891484e-01 4.82195586e-01
5.45379281e-01 -9.85125154e-02 -3.79920840e-01 -7.50059783e-01
-4.26297218e-01 6.42632782e-01 -2.13730827e-01 -2.46551663e-01
1.13612258e+00 -5.15292168e-01 -1.70196325e-01 5.73377609e-01
1.05554783e+00 -2.70043146e-02 -1.20879531e+00 -7.83033669e-01
-2.26996168e-01 -5.10593891e-01 -4.68308896e-01 -7.06695855e-01
-8.01851034e-01 1.02727914e+00 6.08761847e-01 2.28518605e-01
1.60573936e+00 9.39135924e-02 7.66345620e-01 2.03873277e-01
3.12255900e-02 -1.22687376e+00 4.63341296e-01 5.72214067e-01
6.78867221e-01 -1.52376926e+00 -1.17773741e-01 -3.64700198e-01
-1.18586648e+00 1.21350145e+00 9.85038280e-01 3.99420768e-01
2.72797257e-01 -2.47396320e-01 1.76024660e-01 2.62554616e-01
-3.44430238e-01 -4.92708206e-01 6.22744739e-01 9.86810565e-01
5.29608846e-01 -4.57764156e-02 -2.83960700e-01 1.00794542e+00
1.77279234e-01 -2.06104636e-01 5.91027737e-02 1.04615974e+00
-1.48214355e-01 -1.73234081e+00 -5.36599636e-01 -3.33833992e-02
2.19617128e-01 2.18780994e-01 -5.14079809e-01 6.70690536e-01
2.57106731e-03 8.49851370e-01 -2.22027481e-01 -9.87388492e-01
3.86506021e-01 1.51439250e-01 2.32156262e-01 -5.25252640e-01
-1.32406224e-02 2.09505692e-01 -9.49780494e-02 -1.79895461e-01
-6.85737550e-01 -7.36492515e-01 -1.01092875e+00 -1.75479993e-01
-8.54704902e-02 -4.32115830e-02 3.35222244e-01 1.09846210e+00
3.71422976e-01 3.41635346e-01 1.03389740e+00 -7.96293676e-01
-4.68244255e-01 -8.83559108e-01 -4.06214029e-01 7.76070416e-01
7.23024458e-02 -9.28535283e-01 -1.22845091e-01 3.06355387e-01] | [10.386143684387207, 3.145720958709717] |
e07b5f69-6288-468d-aad9-31e605ea508f | understanding-the-semantics-of-narratives-of | null | null | https://aclanthology.org/W17-1801 | https://aclanthology.org/W17-1801.pdf | Understanding the Semantics of Narratives of Interpersonal Violence through Reader Annotations and Physiological Reactions | Interpersonal violence (IPV) is a prominent sociological problem that affects people of all demographic backgrounds. By analyzing how readers interpret, perceive, and react to experiences narrated in social media posts, we explore an understudied source for discourse about abuse. We asked readers to annotate Reddit posts about relationships with vs. without IPV for stakeholder roles and emotion, while measuring their galvanic skin response (GSR), pulse, and facial expression. We map annotations to coreference resolution output to obtain a labeled coreference chain for stakeholders in texts, and apply automated semantic role labeling for analyzing IPV discourse. Findings provide insights into how readers process roles and emotion in narratives. For example, abusers tend to be linked with violent actions and certain affect states. We train classifiers to predict stakeholder categories of coreference chains. We also find that subjects{'} GSR noticeably changed for IPV texts, suggesting that co-collected measurement-based data about annotators can be used to support text annotation. | ['er', 'Elizabeth A. Pruett', 'Alex Calderwood', 'Christopher Homan', 'Raymond Ptucha', 'Cecilia Ovesdotter Alm'] | 2017-04-01 | null | null | null | ws-2017-4 | ['text-annotation'] | ['natural-language-processing'] | [ 2.93124795e-01 7.63502479e-01 -7.08073378e-01 -4.52266425e-01
-5.44505000e-01 -1.11520231e+00 8.61757576e-01 8.84207606e-01
-3.57095480e-01 5.86039007e-01 1.67074263e+00 8.31890404e-02
1.29946291e-01 -5.34494519e-01 8.38025939e-03 -3.32228720e-01
3.09087068e-01 1.69235915e-01 -4.35845047e-01 -6.79536700e-01
3.49790573e-01 4.70708638e-01 -8.73784602e-01 9.58405554e-01
4.56553847e-01 3.83061618e-01 -7.64009655e-01 3.34863126e-01
-2.86936283e-01 1.75061929e+00 -8.96400094e-01 -9.77216482e-01
-4.68740851e-01 -6.72253430e-01 -1.36461258e+00 -3.41197997e-01
4.54685032e-01 -2.24242255e-01 -4.10382599e-01 9.60167706e-01
4.52681005e-01 4.75968532e-02 5.47012925e-01 -1.23369920e+00
-5.27817070e-01 1.10319984e+00 -5.87209702e-01 6.85707867e-01
1.13985407e+00 -2.03603402e-01 9.89705980e-01 -1.95558757e-01
1.28543615e+00 1.77830815e+00 8.74711215e-01 9.62992430e-01
-1.12286329e+00 -7.60618806e-01 -1.85596704e-01 2.17359364e-01
-8.25506091e-01 -6.70724392e-01 8.77601564e-01 -9.11761224e-01
7.11224198e-01 7.85655975e-01 7.85504997e-01 1.90942121e+00
-4.49685007e-01 3.91962796e-01 9.29034591e-01 6.41800696e-03
-1.58030212e-01 2.27864191e-01 5.35150170e-01 3.67581546e-01
-8.02865401e-02 -7.27629423e-01 -7.78636634e-01 -9.28785384e-01
5.72156496e-02 -2.27399349e-01 -2.00392559e-01 7.95922399e-01
-8.19178879e-01 1.16695583e+00 4.80106264e-01 6.20479226e-01
-2.83096641e-01 -8.61833170e-02 9.75239396e-01 2.20511571e-01
8.62302542e-01 7.44807124e-01 1.51720539e-01 -7.04878271e-01
-2.20955893e-01 2.70090371e-01 1.01997244e+00 3.08733553e-01
2.48638138e-01 -7.92425513e-01 -4.76285458e-01 1.18889070e+00
-4.47897203e-02 4.39254820e-01 -1.92682981e-01 -1.54016483e+00
5.61479092e-01 8.44414830e-01 8.13321397e-02 -1.76104212e+00
-8.85062635e-01 2.52179146e-01 -4.20691371e-01 -5.51534355e-01
4.45385814e-01 -5.17367005e-01 2.25887060e-01 1.90794814e+00
4.64170814e-01 -2.34949872e-01 2.59587355e-02 9.80139613e-01
1.19144642e+00 5.46807110e-01 7.81526864e-01 -6.97052658e-01
1.74136579e+00 -1.25110552e-01 -1.32851934e+00 -2.04425424e-01
9.77315187e-01 -5.42854965e-01 5.29968798e-01 -2.15326212e-02
-1.12443399e+00 4.51372504e-01 -5.41686654e-01 -3.25426400e-01
8.45278800e-02 -7.95263231e-01 3.58323187e-01 5.29835939e-01
-5.69341362e-01 5.84042668e-01 -3.09386015e-01 -8.48455071e-01
6.48995876e-01 -2.83881485e-01 -5.40478647e-01 4.60817724e-01
-1.47949445e+00 1.23654854e+00 -4.90471832e-02 -2.59479225e-01
-1.13262922e-01 -7.17329383e-01 -8.22799623e-01 -5.68301737e-01
2.74287045e-01 -1.54004216e-01 1.22227561e+00 -1.22425580e+00
-8.47126424e-01 1.65553331e+00 -3.25212181e-01 -1.96147546e-01
2.61739194e-01 -3.25562268e-01 -7.06657171e-01 6.98067546e-01
3.48193884e-01 2.17197672e-01 4.27374572e-01 -1.14291620e+00
-3.71695548e-01 -6.10299587e-01 4.15485501e-01 4.35746729e-01
-7.01293826e-01 1.34484565e+00 5.30298769e-01 -3.99232149e-01
-9.72736627e-02 -6.12553596e-01 3.17458481e-01 7.95746222e-02
-4.91668105e-01 -5.55266678e-01 9.91840720e-01 -1.06918097e+00
1.58787084e+00 -2.12424517e+00 1.61573306e-01 6.37208074e-02
8.14904392e-01 -1.47137016e-01 3.26096475e-01 8.07428300e-01
4.33670096e-02 5.65585375e-01 1.61269922e-02 -1.09186739e-01
-2.10956812e-01 3.30252647e-01 -3.72319579e-01 8.90825212e-01
-1.56024560e-01 6.35458469e-01 -1.40317559e+00 -7.65153825e-01
-2.03241065e-01 4.20960337e-01 -3.47145796e-01 8.18163157e-02
1.23467240e-02 5.54988444e-01 -4.87487227e-01 5.38895488e-01
2.35028639e-01 -1.11299390e-02 5.13191342e-01 -3.99139851e-01
-3.36644381e-01 4.44578439e-01 -1.01275638e-01 1.06689250e+00
-1.19504847e-01 1.18505812e+00 6.57101810e-01 -6.82213306e-01
8.74812901e-01 5.25972962e-01 4.57384467e-01 -6.75326824e-01
6.27711236e-01 -1.89243540e-01 1.47593254e-02 -1.07638705e+00
2.87426174e-01 -3.51453692e-01 -7.88401067e-01 6.98210418e-01
-4.63867307e-01 1.94150493e-01 -2.83795506e-01 4.49862957e-01
1.12080932e+00 -4.60864991e-01 3.56407285e-01 -2.11694747e-01
1.83126912e-01 2.06775054e-01 5.28722048e-01 3.26070726e-01
-4.34586287e-01 2.45077491e-01 1.28317404e+00 -4.27171499e-01
-7.92874098e-01 -6.66281104e-01 -2.18467638e-01 1.51608360e+00
3.29054475e-01 -3.47939700e-01 -8.72231245e-01 -3.05723250e-01
-3.35716873e-01 1.02928889e+00 -8.78555536e-01 -1.37012154e-01
-7.83691704e-01 -5.42187035e-01 1.05731010e+00 3.22190583e-01
2.84523129e-01 -1.11418462e+00 -8.88272762e-01 1.74492784e-02
-9.11357284e-01 -1.17961645e+00 4.10897359e-02 -5.33833027e-01
4.90564061e-03 -1.45188355e+00 -5.70870824e-02 -3.88974458e-01
3.80985051e-01 -1.48467988e-01 8.20701063e-01 3.72460485e-01
-9.37234610e-02 4.22907233e-01 -6.40133321e-01 -2.99953371e-01
-7.79460013e-01 -8.52035359e-02 -4.51598270e-03 6.55760169e-02
5.15465438e-01 -4.28777337e-01 -4.18007553e-01 -1.22464307e-01
-5.73834240e-01 -2.45262571e-02 -6.53812528e-01 7.76700675e-02
-5.74189305e-01 -8.99027169e-01 2.43010953e-01 -1.43473077e+00
9.95291412e-01 -1.44384778e+00 6.01650715e-01 -4.19958197e-02
-1.85983162e-02 -5.86005211e-01 1.08851850e-01 -3.74374241e-01
-1.23729503e+00 -9.99529421e-01 -1.13776818e-01 2.66016293e-02
-1.77027225e-01 2.78319001e-01 2.43598163e-01 5.47702074e-01
1.15040135e+00 -9.43323493e-01 9.37329158e-02 -8.00431520e-02
1.36071578e-01 1.05581784e+00 7.19295681e-01 -5.78359962e-01
5.23128569e-01 7.88111985e-01 -2.08557069e-01 -1.01450372e+00
-1.45538211e+00 -6.04695380e-01 -9.58390832e-02 -7.60440648e-01
1.29149210e+00 -8.50830376e-01 -1.29774940e+00 2.55893677e-01
-1.67765903e+00 -1.79305702e-01 1.03287779e-01 9.40001905e-02
-1.08647361e-01 2.58319318e-01 -1.08759999e+00 -1.04975152e+00
-6.43349528e-01 -2.07092836e-01 5.45342505e-01 2.32233852e-01
-1.70300090e+00 -1.18830895e+00 3.68374348e-01 8.86591554e-01
1.79187074e-01 1.07086360e+00 8.92676532e-01 -1.17053306e+00
9.91126537e-01 1.92215413e-01 -3.55816811e-01 -4.95739251e-01
5.90705536e-02 5.39297611e-03 -9.28501725e-01 1.71716943e-01
5.60677536e-02 -7.13219941e-01 -5.70452511e-02 -3.41185272e-01
6.85313761e-01 -1.25438273e+00 -5.46092570e-01 -2.81533040e-02
7.44253039e-01 3.90819088e-02 6.07439756e-01 5.17380297e-01
7.74844408e-01 1.35776937e+00 4.47898090e-01 8.08319211e-01
3.52646768e-01 4.90596801e-01 2.09725618e-01 1.90943643e-01
2.67105967e-01 -3.49010557e-01 5.53314149e-01 1.48056243e-02
-2.19163805e-01 -2.78746009e-01 -1.18119848e+00 5.54166675e-01
-1.76745546e+00 -1.58059740e+00 -7.87307799e-01 1.04906392e+00
1.06918538e+00 -2.26028517e-01 2.67704129e-01 6.32809326e-02
1.05070961e+00 6.31796777e-01 -2.68525720e-01 -9.69774365e-01
-1.49303943e-01 7.63203427e-02 1.11732140e-01 6.55952275e-01
-7.76893258e-01 6.42369509e-01 6.03001785e+00 2.51972169e-01
-8.44830334e-01 6.06676102e-01 6.17589712e-01 -3.87744129e-01
-7.16130674e-01 -1.80097550e-01 -1.11590333e-01 3.27057838e-01
9.52219069e-01 -1.03387713e-01 3.71540636e-01 4.90167767e-01
3.75538975e-01 8.49791169e-02 -9.09511209e-01 8.58158529e-01
4.25165772e-01 -1.25622058e+00 -6.15870237e-01 -4.13191438e-01
4.17259336e-01 -3.87931645e-01 -2.32750982e-01 -7.23637715e-02
1.99719980e-01 -1.11452961e+00 8.32549632e-01 5.43442726e-01
9.87429857e-01 -5.16404033e-01 6.46709323e-01 9.77224931e-02
-5.11715174e-01 -3.01717758e-01 2.52819657e-01 -5.01338780e-01
5.46870887e-01 9.05840397e-02 -6.96217954e-01 -2.73455322e-01
7.27838516e-01 7.95037031e-01 -4.73406054e-02 -2.36533564e-02
-3.74364197e-01 7.02604115e-01 -1.82159498e-01 -2.01202616e-01
-5.42855412e-02 -3.87829356e-02 8.71065915e-01 1.52772784e+00
-1.31237030e-01 1.06287682e+00 -2.49739930e-01 6.92959309e-01
-3.31266999e-01 3.71214375e-02 -6.30667806e-01 -3.32573205e-01
1.08179533e+00 1.54866827e+00 -4.06928986e-01 -1.41022801e-01
-2.80663133e-01 8.27202201e-01 5.72543681e-01 1.05764836e-01
-7.49019146e-01 2.16363564e-01 8.95134747e-01 6.30301416e-01
-9.16113734e-01 3.90034884e-01 -3.19897115e-01 -6.74749553e-01
-6.56078577e-01 -8.20356727e-01 8.03290844e-01 -9.22398865e-01
-1.51475430e+00 2.58556634e-01 3.94317687e-01 -4.99767631e-01
-2.10307643e-01 -8.24848749e-03 -8.03046525e-01 4.15294379e-01
-6.10166669e-01 -1.10150075e+00 -5.86382747e-01 4.25776631e-01
1.31370947e-01 3.76894236e-01 8.48066449e-01 9.98312980e-03
-6.53920829e-01 6.08538985e-01 -7.60795057e-01 6.48386300e-01
8.15736949e-01 -6.72008097e-01 -4.65981551e-02 1.49176314e-01
-7.00611889e-01 3.25991958e-01 1.22205770e+00 -8.78231764e-01
-1.06462109e+00 -6.84458673e-01 1.09503317e+00 -8.11143994e-01
1.28804159e+00 -5.43269962e-02 -8.68655264e-01 8.20747018e-01
6.27688468e-01 -5.51111877e-01 1.37876010e+00 3.74262810e-01
-6.53254688e-01 5.56815326e-01 -1.70527518e+00 8.41163576e-01
1.78711200e+00 -8.59173417e-01 -1.24545956e+00 3.99677247e-01
7.05781817e-01 -4.42694575e-01 -1.13100791e+00 -6.90872669e-02
4.34040010e-01 -7.19910383e-01 5.67304969e-01 -9.06857848e-01
8.70736003e-01 5.40515184e-01 6.28875718e-02 -9.49222684e-01
-2.89790183e-01 -9.78902876e-01 1.17274418e-01 1.72359848e+00
2.25703463e-01 -4.61365759e-01 2.18357101e-01 1.51939464e+00
4.89371493e-02 -1.75858125e-01 -8.50876510e-01 3.78365338e-01
1.82811886e-01 -3.08080465e-01 2.56744623e-01 1.97633743e+00
1.15022635e+00 6.14848554e-01 -2.70644993e-01 -1.59094125e-01
3.82393330e-01 -3.77513796e-01 3.23780030e-01 -1.46869552e+00
2.03074530e-01 -4.23990011e-01 -6.88670725e-02 1.89738423e-01
6.19537652e-01 -8.90610158e-01 -4.91546690e-01 -1.46443582e+00
7.56902635e-01 -1.61562428e-01 4.93668705e-01 6.92108035e-01
8.92969444e-02 3.46558809e-01 2.40255401e-01 2.77069360e-01
-4.67865199e-01 -1.52627394e-01 1.00400722e+00 8.99955034e-02
-2.84151047e-01 -8.12798083e-01 -1.27149057e+00 1.05936790e+00
7.31349230e-01 -3.88677686e-01 -8.27858318e-03 -1.39310062e-01
1.01277828e+00 4.18109417e-01 5.29323161e-01 -3.29363525e-01
-1.25367090e-01 -5.97085774e-01 4.20289338e-02 -4.49163392e-02
3.04035753e-01 -5.16839087e-01 3.29030484e-01 4.35239911e-01
-9.18170333e-01 -2.09025666e-01 3.30593884e-02 1.27955154e-02
1.77686438e-01 -4.07586813e-01 6.91002429e-01 -6.26270175e-02
-1.75452068e-01 -2.77767599e-01 -7.94215143e-01 6.64373457e-01
1.08107162e+00 -1.40269816e-01 -1.08046532e+00 -8.17558646e-01
-9.02378857e-01 2.51351625e-01 5.10214210e-01 4.69343603e-01
3.38141620e-01 -1.21488512e+00 -1.03621554e+00 -7.06622779e-01
1.83934167e-01 -5.08913279e-01 3.73310655e-01 9.42880392e-01
-3.58837813e-01 -3.51994365e-01 -1.39905900e-01 1.31882995e-01
-1.47797763e+00 3.37898761e-01 1.39519319e-01 3.22579950e-01
-4.73248929e-01 7.17644215e-01 -2.17565656e-01 -4.76612784e-02
-8.03585351e-02 3.71034980e-01 -7.94858098e-01 1.17767358e+00
8.02941978e-01 7.98235953e-01 -6.92155361e-01 -1.32698452e+00
-3.63513291e-01 2.26344332e-01 1.86840713e-01 -2.46629596e-01
1.22425520e+00 -2.69726813e-01 -6.70279682e-01 5.48502982e-01
1.57273769e+00 2.61245012e-01 -3.59967083e-01 -1.08017130e-02
8.61294419e-02 -4.15196329e-01 -4.68057781e-01 -5.22096515e-01
-4.53335583e-01 4.07909423e-01 -1.80532098e-01 5.94901621e-01
6.09625518e-01 7.64813483e-01 8.33758831e-01 -4.62274812e-02
1.65559612e-02 -1.19645166e+00 2.29815915e-01 3.57022882e-01
1.44587278e+00 -7.41158664e-01 -5.47722690e-02 -7.16886520e-01
-1.09245872e+00 1.05353129e+00 7.70487428e-01 8.67909193e-02
2.35717252e-01 3.94224763e-01 2.43636489e-01 -7.48528063e-01
-5.60835063e-01 3.39891523e-01 -3.07440847e-01 4.93925065e-01
6.01096392e-01 1.26486301e-01 -8.44907880e-01 7.58811474e-01
-6.48424387e-01 -6.33565247e-01 6.86666369e-01 4.99879986e-01
-3.27975601e-01 -5.67854762e-01 -6.89754188e-01 4.48438108e-01
-8.92508090e-01 2.39400253e-01 -1.28837264e+00 4.93460894e-01
1.56847641e-01 1.22907436e+00 5.39570808e-01 -3.42620850e-01
2.57897019e-01 -4.10482250e-02 2.30324194e-01 -6.48131788e-01
-1.25512409e+00 -4.90874618e-01 1.34604001e+00 -6.62815809e-01
-8.62821400e-01 -1.10857666e+00 -1.69507813e+00 -8.30595016e-01
2.92433113e-01 3.15959826e-02 -4.35947813e-03 1.29833078e+00
1.31151527e-01 -1.40948281e-01 4.87200022e-01 -2.04416826e-01
1.48965433e-01 -9.80504096e-01 -4.84659150e-02 1.21277082e+00
3.70222330e-01 -4.77339208e-01 -4.44542110e-01 -3.16705495e-01] | [8.631144523620605, 10.429652214050293] |
ffe3fdf0-4d4b-4090-9dcf-ecb5d8030768 | deep-attention-guided-hashing | 1812.01404 | null | http://arxiv.org/abs/1812.01404v2 | http://arxiv.org/pdf/1812.01404v2.pdf | Deep Attention-guided Hashing | With the rapid growth of multimedia data (e.g., image, audio and video etc.)
on the web, learning-based hashing techniques such as Deep Supervised Hashing
(DSH) have proven to be very efficient for large-scale multimedia search. The
recent successes seen in Learning-based hashing methods are largely due to the
success of deep learning-based hashing methods. However, there are some
limitations to previous learning-based hashing methods (e.g., the learned hash
codes containing repetitive and highly correlated information). In this paper,
we propose a novel learning-based hashing method, named Deep Attention-guided
Hashing (DAgH). DAgH is implemented using two stream frameworks. The core idea
is to use guided hash codes which are generated by the hashing network of the
first stream framework (called first hashing network) to guide the training of
the hashing network of the second stream framework (called second hashing
network). Specifically, in the first network, it leverages an attention network
and hashing network to generate the attention-guided hash codes from the
original images. The loss function we propose contains two components: the
semantic loss and the attention loss. The attention loss is used to punish the
attention network to obtain the salient region from pairs of images; in the
second network, these attention-guided hash codes are used to guide the
training of the second hashing network (i.e., these codes are treated as
supervised labels to train the second network). By doing this, DAgH can make
full use of the most critical information contained in images to guide the
second hashing network in order to learn efficient hash codes in a true
end-to-end fashion. Results from our experiments demonstrate that DAgH can
generate high quality hash codes and it outperforms current state-of-the-art
methods on three benchmark datasets, CIFAR-10, NUS-WIDE, and ImageNet. | ['Jun Long', 'Wuqing Sun', 'Zhan Yang', 'Osolo Ian Raymond'] | 2018-12-04 | null | null | null | null | ['deep-attention', 'deep-attention'] | ['computer-vision', 'natural-language-processing'] | [-1.77918985e-01 -1.14367135e-01 -1.63428262e-01 -2.01540530e-01
-1.14741004e+00 -1.54206976e-01 2.92860270e-01 1.36560306e-01
-3.86913806e-01 3.69207680e-01 3.43434125e-01 1.47090927e-01
2.83449292e-01 -8.71369183e-01 -8.98948729e-01 -7.90911198e-01
-1.53203681e-01 3.65469456e-01 3.93479615e-01 -1.00978367e-01
3.05313200e-01 5.26585393e-02 -1.76622772e+00 2.69028068e-01
3.57527763e-01 1.30868661e+00 4.54183161e-01 3.88087153e-01
1.55309513e-01 8.56538117e-01 -1.56132057e-01 -2.52846301e-01
1.02586865e-01 -4.28652763e-01 -7.96831250e-01 -3.96027088e-01
2.18118653e-01 -8.11746180e-01 -8.97826374e-01 9.38241065e-01
8.12657893e-01 3.35435271e-01 4.08869296e-01 -1.36057198e+00
-7.68370926e-01 6.71499372e-01 -6.15339041e-01 1.46994069e-01
3.36840838e-01 1.15231380e-01 1.53539860e+00 -1.20095396e+00
5.10409951e-01 1.30615032e+00 6.35787070e-01 5.02913535e-01
-7.71058619e-01 -9.31113660e-01 -2.87210852e-01 4.76012170e-01
-1.81576300e+00 -3.46819639e-01 6.68166697e-01 -9.49721634e-02
7.14420378e-01 -2.27571502e-01 6.01991236e-01 7.56113946e-01
1.28922309e-03 1.10464680e+00 6.15778625e-01 -1.96960688e-01
2.79917657e-01 1.21682517e-01 -1.17898993e-01 9.50433433e-01
-1.79247767e-01 2.27697968e-01 -6.87258840e-01 -1.06557734e-01
6.21950865e-01 3.08510542e-01 -1.97166130e-01 -2.24400595e-01
-1.04652858e+00 1.38955593e+00 1.10605252e+00 -5.16119972e-03
-5.20138681e-01 3.27548712e-01 5.73218465e-01 -1.81750171e-02
3.33945528e-02 1.27186477e-01 2.84693874e-02 2.10075453e-02
-8.66738915e-01 3.63562912e-01 5.51769257e-01 8.15028369e-01
1.11728549e+00 -4.17230368e-01 -4.62565005e-01 9.30496752e-01
5.81652045e-01 4.37468529e-01 7.09575534e-01 -7.13630855e-01
3.89184237e-01 3.35848689e-01 -3.08543622e-01 -1.13249660e+00
-7.99303502e-02 -1.01415522e-01 -7.21145928e-01 -3.58654201e-01
-7.03496486e-02 1.87915027e-01 -1.00339174e+00 1.80127859e+00
3.79712522e-01 3.44689369e-01 -8.32289681e-02 1.10945356e+00
8.24388981e-01 1.07642663e+00 1.04932427e-01 3.29766244e-01
1.37894654e+00 -1.25276971e+00 -3.82277638e-01 -2.58364528e-02
5.97560406e-01 -6.39204025e-01 1.04441655e+00 1.58810556e-01
-1.17245245e+00 -6.86904371e-01 -1.01410794e+00 -3.02607924e-01
-4.60383207e-01 -3.16903472e-01 2.36328796e-01 9.34412777e-02
-1.18542922e+00 3.81743133e-01 -5.11934638e-01 -3.60033661e-01
5.86532533e-01 4.00311172e-01 -8.78143683e-02 -4.40377891e-01
-1.62330103e+00 3.37098032e-01 5.84580123e-01 -3.31153572e-01
-1.29766989e+00 -5.00576377e-01 -1.10451651e+00 5.46911776e-01
-5.45163825e-02 -3.93250138e-01 1.27544665e+00 -5.97845137e-01
-9.26504672e-01 7.46756494e-01 1.41210020e-01 -6.94210470e-01
4.69631031e-02 -7.31426403e-02 1.33709937e-01 6.36021018e-01
4.20832813e-01 1.50179327e+00 7.96159327e-01 -1.08039820e+00
-8.91914010e-01 -1.58409867e-02 -1.05309464e-01 3.04297447e-01
-7.28905857e-01 -1.18940055e-01 -8.15928698e-01 -6.92875385e-01
-1.22809179e-01 -1.11180067e+00 1.26628026e-01 1.75024986e-01
-3.91581953e-01 -3.50554079e-01 1.13259411e+00 -5.85464895e-01
1.26961541e+00 -2.53541207e+00 1.17645144e-01 1.98850915e-01
1.04115263e-01 2.05814198e-01 -2.38976359e-01 4.37442511e-01
3.70076708e-02 -7.65396804e-02 -1.42922744e-01 -6.64983273e-01
1.50388941e-01 2.57382393e-01 -5.11418641e-01 3.10924351e-01
9.65053961e-02 9.13318634e-01 -1.10604918e+00 -6.75598860e-01
3.61975841e-02 8.22840869e-01 -7.94605017e-01 6.20702684e-01
5.43523841e-02 -8.25645104e-02 -3.01786304e-01 7.19316542e-01
3.88293028e-01 -4.43149149e-01 -3.96980047e-01 -2.73078442e-01
-2.38383841e-02 4.41076726e-01 -8.40759277e-01 1.75595903e+00
-3.66612911e-01 4.57240850e-01 -3.18051070e-01 -8.17599833e-01
7.94113100e-01 4.01347339e-01 4.02593851e-01 -9.10146058e-01
-3.06874141e-02 1.25057682e-01 -4.69505459e-01 -3.24648857e-01
4.38026637e-01 -3.48919891e-02 -2.02616856e-01 5.92037976e-01
2.19454467e-01 1.90118253e-01 3.61556682e-04 4.56139237e-01
9.90682662e-01 -3.11887264e-01 2.29114369e-02 2.16464370e-01
5.51172197e-01 -2.32858941e-01 4.38942403e-01 7.45815277e-01
-2.53411531e-01 9.52177823e-01 8.44953433e-02 -4.88701791e-01
-1.27241743e+00 -9.01820540e-01 1.06537500e-02 1.25657129e+00
2.89961696e-01 -4.32546824e-01 -7.00156391e-01 -7.40690410e-01
1.55988649e-01 6.58941120e-02 -7.06520021e-01 -6.56731784e-01
-3.88709217e-01 -4.02252436e-01 5.97149134e-01 5.28839946e-01
8.87664557e-01 -1.51233983e+00 -6.09274685e-01 3.25438052e-01
-3.92615676e-01 -1.06634700e+00 -8.79193425e-01 1.67775303e-01
-4.35052276e-01 -8.15372944e-01 -8.88597071e-01 -1.11171317e+00
4.31798190e-01 6.84178829e-01 9.45017755e-01 5.54514647e-01
-3.83144945e-01 3.38887006e-01 -5.83983481e-01 -7.99198672e-02
7.86600914e-03 4.23656046e-01 -1.52127996e-01 2.11199194e-01
3.73808503e-01 -4.04120475e-01 -1.01198018e+00 3.45178545e-01
-1.15763390e+00 -2.08972216e-01 7.64723361e-01 1.08928275e+00
7.24076807e-01 1.14832290e-01 5.06672144e-01 -4.78680164e-01
2.89739877e-01 -9.06319678e-01 -4.02944624e-01 -1.09699875e-01
-4.96705443e-01 1.24334857e-01 7.10800588e-01 -3.37325573e-01
-3.15621495e-01 4.87600826e-02 -2.61455774e-01 -7.79736936e-01
1.28611907e-01 6.94686830e-01 1.48790643e-01 -1.47425517e-01
1.81410521e-01 5.57879686e-01 -3.58610898e-02 -2.48325914e-01
6.27099201e-02 7.65783429e-01 5.49722075e-01 -3.76691043e-01
8.84459972e-01 4.15443867e-01 -3.39813352e-01 -4.76281792e-01
-9.40926194e-01 -6.93512738e-01 -1.32368058e-01 -9.50203165e-02
1.05847979e+00 -1.11544502e+00 -5.23962438e-01 4.16165680e-01
-1.01448190e+00 -1.51662722e-01 -1.70778185e-01 3.33740562e-01
-5.45368373e-01 1.70362845e-01 -9.17775452e-01 -6.63471937e-01
-5.96031070e-01 -1.42380857e+00 1.50682199e+00 4.18711841e-01
3.74181010e-02 -8.66631687e-01 2.04299867e-01 2.63564318e-01
3.42851669e-01 -1.69263855e-01 9.70849872e-01 -6.80841923e-01
-7.94667721e-01 -2.71780491e-01 -5.19042552e-01 1.72661394e-01
-2.03106895e-01 -3.19826901e-01 -1.12812388e+00 -5.93879998e-01
-3.20508480e-01 -9.17419612e-01 9.14731979e-01 1.55400798e-01
1.33105099e+00 -5.19984245e-01 -1.25463679e-01 7.59441376e-01
1.56424737e+00 1.85776010e-01 8.25881243e-01 4.81113642e-01
6.85236812e-01 4.28873658e-01 6.74106121e-01 5.05844772e-01
7.47462392e-01 8.53615105e-01 5.64400256e-01 -1.67908117e-01
-1.94957748e-01 -6.28563344e-01 4.83516395e-01 8.58412683e-01
6.47305667e-01 7.06439316e-02 -7.70002782e-01 7.12564588e-01
-1.92359281e+00 -1.01570499e+00 6.18358612e-01 2.28394985e+00
8.84854019e-01 1.22134909e-01 3.47290397e-01 1.49629623e-01
8.29660535e-01 1.66883245e-01 -5.40817201e-01 -2.04531223e-01
2.07598835e-01 4.55834381e-02 1.30865395e-01 2.96308100e-01
-1.25716662e+00 7.91611195e-01 5.06329775e+00 7.42833436e-01
-1.20417213e+00 4.23586294e-02 7.06449687e-01 5.64576834e-02
-1.65001214e-01 -1.26909003e-01 -1.05025387e+00 7.40787864e-01
1.01780641e+00 4.27278318e-02 5.00351548e-01 1.02832699e+00
-1.40586808e-01 1.93799689e-01 -9.25273538e-01 1.19450629e+00
1.98695883e-01 -1.25156307e+00 3.40133399e-01 1.01107761e-01
5.03784716e-01 2.32680842e-01 4.51204598e-01 6.52994752e-01
7.07187802e-02 -8.09659600e-01 7.78800845e-01 -6.61252439e-02
7.18833864e-01 -9.75827873e-01 9.22149003e-01 2.91263852e-02
-1.48042428e+00 -3.11807454e-01 -5.68892777e-01 4.90817517e-01
3.92967276e-02 4.60239261e-01 -8.27761889e-01 1.28371447e-01
1.03909731e+00 8.27849090e-01 -4.11279947e-01 1.25895190e+00
-6.40452877e-02 4.55047131e-01 -2.19959810e-01 1.42434940e-01
8.35163414e-01 1.85597196e-01 1.53021872e-01 1.23247027e+00
3.30156952e-01 -1.81051902e-02 3.84378076e-01 7.37458050e-01
-5.32852232e-01 1.11994967e-04 -6.82215631e-01 1.09326921e-01
7.89358675e-01 1.30241776e+00 -5.98958969e-01 -3.56541365e-01
-6.33198023e-01 1.00317466e+00 4.41609353e-01 1.79669201e-01
-1.07461047e+00 -5.89000285e-01 4.05313671e-01 5.35128415e-02
7.66508639e-01 1.88325509e-01 2.42861494e-01 -8.18475544e-01
-1.36267275e-01 -7.96734095e-01 6.86999202e-01 -8.21663737e-01
-1.25361621e+00 5.85129082e-01 -2.04148725e-01 -1.08925223e+00
-5.82867861e-02 -3.34068798e-02 -5.44092119e-01 7.27469802e-01
-2.01641202e+00 -9.99282897e-01 -4.86886412e-01 8.27151060e-01
5.60382485e-01 -8.65479466e-03 6.50365949e-01 7.28782058e-01
-4.84785646e-01 9.32008266e-01 -8.73714387e-02 4.86614019e-01
7.90088236e-01 -1.12733316e+00 2.85236508e-01 4.12492901e-01
6.44663498e-02 4.51497763e-01 2.23676786e-01 -3.74066353e-01
-1.41996157e+00 -1.29840469e+00 8.59552264e-01 2.23810181e-01
5.06176233e-01 -3.94797444e-01 -1.06302392e+00 3.40682894e-01
2.98644155e-02 3.54930222e-01 6.43647432e-01 -4.78956193e-01
-8.00266564e-01 -3.02797675e-01 -1.10884464e+00 2.65378892e-01
4.36665505e-01 -8.92280281e-01 -3.57939929e-01 3.81966233e-01
1.00371623e+00 -1.76235572e-01 -7.25506783e-01 1.95053384e-01
6.77853107e-01 -9.17314827e-01 1.07530689e+00 -2.07900375e-01
6.13700807e-01 -1.76256016e-01 -3.79292399e-01 -9.27394331e-01
-4.40969437e-01 -4.54382062e-01 -1.81553811e-01 1.07459760e+00
1.91394791e-01 -3.25363278e-01 7.61013508e-01 2.36868545e-01
-1.29604131e-01 -9.83337879e-01 -8.52755129e-01 -5.48112094e-01
-1.99221238e-01 2.00156599e-01 5.74092329e-01 8.29099119e-01
-1.06781341e-01 3.98937255e-01 -3.68127853e-01 1.54368505e-01
7.67713308e-01 1.38010129e-01 5.00155330e-01 -9.50780034e-01
-3.02028507e-01 -2.82571912e-01 -4.75610316e-01 -1.07696438e+00
6.73121214e-02 -7.97044754e-01 3.80617887e-01 -1.14903486e+00
6.21072471e-01 -6.83990002e-01 -6.13056779e-01 7.29254842e-01
-3.29821765e-01 5.75365305e-01 4.11374271e-01 5.67000091e-01
-9.69315767e-01 8.60998571e-01 8.49683523e-01 -2.17512414e-01
-2.17453707e-02 -3.69539499e-01 -6.38106525e-01 2.56617248e-01
5.63027918e-01 -8.19230855e-01 -4.09212977e-01 -2.74775058e-01
1.30930886e-01 2.28790231e-02 5.24902880e-01 -1.18048513e+00
5.71108222e-01 4.59107250e-01 2.15502426e-01 -6.33562088e-01
2.64576524e-01 -7.64601767e-01 -3.24508101e-01 5.38310468e-01
-5.73754489e-01 2.11151496e-01 -1.25133470e-01 4.77172375e-01
-5.61058939e-01 -1.57054067e-01 9.66512442e-01 -4.11877222e-02
-6.83698118e-01 7.50438869e-01 7.12082833e-02 1.49127677e-01
8.97423446e-01 -6.45275926e-03 -6.42674267e-02 -6.32435560e-01
-3.20009947e-01 3.86112034e-01 4.22185153e-01 3.80219460e-01
1.04129791e+00 -1.68657815e+00 -6.02443218e-01 2.62895525e-01
2.10539266e-01 2.53842473e-01 4.03083041e-02 6.33292079e-01
-3.96638751e-01 4.78355139e-01 -9.08777639e-02 -5.58828294e-01
-9.64641213e-01 8.66371095e-01 5.50421290e-02 -2.55587369e-01
-5.88333428e-01 8.92832160e-01 4.30055678e-01 -1.94787145e-01
6.95838749e-01 7.16577843e-02 -2.62175873e-03 4.72104661e-02
8.89376640e-01 3.73162143e-02 4.41364991e-03 -5.66839695e-01
-2.69086063e-01 4.62938219e-01 -5.52468836e-01 -1.35647878e-01
1.36994362e+00 -7.78111592e-02 -1.51395453e-02 9.74736959e-02
1.93765461e+00 -6.43289804e-01 -1.29871929e+00 -3.80714685e-01
-2.13460475e-01 -3.40719312e-01 2.36946851e-01 -2.98657864e-01
-1.44617605e+00 1.01065779e+00 6.76988661e-01 6.63566515e-02
1.20712388e+00 1.89951099e-02 1.59553421e+00 2.71272302e-01
3.32275957e-01 -9.41578865e-01 4.87990320e-01 5.43145001e-01
5.72606564e-01 -1.17156363e+00 -3.38898689e-01 9.59785804e-02
-6.17604971e-01 8.76311958e-01 6.18942678e-01 -1.30715311e-01
4.83275026e-01 2.53718570e-02 -1.89123362e-01 -2.51575828e-01
-7.59841681e-01 -2.00800136e-01 1.53831765e-01 2.93500930e-01
1.73448056e-01 -5.20209253e-01 2.52378583e-01 2.41568923e-01
1.34476110e-01 -5.83304279e-02 6.29613772e-02 1.04832709e+00
-7.36047924e-01 -7.65579760e-01 -4.16196048e-01 2.40439743e-01
-3.35960746e-01 -3.44458610e-01 -1.12358764e-01 4.72367913e-01
-8.34486112e-02 5.74476242e-01 1.27256662e-01 -6.09480560e-01
1.03607096e-01 -8.80555287e-02 -2.73633748e-02 -6.96503162e-01
-6.20640397e-01 -2.78123897e-02 -4.77640927e-01 -7.02655315e-01
-1.48141056e-01 -4.95036215e-01 -1.45194995e+00 -2.47024357e-01
-2.07333267e-01 4.78936374e-01 5.13813019e-01 6.20201707e-01
4.05784369e-01 6.63623363e-02 1.09316158e+00 -1.00846744e+00
-6.61669612e-01 -5.99561989e-01 -4.36855495e-01 6.97074950e-01
5.52634180e-01 -6.71786129e-01 -4.81988311e-01 -8.15840624e-03] | [11.282415390014648, 0.9500383138656616] |
12c68deb-3403-4e92-b69d-b9024bcc764b | consistent-cross-view-matching-for | 1908.10486 | null | https://arxiv.org/abs/1908.10486v3 | https://arxiv.org/pdf/1908.10486v3.pdf | Exploiting Global Camera Network Constraints for Unsupervised Video Person Re-identification | Many unsupervised approaches have been proposed recently for the video-based re-identification problem since annotations of samples across cameras are time-consuming. However, higher-order relationships across the entire camera network are ignored by these methods, leading to contradictory outputs when matching results from different camera pairs are combined. In this paper, we address the problem of unsupervised video-based re-identification by proposing a consistent cross-view matching (CCM) framework, in which global camera network constraints are exploited to guarantee the matched pairs are with consistency. Specifically, we first propose to utilize the first neighbor of each sample to discover relations among samples and find the groups in each camera. Additionally, a cross-view matching strategy followed by global camera network constraints is proposed to explore the matching relationships across the entire camera network. Finally, we learn metric models for camera pairs progressively by alternatively mining consistent cross-view matching pairs and updating metric models using these obtained matches. Rigorous experiments on two widely-used benchmarks for video re-identification demonstrate the superiority of the proposed method over current state-of-the-art unsupervised methods; for example, on the MARS dataset, our method achieves an improvement of 4.2\% over unsupervised methods, and even 2.5\% over one-shot supervision-based methods for rank-1 accuracy. | ['Amit K. Roy-Chowdhury', 'Xueping Wang', 'Min Liu', 'Yaonan Wang', 'Rameswar Panda'] | 2019-08-27 | null | null | null | null | ['unsupervised-person-re-identification'] | ['computer-vision'] | [ 2.33367473e-01 -4.00464356e-01 -5.49854696e-01 -3.92955273e-01
-5.47327638e-01 -5.14532268e-01 3.73029232e-01 1.10078841e-01
-2.72384018e-01 2.75688767e-01 8.98394138e-02 3.50346148e-01
-4.36322004e-01 -4.16145235e-01 -5.53131461e-01 -6.00464702e-01
7.47987181e-02 2.81502038e-01 2.69081652e-01 2.61577606e-01
4.17358994e-01 3.34550709e-01 -1.68902373e+00 1.19057789e-01
5.59497356e-01 1.00877512e+00 1.20197706e-01 3.97953749e-01
1.12120837e-01 6.53215408e-01 -1.79410115e-01 -5.29737294e-01
5.22697330e-01 -3.94677103e-01 -7.20298171e-01 5.77858210e-01
8.27165484e-01 -3.49894971e-01 -6.76011205e-01 1.41108263e+00
1.26479581e-01 3.70670378e-01 2.83662528e-01 -1.46626222e+00
-2.66520858e-01 6.57441795e-01 -9.55515563e-01 1.77968502e-01
6.26689255e-01 -1.99121192e-01 1.25235546e+00 -9.00859058e-01
6.99392021e-01 9.40221190e-01 5.51274657e-01 3.75281185e-01
-1.19936287e+00 -9.19558823e-01 2.61104345e-01 6.02915943e-01
-1.78869557e+00 -6.01820350e-01 8.55174839e-01 -4.71789092e-01
5.58529913e-01 2.93316007e-01 4.92237508e-01 5.91932654e-01
-4.49402034e-01 6.24771953e-01 7.45337784e-01 -3.13079923e-01
-1.48651123e-01 7.53908232e-02 1.18993446e-01 8.40411127e-01
2.99531907e-01 -9.00891498e-02 -8.76947880e-01 -2.10341170e-01
4.79056597e-01 3.75193775e-01 -3.21611524e-01 -7.13706970e-01
-1.44594979e+00 4.79360014e-01 2.00223893e-01 1.93790272e-01
-1.03721144e-02 -7.81857446e-02 5.24736047e-01 2.74892062e-01
2.21113205e-01 4.69601125e-01 -2.75555998e-02 6.11819467e-03
-1.10570872e+00 -1.29245430e-01 5.29984295e-01 1.38020039e+00
9.38237846e-01 -4.16217268e-01 6.45443425e-02 1.02401590e+00
2.31656477e-01 1.88890956e-02 3.36218894e-01 -1.03082025e+00
6.92591369e-01 7.79338777e-01 3.65325436e-02 -1.69161749e+00
-3.16465832e-02 -1.75504997e-01 -9.97782767e-01 -4.51649278e-01
2.60637611e-01 3.34609509e-01 -4.09647912e-01 1.60237038e+00
2.98324347e-01 6.42008543e-01 -6.08567446e-02 9.35432434e-01
6.98190093e-01 2.96012551e-01 -2.53514677e-01 -4.19586539e-01
1.09255087e+00 -1.13031960e+00 -3.73459786e-01 1.07612655e-01
6.27805054e-01 -7.69479871e-01 4.34210718e-01 3.90604526e-01
-9.12172914e-01 -8.09323847e-01 -1.09110498e+00 4.82348472e-01
-2.48979200e-02 1.97388738e-01 2.66607761e-01 3.47176433e-01
-8.39078128e-01 6.69235885e-01 -4.98309284e-01 -7.17266798e-01
2.35610992e-01 4.32034582e-01 -7.82691598e-01 -4.94572639e-01
-8.51021409e-01 2.82931507e-01 5.66838205e-01 1.59773827e-01
-8.24494123e-01 -5.02430856e-01 -8.14456165e-01 -1.30735338e-01
7.46361732e-01 -4.46549833e-01 7.11100996e-01 -9.36233222e-01
-1.13933551e+00 9.35961843e-01 -3.26226592e-01 -2.31128544e-01
4.16772783e-01 -1.20910816e-01 -7.14490592e-01 3.52748007e-01
2.70788610e-01 5.10699570e-01 6.80361509e-01 -1.32300174e+00
-9.48954403e-01 -2.76292354e-01 1.81524873e-01 1.97867736e-01
-7.60301948e-01 1.44293144e-01 -1.18763041e+00 -5.98625839e-01
5.32159269e-01 -1.15221226e+00 -8.37219581e-02 -1.12145327e-01
-6.12217307e-01 -2.36108243e-01 7.61779308e-01 -5.95806837e-01
1.55923915e+00 -2.22800732e+00 2.67082661e-01 5.49191892e-01
3.73908579e-01 -5.17045148e-03 -1.45592391e-01 3.73116046e-01
-2.39256591e-01 4.74201106e-02 -1.56450167e-01 -5.01331329e-01
-2.45531589e-01 7.44281262e-02 7.38203749e-02 7.28134811e-01
-1.73767284e-01 3.14910144e-01 -9.78220999e-01 -6.91620946e-01
2.50125885e-01 -7.50245973e-02 -4.83187377e-01 3.52671474e-01
3.22656542e-01 3.55980456e-01 -1.34695053e-01 8.73569727e-01
6.74923599e-01 -3.88839245e-01 5.95346034e-01 -7.35696316e-01
2.30155755e-02 -3.54191303e-01 -1.47885168e+00 1.97240698e+00
-5.28011769e-02 5.54200828e-01 -2.09303766e-01 -1.26283038e+00
8.89444232e-01 1.13824278e-01 8.53454411e-01 -3.07637960e-01
-7.66282901e-02 1.06697053e-01 -2.40296602e-01 -4.90540028e-01
5.59383929e-01 3.64582360e-01 5.78960124e-03 4.57771152e-01
-7.68599883e-02 4.16117966e-01 4.39405203e-01 4.47774708e-01
9.31780994e-01 -4.36204262e-02 2.98795164e-01 -1.92022577e-01
9.39956546e-01 -1.02138355e-01 9.59235311e-01 7.42053151e-01
-1.34210974e-01 6.81214213e-01 2.40173668e-01 -4.96900201e-01
-9.78645027e-01 -7.39930272e-01 1.09753676e-01 7.67835498e-01
8.88764381e-01 -8.82726371e-01 -6.55173361e-01 -8.62533033e-01
-4.15301584e-02 1.08099297e-01 -4.70050424e-01 -1.06130198e-01
-4.91937786e-01 -4.25325066e-01 4.69875842e-01 5.02758086e-01
7.49072194e-01 -2.95619071e-01 -2.47666195e-01 -1.93346087e-02
-4.38682109e-01 -1.57299614e+00 -9.43760514e-01 -3.29584450e-01
-9.05301273e-01 -1.66770029e+00 -4.05903518e-01 -8.69934916e-01
1.12550318e+00 8.88027370e-01 8.40092540e-01 4.26874727e-01
-1.78703979e-01 6.45236731e-01 -5.21060050e-01 5.79325736e-01
-1.06239572e-01 -1.40220210e-01 5.09406388e-01 4.94632393e-01
4.69236821e-01 -3.95441681e-01 -5.82481086e-01 1.00067484e+00
-7.23406792e-01 5.73274530e-02 3.03918898e-01 9.24402356e-01
8.78671408e-01 2.74470687e-01 1.39726296e-01 -8.60051811e-01
-1.40385572e-02 -4.93804634e-01 -7.06329584e-01 6.15768611e-01
-6.82179451e-01 -1.89905703e-01 5.50221145e-01 -4.82595474e-01
-7.94719279e-01 3.98873270e-01 7.39717424e-01 -9.81963158e-01
1.86336506e-02 3.38181734e-01 -4.42559838e-01 -2.27357596e-01
7.08352476e-02 2.66938239e-01 -1.73297778e-01 -3.66995156e-01
1.60670012e-01 6.79796815e-01 7.68495440e-01 -4.64348435e-01
1.01945269e+00 6.02304876e-01 1.55509412e-02 -5.35547733e-01
-7.95712173e-01 -1.05369318e+00 -9.96300519e-01 -4.01281595e-01
8.18145096e-01 -1.10732591e+00 -7.21093357e-01 3.38206589e-01
-8.95045638e-01 2.69290000e-01 1.43245965e-01 6.68539643e-01
-4.45623130e-01 9.67812181e-01 -3.27423215e-01 -5.60367405e-01
-5.23990691e-02 -1.25942779e+00 6.99259877e-01 2.58880109e-01
-2.18021408e-01 -7.68554568e-01 -3.73988152e-02 6.98772728e-01
-1.97929174e-01 3.67132276e-02 5.21985710e-01 -6.53312743e-01
-6.23908341e-01 -2.91072726e-01 -4.06672955e-01 3.09748650e-01
4.41483408e-01 1.16080157e-01 -7.05402195e-01 -5.14377534e-01
-3.60385895e-01 9.80297849e-02 7.47952700e-01 -1.14667349e-01
1.36103153e+00 -2.12611824e-01 -6.69263124e-01 8.65190685e-01
1.39721620e+00 1.30270153e-01 3.39676440e-01 4.51162308e-01
1.10589719e+00 6.81360424e-01 9.97271538e-01 6.49273336e-01
4.08598959e-01 8.87362301e-01 4.07127827e-01 2.62927860e-01
2.41090000e-01 -3.88121754e-01 3.23402554e-01 1.07675457e+00
-1.46753848e-01 -8.53409097e-02 -6.57302558e-01 6.90501273e-01
-2.09537077e+00 -1.12127149e+00 -1.02162786e-01 2.53460813e+00
3.43790263e-01 -6.23556860e-02 2.78335065e-01 1.62489593e-01
1.35698509e+00 1.41627178e-01 -5.11433899e-01 1.86674163e-01
-1.16301052e-01 -4.32803482e-01 7.26192653e-01 1.00267775e-01
-1.31404924e+00 7.38193512e-01 5.72924614e+00 7.43188918e-01
-5.78549385e-01 -5.32624796e-02 4.83120382e-01 -7.63805211e-02
-4.88388725e-02 4.78402197e-01 -8.63844991e-01 5.06782055e-01
3.62540305e-01 -1.74182534e-01 4.96456236e-01 7.83131659e-01
8.22660476e-02 -2.02325687e-01 -1.43376553e+00 1.56207466e+00
6.28064871e-01 -1.15934289e+00 8.32932666e-02 -2.66485196e-02
1.02015722e+00 -3.22319001e-01 -2.82806724e-01 -2.75435537e-01
7.56706297e-02 -4.84365970e-01 4.57372189e-01 6.19450212e-01
8.79521608e-01 -8.92110825e-01 6.24433517e-01 1.50173664e-01
-1.67031085e+00 -1.38323918e-01 -4.60202396e-01 3.40539575e-01
2.23723371e-02 5.74757218e-01 -4.68503803e-01 9.57122386e-01
9.33311582e-01 1.43112469e+00 -6.79214656e-01 1.18456185e+00
1.51244417e-01 2.46240705e-01 -7.55617097e-02 4.68062252e-01
-1.83171760e-02 -3.91022891e-01 6.33043647e-01 1.07581043e+00
4.10517067e-01 8.78633559e-02 4.69553024e-01 3.53300780e-01
-2.77634263e-01 8.10090173e-03 -5.73642433e-01 1.61645994e-01
6.86963439e-01 1.27694070e+00 -6.57997012e-01 -4.27554488e-01
-5.91915071e-01 1.15189254e+00 2.25171953e-01 1.63717493e-01
-9.28900123e-01 -3.32527369e-01 7.59213030e-01 -2.72231596e-03
2.84907848e-01 -1.49840683e-01 3.37729335e-01 -1.38643658e+00
7.77099952e-02 -8.78104925e-01 9.12779272e-01 -5.19030094e-01
-1.44669282e+00 3.99791121e-01 8.23346153e-02 -1.88054752e+00
-3.42474654e-02 -2.94470549e-01 -4.00921077e-01 2.60583490e-01
-1.42211699e+00 -9.23789740e-01 -7.69655526e-01 8.82871687e-01
5.86376786e-01 -4.87197578e-01 4.64053601e-01 6.13421917e-01
-8.96383405e-01 9.41464603e-01 1.68754518e-01 3.96688879e-01
9.66331542e-01 -8.21377993e-01 -3.67047004e-02 1.13378906e+00
4.88716274e-01 6.97732568e-01 2.60785758e-01 -5.96302032e-01
-1.65119874e+00 -1.11646831e+00 6.34275854e-01 -2.51708448e-01
6.89076006e-01 -1.57505289e-01 -7.09729791e-01 5.57053924e-01
-8.40969607e-02 1.71712354e-01 8.96911025e-01 -2.15865895e-02
-7.14077413e-01 -5.59050739e-01 -8.54018271e-01 4.78233069e-01
1.54672027e+00 -7.21089900e-01 -2.44670048e-01 2.43398800e-01
4.58591640e-01 -2.07128525e-01 -1.10070014e+00 4.62847173e-01
6.05337143e-01 -1.06464875e+00 8.45847964e-01 -4.74663436e-01
2.42777899e-01 -8.22715878e-01 -4.65664893e-01 -9.03036475e-01
-2.87714571e-01 -6.15188301e-01 -8.88651460e-02 1.58866692e+00
4.96188588e-02 -3.40911388e-01 8.52444708e-01 6.05058074e-01
9.70033631e-02 -2.27656692e-01 -8.37633848e-01 -9.32498395e-01
-8.46604586e-01 -3.48766774e-01 5.07837474e-01 1.27208185e+00
2.16422766e-01 -7.19532669e-02 -7.73875117e-01 3.66204947e-01
9.64246869e-01 3.29924136e-01 1.02388120e+00 -1.13537347e+00
-4.04882193e-01 -3.30205083e-01 -7.90561140e-01 -1.19419217e+00
3.31771046e-01 -8.63051534e-01 -1.53460484e-02 -1.06756449e+00
8.30485225e-01 -5.35264373e-01 -5.56605756e-01 1.42759204e-01
-9.99980941e-02 3.99482757e-01 3.46760511e-01 8.40041697e-01
-1.16871297e+00 2.03639328e-01 8.02257359e-01 -3.84838611e-01
-5.53240255e-02 -1.01881377e-01 -5.62959969e-01 7.84671605e-01
4.97000575e-01 -5.56978703e-01 -3.64541024e-01 -4.95491177e-01
1.28378540e-01 2.80059054e-02 3.30755144e-01 -1.20752442e+00
8.69637847e-01 -1.88826159e-01 2.55377263e-01 -6.12209141e-01
1.61556870e-01 -1.14869165e+00 5.68048894e-01 1.69237792e-01
-2.63385803e-01 2.78513461e-01 -2.68806368e-01 1.02743757e+00
-5.91991603e-01 -2.08157465e-01 6.85918272e-01 -8.34869593e-02
-9.69629228e-01 5.66388071e-01 1.08349383e-01 -7.83223361e-02
1.24738407e+00 -5.63504457e-01 4.47436683e-02 -2.47095197e-01
-6.52726233e-01 4.34669644e-01 8.23361754e-01 5.13679087e-01
6.97829723e-01 -1.56444418e+00 -5.90528965e-01 3.05725727e-02
6.41912758e-01 -9.43408161e-02 3.40818733e-01 1.00595427e+00
-2.83807784e-01 1.32976264e-01 -2.54742075e-02 -8.68786573e-01
-1.74456859e+00 6.65512383e-01 1.38431981e-01 -1.26451299e-01
-4.66809064e-01 6.46442950e-01 1.37309760e-01 -3.88582759e-02
3.04266363e-01 2.36790672e-01 -2.18453571e-01 2.75201499e-01
5.12884200e-01 6.12377465e-01 -2.58606911e-01 -9.42455471e-01
-5.63022256e-01 1.12130070e+00 -1.94387227e-01 3.54162544e-01
1.12376463e+00 -4.18517947e-01 -3.17626506e-01 2.97100127e-01
1.37848508e+00 3.06591112e-03 -1.05860853e+00 -6.24425650e-01
2.45360687e-01 -1.01052153e+00 -3.47445011e-01 -9.43941474e-02
-1.43403351e+00 3.93156499e-01 3.90785456e-01 -1.28114149e-01
1.31205606e+00 -8.55198056e-02 6.15461826e-01 4.95950669e-01
4.61948752e-01 -1.30227923e+00 1.53527051e-01 8.65155235e-02
2.74655074e-01 -1.47622108e+00 3.03460628e-01 -5.60083985e-01
-5.37503541e-01 1.14375973e+00 7.82499075e-01 8.05785134e-03
4.89498705e-01 -2.56637633e-01 -4.44405794e-01 -3.54622081e-02
-4.39065844e-01 -2.48588007e-02 3.17037493e-01 2.96863675e-01
8.12335089e-02 -5.69153577e-02 -4.77387384e-02 2.61882156e-01
2.93331891e-01 -2.06025004e-01 4.12588894e-01 6.98329091e-01
-5.97216636e-02 -1.20001984e+00 -3.51541370e-01 4.43555176e-01
-1.36159807e-01 7.96208382e-02 -4.54727918e-01 6.49790108e-01
9.24741775e-02 1.19251859e+00 1.13386236e-01 -9.92469609e-01
2.54518926e-01 -2.78207451e-01 3.59420121e-01 -5.19059658e-01
-4.14913684e-01 7.34530091e-02 7.58436546e-02 -7.17018843e-01
-1.05360508e+00 -8.21491301e-01 -8.58058393e-01 -4.48692590e-01
-6.69653118e-01 2.50375539e-01 2.00978816e-01 7.78786063e-01
5.12148261e-01 3.98016647e-02 1.19988203e+00 -5.71679533e-01
-2.50827909e-01 -5.98977268e-01 -6.23186231e-01 7.42158055e-01
-3.67096439e-03 -7.50280201e-01 -5.96500754e-01 3.88831824e-01] | [14.7605619430542, 1.0264275074005127] |
12ff0ae6-a4f6-4713-a746-d867f2061134 | consistent-representation-learning-for | 2203.02721 | null | https://arxiv.org/abs/2203.02721v2 | https://arxiv.org/pdf/2203.02721v2.pdf | Consistent Representation Learning for Continual Relation Extraction | Continual relation extraction (CRE) aims to continuously train a model on data with new relations while avoiding forgetting old ones. Some previous work has proved that storing a few typical samples of old relations and replaying them when learning new relations can effectively avoid forgetting. However, these memory-based methods tend to overfit the memory samples and perform poorly on imbalanced datasets. To solve these challenges, a consistent representation learning method is proposed, which maintains the stability of the relation embedding by adopting contrastive learning and knowledge distillation when replaying memory. Specifically, supervised contrastive learning based on a memory bank is first used to train each new task so that the model can effectively learn the relation representation. Then, contrastive replay is conducted of the samples in memory and makes the model retain the knowledge of historical relations through memory knowledge distillation to prevent the catastrophic forgetting of the old task. The proposed method can better learn consistent representations to alleviate forgetting effectively. Extensive experiments on FewRel and TACRED datasets show that our method significantly outperforms state-of-the-art baselines and yield strong robustness on the imbalanced dataset. | ['Kai Gao', 'Jiangong Yang', 'Hua Xu', 'Kang Zhao'] | 2022-03-05 | null | https://aclanthology.org/2022.findings-acl.268 | https://aclanthology.org/2022.findings-acl.268.pdf | findings-acl-2022-5 | ['continual-relation-extraction'] | ['natural-language-processing'] | [-6.46033883e-02 3.55419397e-01 -5.88162720e-01 -3.09376001e-01
-2.67313540e-01 1.41142756e-01 4.79649812e-01 4.37094241e-01
-3.61111581e-01 1.08701515e+00 1.91011146e-01 -1.55194595e-01
-1.25848293e-01 -1.21348679e+00 -7.80869484e-01 -7.06208706e-01
5.70264049e-02 6.90560937e-01 3.75487059e-01 -4.31083202e-01
4.95520160e-02 4.70322698e-01 -1.39897084e+00 2.96327263e-01
9.40415144e-01 9.01332080e-01 7.24577205e-03 8.65529254e-02
-1.87889367e-01 1.54826009e+00 -7.29661107e-01 -5.58022320e-01
-1.27252534e-01 -2.71799266e-01 -1.15512705e+00 -2.13961080e-01
7.62627870e-02 -4.09119844e-01 -1.09869659e+00 6.77210450e-01
4.32659358e-01 4.73113060e-01 3.13783795e-01 -9.75540638e-01
-1.30032313e+00 8.78780186e-01 -5.53063393e-01 7.28250802e-01
4.67355847e-02 -6.45088479e-02 8.04120004e-01 -1.02279544e+00
6.62072718e-01 1.17881262e+00 5.79964757e-01 4.32756394e-01
-1.11826706e+00 -9.85900760e-01 5.10577381e-01 6.50136650e-01
-1.41796374e+00 -7.00666904e-01 7.89542913e-01 1.21968202e-01
1.42172980e+00 5.96781857e-02 7.32170403e-01 9.41815317e-01
4.25413936e-01 8.70346904e-01 5.11280358e-01 -3.85955453e-01
6.48952127e-02 3.49093042e-02 7.65240729e-01 5.08422315e-01
6.33145630e-01 2.52088085e-02 -1.07680631e+00 5.23998030e-02
3.37796509e-01 3.82388234e-01 -3.46761167e-01 -1.90839484e-01
-9.35933888e-01 3.69748414e-01 8.58900666e-01 5.00626385e-01
-4.42889422e-01 -2.73355097e-01 5.56383371e-01 6.41533852e-01
7.21260190e-01 4.34694082e-01 -6.76626682e-01 2.63033450e-01
-6.97936893e-01 7.16649890e-02 5.06361723e-01 9.67981339e-01
1.07327652e+00 -4.88514043e-02 -2.46193692e-01 8.68132889e-01
-1.58375561e-01 1.84775651e-01 9.26216841e-01 -4.21490192e-01
7.35263884e-01 1.14788783e+00 -3.51492822e-01 -1.15118015e+00
-4.35406297e-01 -6.94858909e-01 -1.20840621e+00 -3.33548456e-01
-9.81882587e-02 3.21986049e-01 -8.89352977e-01 1.56266129e+00
1.60247311e-01 3.69634539e-01 3.56902659e-01 4.18406010e-01
7.41695285e-01 7.08210945e-01 1.03594653e-01 -8.06518137e-01
1.10335684e+00 -1.10859084e+00 -1.15148795e+00 -6.96886122e-01
7.35433698e-01 -2.73114026e-01 9.22281027e-01 2.83390433e-01
-1.00964320e+00 -7.97833204e-01 -1.51694834e+00 -4.02362078e-01
-5.82296669e-01 -2.82862931e-01 7.13681519e-01 1.84218809e-01
-5.97779512e-01 7.34392703e-01 -9.66165841e-01 -9.12095830e-02
7.22931683e-01 3.31088096e-01 -4.35559720e-01 -3.89547020e-01
-1.56670034e+00 1.42883515e+00 9.05040562e-01 3.26201320e-01
-5.53021669e-01 -8.00842404e-01 -8.86475086e-01 9.10481662e-02
6.55439854e-01 -5.87822914e-01 9.24637556e-01 -5.03605485e-01
-1.07946527e+00 7.53370821e-01 -3.69284391e-01 -7.43103683e-01
4.69958857e-02 -8.25569630e-01 -8.37469280e-01 -1.90802589e-01
-2.10467517e-01 6.60625994e-02 7.63239563e-01 -1.02915752e+00
-2.26463944e-01 -4.79251444e-01 -2.01672107e-01 2.25595668e-01
-6.35617912e-01 -5.30295134e-01 -1.75045982e-01 -7.08748996e-01
5.79355955e-01 -5.21449685e-01 2.52015054e-01 -3.69756937e-01
-2.04001755e-01 -3.09056133e-01 9.43334699e-01 -7.29797184e-01
1.73379123e+00 -2.02407098e+00 3.37711960e-01 -1.66602105e-01
4.04267222e-01 6.91873312e-01 -2.14852527e-01 1.53265581e-01
-3.86084884e-01 -2.47630373e-01 -3.69294025e-02 -3.20156783e-01
-6.53884053e-01 6.19404256e-01 -6.91481173e-01 2.38062218e-01
4.15811390e-01 1.18363464e+00 -1.10113358e+00 -3.69897157e-01
-1.56642556e-01 3.87436569e-01 -6.21208102e-02 5.12097657e-01
-9.22596455e-02 3.90719697e-02 6.27220720e-02 5.28691351e-01
8.74056756e-01 -2.84475535e-01 5.68096936e-01 -3.54316592e-01
5.10141611e-01 7.94032872e-01 -9.08264756e-01 1.74553132e+00
-4.74735647e-01 1.99664503e-01 -7.52472520e-01 -1.14628839e+00
1.32648718e+00 1.09803475e-01 2.87796315e-02 -1.10191166e+00
-3.89350504e-02 6.15616590e-02 2.92206351e-02 -4.11599249e-01
5.94184041e-01 -1.61781505e-01 2.49921441e-01 6.16939306e-01
2.94442594e-01 1.08616821e-01 -1.03348158e-01 4.03776169e-01
1.14714193e+00 9.57288146e-02 4.58233625e-01 1.59489453e-01
5.19696176e-01 -3.60393494e-01 9.73551452e-01 4.36379492e-01
-1.05061792e-01 1.35107234e-01 2.89466321e-01 -9.82056916e-01
-5.89189649e-01 -1.06108212e+00 2.08745599e-01 1.08745503e+00
1.74850002e-01 -6.28610194e-01 -4.30765599e-02 -9.81058359e-01
4.51349244e-02 8.91961634e-01 -7.69717157e-01 -1.29687607e+00
-1.06034160e+00 -1.11996424e+00 7.86732659e-02 6.66442633e-01
9.18672740e-01 -1.14442623e+00 -2.39358962e-01 4.80653107e-01
-2.50046343e-01 -5.79552770e-01 -2.02629268e-01 5.65118611e-01
-1.20081389e+00 -1.05536652e+00 -1.59606174e-01 -8.40050459e-01
7.00903594e-01 5.24525940e-01 1.36266053e+00 4.74501967e-01
1.65202320e-01 -3.62352252e-01 -3.83666724e-01 -2.42070877e-03
-1.98524147e-01 6.82435513e-01 2.62360156e-01 -1.77761674e-01
7.47586608e-01 -9.79908049e-01 -3.65825117e-01 -2.47073453e-02
-7.90666819e-01 -1.47881135e-02 6.57365620e-01 1.29510677e+00
6.01490974e-01 4.86720920e-01 1.00936484e+00 -1.34689879e+00
4.97475266e-01 -6.86335266e-01 3.53203304e-02 5.80073118e-01
-1.09879625e+00 4.48715329e-01 5.05641222e-01 -8.50420475e-01
-1.24956930e+00 -4.48864400e-01 2.96041489e-01 -3.64016473e-01
6.78178906e-01 4.54403192e-01 -4.27559644e-01 4.12114650e-01
7.41823137e-01 3.13125253e-01 -2.10688189e-01 -4.24975902e-01
5.09072721e-01 3.07380676e-01 7.47809947e-01 -5.60614586e-01
7.41481900e-01 1.55593097e-01 -2.59190530e-01 -1.40064299e-01
-1.46023202e+00 1.25522450e-01 -8.48936021e-01 2.09921092e-01
1.24186568e-01 -9.59367454e-01 -3.63872796e-01 6.26854897e-01
-1.09861052e+00 -1.38261333e-01 -5.41577220e-01 1.44587666e-01
-2.00964034e-01 -1.43486187e-01 -9.11099851e-01 -7.45873690e-01
-5.43236911e-01 -4.70635116e-01 4.59691852e-01 3.95722806e-01
-3.79506588e-01 -7.30107009e-01 1.85743406e-01 1.57785788e-01
4.20199722e-01 -9.06998664e-02 1.29813528e+00 -6.06814682e-01
-5.60131609e-01 -1.14662141e-01 -2.15639994e-01 3.68267119e-01
5.19394815e-01 -4.61141467e-01 -1.13259149e+00 -3.50353718e-01
2.26128057e-01 -4.80178088e-01 1.48300731e+00 -4.45760846e-01
9.55043674e-01 -4.91600782e-01 -6.22300625e-01 2.97727972e-01
1.02508116e+00 3.10565710e-01 9.12279487e-01 3.81442010e-01
7.89867759e-01 3.77740026e-01 8.10069442e-01 3.58889028e-02
7.19142139e-01 3.00542384e-01 5.68094514e-02 2.96913445e-01
-4.18443859e-01 -6.13042891e-01 1.66452304e-01 1.29661810e+00
1.18735626e-01 1.57058820e-01 -7.20870137e-01 5.89918256e-01
-2.13231421e+00 -8.17897677e-01 5.09393156e-01 2.10644078e+00
1.66853344e+00 6.86166942e-01 -3.37285876e-01 6.44641578e-01
5.26764512e-01 3.53874534e-01 -8.38577449e-01 -1.39344931e-01
-3.95609319e-01 5.46112835e-01 8.76315404e-03 4.68783498e-01
-8.76886547e-01 1.21792567e+00 5.65462685e+00 5.84750712e-01
-9.30234611e-01 3.00832450e-01 8.01760554e-01 -1.36734635e-01
-1.44206882e-01 1.55568436e-01 -9.28521693e-01 1.87274233e-01
1.05251157e+00 -2.01197267e-01 3.89023602e-01 6.57943785e-01
-5.36006391e-01 -3.33605349e-01 -1.20947111e+00 9.41522896e-01
8.49786177e-02 -1.38123190e+00 3.04998517e-01 -4.69103485e-01
4.93679583e-01 -5.81961095e-01 1.12770051e-01 9.13050830e-01
2.15858951e-01 -9.46232140e-01 3.11791122e-01 1.09982014e+00
4.29384500e-01 -9.73032117e-01 9.44239974e-01 2.73702025e-01
-9.96587396e-01 -1.77717447e-01 -6.18749261e-01 -5.18942952e-01
-2.16636375e-01 9.93030012e-01 -6.96231425e-01 7.01183438e-01
5.76669514e-01 1.01591575e+00 -1.06593084e+00 4.69208866e-01
-4.57118183e-01 3.86831254e-01 1.03097878e-01 3.39204401e-01
-6.61816120e-01 2.59440601e-01 1.15789780e-02 7.40960896e-01
5.76357096e-02 2.25770548e-01 -1.54075950e-01 4.45249349e-01
-3.74643058e-01 -1.32129326e-01 -5.62476039e-01 1.73768383e-02
1.03131163e+00 8.56687903e-01 -3.38502079e-01 -4.96227682e-01
-1.02727368e-01 9.72562432e-01 1.26642871e+00 2.63801098e-01
-4.21073943e-01 -3.87634069e-01 4.28324640e-01 1.27748355e-01
1.64686471e-01 -2.35503048e-01 -5.01080096e-01 -1.36317861e+00
2.04116389e-01 -7.18110681e-01 7.19263554e-01 -6.22500062e-01
-1.27099907e+00 7.56333590e-01 -2.38795578e-01 -7.33348787e-01
-1.15326524e-01 -2.72029918e-02 -4.28659648e-01 7.30905533e-01
-1.58418262e+00 -1.04993761e+00 -1.58113077e-01 3.74004930e-01
3.81199390e-01 -3.10950577e-01 9.50343251e-01 3.54600698e-01
-1.05390179e+00 8.65825236e-01 -4.77355719e-01 4.39638197e-02
9.20915306e-01 -8.44961941e-01 3.31596047e-01 7.87551880e-01
6.80823848e-02 1.05353296e+00 3.85805398e-01 -9.98853147e-01
-1.35706079e+00 -1.24365580e+00 1.24728644e+00 -3.99723232e-01
5.75907707e-01 -2.98656821e-01 -1.82393396e+00 1.00744629e+00
8.78057703e-02 3.27401578e-01 5.02651453e-01 4.54012573e-01
-9.51137781e-01 -7.88583636e-01 -9.14377332e-01 4.21807230e-01
1.27800941e+00 -6.54706478e-01 -1.28721440e+00 2.43749097e-01
1.13456893e+00 -3.27487350e-01 -8.52585137e-01 5.61817586e-01
3.28459769e-01 -7.11078167e-01 8.56667340e-01 -7.95270741e-01
1.31375983e-01 -3.52033198e-01 1.11581229e-01 -1.41055620e+00
-4.03772444e-01 -4.65564996e-01 -1.29456615e+00 1.59655476e+00
3.19279611e-01 -7.39985645e-01 4.80646312e-01 5.16978383e-01
2.21282169e-01 -9.65375066e-01 -8.06999147e-01 -8.23509634e-01
1.75581634e-01 5.12838215e-02 9.27177131e-01 1.06789052e+00
1.05869889e-01 9.28253829e-01 -5.70080578e-01 -9.03854426e-03
2.99868107e-01 1.53044567e-01 5.62174976e-01 -1.23359334e+00
-3.52996625e-02 2.71292422e-02 -2.71353096e-01 -7.91403294e-01
4.38495755e-01 -9.50598955e-01 -2.70376384e-01 -1.19679463e+00
2.83308327e-01 -6.03904307e-01 -8.05244505e-01 8.68278146e-01
-7.59444058e-01 -1.60090670e-01 -1.69746906e-01 3.90505522e-01
-6.76031411e-01 8.30596924e-01 1.21004534e+00 -4.46503997e-01
-3.59495550e-01 -1.57402068e-01 -1.04341602e+00 2.24804074e-01
7.15963483e-01 -6.25915110e-01 -7.02232540e-01 -2.81810910e-01
4.60556418e-01 -3.33038360e-01 -4.31825668e-02 -1.00554240e+00
4.82420444e-01 6.91887215e-02 7.39092469e-01 -6.95186913e-01
2.93514550e-01 -6.23392701e-01 1.03490420e-01 5.68256736e-01
-2.79703557e-01 6.12913584e-03 2.10670471e-01 5.44158340e-01
-1.99162170e-01 1.27052739e-02 7.74035871e-01 2.09521577e-02
-6.04634285e-01 3.16881567e-01 2.42469892e-01 1.24088749e-01
8.70005250e-01 3.29314321e-01 -7.83272266e-01 1.53664360e-02
-8.26109529e-01 3.57323766e-01 4.89925332e-02 7.02965319e-01
9.64018941e-01 -1.66298604e+00 -5.68920076e-01 4.24584895e-01
8.77084211e-02 4.20429856e-01 3.33610505e-01 5.50544024e-01
1.15457131e-02 -4.64573354e-02 -1.45117417e-01 -6.83368444e-02
-7.88804293e-01 1.27210319e+00 2.41195247e-01 -7.14597762e-01
-7.33133733e-01 9.42494154e-01 -2.38932550e-01 -1.39303997e-01
2.05901772e-01 -3.17551732e-01 -1.48791268e-01 4.61940616e-01
9.44918692e-01 3.13410163e-01 5.73498487e-01 -2.00171411e-01
-4.64176804e-01 1.13305114e-01 -1.03998220e+00 4.19518530e-01
1.47269571e+00 -2.52296537e-01 -4.56625432e-01 9.43276823e-01
1.02248776e+00 -3.11375827e-01 -9.85124648e-01 -9.45106983e-01
3.76693398e-01 -2.91081339e-01 -9.56623033e-02 -7.20628738e-01
-1.22539937e+00 7.82814920e-01 3.12001586e-01 -5.24941385e-02
1.34348047e+00 -1.25431836e-01 1.10482430e+00 7.86262631e-01
4.65129435e-01 -1.05787456e+00 4.33124453e-01 8.13847899e-01
8.91675055e-01 -1.00760758e+00 4.85161394e-01 -3.62906992e-01
-3.93540204e-01 1.01214242e+00 1.06785429e+00 -2.01651901e-01
8.02524865e-01 2.96291381e-01 -1.90555409e-01 -8.99695698e-03
-1.13680446e+00 3.19711268e-02 2.80063376e-02 6.74109399e-01
2.05265820e-01 -1.90036327e-01 -1.46615475e-01 1.03627527e+00
-1.74788073e-01 5.69151230e-02 2.63912648e-01 1.27396977e+00
-4.17447686e-01 -1.08426809e+00 2.68248804e-02 5.33918262e-01
1.28071308e-01 -2.67934114e-01 -2.41794318e-01 4.75991815e-01
2.76064634e-01 7.72507787e-01 2.03202907e-02 -7.62289584e-01
5.29885054e-01 6.23781025e-01 4.80668634e-01 -7.55930901e-01
-5.36878884e-01 -5.52260220e-01 1.63293574e-02 -4.55017120e-01
-2.80757099e-01 -4.21863526e-01 -1.28914070e+00 -4.61895913e-01
-5.61092794e-01 1.03651792e-01 -4.08796221e-01 1.01108670e+00
4.36648160e-01 9.00716424e-01 5.36776364e-01 -3.10597688e-01
-3.98960888e-01 -1.16598487e+00 -5.91333866e-01 3.01811010e-01
4.59461004e-01 -1.03464365e+00 -1.73793256e-01 -1.07650660e-01] | [9.158329010009766, 8.509424209594727] |
72328f5c-fedb-4e7c-91f5-7d582fda19c9 | statistical-estimation-for-covariance | 2305.11282 | null | https://arxiv.org/abs/2305.11282v1 | https://arxiv.org/pdf/2305.11282v1.pdf | Statistical Estimation for Covariance Structures with Tail Estimates using Nodewise Quantile Predictive Regression Models | This paper considers the specification of covariance structures with tail estimates. We focus on two aspects: (i) the estimation of the VaR-CoVaR risk matrix in the case of larger number of time series observations than assets in a portfolio using quantile predictive regression models without assuming the presence of nonstationary regressors and; (ii) the construction of a novel variable selection algorithm, so-called, Feature Ordering by Centrality Exclusion (FOCE), which is based on an assumption-lean regression framework, has no tuning parameters and is proved to be consistent under general sparsity assumptions. We illustrate the usefulness of our proposed methodology with numerical studies of real and simulated datasets when modelling systemic risk in a network. | ['Christis Katsouris'] | 2023-05-18 | null | null | null | null | ['variable-selection'] | ['methodology'] | [ 1.30570263e-01 -1.94334373e-01 1.50445811e-02 -1.33242950e-01
-1.28484219e-01 -5.41736662e-01 4.43696320e-01 1.23404130e-01
-1.58872292e-03 7.93964028e-01 5.39238639e-02 -8.06718409e-01
-1.13889229e+00 -9.06146109e-01 -2.47351304e-01 -5.96489608e-01
-5.87922096e-01 4.77944344e-01 -4.75234725e-02 9.50857177e-02
3.76988024e-01 4.84763652e-01 -9.37634170e-01 -7.02391088e-01
8.08652163e-01 1.07754731e+00 -4.37066168e-01 4.96766984e-01
2.29067057e-01 6.58156455e-01 -2.11972401e-01 -7.75272369e-01
6.99350536e-01 -3.24936926e-01 -2.66398937e-01 -2.51476131e-02
-1.59190401e-01 9.83647034e-02 1.84588790e-01 9.60005760e-01
2.02622652e-01 2.14634031e-01 1.17100990e+00 -1.16592038e+00
-2.66612291e-01 7.04908609e-01 -8.76736283e-01 6.79479599e-01
-3.63996029e-02 -3.83665591e-01 1.09285069e+00 -5.39861381e-01
3.73040259e-01 9.48031604e-01 7.15719581e-01 -3.61648351e-01
-1.40913725e+00 -6.36766136e-01 -2.25489903e-02 -9.17558968e-02
-1.25600719e+00 -9.19781849e-02 6.45397842e-01 -8.38770986e-01
7.38526046e-01 1.91513151e-01 5.65263391e-01 7.09943950e-01
6.77488208e-01 -9.47255269e-02 1.24499702e+00 -2.94931233e-01
3.87202173e-01 1.50631517e-01 5.66355467e-01 2.57506281e-01
8.95881951e-01 4.61939931e-01 -2.71539599e-01 -7.57220387e-01
8.64454687e-01 1.43709898e-01 2.19829716e-02 -3.44589174e-01
-9.41356897e-01 1.20956349e+00 -4.29295808e-01 3.89262289e-02
-8.33517373e-01 -7.65123442e-02 5.17510474e-01 8.52021456e-01
6.95165753e-01 4.50244173e-02 -4.21121627e-01 3.16071399e-02
-9.96541023e-01 -1.01144932e-01 1.15463448e+00 1.06850803e+00
2.39812452e-02 6.78828597e-01 -5.38156331e-02 2.76760757e-01
3.10764611e-01 7.55770624e-01 1.31946266e-01 -5.61089694e-01
5.77551007e-01 1.68680266e-01 1.36941537e-01 -1.01151168e+00
-5.79135418e-01 -1.03157926e+00 -1.08584249e+00 1.20574169e-01
2.81839877e-01 -8.49648774e-01 -2.48104990e-01 1.63618433e+00
2.17392817e-01 4.28515166e-01 -6.46077320e-02 4.61055905e-01
1.19579613e-01 2.50342995e-01 -5.74533530e-02 -9.63543296e-01
8.50688398e-01 -3.32388282e-01 -7.20032632e-01 6.12180471e-01
9.94307995e-02 -6.12552106e-01 1.22286357e-01 6.77690089e-01
-9.48366880e-01 -1.96612179e-01 -7.47111082e-01 1.09199440e+00
-8.89244117e-03 -1.86744511e-01 8.55674982e-01 1.15051150e+00
-8.06517363e-01 6.81489706e-01 -5.40809155e-01 -1.93694398e-01
-3.66476104e-02 3.93690079e-01 -1.77624643e-01 5.92112303e-01
-1.07276213e+00 6.33855700e-01 1.08115254e-02 1.36746451e-01
-7.88944840e-01 -7.26452351e-01 -2.92843014e-01 5.65047681e-01
5.49628913e-01 -8.53926182e-01 4.00510252e-01 -1.32447791e+00
-1.23690391e+00 1.20967906e-02 2.24676773e-01 -5.12763202e-01
6.54538870e-01 -1.86890319e-01 -6.80797577e-01 -9.86838490e-02
-1.83299959e-01 -9.59428906e-01 1.20836198e+00 -6.37913764e-01
-2.57692039e-01 -3.31247896e-01 -4.19626981e-01 -2.40757182e-01
4.65998938e-03 3.01515728e-01 4.40480202e-01 -1.03168726e+00
8.26541614e-03 -7.75023222e-01 -4.55302358e-01 -6.12446666e-01
-4.49476808e-01 2.19825834e-01 1.16193958e-01 -6.94604754e-01
1.49316192e+00 -1.68196011e+00 1.30814672e-01 1.39840186e+00
-3.88219282e-02 -3.64071250e-01 2.69458354e-01 1.01577878e+00
-6.07774556e-01 1.06989674e-01 -3.31524163e-02 2.89490879e-01
-1.69483930e-01 -3.63035411e-01 -6.33903801e-01 1.01635194e+00
-1.80373088e-01 4.48577804e-03 -3.19413632e-01 -4.39457715e-01
1.81625545e-01 8.42017084e-02 -3.55958104e-01 2.16937717e-02
4.06311691e-01 2.04478800e-01 -7.21941113e-01 5.70844710e-01
7.39827752e-01 -2.96476930e-01 2.59169400e-01 1.27401993e-01
-3.25488299e-01 -3.61415237e-01 -1.79865670e+00 5.51457942e-01
-2.19018891e-01 2.41692051e-01 -1.54941261e-01 -1.18089020e+00
1.03327453e+00 4.59210545e-01 5.49462080e-01 1.00403123e-01
5.20796835e-01 -8.53938311e-02 1.99251786e-01 -2.13836998e-01
-6.22159056e-02 -3.41101885e-01 -8.27280711e-03 6.21914387e-01
2.98487872e-01 6.21202886e-01 2.10031047e-01 7.77186230e-02
1.09897244e+00 -2.08061755e-01 7.36196160e-01 -6.90280318e-01
5.75591505e-01 -4.16520953e-01 7.97096908e-01 8.90701592e-01
6.50160089e-02 2.60984246e-02 1.01759541e+00 -5.21463566e-02
-9.30945456e-01 -1.15271378e+00 -5.43943763e-01 7.53229439e-01
-3.30524474e-01 1.88795049e-02 -1.47547111e-01 -3.31075788e-01
3.70104373e-01 6.20294154e-01 -8.17524374e-01 2.55427063e-01
-8.57846588e-02 -1.05845940e+00 1.09178394e-01 2.72435457e-01
-6.75158128e-02 -4.96154070e-01 -4.85709399e-01 2.68927664e-01
5.75434327e-01 -5.49551010e-01 -1.20608583e-01 3.49920243e-01
-1.04606342e+00 -1.22246039e+00 -5.29064775e-01 -1.09298825e-01
4.70159531e-01 -6.02818094e-02 1.20263040e+00 -1.35794684e-01
-1.38056308e-01 6.58006489e-01 -4.48823631e-01 -5.21618962e-01
-7.04639256e-02 -3.01497549e-01 1.14476308e-01 3.88436854e-01
1.80648431e-01 -9.43156004e-01 -5.31928360e-01 1.97953716e-01
-4.61346179e-01 -5.97349048e-01 7.18707025e-01 9.92673635e-01
3.51675212e-01 3.94819558e-01 1.14807320e+00 -1.31704748e+00
7.53917575e-01 -1.00755870e+00 -9.91142750e-01 4.92361844e-01
-1.03070498e+00 -1.57823816e-01 4.78499979e-01 -5.35561085e-01
-1.16957879e+00 -2.24768028e-01 6.75163150e-01 -2.60795206e-01
2.93675750e-01 1.00409341e+00 2.99693882e-01 -7.89770484e-03
1.59089535e-01 2.03308966e-02 -2.86196079e-02 -5.05525291e-01
2.31681332e-01 1.70854330e-01 1.68879092e-01 -4.26703870e-01
1.03969979e+00 3.56297225e-01 9.06211376e-01 -7.21844614e-01
-1.32405013e-01 -5.55434287e-01 -6.45181656e-01 -1.90525472e-01
4.75976944e-01 -8.76679957e-01 -8.73572946e-01 1.42676935e-01
-4.30254340e-01 2.46346921e-01 1.80835847e-03 9.91314530e-01
-4.98516649e-01 3.20055425e-01 -4.48440611e-01 -1.43869579e+00
-5.83364248e-01 -3.47840786e-01 1.35920584e-01 -1.28328716e-02
-1.76605150e-01 -1.39224851e+00 3.84105772e-01 -1.45139277e-01
2.90364355e-01 6.52080536e-01 1.04755306e+00 -1.23859346e+00
-3.28450948e-01 -2.78150797e-01 -2.53038555e-02 1.86140180e-01
-1.44462511e-01 5.03846109e-01 -2.93109983e-01 -4.71512586e-01
1.52144581e-01 3.69002044e-01 4.02255565e-01 6.94594920e-01
6.93277836e-01 -2.90837497e-01 -1.25396162e-01 6.29198968e-01
1.82149339e+00 1.06230065e-01 1.29623920e-01 2.24362820e-01
-3.91029008e-02 6.34950221e-01 7.08842218e-01 1.29638767e+00
-1.98839560e-01 2.31053457e-01 3.87337178e-01 7.59067759e-02
6.75185263e-01 1.25528291e-01 1.51934117e-01 9.51082468e-01
-3.30233186e-01 -2.75386184e-01 -7.66502976e-01 5.63147664e-01
-1.74975193e+00 -1.22357428e+00 -5.55243552e-01 2.33187270e+00
3.19755107e-01 1.84780002e-01 5.32296419e-01 -1.29578441e-01
8.96502793e-01 -1.10908069e-01 -3.80617112e-01 -5.37910819e-01
-6.66483566e-02 4.10036922e-01 1.16675043e+00 3.15658629e-01
-1.02352107e+00 1.89088121e-01 6.57744074e+00 6.45122707e-01
-7.08107889e-01 1.13898665e-01 4.14268225e-01 -2.74467934e-02
-3.25174958e-01 3.37558508e-01 -6.57520592e-01 6.60121083e-01
1.09124696e+00 -5.90525091e-01 6.73658550e-02 8.53324056e-01
3.94404888e-01 -1.27556711e-01 -6.23085082e-01 5.00296652e-01
-1.12007007e-01 -7.00382292e-01 -1.84350535e-01 1.48040444e-01
8.40885997e-01 -3.40192437e-01 1.80848077e-01 8.49922523e-02
5.92557132e-01 -1.00034237e+00 4.43914741e-01 1.15264940e+00
3.47636729e-01 -1.20997381e+00 1.00992036e+00 1.00032181e-01
-9.08004642e-01 -3.73387843e-01 -4.01378661e-01 -2.63466407e-02
2.81411678e-01 9.52015102e-01 -4.06870455e-01 8.59963834e-01
4.06040072e-01 6.08412266e-01 -4.25471336e-01 1.38855791e+00
1.18400507e-01 1.15190268e+00 -2.96104461e-01 -1.97740849e-02
-1.13964595e-01 -7.06809163e-01 7.44935632e-01 9.31488693e-01
6.72978222e-01 -3.16984430e-02 -2.01555327e-01 6.53268397e-01
5.59471428e-01 5.35870671e-01 -5.76092780e-01 2.86361694e-01
6.35729969e-01 1.15951943e+00 -7.49100149e-01 5.20755760e-02
-6.61770761e-01 1.62236802e-02 -2.32283086e-01 5.29329419e-01
-5.94637096e-01 -3.54425788e-01 2.23148644e-01 9.54326242e-02
7.10651398e-01 -1.95385396e-01 -5.87455451e-01 -9.68139231e-01
-2.73949027e-01 -7.23418951e-01 7.94858634e-01 -2.04017580e-01
-1.72459507e+00 3.69609326e-01 2.88943946e-01 -1.41283405e+00
-4.43495393e-01 -3.80060196e-01 -1.01914525e+00 1.12440944e+00
-1.04525888e+00 -7.64690399e-01 1.56436488e-01 8.10438097e-01
-3.18654254e-02 -8.38953495e-01 5.07937253e-01 6.46374226e-02
-6.41311646e-01 4.59637582e-01 7.39691198e-01 -1.64683163e-01
4.76733208e-01 -1.17023885e+00 -2.45077223e-01 8.93610775e-01
-2.93412507e-01 8.31326902e-01 1.03540337e+00 -1.08167136e+00
-1.14167488e+00 -8.78675342e-01 5.38132250e-01 -1.43999904e-01
1.17422545e+00 -8.18187147e-02 -4.96471226e-01 9.27116454e-01
3.13867509e-01 -3.73407930e-01 1.10399902e+00 4.12865549e-01
3.27881537e-02 -2.47266054e-01 -9.76694763e-01 1.83122694e-01
4.30800676e-01 3.00393463e-03 -5.07408619e-01 3.58709753e-01
2.40315109e-01 2.70503193e-01 -1.46038842e+00 4.11929369e-01
5.33067167e-01 -9.00244355e-01 9.68424559e-01 -7.19756544e-01
-1.77375361e-01 5.95229736e-04 -6.81437254e-02 -9.11632240e-01
-6.30795360e-01 -9.37700391e-01 1.04341861e-02 1.58179927e+00
3.39855701e-01 -1.00106919e+00 4.05976653e-01 4.59123880e-01
6.24882460e-01 -4.80683565e-01 -1.16361082e+00 -1.01777387e+00
1.02112219e-01 -3.42657082e-02 5.17542958e-01 1.02264690e+00
-1.27254859e-01 2.34499097e-01 -6.71879351e-01 1.97474942e-01
9.63939786e-01 1.83394819e-01 6.17647290e-01 -1.62375546e+00
-6.34712994e-01 -2.74865985e-01 -4.63912845e-01 -2.43606731e-01
1.70316279e-01 -5.29139519e-01 -5.98286986e-01 -6.56558037e-01
2.29135439e-01 -5.13361931e-01 -6.04796052e-01 -2.07028538e-01
1.56652182e-01 -5.38048923e-01 -9.94626358e-02 2.77583122e-01
-2.36075550e-01 4.24491078e-01 4.94956255e-01 5.83587587e-01
-2.56464840e-03 7.97608197e-01 -5.66643357e-01 5.95841229e-01
5.92510223e-01 -6.51241899e-01 -7.66017854e-01 6.00241482e-01
5.94841719e-01 8.93312573e-01 3.55031699e-01 -5.77159584e-01
-2.13914700e-02 -4.78588462e-01 3.51987958e-01 -8.47488105e-01
-2.12409839e-01 -1.15224361e+00 8.11646223e-01 4.38624114e-01
-4.04495686e-01 5.74642956e-01 -3.47384512e-01 1.06149518e+00
3.17311473e-02 -7.15501666e-01 4.27006245e-01 2.02772141e-01
-9.67224091e-02 1.45202372e-02 -4.25830066e-01 3.27626872e-03
1.21138489e+00 8.73717442e-02 -2.80706704e-01 -6.44540906e-01
-8.67132962e-01 1.35435984e-01 2.27844454e-02 -1.78694129e-01
1.61876813e-01 -1.14972866e+00 -9.49175656e-01 -2.07770392e-02
-5.44542730e-01 -7.95316577e-01 3.15576851e-01 1.24878693e+00
-3.62944901e-01 4.04598832e-01 -1.27610266e-01 -2.11919606e-01
-1.31601059e+00 8.29892635e-01 9.60064679e-02 -5.06689489e-01
-5.08993387e-01 2.63098508e-01 1.08082041e-01 1.94647953e-01
1.23699186e-02 1.36973068e-01 -7.08181620e-01 3.05619210e-01
1.36438876e-01 8.35932434e-01 -8.68006051e-02 -4.79416043e-01
-3.00910830e-01 4.93072182e-01 3.47403467e-01 -1.45794138e-01
1.66945827e+00 -3.92364949e-01 -3.14176530e-01 5.30689299e-01
7.57755578e-01 2.59829521e-01 -9.04398024e-01 -1.93282381e-01
3.33027959e-01 -6.09847248e-01 -1.70562044e-02 -5.37086070e-01
-1.15946865e+00 3.84428918e-01 6.36292398e-01 5.65470695e-01
1.07551217e+00 -5.02278566e-01 4.64957133e-02 2.39963561e-01
2.70986319e-01 -9.93530154e-01 -3.75861973e-01 3.10284495e-01
7.36699939e-01 -6.98118031e-01 4.98962015e-01 -4.60055858e-01
-7.22959757e-01 1.29232717e+00 3.44271362e-02 -7.69032598e-01
1.49070156e+00 1.63861185e-01 -3.89435232e-01 -2.35058784e-01
-1.12989783e+00 -1.35860205e-01 3.61551374e-01 6.95737839e-01
2.62796968e-01 5.56003675e-02 -8.90999854e-01 1.05451107e+00
-2.52325326e-01 -1.26539394e-01 8.64909172e-01 6.46028996e-01
-1.85822591e-01 -7.05396056e-01 -3.61072510e-01 8.69551957e-01
-1.07174873e+00 -2.87448764e-01 1.23130031e-01 1.12025785e+00
-3.19710940e-01 9.46250558e-01 -1.80370390e-01 1.40260821e-02
2.26035342e-01 -5.05224504e-02 6.52950723e-03 -3.59496564e-01
-8.22652102e-01 4.90660489e-01 4.04876560e-01 -2.60309894e-02
-4.32626873e-01 -1.26916730e+00 -3.08061540e-01 -4.11258936e-01
-8.20895672e-01 3.07211995e-01 4.00682628e-01 7.95693040e-01
1.02553159e-01 5.88178992e-01 1.08971977e+00 -1.24678075e-01
-1.18912876e+00 -9.85689759e-01 -1.43582690e+00 1.53416887e-01
-7.35811517e-02 -8.97186339e-01 -7.58123517e-01 -1.02242529e-01] | [5.168966770172119, 4.061452388763428] |
23b99097-6c48-4aa2-a5fe-4e61ce1864da | compound-probabilistic-context-free-grammars | 1906.10225 | null | https://arxiv.org/abs/1906.10225v9 | https://arxiv.org/pdf/1906.10225v9.pdf | Compound Probabilistic Context-Free Grammars for Grammar Induction | We study a formalization of the grammar induction problem that models sentences as being generated by a compound probabilistic context-free grammar. In contrast to traditional formulations which learn a single stochastic grammar, our grammar's rule probabilities are modulated by a per-sentence continuous latent variable, which induces marginal dependencies beyond the traditional context-free assumptions. Inference in this grammar is performed by collapsed variational inference, in which an amortized variational posterior is placed on the continuous variable, and the latent trees are marginalized out with dynamic programming. Experiments on English and Chinese show the effectiveness of our approach compared to recent state-of-the-art methods when evaluated on unsupervised parsing. | ['Yoon Kim', 'Alexander M. Rush', 'Chris Dyer'] | 2019-06-24 | compound-probabilistic-context-free-grammars-1 | https://aclanthology.org/P19-1228 | https://aclanthology.org/P19-1228.pdf | acl-2019-7 | ['constituency-grammar-induction'] | ['natural-language-processing'] | [ 3.58141422e-01 7.17688143e-01 -1.61610842e-01 -7.69495249e-01
-1.17454660e+00 -5.94089687e-01 7.48800337e-01 -5.26386276e-02
-3.80235940e-01 8.95830929e-01 2.76382446e-01 -6.93137586e-01
1.44614175e-01 -8.11554611e-01 -9.18125808e-01 -8.99485588e-01
-6.10207021e-02 1.06537759e+00 1.38385892e-01 9.66430083e-02
9.07612517e-02 -2.23755598e-01 -1.21456921e+00 9.39925462e-02
9.85486269e-01 1.22414514e-01 6.19415283e-01 9.19226527e-01
-5.38457513e-01 7.66391873e-01 -5.05762339e-01 -5.19108772e-01
-4.40265417e-01 -4.03621614e-01 -8.17126572e-01 1.11310430e-01
1.29703566e-01 -1.53374657e-01 1.48168474e-01 1.38869417e+00
-1.31483287e-01 1.18004397e-01 5.94960332e-01 -6.51618361e-01
-7.17740774e-01 1.30854154e+00 -2.86190957e-01 1.61110148e-01
3.21605414e-01 -2.19236732e-01 1.46513247e+00 -7.18755066e-01
7.31907666e-01 1.97772038e+00 1.25553727e-01 5.76655686e-01
-1.98826969e+00 -2.41392449e-01 7.81713486e-01 -3.53031039e-01
-1.03560257e+00 -1.30013704e-01 7.64049530e-01 -2.53556877e-01
1.16190863e+00 3.91270220e-02 2.29554757e-01 1.28115082e+00
5.76541781e-01 1.01490402e+00 1.01138127e+00 -9.46620286e-01
4.80691999e-01 -1.14650369e-01 4.62293595e-01 7.13794529e-01
1.11563556e-01 -1.88381761e-01 -5.04579663e-01 -4.62076485e-01
3.28995824e-01 -3.38919640e-01 2.30098099e-01 -1.80916354e-01
-8.20724249e-01 1.24869430e+00 -4.61208582e-01 -5.75343985e-03
-3.76121968e-01 4.65740532e-01 8.66945758e-02 -2.08730161e-01
7.49791145e-01 -2.89474458e-01 -7.20982492e-01 -3.97615612e-01
-1.01276755e+00 4.63987768e-01 1.04996276e+00 1.05885887e+00
8.53140175e-01 -1.54440433e-01 -1.85788438e-01 6.48678541e-01
9.65871811e-01 7.78547347e-01 -3.85824665e-02 -1.08638501e+00
6.41909897e-01 7.78455436e-02 3.07521243e-02 -2.12528348e-01
4.21582386e-02 -5.42316921e-02 -3.47509354e-01 -2.43095327e-02
4.04216021e-01 -3.20446491e-01 -1.16016757e+00 2.15464616e+00
2.61270106e-01 -3.59185934e-02 4.60820571e-02 2.78279930e-01
1.74455777e-01 8.57359469e-01 5.72520554e-01 -6.75997853e-01
1.28508973e+00 -5.12994289e-01 -1.08103073e+00 -4.20616895e-01
2.79193819e-01 -4.68362033e-01 1.11373878e+00 4.17112559e-01
-1.31874466e+00 -3.19471151e-01 -5.60047686e-01 -1.67208761e-01
-2.57398672e-02 -3.40996355e-01 8.08437765e-01 9.40630615e-01
-1.12254500e+00 4.60642338e-01 -1.57238460e+00 1.32129252e-01
8.09936300e-02 3.37887943e-01 -6.53655380e-02 2.12136880e-01
-1.04361308e+00 5.27458787e-01 5.58630288e-01 9.73837003e-02
-9.99242425e-01 -3.71762007e-01 -1.03005707e+00 -5.15017360e-02
3.45220596e-01 -6.30359292e-01 1.61652625e+00 -5.78982651e-01
-1.93912244e+00 8.53410065e-01 -8.69783938e-01 -4.22983646e-01
1.00141607e-01 -2.67389446e-01 5.81267662e-02 6.07864670e-02
3.01832765e-01 5.05362153e-01 8.68200481e-01 -1.37885618e+00
-5.51322460e-01 -3.16131979e-01 1.11889705e-01 1.19033642e-01
4.29757118e-01 4.19554085e-01 -6.23825431e-01 -4.92130905e-01
3.59005064e-01 -1.07917488e+00 -3.22464943e-01 -9.56061959e-01
-7.17156589e-01 -5.48718929e-01 3.86385381e-01 -7.82177091e-01
1.42575550e+00 -1.90494275e+00 5.98322213e-01 -4.71026152e-02
-1.21795826e-01 -5.34333766e-01 2.17076689e-01 3.57537746e-01
1.18798979e-01 5.45752168e-01 -8.15709889e-01 -1.02988183e+00
3.85095686e-01 8.64003301e-01 -5.88422775e-01 2.02437788e-02
2.86058336e-01 7.35886097e-01 -9.25797105e-01 -7.41466463e-01
3.10207233e-02 4.00786996e-01 -7.66592801e-01 9.57505107e-02
-8.84673059e-01 2.92148679e-01 -7.31923342e-01 5.07171154e-01
6.44544601e-01 -6.07477874e-02 9.39440072e-01 8.15993071e-01
-1.32131889e-01 8.11549306e-01 -1.02049518e+00 1.73375344e+00
-3.88123572e-01 6.11037500e-02 2.40415767e-01 -8.66165280e-01
5.69459379e-01 2.27952465e-01 -2.00090915e-01 1.36901423e-01
-1.30102560e-01 -2.59287596e-01 -1.83407396e-01 -3.11203688e-01
4.32015568e-01 -4.86077160e-01 -5.73180377e-01 4.61092114e-01
2.86688715e-01 -3.77913535e-01 3.52364153e-01 6.11530662e-01
8.08211923e-01 7.94692576e-01 -4.73459512e-02 -5.01422048e-01
1.14659034e-01 -1.15849696e-01 9.45155621e-01 1.22298467e+00
3.55286598e-02 3.12878847e-01 1.05934548e+00 -7.19679669e-02
-5.74350476e-01 -1.64906311e+00 -3.64895284e-01 1.38323069e+00
-3.04126650e-01 -5.50984979e-01 -1.10102892e+00 -8.70762825e-01
-2.71128327e-01 1.48171961e+00 -7.48143494e-01 5.44543564e-01
-7.01983929e-01 -1.31696725e+00 1.45790935e-01 3.55085641e-01
-5.99166229e-02 -1.25620472e+00 -3.69638175e-01 6.41789854e-01
-3.49443704e-01 -1.22400498e+00 -1.81816757e-01 5.64326525e-01
-1.17587483e+00 -5.56580961e-01 9.64736491e-02 -6.36360645e-01
5.94093680e-01 -5.69798410e-01 1.57160115e+00 -2.45202273e-01
2.24264581e-02 2.81779528e-01 -1.73749626e-01 -5.23293912e-01
-9.01488304e-01 -3.09484755e-03 -1.71285972e-01 -2.11163104e-01
6.37316525e-01 -6.43005967e-01 9.67465565e-02 -5.91454148e-01
-7.44657099e-01 -3.29035422e-04 3.28897268e-01 8.89097333e-01
9.04734731e-01 -7.83056468e-02 6.66888282e-02 -1.47958028e+00
4.43855971e-01 -5.27644932e-01 -9.48546648e-01 3.26590478e-01
-3.27621430e-01 4.92160290e-01 2.70065606e-01 -2.61113077e-01
-1.79695559e+00 1.43021509e-01 -5.25583960e-02 1.01455905e-01
-3.80136192e-01 6.86959147e-01 -4.88414317e-01 8.33128333e-01
5.66683747e-02 2.56376207e-01 -1.85997322e-01 -4.80557680e-01
6.17832363e-01 3.41837645e-01 4.72402126e-01 -1.16061878e+00
6.11388028e-01 3.40160042e-01 3.22487578e-02 -7.97871053e-01
-9.46007073e-01 9.82102007e-02 -8.75358224e-01 1.64379120e-01
1.38320589e+00 -8.04522932e-01 -4.36604559e-01 1.73924550e-01
-1.38858104e+00 -5.53984344e-01 -1.88202679e-01 6.22956693e-01
-6.72646224e-01 5.77074111e-01 -8.83208513e-01 -1.38566434e+00
4.34515774e-02 -1.13713074e+00 1.37286448e+00 1.52429864e-01
-2.39474773e-01 -1.46893430e+00 3.50209653e-01 2.09470659e-01
-1.59300759e-01 1.68607396e-03 1.25843310e+00 -3.00999999e-01
-7.82858789e-01 8.92562270e-02 3.32251132e-01 2.90153176e-01
-1.14533439e-01 5.05195022e-01 -8.60129654e-01 1.58746496e-01
3.79019156e-02 -1.72955897e-02 8.88966084e-01 8.56416702e-01
8.39821041e-01 -9.62635204e-02 -4.66132581e-01 3.95501733e-01
1.43736029e+00 9.65903699e-02 3.21436137e-01 -7.41356388e-02
5.12629747e-01 5.47131896e-01 3.30645680e-01 2.83289284e-01
5.77640891e-01 1.84094161e-01 3.92449766e-01 5.33909202e-01
4.80637401e-01 -5.19464195e-01 6.68386102e-01 9.12952960e-01
-2.15934049e-02 -3.84865612e-01 -8.64111304e-01 5.55038810e-01
-1.82460594e+00 -9.17861462e-01 -2.44850069e-01 1.91004062e+00
1.10839438e+00 5.76814175e-01 -2.11169064e-01 -2.89275259e-01
7.38515854e-01 2.46787801e-01 -1.27887860e-01 -7.82237351e-01
-2.25661263e-01 5.87862253e-01 2.68226296e-01 1.30845749e+00
-1.09388506e+00 1.42477548e+00 7.94377279e+00 4.88197088e-01
-4.19480354e-01 3.09508950e-01 4.33222592e-01 -1.41151166e-02
-7.54082263e-01 7.87890196e-01 -1.31282806e+00 4.60895866e-01
1.23240232e+00 1.45247251e-01 4.66389865e-01 8.28506589e-01
2.47153178e-01 -4.43353742e-01 -9.26236331e-01 2.03419730e-01
-1.49339244e-01 -1.00247157e+00 5.76405488e-02 3.63690615e-01
6.00507140e-01 1.06948070e-01 -5.97115718e-02 5.11704743e-01
1.21959269e+00 -6.98980331e-01 8.76055896e-01 5.29238999e-01
3.89915794e-01 -6.73527896e-01 3.47587645e-01 6.52578413e-01
-7.54792392e-01 1.34142324e-01 -4.64062244e-01 -2.85950422e-01
7.32989550e-01 7.27712512e-01 -4.82061476e-01 2.83223540e-01
6.28404498e-01 1.00548513e-01 -8.45697299e-02 -1.40708759e-01
-9.99003828e-01 1.42658889e+00 -4.58696276e-01 -1.80634052e-01
2.59109825e-01 -7.24873304e-01 7.06332803e-01 1.48272252e+00
-1.94710970e-01 2.46890962e-01 4.41141069e-01 1.16945982e+00
5.56751937e-02 -1.23548508e-01 -2.94455022e-01 1.06971301e-01
4.31325585e-01 9.48156536e-01 -8.77750099e-01 -6.25072539e-01
-4.27210063e-01 7.85289228e-01 5.22594512e-01 5.18156230e-01
-8.54126453e-01 -7.71437660e-02 3.75812531e-01 -3.33575010e-01
7.26285458e-01 -4.86949116e-01 -3.75839442e-01 -1.21793902e+00
1.00965105e-01 -4.81912762e-01 4.18130666e-01 -4.54186827e-01
-1.20024943e+00 4.55605507e-01 7.41847336e-01 -1.76817730e-01
-8.68171871e-01 -6.96603179e-01 -7.34517872e-01 1.28415048e+00
-1.25052583e+00 -1.10547638e+00 5.41041672e-01 8.89617279e-02
6.80138886e-01 9.90836620e-02 1.26598704e+00 -5.53748190e-01
-7.60684550e-01 1.66301414e-01 3.78253907e-02 -4.68305498e-02
-2.81641334e-02 -1.92454326e+00 7.53940165e-01 1.28872669e+00
4.50569600e-01 1.03073943e+00 1.11634696e+00 -9.28328991e-01
-1.42421103e+00 -6.92819059e-01 1.34531820e+00 -9.31132317e-01
7.91635811e-01 -8.95668685e-01 -1.04966938e+00 1.22289920e+00
4.63863075e-01 -1.40282467e-01 7.75991917e-01 6.36984766e-01
-2.56305873e-01 5.10704458e-01 -8.37290883e-01 3.42971534e-01
7.70165861e-01 -4.52411532e-01 -1.13424158e+00 3.89904708e-01
8.42818856e-01 -3.00822109e-01 -6.69333696e-01 7.44800270e-02
3.80630612e-01 -6.15557730e-01 4.75960463e-01 -7.79314041e-01
3.24193686e-01 -1.14616714e-01 -5.07174015e-01 -9.95188475e-01
-4.31607544e-01 -1.06194091e+00 -2.50997096e-01 1.26617241e+00
6.57752335e-01 -4.01084691e-01 8.63201857e-01 8.11259031e-01
-1.87308475e-01 -3.56664628e-01 -1.07833111e+00 -5.96198082e-01
5.80688834e-01 -8.12194049e-01 3.69621962e-01 5.63079238e-01
3.94634455e-02 6.62678599e-01 2.45930757e-02 4.94381011e-01
1.10419667e+00 9.51571390e-02 4.03499097e-01 -1.13789272e+00
-7.74590492e-01 -1.00313812e-01 1.87166959e-01 -1.03176796e+00
8.42540860e-01 -8.26042831e-01 6.38652265e-01 -1.46265912e+00
3.85993630e-01 -1.57516852e-01 4.81926138e-03 3.73791873e-01
-6.72431350e-01 -4.66881603e-01 -5.22049107e-02 -2.76031107e-01
-5.46719909e-01 5.42331874e-01 6.56011641e-01 1.02096774e-01
-2.88716555e-01 -6.91509694e-02 -6.04751348e-01 9.35138941e-01
5.47373831e-01 -9.16703105e-01 -4.54717904e-01 -5.00851870e-01
3.04465890e-01 3.77931237e-01 3.15464512e-02 -1.59913540e-01
6.33494556e-03 -4.34387952e-01 -1.41074002e-01 -8.28886807e-01
1.65143549e-01 2.35100426e-02 -4.15381528e-02 1.20777555e-01
-1.76046312e-01 2.45527904e-02 3.92486453e-02 7.75436699e-01
-2.10573494e-01 -4.18205321e-01 5.18574178e-01 -3.45879972e-01
-2.97535479e-01 1.50517952e-02 -8.24556947e-01 3.27488959e-01
5.55496693e-01 1.92808792e-01 1.85429648e-01 -1.61306217e-01
-1.19259536e+00 -6.03228547e-02 9.33663920e-02 1.37360334e-01
3.88797760e-01 -6.94330990e-01 -8.95823121e-01 5.81129715e-02
-3.55515599e-01 3.29130739e-01 2.33989090e-01 3.06489974e-01
-3.29008937e-01 3.46314847e-01 5.89194119e-01 -8.11621904e-01
-9.70642745e-01 6.14962816e-01 -6.80103665e-04 -6.81621730e-01
-5.63809037e-01 1.01431298e+00 3.24366570e-01 -5.31821728e-01
3.95333730e-02 -4.83731270e-01 1.19057752e-01 -2.80511200e-01
3.19383353e-01 -1.45932987e-01 -3.39938790e-01 -4.31443900e-01
-3.68015200e-01 2.44571805e-01 -1.88827902e-01 -8.58165026e-01
1.21809494e+00 -2.55125850e-01 -5.61321378e-01 1.08639431e+00
5.60287833e-01 2.99461007e-01 -1.43044364e+00 -5.90626821e-02
2.26579070e-01 -6.02701455e-02 7.18080029e-02 -6.91528678e-01
-2.34926775e-01 8.74141514e-01 5.00310101e-02 3.34890515e-01
6.88716590e-01 3.03741336e-01 1.90573886e-01 1.91316128e-01
4.19041395e-01 -1.17911744e+00 -5.52440226e-01 8.50733638e-01
5.33448100e-01 -1.06082332e+00 -1.78705364e-01 -6.89190269e-01
-6.02544308e-01 8.82395387e-01 -9.27904621e-03 -2.12904409e-01
8.94095004e-01 7.74803579e-01 -6.07348420e-02 8.90533924e-02
-1.22018135e+00 -5.25730401e-02 -1.30177364e-01 4.48490024e-01
7.20676899e-01 5.55398226e-01 -2.88365364e-01 8.55718911e-01
-4.30107683e-01 -3.23839873e-01 6.26383126e-01 1.21841073e+00
-4.76437598e-01 -1.61772788e+00 -3.19919169e-01 8.27712491e-02
-1.05298686e+00 -5.00137806e-01 7.44669279e-03 7.69058347e-01
-2.74560805e-02 1.09253192e+00 3.21023852e-01 3.42007428e-01
-1.16494574e-01 7.35983193e-01 6.78913295e-01 -1.25510025e+00
-1.36878222e-01 4.57270265e-01 1.63212612e-01 -3.29494387e-01
-2.58239925e-01 -1.38384235e+00 -1.40516114e+00 2.69240022e-01
-2.53200650e-01 4.52686250e-01 9.55419719e-01 1.23438656e+00
-6.39223531e-02 4.79753405e-01 3.74715954e-01 -6.27375722e-01
-8.37572396e-01 -1.25263751e+00 -5.43642461e-01 1.26569271e-01
1.50554165e-01 -4.56103027e-01 -5.67440450e-01 4.32766408e-01] | [10.426616668701172, 9.649282455444336] |
ec30a64a-ad22-4a96-98a1-91d8e8633cff | grad-cam-improved-visual-explanations-for | 1710.11063 | null | http://arxiv.org/abs/1710.11063v3 | http://arxiv.org/pdf/1710.11063v3.pdf | Grad-CAM++: Improved Visual Explanations for Deep Convolutional Networks | Over the last decade, Convolutional Neural Network (CNN) models have been
highly successful in solving complex vision problems. However, these deep
models are perceived as "black box" methods considering the lack of
understanding of their internal functioning. There has been a significant
recent interest in developing explainable deep learning models, and this paper
is an effort in this direction. Building on a recently proposed method called
Grad-CAM, we propose a generalized method called Grad-CAM++ that can provide
better visual explanations of CNN model predictions, in terms of better object
localization as well as explaining occurrences of multiple object instances in
a single image, when compared to state-of-the-art. We provide a mathematical
derivation for the proposed method, which uses a weighted combination of the
positive partial derivatives of the last convolutional layer feature maps with
respect to a specific class score as weights to generate a visual explanation
for the corresponding class label. Our extensive experiments and evaluations,
both subjective and objective, on standard datasets showed that Grad-CAM++
provides promising human-interpretable visual explanations for a given CNN
architecture across multiple tasks including classification, image caption
generation and 3D action recognition; as well as in new settings such as
knowledge distillation. | ['Vineeth N. Balasubramanian', 'Anirban Sarkar', 'Prantik Howlader', 'Aditya Chattopadhyay'] | 2017-10-30 | null | null | null | null | ['3d-human-action-recognition'] | ['computer-vision'] | [ 2.81234115e-01 6.57787681e-01 -1.44709632e-01 -5.90377986e-01
-9.88169387e-02 -2.29690731e-01 8.30178857e-01 8.31183493e-02
5.05726524e-02 5.53295016e-01 3.25579256e-01 -3.12970996e-01
-1.71569198e-01 -5.54773450e-01 -9.39501345e-01 -5.09591758e-01
2.72146136e-01 3.50146174e-01 -1.42032339e-03 -6.51824474e-02
3.96511495e-01 5.97141802e-01 -1.60608757e+00 7.44186521e-01
6.13549829e-01 1.10383642e+00 1.10870190e-01 5.47402143e-01
-1.90054104e-01 9.32609022e-01 -6.67811096e-01 -5.72928131e-01
-2.12257113e-02 -4.43285525e-01 -9.20286477e-01 4.33528304e-01
6.06213570e-01 -1.98819667e-01 -1.79751128e-01 8.74804437e-01
1.40039131e-01 -3.68405171e-02 6.89723194e-01 -1.63697827e+00
-1.43058765e+00 3.20367932e-01 -2.25647062e-01 1.33385956e-01
1.58452004e-01 2.97058880e-01 1.09119749e+00 -1.08027434e+00
4.51096624e-01 1.48700702e+00 4.47654426e-01 9.60967898e-01
-1.24896264e+00 -3.29813123e-01 3.82972956e-01 6.39617503e-01
-9.31712091e-01 1.97291896e-02 7.07115591e-01 -4.47975039e-01
1.30243087e+00 2.70179957e-01 8.27079058e-01 1.07677805e+00
2.27614731e-01 1.03221655e+00 1.04256153e+00 -5.62044382e-01
2.33386248e-01 2.31039092e-01 1.24423936e-01 7.35142589e-01
3.21126789e-01 1.76569015e-01 -5.52065194e-01 2.35967219e-01
9.65771854e-01 1.33453444e-01 -2.33473346e-01 -5.28820157e-01
-1.15357542e+00 8.82410586e-01 1.08447957e+00 1.43956557e-01
-5.47948539e-01 4.52278405e-01 1.18685644e-02 -1.81065515e-01
5.75666845e-01 6.17765963e-01 -5.55900753e-01 2.37361953e-01
-4.10595149e-01 5.33466637e-01 4.31049973e-01 7.34237432e-01
6.94425166e-01 3.47521096e-01 -4.01876926e-01 5.27920604e-01
3.82433772e-01 2.12980598e-01 4.59840864e-01 -8.85741591e-01
3.10319096e-01 9.03713048e-01 6.15295321e-02 -1.18339074e+00
-4.21623856e-01 -5.49874663e-01 -1.03156805e+00 6.49143040e-01
2.53423870e-01 2.14511842e-01 -1.33078706e+00 1.79435349e+00
-4.05655568e-03 1.79102048e-01 1.93750590e-01 1.14521801e+00
1.03703415e+00 5.60406327e-01 2.61140108e-01 2.64096797e-01
1.21075368e+00 -1.14338517e+00 -7.18876541e-01 -5.89393318e-01
2.95081049e-01 -4.09702063e-01 1.04680049e+00 2.87808985e-01
-1.08889270e+00 -9.01687741e-01 -9.92534280e-01 -2.78529853e-01
-5.29324651e-01 2.67044604e-01 8.46266270e-01 3.12919527e-01
-1.20215189e+00 5.72822154e-01 -7.52228260e-01 -4.51739579e-01
6.91491663e-01 2.60162055e-01 -4.08192575e-01 4.07154039e-02
-7.82261789e-01 1.18087649e+00 4.33093250e-01 3.70151043e-01
-8.59098613e-01 -5.09910285e-01 -8.69136870e-01 4.01235938e-01
1.46864384e-01 -1.03681982e+00 1.22508132e+00 -1.17165613e+00
-1.02135086e+00 8.15374911e-01 -2.21877158e-01 -7.97303259e-01
3.13786417e-01 -3.11294198e-01 -2.05735981e-01 -1.06076732e-01
-1.26967311e-01 1.34582376e+00 7.48237908e-01 -1.50264537e+00
-2.95173287e-01 -2.81160235e-01 5.03363907e-01 1.12904370e-01
-1.60838902e-01 -4.68673497e-01 -1.49523824e-01 -6.24794960e-01
2.08611518e-01 -9.31483448e-01 -3.29736471e-01 4.34590220e-01
-6.15811110e-01 -2.51155674e-01 9.38226700e-01 -4.66290742e-01
5.44533968e-01 -1.92067635e+00 2.51245856e-01 -1.04899079e-01
3.43778074e-01 5.79795301e-01 -2.07540929e-01 3.57316397e-02
-4.95310456e-01 2.91884005e-01 -1.85007542e-01 -4.42040533e-01
2.85657831e-02 4.81330812e-01 -4.42674011e-01 -8.59754626e-03
8.13207209e-01 1.22324932e+00 -8.40827405e-01 -1.04014762e-01
6.12459600e-01 6.77213371e-01 -5.05738378e-01 2.37006053e-01
-5.49496412e-01 4.73532259e-01 -2.34421566e-01 5.11790514e-01
4.26655650e-01 -4.83594745e-01 -1.44286558e-01 -2.05944896e-01
2.47328907e-01 -9.78627824e-04 -9.06227171e-01 1.37355530e+00
-2.51721531e-01 1.10594761e+00 -7.03950822e-01 -1.03569424e+00
1.10648334e+00 3.00676376e-01 -6.51782900e-02 -7.27665365e-01
1.04156345e-01 7.84799531e-02 3.48680392e-02 -6.88882470e-01
4.62525249e-01 -2.22124100e-01 4.03293580e-01 1.28393948e-01
1.80236213e-02 -3.81073505e-02 -8.54629874e-02 7.52601624e-02
7.39990771e-01 2.69684106e-01 5.02652586e-01 5.34174219e-02
5.48688352e-01 1.29203916e-01 2.06232741e-01 7.41718173e-01
-1.93448409e-01 8.93441439e-01 6.40532017e-01 -1.05817759e+00
-1.02822971e+00 -8.23768795e-01 1.82196364e-01 4.80727822e-01
1.75505087e-01 -1.40847251e-01 -8.17354023e-01 -6.92390621e-01
5.86201213e-02 9.58397448e-01 -1.00190389e+00 -3.27683568e-01
-3.24050456e-01 -4.28526253e-01 2.44130597e-01 9.97670710e-01
6.23714805e-01 -1.58735776e+00 -8.27710509e-01 -4.77850586e-02
-6.16323203e-02 -1.14856362e+00 7.90084526e-02 2.21978083e-01
-9.81100678e-01 -1.14412916e+00 -7.35944569e-01 -6.47356570e-01
1.01268542e+00 4.35012877e-01 1.13261056e+00 4.22808349e-01
-5.20251989e-01 3.22972715e-01 -3.53378147e-01 -7.46025026e-01
-3.00935477e-01 -3.38857502e-01 -1.91862956e-01 7.74928778e-02
3.88729393e-01 -1.66789398e-01 -7.06907034e-01 7.73295090e-02
-1.05970120e+00 5.88642895e-01 8.36260259e-01 8.24264884e-01
5.69127023e-01 -3.82181555e-01 4.07768279e-01 -6.84853077e-01
6.19293809e-01 -3.40746760e-01 -3.14215004e-01 2.45671511e-01
-6.49565458e-01 2.99462348e-01 4.88238215e-01 -3.86672914e-01
-9.72932994e-01 1.17439933e-01 -1.47047145e-02 -5.61520576e-01
-4.68276888e-01 4.57761139e-01 5.98691590e-02 -4.96620312e-02
7.73789346e-01 1.30121574e-01 -8.89816359e-02 -4.12155926e-01
5.61317742e-01 3.57158661e-01 6.18369579e-01 -8.03286880e-02
6.42283976e-01 4.71562713e-01 1.89899787e-01 -4.84076470e-01
-1.22948301e+00 3.66450809e-02 -6.62916660e-01 -3.64957660e-01
1.13679504e+00 -5.68530977e-01 -7.59423375e-01 2.73690879e-01
-1.73727405e+00 -1.22029305e-01 -2.45653450e-01 3.79563004e-01
-5.56510270e-01 -2.78358776e-02 -9.70176905e-02 -6.87144697e-01
-1.02750666e-01 -1.20853901e+00 1.05849624e+00 5.14584363e-01
-4.23145205e-01 -1.08857489e+00 -2.96100587e-01 4.41075891e-01
4.66434807e-01 4.35640544e-01 1.26623762e+00 -6.48087740e-01
-8.79359066e-01 -2.39704087e-01 -6.89936161e-01 5.60381770e-01
-1.08183339e-01 -1.10283442e-01 -1.14051187e+00 6.94576353e-02
-3.54335994e-01 -3.74267310e-01 1.01216054e+00 5.38546562e-01
1.62364280e+00 -4.02331740e-01 -2.54281789e-01 3.64395440e-01
1.29343879e+00 1.67897820e-01 8.24936032e-01 4.18067127e-01
6.56859994e-01 5.69818318e-01 4.68090266e-01 2.63855428e-01
2.34916985e-01 8.27148974e-01 1.22054076e+00 -5.61682045e-01
-4.22866702e-01 -2.69663066e-01 -8.55313540e-02 -3.63772064e-02
-3.32537770e-01 -4.33984399e-01 -8.06031942e-01 5.97186327e-01
-2.15650249e+00 -8.88410568e-01 -4.19647008e-01 1.79868078e+00
2.40204930e-01 2.05744654e-01 -2.82812029e-01 1.77584916e-01
5.41220069e-01 1.30761847e-01 -6.55092776e-01 -7.09768295e-01
-2.39572972e-01 2.23680660e-01 1.23798035e-01 3.71507108e-01
-8.80513310e-01 8.91583800e-01 6.37313032e+00 2.88956642e-01
-1.13610542e+00 -2.16441467e-01 7.80784547e-01 3.44516724e-01
-2.88240016e-01 1.73992645e-02 -5.38726568e-01 2.68492736e-02
4.92759407e-01 1.64391585e-02 9.57158953e-02 1.12301672e+00
1.88687801e-01 8.08733255e-02 -1.29545605e+00 9.74932969e-01
3.30949426e-01 -1.76075995e+00 6.67106032e-01 1.50439646e-02
7.56715894e-01 -3.94814223e-01 3.62133235e-01 1.13323927e-01
-4.97896373e-02 -1.41058373e+00 9.27385271e-01 6.26449823e-01
2.60599494e-01 -5.18559456e-01 9.63630855e-01 3.04279417e-01
-7.93143094e-01 -1.14880629e-01 -5.34082413e-01 -4.57053632e-01
9.44179669e-03 2.95642436e-01 -1.19036746e+00 3.24720472e-01
6.70340657e-01 8.40922177e-01 -7.94904053e-01 1.16591525e+00
-6.45454109e-01 3.45764577e-01 3.65725696e-01 -1.14098065e-01
4.31724310e-01 3.11946094e-01 2.97899872e-01 9.56310928e-01
3.79991561e-01 8.81611630e-02 -2.24946231e-01 1.39304471e+00
5.56210168e-02 -3.13871056e-01 -5.15072525e-01 -3.23672802e-03
-1.34266084e-02 1.15693998e+00 -7.23528028e-01 -4.26845074e-01
-2.45106250e-01 9.55995560e-01 4.67765987e-01 5.13260186e-01
-8.72009754e-01 -6.22908957e-02 7.05079854e-01 5.81858680e-02
3.26878488e-01 -2.33340815e-01 -6.52882874e-01 -9.41085815e-01
-1.23489546e-02 -6.34692311e-01 1.06932677e-01 -1.56013048e+00
-1.05459893e+00 8.40072036e-01 6.92720637e-02 -1.26760519e+00
-2.82760412e-01 -1.08619881e+00 -6.14314139e-01 8.55354130e-01
-1.63780415e+00 -1.25058579e+00 -7.37424552e-01 3.99382025e-01
7.24694550e-01 -1.88028380e-01 9.97737467e-01 -1.37872249e-01
-2.49503016e-01 2.79349804e-01 -5.26805460e-01 -8.03063586e-02
2.78400540e-01 -1.28481340e+00 5.75815439e-01 5.89271128e-01
5.14389694e-01 4.35645252e-01 9.96913791e-01 -3.62429112e-01
-7.97745109e-01 -1.03748143e+00 1.00303316e+00 -5.16437709e-01
1.81549385e-01 -1.54156104e-01 -1.01221299e+00 8.04782748e-01
3.29348356e-01 1.03494890e-01 4.03478801e-01 -7.42351934e-02
-3.90476018e-01 8.08784589e-02 -1.02201962e+00 7.22995281e-01
9.38448787e-01 -1.82893246e-01 -6.44015789e-01 4.64872390e-01
7.49999881e-01 -5.30392706e-01 -2.93271899e-01 4.69041586e-01
4.50126857e-01 -1.25478649e+00 1.15005720e+00 -1.08628094e+00
8.74103665e-01 -3.94396394e-01 1.16201090e-02 -1.38250375e+00
-4.54867601e-01 -3.90524454e-02 -1.54277265e-01 7.45119989e-01
5.00397682e-01 -2.85471946e-01 9.02052760e-01 7.28405833e-01
-3.95899951e-01 -9.12334263e-01 -7.88183987e-01 -6.09674692e-01
-2.66530037e-01 -4.76833194e-01 5.16754687e-01 5.69993258e-01
-3.14179361e-01 2.37717047e-01 -5.43516815e-01 2.89604366e-01
4.67559785e-01 -3.05298064e-02 8.73683631e-01 -1.23572123e+00
-1.87057137e-01 -4.53639865e-01 -9.44194376e-01 -9.89302456e-01
1.81217447e-01 -7.31587589e-01 -3.38768996e-02 -1.99833846e+00
4.17168230e-01 6.71629161e-02 -2.72668719e-01 9.38602984e-01
-9.20034647e-02 4.57586229e-01 4.96288836e-01 1.15731046e-01
-3.56532365e-01 5.70667565e-01 1.56662357e+00 -3.01457286e-01
1.00400090e-01 -3.75492722e-02 -9.33261037e-01 8.87815773e-01
7.28608847e-01 -3.55116308e-01 -3.17661077e-01 -7.40624785e-01
1.56046808e-01 -5.29482849e-02 1.16461360e+00 -1.03887153e+00
-8.87911916e-02 -2.29087457e-01 7.58540452e-01 -4.85510975e-01
5.90700030e-01 -8.93982470e-01 9.46604684e-02 5.92473090e-01
-7.35717177e-01 8.53304043e-02 4.22407866e-01 7.19208300e-01
-4.25945222e-01 -1.55108720e-01 6.18945897e-01 -1.91812277e-01
-1.13176155e+00 1.01934083e-01 -2.45736092e-02 -4.13154781e-01
1.04808700e+00 -4.03512716e-01 -6.25847161e-01 -6.77668750e-01
-9.19655085e-01 2.53324509e-02 9.99868587e-02 8.60099494e-01
1.00168240e+00 -1.45819116e+00 -5.56769669e-01 1.78332478e-01
3.47462296e-01 -1.33575410e-01 3.10984224e-01 4.69789088e-01
-4.49679673e-01 8.90516162e-01 -5.65117717e-01 -7.26919651e-01
-1.33534479e+00 4.30004656e-01 4.45811152e-01 -5.49454764e-02
-6.28511727e-01 9.20703113e-01 5.31584442e-01 -1.83501363e-01
2.99325824e-01 -6.57971799e-01 -3.57413530e-01 -4.90134925e-01
5.30969262e-01 6.12150133e-02 -1.12922482e-01 -5.72172344e-01
-3.63147229e-01 5.28446615e-01 1.14112660e-01 1.05238535e-01
1.29480612e+00 9.62037668e-02 2.36681715e-01 2.51736343e-01
8.98154378e-01 -8.43676686e-01 -1.49614322e+00 -1.08518042e-01
-1.25703022e-01 -6.08684123e-01 -1.07823260e-01 -1.19988549e+00
-1.17746305e+00 1.35901749e+00 6.19217992e-01 3.63140941e-01
9.77546334e-01 2.86217421e-01 4.48987573e-01 3.21332544e-01
5.73813133e-02 -5.57728708e-01 6.03662550e-01 3.86176825e-01
1.31427729e+00 -1.47396970e+00 -1.35783017e-01 -3.42638165e-01
-8.86541367e-01 1.24436319e+00 9.79007006e-01 -1.10336810e-01
2.09281176e-01 -4.06067431e-01 1.09830789e-01 -3.27410370e-01
-7.91619480e-01 -1.83060452e-01 7.06948876e-01 7.74097979e-01
4.47840303e-01 4.53086570e-02 2.36676037e-02 5.56055069e-01
5.29134180e-04 1.28584564e-01 6.36096716e-01 6.73265159e-01
-5.12552977e-01 -8.18408489e-01 -2.95002788e-01 1.69367239e-01
-4.63903882e-02 1.71238501e-02 -7.15651214e-01 9.47528958e-01
3.23109448e-01 8.30645025e-01 8.73112530e-02 -3.90496373e-01
4.32694048e-01 6.97418526e-02 4.12758917e-01 -6.97211385e-01
-2.92471349e-01 -4.14062440e-01 -1.42449975e-01 -5.65207899e-01
-6.95499241e-01 -2.88273841e-01 -1.36801767e+00 1.84292793e-02
-1.38683408e-01 -2.48979405e-01 9.04887378e-01 1.32465649e+00
3.60777825e-01 8.39133978e-01 1.31062567e-01 -1.16101587e+00
-3.77759367e-01 -8.60403061e-01 -2.53789306e-01 6.74551845e-01
3.61931831e-01 -8.70079637e-01 -3.98422480e-01 2.77275383e-01] | [8.991789817810059, 5.420103549957275] |
4bbbe4f0-ae04-4109-8caa-0b9ced5a2850 | an-accelerated-pipeline-for-multi-label-renal | 2305.14566 | null | https://arxiv.org/abs/2305.14566v1 | https://arxiv.org/pdf/2305.14566v1.pdf | An Accelerated Pipeline for Multi-label Renal Pathology Image Segmentation at the Whole Slide Image Level | Deep-learning techniques have been used widely to alleviate the labour-intensive and time-consuming manual annotation required for pixel-level tissue characterization. Our previous study introduced an efficient single dynamic network - Omni-Seg - that achieved multi-class multi-scale pathological segmentation with less computational complexity. However, the patch-wise segmentation paradigm still applies to Omni-Seg, and the pipeline is time-consuming when providing segmentation for Whole Slide Images (WSIs). In this paper, we propose an enhanced version of the Omni-Seg pipeline in order to reduce the repetitive computing processes and utilize a GPU to accelerate the model's prediction for both better model performance and faster speed. Our proposed method's innovative contribution is two-fold: (1) a Docker is released for an end-to-end slide-wise multi-tissue segmentation for WSIs; and (2) the pipeline is deployed on a GPU to accelerate the prediction, achieving better segmentation quality in less time. The proposed accelerated implementation reduced the average processing time (at the testing stage) on a standard needle biopsy WSI from 2.3 hours to 22 minutes, using 35 WSIs from the Kidney Tissue Atlas (KPMP) Datasets. The source code and the Docker have been made publicly available at https://github.com/ddrrnn123/Omni-Seg. | ['Yuankai Huo', 'Lipeng Wan', 'Haichun Yang', 'R. Michael Womick', 'Zuhayr Asad', 'Ruining Deng', 'Haoju Leng'] | 2023-05-23 | null | null | null | null | ['whole-slide-images'] | ['computer-vision'] | [ 3.23938429e-01 7.58056901e-03 5.14861057e-03 -3.57504964e-01
-1.07610667e+00 -4.36969161e-01 -1.92965213e-02 5.33620954e-01
-5.38407624e-01 4.68917251e-01 -4.35717046e-01 -5.31412959e-01
-1.60438776e-01 -8.59117627e-01 -4.52640444e-01 -8.93070400e-01
1.50000649e-02 6.22526765e-01 5.96515596e-01 3.05886596e-01
1.91776469e-01 7.21821308e-01 -1.04757404e+00 3.21704119e-01
9.74642277e-01 8.99491370e-01 6.52492762e-01 8.27941000e-01
-8.82915854e-02 2.25243896e-01 -1.33108437e-01 -1.86576813e-01
2.44560972e-01 -2.17839494e-01 -7.24830151e-01 -1.25073954e-01
2.45778009e-01 -1.45049915e-01 8.78969133e-02 9.57003534e-01
7.33338475e-01 -2.42755935e-01 5.15104651e-01 -9.10086751e-01
2.20697239e-01 2.78047889e-01 -9.21762049e-01 2.55191982e-01
-3.57836485e-01 3.27683777e-01 5.16167521e-01 -3.90629947e-01
7.22610354e-01 6.28088593e-01 8.91236007e-01 4.86295313e-01
-1.01023257e+00 -7.41194546e-01 -5.28046310e-01 2.00135604e-01
-1.48745656e+00 3.41101475e-02 3.14811558e-01 -4.72286016e-01
7.63269722e-01 5.07545471e-01 9.65737104e-01 5.60165286e-01
5.98074913e-01 7.55439103e-01 1.19130027e+00 -1.51549667e-01
2.75006324e-01 -2.43226677e-01 2.79911280e-01 7.22981632e-01
2.61466205e-01 -4.43803489e-01 1.40687078e-01 -2.91248679e-01
8.73740256e-01 2.04536736e-01 -1.29320264e-01 -5.37190065e-02
-1.13266051e+00 5.76924741e-01 4.65860933e-01 5.45772791e-01
-4.76126254e-01 -6.58742571e-03 7.76223183e-01 -9.81560722e-02
5.40013731e-01 -1.18901907e-02 -4.20697510e-01 -2.22457424e-01
-1.13712239e+00 1.87614828e-01 5.40607870e-01 5.11563361e-01
5.60752511e-01 -5.34337103e-01 -9.20927078e-02 7.94749141e-01
2.28328794e-01 1.07565440e-01 7.91677535e-01 -5.61780632e-01
-2.55984664e-02 8.23873281e-01 -1.73239604e-01 -7.74186313e-01
-9.69692230e-01 -7.14671493e-01 -1.18060160e+00 1.93941712e-01
6.03768229e-01 -1.27035482e-02 -1.09885025e+00 1.28540730e+00
6.63041651e-01 2.87259072e-01 -3.17728966e-01 1.04678893e+00
7.36227572e-01 3.67064804e-01 3.00461054e-01 -1.07723936e-01
1.68193209e+00 -1.07619667e+00 -3.67197782e-01 1.48291066e-01
1.22181046e+00 -7.39125192e-01 1.04210520e+00 2.73768425e-01
-9.55525339e-01 -2.11676255e-01 -8.65377545e-01 3.48750874e-02
-2.40599558e-01 3.38848293e-01 7.00240195e-01 5.21809876e-01
-1.16562927e+00 5.30244648e-01 -1.37429953e+00 -5.57076693e-01
9.94878054e-01 5.33360839e-01 -4.36397344e-01 1.01419911e-02
-6.88171685e-01 6.04337454e-01 2.56675094e-01 2.65643746e-01
-7.80413389e-01 -1.14194131e+00 -3.96855086e-01 -1.15249455e-01
2.03129977e-01 -8.96991432e-01 1.05525362e+00 -6.90024912e-01
-1.43628311e+00 1.12172735e+00 -2.29223073e-01 -2.90065408e-01
9.01979327e-01 2.94984162e-01 1.76020563e-01 2.76601076e-01
9.65638757e-02 4.44165766e-01 -2.11786502e-03 -8.27138782e-01
-5.32637656e-01 -7.41181970e-01 -3.77110809e-01 1.25371635e-01
-1.42290160e-01 -4.60758340e-03 -7.44128823e-01 -3.78268272e-01
1.77230865e-01 -1.10624874e+00 -5.94229579e-01 2.89114207e-01
-4.98673350e-01 -8.49108696e-02 7.08953679e-01 -7.38947868e-01
9.64353323e-01 -1.98475921e+00 -2.04229906e-01 2.97861278e-01
2.40336299e-01 5.72440505e-01 7.60231987e-02 1.52168214e-01
-1.25615582e-01 9.40390229e-02 -3.39463770e-01 -4.33679342e-01
-5.53339422e-01 -2.71234568e-02 5.94836235e-01 7.42886364e-01
-3.21054190e-01 8.70330155e-01 -5.59682131e-01 -8.83432209e-01
2.97543168e-01 4.55098599e-01 -4.24426228e-01 5.67680933e-02
6.86719865e-02 5.90339959e-01 -3.33957255e-01 7.98851132e-01
1.05302083e+00 -3.55277121e-01 1.55240029e-01 -3.43143731e-01
-2.17552409e-01 -2.97104478e-01 -1.07749081e+00 2.01563787e+00
-3.95818859e-01 7.63724372e-02 4.76089358e-01 -8.94229710e-01
6.75064147e-01 2.72621214e-01 8.51405025e-01 -4.26926851e-01
2.80202568e-01 4.21998888e-01 7.69000649e-02 -6.54734313e-01
-9.52878743e-02 -1.81868449e-01 2.24951699e-01 1.97686136e-01
-2.93757886e-01 -1.26746641e-02 3.86823893e-01 -6.03652745e-03
1.19781375e+00 2.08648741e-01 2.72223711e-01 -5.86888552e-01
6.15378618e-01 4.56551880e-01 7.13396907e-01 4.31176007e-01
-4.22100693e-01 5.49524367e-01 5.00218749e-01 -5.02391219e-01
-9.38431919e-01 -7.29401469e-01 -4.12993997e-01 5.88081956e-01
4.79631647e-02 -1.36362076e-01 -8.37276936e-01 -4.26232457e-01
-6.02682680e-02 3.37762803e-01 -5.79015553e-01 4.83354330e-01
-4.86609519e-01 -1.18053436e+00 7.63481915e-01 2.88307041e-01
3.10688317e-01 -6.61246657e-01 -7.93528795e-01 3.54659557e-01
-6.85570464e-02 -7.25469828e-01 -2.76379138e-01 1.52599022e-01
-1.11763859e+00 -1.12250292e+00 -9.93628263e-01 -8.16840470e-01
9.67499435e-01 9.83170941e-02 4.47940350e-01 2.56225288e-01
-9.97463703e-01 -3.19138825e-01 -1.23726845e-01 -4.73820150e-01
-2.58672327e-01 2.08938837e-01 -5.26262224e-01 -1.99810416e-01
1.89717904e-01 -6.04517043e-01 -1.05248761e+00 2.34450668e-01
-9.34896111e-01 6.43420100e-01 7.35735953e-01 8.42137754e-01
1.15894961e+00 -1.41737938e-01 4.59208846e-01 -1.22160542e+00
2.62019396e-01 -4.00446236e-01 -6.25878274e-01 2.05685794e-01
-4.60270077e-01 -6.17959321e-01 6.05883598e-01 -4.27421741e-02
-1.00100362e+00 2.74390012e-01 -5.49677968e-01 -1.05108760e-01
-2.44570315e-01 6.34919107e-01 2.67256230e-01 -3.26675534e-01
4.15450722e-01 2.44862407e-01 4.01824623e-01 -3.72991472e-01
-2.26781815e-01 7.12993443e-01 2.72784829e-01 -3.10152203e-01
1.85358837e-01 6.10503912e-01 2.96508282e-01 -6.10607564e-01
-5.09417951e-01 -8.11967134e-01 -5.68285763e-01 -1.88440204e-01
8.57961655e-01 -7.31020808e-01 -8.75031590e-01 8.46020460e-01
-8.48687530e-01 -6.30019248e-01 2.04193100e-01 4.73105878e-01
-3.14178497e-01 5.29985666e-01 -1.05839825e+00 -4.60256994e-01
-1.10955024e+00 -1.41843688e+00 1.02712715e+00 4.45943981e-01
-8.94799829e-02 -8.60375464e-01 1.51505917e-01 6.11068308e-01
6.63588881e-01 6.43152833e-01 8.12523901e-01 -5.73780179e-01
-4.66479272e-01 -4.07495290e-01 -5.27556419e-01 5.07270582e-02
-3.57569233e-02 2.34909087e-01 -6.78566515e-01 -3.46726835e-01
-1.34450138e-01 -6.26814514e-02 5.94636381e-01 7.04093337e-01
1.59707975e+00 2.23860625e-04 -6.54970348e-01 8.86468530e-01
1.96963930e+00 2.52163708e-01 5.68544507e-01 5.77134967e-01
6.10793650e-01 4.02830899e-01 8.18957806e-01 3.62885296e-01
3.31056595e-01 5.05497158e-01 5.50744951e-01 -7.53653407e-01
-1.41851291e-01 1.78829670e-01 -4.71259445e-01 7.00266063e-01
-2.26738393e-01 -1.24507591e-01 -1.30613589e+00 8.61936212e-01
-1.74065208e+00 -4.93643701e-01 -5.99809945e-01 1.88768494e+00
8.57885897e-01 -1.32381991e-01 2.97229104e-02 -1.01805873e-01
5.51621258e-01 -2.67397612e-01 -4.53919023e-01 -2.05296502e-01
2.76167244e-01 3.86227965e-01 6.02154613e-01 3.98629904e-01
-9.87792313e-01 6.74752951e-01 4.60729742e+00 1.19931662e+00
-1.41960585e+00 4.70911980e-01 9.50993657e-01 -2.70107418e-01
2.04256892e-01 -1.43819660e-01 -6.26456380e-01 5.35720527e-01
8.30410004e-01 -1.62931263e-01 -2.04458639e-01 8.79798293e-01
5.78163326e-01 -5.35585761e-01 -6.45578504e-01 7.69845247e-01
-4.60552812e-01 -1.60713995e+00 -2.82445610e-01 2.82989651e-01
4.13524687e-01 4.47870612e-01 -4.42800760e-01 -6.84050098e-02
-2.06830785e-01 -6.70947373e-01 1.33560687e-01 5.32767177e-01
8.78151834e-01 -6.82264328e-01 1.24057436e+00 4.32122052e-01
-1.02139664e+00 3.81953150e-01 -3.12054127e-01 2.50612080e-01
5.10119833e-02 1.04285324e+00 -1.27029228e+00 8.06728363e-01
6.30639672e-01 3.48653883e-01 -5.79250813e-01 1.36868048e+00
3.56157631e-01 6.19163990e-01 -4.24344838e-01 -6.24057129e-02
2.70553261e-01 -3.00030649e-01 2.85703629e-01 1.62748575e+00
3.61830980e-01 -4.38490845e-02 1.82889402e-01 5.42979121e-01
3.04718435e-01 4.22928929e-01 1.47970393e-01 2.79265821e-01
2.04042405e-01 1.89122367e+00 -1.31548178e+00 -3.18449527e-01
-2.20688775e-01 7.62884378e-01 8.90450329e-02 -4.43700217e-02
-8.81904006e-01 -4.25315440e-01 3.60601246e-01 3.76841873e-01
-3.11941467e-02 1.61763579e-02 -6.55431986e-01 -7.25095510e-01
-1.70048028e-01 -5.29847026e-01 5.19903362e-01 -3.90457422e-01
-9.62219000e-01 5.41051030e-01 -2.98969656e-01 -9.68215466e-01
3.11188310e-01 -4.49533969e-01 -6.81111157e-01 9.88256395e-01
-1.55659437e+00 -1.37421453e+00 -7.26188660e-01 3.43129218e-01
3.54769737e-01 3.61675441e-01 1.00591433e+00 6.50383234e-01
-8.13477457e-01 6.73050463e-01 2.32988819e-01 -1.17031047e-02
5.11869192e-01 -1.11037207e+00 -4.00536805e-02 6.31754577e-01
-5.78098238e-01 5.77232718e-01 5.22339404e-01 -6.20945632e-01
-1.31431341e+00 -1.18318677e+00 6.40148997e-01 1.63519725e-01
6.92471147e-01 -1.23999722e-01 -8.44865143e-01 4.89103705e-01
-9.20949280e-02 2.45567843e-01 1.11738968e+00 -2.87270337e-01
4.26522642e-01 -2.61882544e-01 -1.60326290e+00 5.09433627e-01
5.77681363e-01 3.50002758e-02 2.05186382e-01 5.48593760e-01
1.76565930e-01 -8.48142982e-01 -1.35566568e+00 4.02124524e-01
4.94334430e-01 -7.88421094e-01 5.54036677e-01 2.10786864e-01
2.80160338e-01 -5.91315806e-01 3.76754850e-01 -9.08148944e-01
-3.47603440e-01 -2.56358832e-01 3.00880313e-01 9.24581409e-01
3.36350858e-01 -7.54731238e-01 1.08806694e+00 5.68243325e-01
-5.35108209e-01 -1.61957312e+00 -1.25518513e+00 -2.76832998e-01
-7.10532516e-02 -3.04238915e-01 2.70746887e-01 7.48915792e-01
-9.97078344e-02 -2.98590422e-01 2.01236397e-01 -4.36425917e-02
6.47650361e-01 4.74181138e-02 6.92333817e-01 -1.02955854e+00
-3.06770623e-01 -4.47957158e-01 -6.37125552e-01 -5.17655969e-01
-4.24945414e-01 -1.15055764e+00 -2.30386972e-01 -1.74781787e+00
6.63156092e-01 -6.87978089e-01 -2.33379051e-01 7.90319026e-01
-1.26088709e-01 4.79852796e-01 -3.83008597e-03 4.03679401e-01
-4.13108975e-01 -5.36677614e-02 1.59144938e+00 5.90782650e-02
2.96094902e-02 1.92293569e-01 -6.18453801e-01 6.84189558e-01
8.82161677e-01 -5.68586648e-01 -1.91365302e-01 -2.25478694e-01
-1.19320922e-01 2.16769770e-01 3.65334630e-01 -1.12218022e+00
4.50474709e-01 3.15872803e-02 2.95361072e-01 -8.33875775e-01
1.06765129e-01 -7.18515635e-01 5.23226321e-01 9.49894905e-01
-3.39242443e-02 -3.49004388e-01 4.63572860e-01 3.30068827e-01
-1.48478523e-01 -3.63145828e-01 1.09117532e+00 -2.37524718e-01
-2.03039899e-01 5.60613036e-01 -3.48860681e-01 -5.26209295e-01
1.51681113e+00 -4.08206344e-01 -3.18328470e-01 4.04112130e-01
-7.97114491e-01 2.97918141e-01 5.04800856e-01 -2.71149397e-01
2.06635281e-01 -7.62283385e-01 -7.80276000e-01 -8.11784714e-02
-8.82663354e-02 4.88298506e-01 1.03174937e+00 1.53679013e+00
-1.38870943e+00 4.08618271e-01 -2.91516215e-01 -9.76861894e-01
-1.53121150e+00 8.89262334e-02 4.74594444e-01 -8.89974594e-01
-8.98269534e-01 1.12654757e+00 1.68190286e-01 -4.19310004e-01
-7.15242550e-02 -2.97658563e-01 9.72964838e-02 -1.74500346e-01
3.83224159e-01 3.50246340e-01 4.10091400e-01 -3.47675055e-01
-5.51556885e-01 6.15425110e-01 -4.48712349e-01 3.50130945e-01
1.49405050e+00 8.19910839e-02 -4.32266086e-01 1.37611926e-01
1.30825508e+00 -2.32660800e-01 -1.05366147e+00 3.71565074e-02
-1.63928300e-01 -3.07637066e-01 2.82106400e-01 -1.02655363e+00
-1.19074285e+00 6.36557102e-01 9.98485386e-01 -2.34576181e-01
1.30428219e+00 -2.12672368e-01 1.15558517e+00 -2.28430122e-01
4.77699429e-01 -8.51592064e-01 -6.56351686e-01 7.15032071e-02
5.11879027e-01 -1.09326196e+00 1.76745832e-01 -6.66489542e-01
-4.07742739e-01 1.30014002e+00 6.43607855e-01 8.98345653e-03
6.35859311e-01 5.47244728e-01 1.76986530e-01 -3.18596750e-01
-5.25611877e-01 3.60244274e-01 -8.98300260e-02 2.61998981e-01
7.49240518e-01 2.90248126e-01 -8.44174445e-01 6.52057528e-01
1.10705666e-01 4.22058105e-01 3.72045457e-01 9.23897088e-01
-1.92016661e-01 -9.45371389e-01 -5.84310815e-02 6.38103545e-01
-8.03454816e-01 1.37135237e-01 1.03692979e-01 6.95073366e-01
1.79149702e-01 4.68603164e-01 -1.49727359e-01 -1.15388446e-02
1.46955192e-01 -2.20904365e-01 2.61379600e-01 -4.66035336e-01
-9.21024859e-01 3.03961456e-01 -1.87094092e-01 -6.79181993e-01
-2.67348796e-01 -5.73639154e-01 -1.60082138e+00 -2.54262984e-01
-2.76882499e-01 -2.05135420e-02 1.12456274e+00 6.82829082e-01
4.99201715e-01 7.57980347e-01 3.44400853e-01 -8.78109217e-01
-1.15711920e-01 -1.07536888e+00 -5.87697268e-01 1.02297693e-01
-2.55726159e-01 -2.93377906e-01 -2.17511222e-01 -1.27796531e-01] | [14.952760696411133, -2.9504849910736084] |
9ed249a6-54cb-4832-9e59-760abb98727d | iterative-loop-learning-combining-self | 2301.13361 | null | https://arxiv.org/abs/2301.13361v4 | https://arxiv.org/pdf/2301.13361v4.pdf | Iterative Loop Method Combining Active and Semi-Supervised Learning for Domain Adaptive Semantic Segmentation | Semantic segmentation is an important technique for environment perception in intelligent transportation systems. With the rapid development of convolutional neural networks (CNNs), road scene analysis can usually achieve satisfactory results in the source domain. However, guaranteeing good generalization to different target domain scenarios remains a significant challenge. Recently, semi-supervised learning and active learning have been proposed to alleviate this problem. Semisupervised learning can improve model accuracy with massive unlabeled data, but some pseudo labels containing noise would be generated with limited or imbalanced training data. And there will be suboptimal models if human guidance is absent. Active learning can select more effective data to intervene, while the model accuracy can not be improved because the massive unlabeled data are not used. And the probability of querying sub-optimal samples will increase when the domain difference is too large, increasing annotation cost. This paper proposes an iterative loop method combining active and semisupervised learning for domain adaptive semantic segmentation. The method first uses semi-supervised to learn massive unlabeled data to improve model accuracy and provide more accurate selection models for active learning. Secondly, combined with the predictive uncertainty sample selection strategy of active learning, manual intervention is used to correct the pseudo-labels. Finally, flexible iterative loops achieve the best performance with minimal labeling cost. Extensive experiments show that our method establishes state-of-the-art performance on tasks of GTAV to Cityscapes, SYNTHIA to Cityscapes, improving by 4.9% mIoU and 5.2% mIoU, compared to the previous best method, respectively. | ['Xue Yuan', 'Licong Guan'] | 2023-01-31 | null | null | null | null | ['semi-supervised-semantic-segmentation'] | ['computer-vision'] | [ 1.60120875e-01 2.40195215e-01 -7.04931200e-01 -6.11431301e-01
-1.08940578e+00 -3.22798431e-01 3.02790344e-01 1.18440278e-01
-7.22918749e-01 9.52860475e-01 -3.12143624e-01 -2.01768890e-01
-1.32606313e-01 -9.98716891e-01 -6.10046685e-01 -8.09998631e-01
2.38619909e-01 1.00320315e+00 8.92453492e-01 6.14176430e-02
1.33649781e-01 2.95213670e-01 -1.86626053e+00 8.66587460e-02
1.75958335e+00 1.13119197e+00 4.17689145e-01 1.84789732e-01
-6.74552977e-01 7.08617151e-01 -4.91414815e-01 2.72510573e-02
3.56248468e-01 1.33368239e-01 -8.16279948e-01 3.98022771e-01
1.22024663e-01 -1.90948471e-01 2.61061281e-01 1.12185562e+00
4.50203598e-01 4.31871235e-01 7.86406994e-01 -1.34495163e+00
-3.69674824e-02 5.01869619e-01 -6.23899937e-01 -1.50483355e-01
-3.13868821e-01 2.24710211e-01 6.55982435e-01 -9.50219810e-01
2.95273125e-01 1.17639387e+00 3.73117477e-01 4.49216485e-01
-1.01803732e+00 -9.14973557e-01 4.52299982e-01 6.14114523e-01
-1.54965127e+00 -3.89031291e-01 6.61136746e-01 -2.73046404e-01
5.32608747e-01 2.15958267e-01 6.52761459e-01 6.89433575e-01
-5.11863589e-01 1.13338208e+00 8.09731364e-01 -2.80329108e-01
5.55844903e-01 3.81282568e-01 2.26907283e-01 5.43837190e-01
2.36671001e-01 3.96118052e-02 -2.42480561e-01 1.27510846e-01
3.34829241e-01 -7.51793385e-02 8.94065425e-02 -4.12853032e-01
-7.09488988e-01 7.69282281e-01 4.31970179e-01 -3.97746526e-02
-2.23048940e-01 -3.43512177e-01 2.77906209e-01 -4.55904119e-02
7.15314686e-01 2.04768777e-01 -5.89558482e-01 -3.87950614e-02
-9.08086181e-01 1.98357418e-01 3.70021224e-01 1.08082604e+00
1.17149842e+00 -6.71491819e-03 4.08806726e-02 1.10647690e+00
4.37750369e-01 6.02054477e-01 1.60090417e-01 -1.06009412e+00
6.88503385e-01 1.04358423e+00 3.88512939e-01 -6.57821059e-01
-2.85870254e-01 -3.73655885e-01 -6.29750371e-01 4.23293263e-01
5.88256836e-01 -2.12830976e-01 -1.40399444e+00 1.07283020e+00
4.75777358e-01 1.12048574e-01 -7.58261159e-02 8.45879197e-01
7.92318821e-01 6.54231429e-01 3.08239132e-01 -3.69287193e-01
7.41144061e-01 -9.95100021e-01 -7.36609995e-01 -5.64133704e-01
7.90035129e-01 -5.45075834e-01 1.13568747e+00 3.57242227e-01
-6.85801566e-01 -7.23618925e-01 -8.87157381e-01 4.32157874e-01
-5.56823313e-01 1.38943493e-01 5.10512829e-01 5.92012286e-01
-5.12047589e-01 1.05201945e-01 -7.67842591e-01 -1.81219459e-01
9.41200733e-01 4.37620282e-01 7.41253495e-02 -1.67096421e-01
-1.35515451e+00 6.35789275e-01 9.01643455e-01 1.23737089e-01
-7.68222809e-01 -6.65229261e-01 -7.26839244e-01 -1.47257745e-01
9.28651690e-01 -1.11488737e-02 1.22513926e+00 -1.14449024e+00
-1.38154197e+00 4.40614671e-01 -1.64458096e-01 -5.03217697e-01
7.49449134e-01 -1.13727048e-01 -6.26811385e-01 1.90518964e-02
2.21056432e-01 9.87775147e-01 5.32276809e-01 -1.35123932e+00
-1.16894853e+00 -3.04021358e-01 -2.46174578e-02 5.21856308e-01
-2.16704473e-01 -3.43185216e-01 -7.06079245e-01 -7.51323104e-02
4.95428443e-01 -8.61446798e-01 -7.11569846e-01 -6.27775639e-02
-1.12644680e-01 -5.51874518e-01 1.12057638e+00 -2.19412982e-01
1.02342820e+00 -1.89300382e+00 -4.22121972e-01 3.86805654e-01
2.74195056e-02 5.88104963e-01 1.85581237e-01 -1.92034468e-01
3.16573381e-01 1.51707157e-01 -5.46134055e-01 -2.15413654e-03
-2.61953145e-01 4.03214604e-01 -2.92502046e-02 1.00330852e-01
3.10133785e-01 7.06637204e-01 -9.32596207e-01 -9.11090016e-01
4.44313139e-01 2.16544922e-02 -2.39161924e-01 1.76525384e-01
-4.96622443e-01 6.14041686e-01 -5.54364264e-01 6.15656793e-01
8.75810504e-01 -2.67621934e-01 -6.78064153e-02 6.99273348e-02
-1.76807180e-01 1.09793477e-01 -1.59530663e+00 1.48395109e+00
-3.61196995e-01 3.80515248e-01 -1.55390948e-01 -1.31021905e+00
1.09188104e+00 1.64057046e-01 5.62853158e-01 -9.96506512e-01
-1.26053467e-02 4.57985222e-01 -1.50961488e-01 -7.90225327e-01
3.10378224e-01 3.11694264e-01 5.21425195e-02 -9.88172069e-02
-3.33047330e-01 -1.73119530e-01 3.06300312e-01 6.33296221e-02
4.42638904e-01 2.63431430e-01 -3.44794355e-02 -9.76167843e-02
5.44203520e-01 6.39169395e-01 9.28302526e-01 7.19586134e-01
-4.96066839e-01 4.57721621e-01 9.82145816e-02 -3.31282735e-01
-6.86715245e-01 -7.85178304e-01 -2.13259444e-01 9.27546620e-01
6.10056698e-01 9.69232246e-02 -7.14951634e-01 -1.09876120e+00
-3.25379550e-01 8.71152401e-01 -1.69022545e-01 -8.99475813e-02
-4.67036456e-01 -1.08853924e+00 2.29746059e-01 5.32181501e-01
1.11996663e+00 -8.08802247e-01 -3.82718414e-01 1.59047559e-01
-3.48029256e-01 -9.57278550e-01 2.42146805e-01 1.40188962e-01
-8.14064145e-01 -1.07082510e+00 -5.07339358e-01 -6.00446403e-01
7.40777671e-01 3.40318799e-01 8.87292266e-01 -6.50899634e-02
9.26803127e-02 -2.95595944e-01 -4.32943463e-01 -7.54459798e-01
-2.07809001e-01 2.09752113e-01 -1.99589029e-01 3.61871682e-02
5.23648322e-01 3.54434922e-02 -4.88627642e-01 7.59749591e-01
-7.68254042e-01 1.43397555e-01 4.57407355e-01 7.11222351e-01
8.35636854e-01 6.58156276e-01 8.68483901e-01 -1.23977327e+00
-2.79142074e-02 -4.43390489e-01 -8.86813462e-01 2.45901436e-01
-9.66996849e-01 -2.06557959e-01 5.80406547e-01 -4.38279003e-01
-1.59592485e+00 4.35884297e-01 -1.90467194e-01 -1.41499013e-01
-5.18577516e-01 3.59038293e-01 -6.05240345e-01 3.81394401e-02
9.57462668e-01 -1.35735795e-01 -2.05322340e-01 -2.95647979e-01
1.94614723e-01 1.07393026e+00 -4.02708612e-02 -3.27855855e-01
6.50735021e-01 4.73973453e-01 -2.53745228e-01 -6.96175218e-01
-1.04853082e+00 -7.28210628e-01 -7.34253705e-01 -5.18151104e-01
6.96498275e-01 -9.61014211e-01 -1.80103347e-01 6.01073027e-01
-6.66746080e-01 -5.12984335e-01 -1.77243844e-01 6.09015703e-01
-3.40557724e-01 2.72314280e-01 2.01203927e-01 -1.05075014e+00
-2.80182604e-02 -1.33247209e+00 8.14801574e-01 5.27279019e-01
-2.92634498e-03 -8.24939609e-01 -4.18070346e-01 7.03863859e-01
3.10460269e-01 -2.45899796e-01 6.69375062e-01 -7.58299172e-01
-8.72111142e-01 -2.35397741e-01 -1.85759649e-01 3.32258284e-01
6.31474033e-02 -1.29078254e-02 -1.05421245e+00 1.58445507e-01
-3.63215446e-01 -4.57307786e-01 9.57292259e-01 4.42010552e-01
1.41285396e+00 -9.97681543e-02 -5.12639761e-01 2.31326267e-01
1.26200652e+00 6.51729226e-01 7.22224951e-01 3.57586056e-01
7.22692788e-01 8.74296308e-01 1.44188833e+00 8.31480771e-02
3.97704035e-01 5.56097269e-01 5.38880289e-01 -3.24485570e-01
2.25594547e-02 2.87285913e-02 -7.17036277e-02 3.90733302e-01
9.03547928e-02 -2.65103906e-01 -1.32958245e+00 6.16557598e-01
-2.13582063e+00 -7.88440824e-01 -4.81283128e-01 2.37126756e+00
7.33567595e-01 6.89821541e-01 2.45692119e-01 4.77613330e-01
7.23525763e-01 -1.73204020e-01 -8.57049644e-01 2.00612426e-01
1.54378647e-02 -1.74986660e-01 8.83080840e-01 5.29703438e-01
-1.27661622e+00 1.16490138e+00 5.27614737e+00 1.38976479e+00
-1.20310414e+00 7.30967075e-02 9.11117077e-01 1.66525900e-01
-1.26987830e-01 -1.93292405e-02 -1.01426685e+00 5.91253996e-01
6.89332426e-01 3.78023207e-01 -5.62413260e-02 1.02838051e+00
5.00114501e-01 -5.76513290e-01 -4.37197000e-01 9.37948525e-01
-3.55008423e-01 -1.25961912e+00 -2.14635178e-01 -1.24424636e-01
9.60373521e-01 3.26445363e-02 -1.93767294e-01 3.56023759e-01
4.35521811e-01 -7.86079705e-01 6.17489576e-01 3.75154912e-01
6.15302622e-01 -1.02768362e+00 8.43342364e-01 9.34960186e-01
-1.07319140e+00 -2.70680875e-01 -4.87215906e-01 1.67651206e-01
1.46032855e-01 7.54165351e-01 -1.06831682e+00 6.08607531e-01
7.77376831e-01 5.73631406e-01 -5.54894507e-01 1.44654739e+00
-1.62743390e-01 9.76009488e-01 -4.85190213e-01 -1.03462420e-01
1.69355541e-01 -2.77107298e-01 5.25615454e-01 7.59012222e-01
-1.63243897e-02 5.64007647e-02 5.68084836e-01 5.71834207e-01
3.77815276e-01 1.99205801e-01 -3.34584892e-01 2.35058382e-01
6.37789130e-01 1.05614734e+00 -1.03992426e+00 -5.84382713e-01
-2.65425354e-01 5.96067965e-01 1.84627682e-01 5.60889006e-01
-8.86049330e-01 -3.71601105e-01 1.53301850e-01 2.55442739e-01
-4.43956070e-02 -1.50915116e-01 -7.41042018e-01 -7.87266016e-01
-2.40968898e-01 -4.95553255e-01 3.68429631e-01 -5.44708431e-01
-9.31806386e-01 2.60970742e-01 1.87679246e-01 -1.41500437e+00
-2.04223379e-01 -3.47352862e-01 -4.31532085e-01 7.55249858e-01
-1.51334119e+00 -9.07538950e-01 -5.15765190e-01 3.08684796e-01
1.05445182e+00 -2.19446078e-01 5.49480498e-01 5.11356711e-01
-7.42304564e-01 4.13705856e-01 9.63692665e-02 8.25306252e-02
6.03688419e-01 -1.08962297e+00 -1.40098471e-03 9.50686455e-01
-2.01858997e-01 6.26013102e-03 4.33176994e-01 -7.64644861e-01
-6.30105257e-01 -1.58234370e+00 7.31413305e-01 -6.40153214e-02
1.13219388e-01 -2.49844611e-01 -9.54332650e-01 1.98083639e-01
-2.87012875e-01 1.23247448e-02 5.60678601e-01 2.20184606e-02
3.45908403e-02 -4.21583295e-01 -1.38269150e+00 5.66180825e-01
1.05677521e+00 -6.67274790e-03 -2.03884453e-01 3.73951286e-01
7.84822643e-01 -3.54672879e-01 -3.85953039e-01 7.54707456e-01
2.11204529e-01 -8.56100619e-01 8.11447561e-01 -3.92344475e-01
-7.96942413e-02 -6.22288465e-01 1.09179348e-01 -1.11430836e+00
-2.46012628e-01 -6.40238971e-02 3.22309025e-02 1.28207040e+00
9.42175508e-01 -6.30950987e-01 1.13705659e+00 7.43438423e-01
-3.66857260e-01 -7.04121053e-01 -7.75506794e-01 -6.77941680e-01
-2.34378770e-01 -7.98961461e-01 6.83242738e-01 8.97138834e-01
-5.05576611e-01 3.53619099e-01 -6.25500455e-02 2.62636662e-01
7.92235196e-01 -4.31602634e-02 6.80152297e-01 -1.63931751e+00
2.53821492e-01 -1.24086738e-01 -2.03329057e-01 -1.13903415e+00
1.26918718e-01 -5.86585879e-01 3.36915463e-01 -1.81913853e+00
-1.58490524e-01 -1.02884352e+00 -2.56285425e-02 5.45833707e-01
-2.68142998e-01 1.77754015e-01 -2.52192646e-01 2.71914601e-01
-8.72875869e-01 5.00833988e-01 1.33924294e+00 -4.51991379e-01
-5.96894324e-01 4.80507702e-01 -3.89407963e-01 7.66065419e-01
1.09083903e+00 -4.47577149e-01 -9.28874731e-01 -2.72025108e-01
9.89520401e-02 -2.97121108e-01 5.87262437e-02 -1.08239841e+00
2.84627020e-01 -5.92331231e-01 4.40437913e-01 -1.00375259e+00
1.89420298e-01 -1.18842781e+00 4.99967709e-02 2.20014647e-01
-3.22686821e-01 -7.98074424e-01 5.17880954e-02 6.09997153e-01
-1.92345858e-01 -3.62322271e-01 1.04896939e+00 -2.62776166e-01
-1.33199871e+00 5.62485158e-01 -5.46223998e-01 3.96127477e-02
1.10106182e+00 -6.54201090e-01 -1.43727083e-02 -6.16639815e-02
-7.54373372e-01 6.90889239e-01 2.39202753e-01 3.99920911e-01
4.63050336e-01 -1.22022164e+00 -4.59692568e-01 1.68125048e-01
4.00322914e-01 8.84018362e-01 3.47032666e-01 7.43885815e-01
-4.38028812e-01 2.49581277e-01 1.62405327e-01 -1.01981497e+00
-1.12157178e+00 2.48748034e-01 3.60046327e-01 -6.42681643e-02
-2.09078655e-01 7.37701595e-01 4.62137116e-03 -7.80171931e-01
4.13417369e-01 -5.19057252e-02 -6.75006986e-01 3.32263738e-01
3.47066194e-01 6.57599926e-01 3.71068239e-01 -4.79645878e-01
-1.47800297e-01 3.74111712e-01 4.92325686e-02 -3.94248590e-02
1.00863445e+00 -4.69900638e-01 2.63629258e-01 4.68363076e-01
8.79138529e-01 -2.87100971e-01 -1.39486301e+00 -3.71529013e-01
2.51851678e-01 -4.77211058e-01 2.78950870e-01 -1.02747095e+00
-1.00349903e+00 8.80033731e-01 1.01739109e+00 1.90538153e-01
1.08455062e+00 -1.12431474e-01 6.54489815e-01 4.42819774e-01
4.95367259e-01 -1.75371349e+00 -7.56489336e-02 5.46493649e-01
3.78270388e-01 -1.80926120e+00 -2.51449645e-01 -7.97749519e-01
-9.50646043e-01 7.83958316e-01 1.03259659e+00 2.87530512e-01
7.02477217e-01 -6.58529438e-03 3.01068246e-01 1.11827739e-02
-3.95146698e-01 -6.12654030e-01 3.34922016e-01 6.47720337e-01
-2.49494277e-02 1.89687237e-01 -2.24447578e-01 3.48501652e-01
2.86818475e-01 -3.35879587e-02 1.58719271e-01 7.27838874e-01
-1.00704527e+00 -1.08608484e+00 -4.78002578e-01 6.67953670e-01
-4.35043611e-02 2.71571457e-01 -1.57435730e-01 5.98670900e-01
5.84561706e-01 1.36748517e+00 3.21144015e-01 -2.58402020e-01
2.51980782e-01 1.51604190e-01 1.28892124e-01 -7.01589108e-01
4.16094735e-02 2.68206209e-01 3.72092307e-01 -3.50504458e-01
-7.57473409e-01 -4.86438543e-01 -1.72384369e+00 1.26263484e-01
-6.90010786e-01 2.71727413e-01 4.70668197e-01 1.05794382e+00
2.38996610e-01 2.70728558e-01 7.06580400e-01 -4.04698759e-01
-2.16018528e-01 -9.09491777e-01 -3.65096509e-01 1.34563863e-01
2.52772272e-02 -8.67121816e-01 -2.83021569e-01 3.73217463e-02] | [9.507773399353027, 0.7021668553352356] |
1e20544a-072f-48bd-870b-25a9d189c298 | actionable-recourse-via-gans-for-mobile | 2211.06525 | null | https://arxiv.org/abs/2211.06525v1 | https://arxiv.org/pdf/2211.06525v1.pdf | Actionable Recourse via GANs for Mobile Health | Mobile health apps provide a unique means of collecting data that can be used to deliver adaptive interventions.The predicted outcomes considerably influence the selection of such interventions. Recourse via counterfactuals provides tangible mechanisms to modify user predictions. By identifying plausible actions that increase the likelihood of a desired prediction, stakeholders are afforded agency over their predictions. Furthermore, recourse mechanisms enable counterfactual reasoning that can help provide insights into candidates for causal interventional features. We demonstrate the feasibility of GAN-generated recourse for mobile health applications on ensemble-survival-analysis-based prediction of medium-term engagement in the Safe Delivery App, a digital training tool for skilled birth attendants. | ['Lauren Bellhouse', 'Africa Perianez', 'Ana Fernandez del Rio', 'Anna Guitart', 'Jennifer Chien'] | 2022-11-12 | null | null | null | null | ['survival-analysis'] | ['miscellaneous'] | [ 7.49382496e-01 1.03002012e+00 -9.20533359e-01 -5.76960266e-01
-8.51725638e-01 -2.25289062e-01 4.79718626e-01 2.66249001e-01
1.84221819e-01 9.11178350e-01 1.21671903e+00 -9.12989318e-01
-1.82977647e-01 -1.07422888e+00 -8.97704303e-01 -1.38757885e-01
-1.67491198e-01 4.48795408e-02 -6.14616275e-01 -7.46409893e-02
1.77012701e-02 -3.33781801e-02 -1.28574896e+00 1.00159943e+00
1.06931245e+00 4.79696751e-01 4.30495627e-02 5.44612467e-01
-3.02664172e-02 9.72085595e-01 -2.29740605e-01 -6.77506983e-01
-1.37232333e-01 -7.98949301e-01 -3.60950440e-01 -8.63818169e-01
-2.11716130e-01 -5.83628416e-01 -1.99370198e-02 2.30419129e-01
9.50142860e-01 -1.48239534e-03 6.89573109e-01 -1.03803062e+00
-4.70801562e-01 1.09816813e+00 -1.21362470e-01 1.51183426e-01
8.09489012e-01 5.74840248e-01 5.34054875e-01 -1.53220400e-01
5.38678885e-01 1.15935016e+00 1.27379465e+00 9.73538399e-01
-1.37123322e+00 -7.23798454e-01 -1.57467112e-01 -4.04610246e-01
-4.72832114e-01 -8.25327277e-01 5.17016232e-01 -7.59872496e-01
6.66932464e-01 7.88525820e-01 1.11364067e+00 1.43979383e+00
6.73082888e-01 5.72404921e-01 9.71560121e-01 -2.53711045e-01
2.94070929e-01 3.85304056e-02 -3.44907194e-01 6.94940627e-01
4.19874117e-02 5.52749276e-01 -5.04429698e-01 -8.11564386e-01
5.06855011e-01 5.15495300e-01 6.00838177e-02 3.54134679e-01
-7.88642943e-01 9.05254602e-01 2.77392685e-01 -1.75999463e-01
-1.04819834e+00 4.83486086e-01 1.57948837e-01 -4.04933929e-01
6.06406569e-01 7.09532917e-01 -6.86401725e-01 -7.11550772e-01
-5.36748111e-01 1.78102359e-01 4.66940641e-01 5.47637284e-01
2.16569304e-02 -7.83803090e-02 -1.05108917e+00 4.93003458e-01
3.34001243e-01 5.49090087e-01 3.20124120e-01 -1.05529332e+00
1.80345476e-01 4.99132603e-01 2.45603114e-01 -7.39921510e-01
-4.60910022e-01 -2.36783773e-01 -4.18057263e-01 -3.83991629e-01
-8.08693618e-02 -1.09708583e+00 -4.80296791e-01 1.69611907e+00
5.32915056e-01 5.26161611e-01 -8.95301998e-02 8.56372714e-02
6.93119586e-01 3.81787270e-01 7.60438204e-01 -6.56482816e-01
1.06136215e+00 1.89551581e-02 -9.05623138e-01 1.60500199e-01
1.01904666e+00 -2.98027694e-01 1.13550472e+00 -2.00427324e-01
-1.16191208e+00 6.46998882e-02 -4.15850997e-01 7.46472180e-01
1.61686033e-01 -4.60651070e-01 1.08098722e+00 1.31520891e+00
-7.73196340e-01 7.51587033e-01 -7.72140741e-01 -8.06095004e-02
1.27187014e+00 2.49728590e-01 2.51165122e-01 -3.35922907e-03
-1.17448831e+00 6.36852860e-01 4.94451895e-02 -1.61814049e-01
-9.84558105e-01 -1.73385572e+00 -5.77065349e-01 1.35111108e-01
-3.21893282e-02 -1.05068326e+00 1.35726798e+00 -9.48129833e-01
-1.37629092e+00 1.48653999e-01 -1.81114316e-01 -5.18492043e-01
6.39858305e-01 5.36374822e-02 -3.46340865e-01 -3.76666307e-01
1.03587598e-01 3.81887585e-01 5.97284317e-01 -6.51012421e-01
-6.13794208e-01 -2.94776082e-01 -6.46823943e-02 1.38313100e-01
-1.98643744e-01 -2.68841139e-03 4.01172847e-01 -6.92596078e-01
-9.14488077e-01 -7.44097054e-01 -8.47851515e-01 -5.39770663e-01
-3.43814939e-01 2.03099355e-01 4.05512661e-01 -7.86193848e-01
1.50001788e+00 -1.72638214e+00 -7.17891693e-01 1.07267797e-01
-4.09423150e-02 -8.64900351e-02 6.04421720e-02 3.62958908e-01
4.15432975e-02 6.93960488e-01 3.14952224e-01 3.62080067e-01
-4.15523261e-01 -2.36656606e-01 -1.15017883e-01 -1.29638329e-01
1.81242496e-01 1.37636042e+00 -1.00044990e+00 -2.85906345e-01
1.75373033e-01 6.18869185e-01 -1.05947113e+00 3.63244712e-01
-5.12608051e-01 7.28564084e-01 -8.30478966e-01 2.54318058e-01
1.50240332e-01 -8.31759050e-02 5.28869092e-01 3.10081989e-01
4.01228741e-02 6.52275085e-01 -2.98443615e-01 1.14377713e+00
-8.52630854e-01 2.55988359e-01 -6.05030119e-01 -5.45019031e-01
3.62167865e-01 3.62836808e-01 8.82756770e-01 -4.12852526e-01
5.59064336e-02 -3.39030832e-01 -8.18164945e-02 -8.44203293e-01
-6.87787905e-02 -4.50018734e-01 -4.01693918e-02 6.61376059e-01
-4.95054901e-01 3.05890799e-01 -8.45448077e-01 -3.35253961e-02
1.45329916e+00 1.18892603e-01 4.84686077e-01 -1.96393624e-01
-4.57645565e-01 -3.69329825e-02 6.58634305e-01 9.73729730e-01
2.04695880e-01 3.23142052e-01 6.57315612e-01 -4.45992738e-01
-6.11814439e-01 -8.43707800e-01 -1.40296161e-01 1.13727725e+00
-6.55156493e-01 -9.50193331e-02 -7.53833652e-01 -9.52115536e-01
8.22786689e-02 1.36134624e+00 -1.20533991e+00 -5.50756752e-01
-1.28951883e-02 -9.33195412e-01 5.53854823e-01 5.69936156e-01
1.64132472e-02 -1.01600444e+00 -9.00958002e-01 5.00655115e-01
-2.29819357e-01 -2.14469418e-01 -5.97117364e-01 -3.42473924e-01
-1.33426237e+00 -1.06668508e+00 -5.98726332e-01 -3.19932364e-02
4.78541493e-01 -1.79371566e-01 9.60267544e-01 -5.32164425e-02
-5.33042364e-02 6.82710350e-01 -1.47075318e-02 -9.96301532e-01
-9.29545999e-01 -3.58145535e-01 -8.99963826e-02 -1.91631287e-01
-8.32078606e-02 -5.36230624e-01 -1.19669008e+00 1.20869726e-02
-3.06686103e-01 4.19841915e-01 2.20375448e-01 9.52094853e-01
5.14360368e-01 -5.57192326e-01 1.28803980e+00 -1.69953036e+00
7.96870768e-01 -1.19273758e+00 -1.73240621e-02 2.91667908e-01
-1.00049722e+00 -7.22689256e-02 3.16800803e-01 -6.52889729e-01
-1.71619689e+00 5.20882532e-02 -2.63062775e-01 2.33570278e-01
-3.93144749e-02 7.07835019e-01 -1.71524614e-01 5.31213641e-01
1.05588555e+00 -2.97239542e-01 -6.43793717e-02 -1.19178377e-01
5.70020020e-01 8.07533681e-01 -3.11905425e-02 -3.66395026e-01
-3.48300159e-01 7.86765069e-02 -2.24802002e-01 -4.02724355e-01
-4.71214592e-01 2.01727867e-01 3.59183967e-01 -5.70169806e-01
7.80379951e-01 -1.03438509e+00 -1.21165788e+00 -3.54416490e-01
-7.48678684e-01 -9.00181472e-01 -5.51952481e-01 3.35651845e-01
-6.07116282e-01 -7.65052736e-01 -7.64669850e-02 -1.11784434e+00
-4.23933774e-01 -6.75780773e-01 5.87775230e-01 2.39289820e-01
-8.88940692e-01 -1.08903253e+00 1.31233811e-01 5.66849828e-01
5.06334424e-01 7.21020818e-01 1.01866913e+00 -6.15583122e-01
-2.34464973e-01 -3.81690770e-01 4.08264518e-01 -2.98826188e-01
4.89745587e-01 1.51957229e-01 -9.87809718e-01 2.58066028e-01
-3.95828366e-01 2.43489489e-01 1.32159621e-01 1.32643056e+00
1.41873229e+00 -1.22601187e+00 -9.02549446e-01 2.36897320e-01
8.67198765e-01 6.61250174e-01 8.49829853e-01 -2.91621059e-01
4.86380845e-01 5.80629647e-01 6.76557064e-01 8.30591619e-01
6.03631258e-01 4.88176614e-01 1.38432473e-01 -2.01815128e-01
1.83243066e-01 -1.03156102e+00 2.34581456e-01 -2.97871292e-01
-3.30176860e-01 -1.13773137e-01 -1.04463780e+00 4.34919477e-01
-1.71064699e+00 -1.23423851e+00 -1.20415784e-01 2.28943753e+00
9.65257466e-01 -7.25999996e-02 4.49646980e-01 -5.53904951e-01
6.39964044e-01 -3.77700448e-01 -5.80832422e-01 -7.35688388e-01
4.08064365e-01 7.26737380e-02 4.90457058e-01 2.16207460e-01
-4.49420512e-01 4.18105543e-01 7.17169809e+00 5.87788522e-01
-7.30069637e-01 6.45559669e-01 1.63744378e+00 -2.88303912e-01
-1.11486769e+00 -9.85060930e-02 -1.57751411e-01 7.65814126e-01
1.50654936e+00 -2.53773123e-01 2.93287456e-01 6.26643836e-01
1.16614461e+00 -5.50404824e-02 -1.00032878e+00 3.00395787e-01
-6.64863884e-01 -2.31078148e+00 -1.71432540e-01 -3.67979356e-03
1.29125929e+00 -2.32119799e-01 2.94572473e-01 1.21639222e-01
8.60820830e-01 -1.09112573e+00 4.51039255e-01 1.10648954e+00
1.45178521e+00 -9.14839983e-01 4.48947698e-01 3.82738590e-01
-2.54914135e-01 -3.86340946e-01 4.43132401e-01 -3.86242300e-01
1.75116826e-02 6.72785580e-01 -1.71584272e+00 -9.66063365e-02
2.82921582e-01 2.66901523e-01 1.85310859e-02 5.00673831e-01
1.37063041e-01 1.41531372e+00 1.43839762e-01 -3.33093643e-01
-5.56483030e-01 4.33803350e-01 2.18704686e-01 1.26525164e+00
5.79948306e-01 6.51452303e-01 -5.82443774e-01 8.21802258e-01
-9.39636081e-02 4.76108491e-02 -1.10593104e+00 -3.61189604e-01
5.62856913e-01 5.54248571e-01 -3.69762897e-01 -1.06063224e-01
-1.37928829e-01 2.93746769e-01 -3.38881522e-01 2.72697955e-01
-6.21730268e-01 2.79980838e-01 7.42844701e-01 6.49460196e-01
-3.35139751e-01 7.57013679e-01 -9.89096761e-01 -4.90702748e-01
-7.00971425e-01 -9.36097145e-01 5.16891003e-01 -7.89574087e-01
-7.85353005e-01 -8.59465525e-02 -1.66773051e-01 -8.64208043e-01
-5.07695913e-01 2.92073458e-01 -9.51855719e-01 1.00410974e+00
-5.44798911e-01 -1.17903173e+00 1.92589089e-01 5.98399378e-02
4.35162723e-01 -5.48836254e-02 1.03740859e+00 -2.12879315e-01
-5.22798538e-01 7.51060188e-01 9.59730521e-02 -5.04738152e-01
1.47507861e-01 -7.36611724e-01 4.46617991e-01 3.16679597e-01
-4.19850200e-01 7.61849642e-01 7.78303146e-01 -1.36078036e+00
-1.22699010e+00 -1.36395741e+00 5.59019148e-01 -9.78836417e-01
2.19746754e-02 8.58083740e-02 -3.63671273e-01 6.09221339e-01
-2.40231916e-01 -4.68764216e-01 1.34769368e+00 5.56096256e-01
4.14414763e-01 8.94839242e-02 -1.66205680e+00 8.75898659e-01
1.08652008e+00 -2.61359096e-01 -4.51794243e-04 3.02195162e-01
1.13260794e+00 -3.53288323e-01 -8.93340468e-01 3.36804152e-01
9.96064842e-01 -7.32494056e-01 8.14478993e-01 -1.28783166e+00
1.11856878e+00 6.73564911e-01 1.68690905e-01 -1.46819103e+00
-8.94191116e-02 -1.25294590e+00 -3.31038862e-01 1.15498114e+00
9.83707309e-01 -5.85783482e-01 8.84724855e-01 1.46308982e+00
1.48749545e-01 -1.13298059e+00 -5.94478905e-01 7.07092360e-02
-2.55176067e-01 -1.02666914e+00 1.41753757e+00 1.18890941e+00
3.71663719e-01 -1.15132108e-01 -4.20021683e-01 -1.68193523e-02
1.98848918e-01 -3.17941546e-01 4.99713868e-01 -7.83734977e-01
-5.52324116e-01 -1.31225333e-01 3.26008536e-03 -3.38234425e-01
-2.45422989e-01 -5.76382339e-01 -1.03323244e-01 -1.52896810e+00
4.28795040e-01 -9.76350009e-01 -3.52639884e-01 5.97310483e-01
-6.47256494e-01 -2.21568868e-01 -1.12782918e-01 -3.97017181e-01
-4.87571806e-02 3.71521682e-01 1.07870066e+00 1.05971165e-01
-8.51894379e-01 8.16906571e-01 -1.36978889e+00 6.78666294e-01
8.99183571e-01 -6.51247263e-01 -5.60015798e-01 5.84281310e-02
5.74724257e-01 9.72902656e-01 3.08046490e-01 -1.75004512e-01
-4.46062475e-01 -9.06227946e-01 3.60526294e-01 3.15489620e-02
-3.42005849e-01 -5.40614963e-01 9.50349748e-01 7.90803134e-01
-1.02570939e+00 -3.04087788e-01 3.47616792e-01 9.91797209e-01
6.88004792e-01 1.27986640e-01 -2.52759829e-02 3.00705791e-01
1.61664993e-01 2.11625472e-01 -7.95335472e-01 4.88204882e-02
9.90974963e-01 -1.74154520e-01 -1.92288999e-02 -9.25968051e-01
-6.92323267e-01 1.27932325e-01 1.22509915e-02 3.32528204e-01
6.65999591e-01 -1.24061489e+00 -7.14188099e-01 -5.61713986e-02
-1.85466662e-01 -4.85882938e-01 6.64006591e-01 6.83746934e-01
-9.14186016e-02 3.45081873e-02 -4.18628864e-02 -1.00162156e-01
-1.11309588e+00 3.32594752e-01 3.92343760e-01 -2.29789112e-02
-2.73330063e-01 6.51510835e-01 6.90345764e-02 -2.75141001e-01
-2.83124208e-01 -4.59547751e-02 -2.62188822e-01 -1.28339782e-01
7.17462182e-01 8.08842719e-01 -1.48066983e-01 2.10018098e-01
-1.14526808e-01 -6.37004673e-01 5.49704313e-01 -2.37802207e-01
1.73765051e+00 -2.89279353e-02 1.98613420e-01 4.71222699e-01
6.31130576e-01 2.01039195e-01 -1.53822780e+00 4.25507545e-01
-3.16113979e-01 -5.19642174e-01 2.98147034e-02 -1.43589723e+00
-5.09684741e-01 5.37968576e-01 9.50720429e-01 1.52469471e-01
1.25406861e+00 -1.15703046e-01 3.05073678e-01 -1.60565883e-01
2.07280442e-01 -6.02275848e-01 -3.93201977e-01 -3.32767010e-01
8.12201023e-01 -1.21092224e+00 -1.14601694e-01 -1.85850516e-01
-9.07146573e-01 6.40904903e-01 2.95434535e-01 5.53670168e-01
6.20901465e-01 3.20094854e-01 -1.53459907e-01 3.52029651e-02
-1.12356591e+00 4.64675575e-01 3.17795157e-01 1.09157431e+00
7.93420970e-01 8.34922612e-01 -3.54632944e-01 1.30956042e+00
4.37557325e-02 6.10113323e-01 3.74223053e-01 4.57898885e-01
1.15067005e-01 -6.83593392e-01 -1.66833833e-01 1.49496198e+00
-8.23621929e-01 -4.33480710e-01 -5.28586917e-02 7.66351223e-02
3.64027739e-01 9.68012214e-01 -7.37666935e-02 -5.18668294e-01
3.93662542e-01 1.22399434e-01 1.57619759e-01 -6.03329182e-01
-8.54767621e-01 -3.24840754e-01 7.31644988e-01 -6.15973651e-01
7.80136436e-02 -9.70179260e-01 -1.16951799e+00 -3.55713457e-01
-2.84775972e-01 1.16139472e-01 5.48681021e-01 8.93022835e-01
9.96637762e-01 8.15140963e-01 8.55016589e-01 -4.56910193e-01
-1.87822748e-02 -8.10004413e-01 2.38693833e-01 3.00071035e-02
1.46549612e-01 -2.33932316e-01 5.09541750e-01 -4.61322721e-03] | [8.213264465332031, 5.552174091339111] |
e59abea5-8c19-486d-b593-4264cda463c8 | aang-automating-auxiliary-learning | 2205.14082 | null | https://arxiv.org/abs/2205.14082v2 | https://arxiv.org/pdf/2205.14082v2.pdf | AANG: Automating Auxiliary Learning | Auxiliary objectives, supplementary learning signals that are introduced to help aid learning on data-starved or highly complex end-tasks, are commonplace in machine learning. Whilst much work has been done to formulate useful auxiliary objectives, their construction is still an art which proceeds by slow and tedious hand-design. Intuition for how and when these objectives improve end-task performance has also had limited theoretical backing. In this work, we present an approach for automatically generating a suite of auxiliary objectives. We achieve this by deconstructing existing objectives within a novel unified taxonomy, identifying connections between them, and generating new ones based on the uncovered structure. Next, we theoretically formalize widely-held intuitions about how auxiliary learning improves generalization on the end-task. This leads us to a principled and efficient algorithm for searching the space of generated objectives to find those most useful to a specified end-task. With natural language processing (NLP) as our domain of study, we demonstrate that our automated auxiliary learning pipeline leads to strong improvements over competitive baselines across continued training experiments on a pre-trained model on 5 NLP tasks. | ['Ameet Talwalkar', 'Graham Neubig', 'Mikhail Khodak', 'Paul Michel', 'Lucio M. Dery'] | 2022-05-27 | null | null | null | null | ['auxiliary-learning'] | ['methodology'] | [ 7.22953618e-01 4.43165898e-01 -3.25707674e-01 -4.67403769e-01
-1.11802912e+00 -7.62967288e-01 6.78662837e-01 4.95568335e-01
-6.45844579e-01 6.75647676e-01 5.94306409e-01 -6.10519946e-01
-1.85992062e-01 -2.85208642e-01 -6.83541298e-01 -5.77009797e-01
5.90606220e-02 6.38743460e-01 -1.58574097e-02 -2.28512540e-01
3.71642947e-01 2.27600932e-01 -1.30218387e+00 3.62582564e-01
8.77224684e-01 7.67465174e-01 2.32689843e-01 5.95068872e-01
-1.50253147e-01 5.97005069e-01 -4.31780040e-01 -5.89661658e-01
1.47262171e-01 -3.73264253e-01 -1.24534523e+00 -8.07300210e-02
2.74211854e-01 9.70889106e-02 2.92800725e-01 7.50031471e-01
3.14694911e-01 2.87226170e-01 8.32122445e-01 -1.16830540e+00
-7.13859320e-01 6.77881956e-01 -1.49474978e-01 2.89633960e-01
9.34206173e-02 1.89281911e-01 1.52771246e+00 -9.08490181e-01
4.45690632e-01 1.15379667e+00 7.01096237e-01 6.98970854e-01
-1.61607027e+00 -2.86613703e-01 3.93053830e-01 -1.12755515e-01
-7.49274015e-01 -6.63955927e-01 7.18708038e-01 -2.67177463e-01
1.05326188e+00 1.17133491e-01 3.08410078e-01 1.18750739e+00
2.75596902e-02 1.26258004e+00 9.48135018e-01 -6.28230333e-01
3.80410068e-02 4.67532992e-01 6.02192819e-01 5.81216395e-01
2.08679631e-01 5.85856363e-02 -6.26684248e-01 1.74303856e-02
1.98126093e-01 -3.72058392e-01 -2.72350639e-01 -3.62725109e-01
-1.28261602e+00 9.17986155e-01 1.64046407e-01 2.79410690e-01
-3.19277585e-01 -1.35852711e-03 3.74044329e-01 6.01327181e-01
6.63671255e-01 1.13548625e+00 -1.09019804e+00 -1.39155671e-01
-6.39331341e-01 1.71395063e-01 8.06473553e-01 7.27998137e-01
7.27178931e-01 -1.18667416e-01 -5.03971219e-01 8.49348247e-01
2.85317302e-01 -3.44056375e-02 4.15987402e-01 -9.07963872e-01
6.61577940e-01 5.68721056e-01 1.49280816e-01 -6.08129621e-01
-5.16887367e-01 -8.17734957e-01 -5.05481958e-01 1.07368454e-01
7.12016821e-01 -3.27104449e-01 -8.24248314e-01 2.00199485e+00
7.96056241e-02 -3.66931319e-01 2.79910117e-01 5.13268411e-01
4.13804710e-01 4.97032940e-01 3.25190365e-01 -1.60903662e-01
1.25102317e+00 -1.17975688e+00 -5.21901250e-01 -5.26188612e-01
9.57274854e-01 -6.92073047e-01 1.64888573e+00 4.72153842e-01
-1.05350220e+00 -5.00823915e-01 -8.25642407e-01 -3.93383682e-01
-4.67471272e-01 1.25349268e-01 9.44175601e-01 6.14130974e-01
-1.16166914e+00 7.34777331e-01 -7.77656376e-01 -7.15945140e-02
5.72336555e-01 3.80539775e-01 -2.92748332e-01 1.40158266e-01
-1.07412291e+00 1.18274271e+00 5.85436523e-01 -2.46914819e-01
-7.73462057e-01 -9.37155724e-01 -8.74981344e-01 1.38534009e-01
5.23730993e-01 -8.95076156e-01 1.53114665e+00 -1.20844030e+00
-1.40849519e+00 9.99546647e-01 -2.10713163e-01 -6.22945726e-01
1.58126518e-01 -4.62899804e-01 -7.20570907e-02 -3.86447087e-02
2.21932665e-01 9.35635686e-01 8.22134614e-01 -1.09613264e+00
-8.15026939e-01 -2.73903161e-01 2.31397390e-01 4.60440189e-01
-6.24784827e-01 9.64737460e-02 -3.03507835e-01 -7.53811300e-01
-7.41844624e-02 -7.02076137e-01 -3.57956439e-01 -1.32956833e-01
-1.63662151e-01 -5.58349788e-01 3.92359614e-01 -5.56388438e-01
1.06421208e+00 -1.95814598e+00 2.95445770e-01 3.68523188e-02
3.02667379e-01 2.95702338e-01 -4.70402241e-01 2.42012113e-01
-1.58323690e-01 2.12452844e-01 -4.36338544e-01 -5.20564020e-01
2.25402996e-01 5.84563762e-02 -5.61479509e-01 5.69268689e-02
5.97482085e-01 1.21182036e+00 -1.22640705e+00 -3.07440311e-01
-2.18481310e-02 3.14972669e-01 -8.47964883e-01 2.60391414e-01
-4.35740918e-01 3.94407839e-01 -3.99747878e-01 6.11841500e-01
1.05504334e-01 -2.11652637e-01 -2.74641514e-01 -2.72330493e-02
9.00090113e-02 9.55276370e-01 -8.17199111e-01 1.71720898e+00
-5.94266117e-01 6.30391896e-01 1.21838711e-01 -1.32479179e+00
7.59370327e-01 1.35768294e-01 3.29435527e-01 -4.86872345e-01
6.65074959e-02 7.75054246e-02 2.65145361e-01 -4.16011781e-01
4.26025331e-01 -2.72110373e-01 -5.39413840e-02 8.26938570e-01
4.65530068e-01 -2.38509461e-01 3.97104710e-01 9.79287997e-02
1.11183286e+00 3.04125935e-01 4.57111239e-01 -4.42638189e-01
5.23717999e-01 2.73284912e-01 1.45909891e-01 9.42320466e-01
-2.02735797e-01 3.97651047e-01 6.02441072e-01 -4.51748729e-01
-8.98063242e-01 -1.12580574e+00 -1.87460687e-02 1.63158941e+00
-4.10956085e-01 -5.57452857e-01 -3.10103714e-01 -9.94461060e-01
-2.60445505e-01 9.64812815e-01 -5.96817374e-01 -2.28584692e-01
-3.64160091e-01 -8.44704986e-01 4.01055247e-01 6.66525304e-01
1.23140335e-01 -1.32168555e+00 -5.68350732e-01 2.45340362e-01
-2.43173867e-01 -9.37772632e-01 -4.70187068e-01 6.63991094e-01
-1.06144977e+00 -8.94395411e-01 -6.02538705e-01 -9.33509588e-01
7.17209399e-01 2.27811038e-01 1.62028551e+00 -6.38434365e-02
-3.79288048e-02 4.98250455e-01 -3.66276890e-01 -8.66527379e-01
-4.92376506e-01 3.81687135e-01 -1.15001284e-01 -2.62716681e-01
6.09420121e-01 -5.75914502e-01 -2.06298769e-01 1.08631030e-02
-9.04963017e-01 1.26029283e-01 9.70857739e-01 9.13560927e-01
4.87762511e-01 -2.11727038e-01 6.33509874e-01 -9.39032435e-01
1.17059720e+00 -3.97294670e-01 -4.53272045e-01 2.13235527e-01
-6.65172517e-01 7.19177246e-01 5.94507337e-01 -3.56614202e-01
-1.00598061e+00 2.62511283e-01 -3.25114548e-01 -7.67710805e-02
-2.53065288e-01 7.41375446e-01 -1.71575978e-01 3.17739487e-01
9.12416041e-01 2.44247600e-01 1.56554244e-02 -4.18791980e-01
6.58521831e-01 4.29934651e-01 4.18142945e-01 -1.01716769e+00
8.48760188e-01 1.63500115e-01 -1.23998024e-01 -7.91889310e-01
-1.56428981e+00 -5.95686913e-01 -7.19510913e-01 2.34233499e-01
5.49889565e-01 -7.21230865e-01 -4.22185987e-01 -4.08439524e-02
-1.15137339e+00 -6.46673560e-01 -4.76768762e-01 4.24568295e-01
-6.08474731e-01 1.79848656e-01 -8.48886371e-02 -5.95924497e-01
-5.21289110e-01 -1.01843262e+00 1.13286746e+00 -3.02331418e-01
-6.51183903e-01 -1.21273351e+00 1.28609717e-01 3.54070306e-01
4.52147931e-01 -5.76550663e-02 1.09811699e+00 -9.52023447e-01
-3.00868601e-01 3.28701511e-02 -1.70960203e-01 5.42991459e-01
2.14298159e-01 -4.26048994e-01 -9.24274087e-01 -1.82485774e-01
2.18686357e-01 -6.57034755e-01 9.98294175e-01 2.59579957e-01
1.28395188e+00 -5.44056594e-01 -2.53482193e-01 5.99871516e-01
1.04486239e+00 -2.08280563e-01 2.08829626e-01 4.82258767e-01
3.78236860e-01 9.81960118e-01 5.69831967e-01 -9.84398872e-02
2.28244424e-01 4.50522512e-01 1.47105992e-01 -8.59535560e-02
7.97495097e-02 -1.76800847e-01 6.19973123e-01 4.15547073e-01
7.12833107e-02 -1.29298836e-01 -1.05012774e+00 8.55942369e-01
-1.70413542e+00 -7.77671456e-01 9.84843299e-02 2.01572156e+00
1.16745067e+00 5.38911223e-01 1.59330204e-01 1.69100881e-01
3.41643631e-01 -1.58579387e-02 -4.52759296e-01 -4.38064307e-01
8.98675323e-02 4.72855091e-01 3.14119071e-01 6.40129745e-01
-1.17936397e+00 1.06276810e+00 6.26087284e+00 5.45369446e-01
-9.50631380e-01 2.05447059e-03 6.99636817e-01 1.19852694e-02
-4.81366098e-01 4.39661555e-02 -9.41242039e-01 6.23677000e-02
9.16450381e-01 -2.44774163e-01 3.25319827e-01 8.13066363e-01
1.67908832e-01 2.48290241e-01 -1.59147108e+00 5.41729927e-01
2.65237782e-02 -1.23399305e+00 2.38462254e-01 -7.27877915e-02
6.04614735e-01 1.31118685e-01 2.25283295e-01 6.32786572e-01
7.19157755e-01 -1.14177251e+00 5.26190579e-01 -4.22585495e-02
5.14223278e-01 -5.10906577e-01 5.53924024e-01 5.18401206e-01
-6.96584225e-01 -1.15654565e-01 -1.99852541e-01 -3.13140154e-01
-1.11902304e-01 4.32514995e-01 -9.71021414e-01 2.18034700e-01
3.29059720e-01 5.68288326e-01 -6.90759420e-01 7.23190188e-01
-7.52650797e-01 8.59538436e-01 -1.96174726e-01 -1.13669939e-01
6.53372169e-01 9.43733193e-03 6.91280603e-01 1.30811274e+00
9.09188355e-04 -1.63568601e-01 2.40791403e-02 8.31766605e-01
-2.19414249e-01 2.19462171e-01 -6.86086893e-01 -3.17277402e-01
9.23675746e-02 1.30854392e+00 -7.03827858e-01 -1.55608818e-01
-4.07599270e-01 6.82987988e-01 6.57833040e-01 3.62600923e-01
-6.07381463e-01 -4.97681469e-01 6.08664870e-01 3.61027271e-02
6.83062896e-02 -2.30864480e-01 -5.00901222e-01 -1.28359354e+00
9.60236937e-02 -1.03879344e+00 6.09623671e-01 -5.63022256e-01
-1.29587030e+00 5.10512769e-01 -2.30199266e-02 -8.58220696e-01
-4.35190290e-01 -8.55390966e-01 -6.34420574e-01 1.01527977e+00
-1.59488773e+00 -9.60905910e-01 2.92190798e-02 4.87125754e-01
9.13020372e-01 -2.90029258e-01 9.57066774e-01 -1.52784705e-01
-1.92690909e-01 7.27835298e-01 -2.71266736e-02 1.29227623e-01
7.41679549e-01 -1.43924999e+00 4.71295685e-01 1.06786740e+00
4.96429116e-01 7.84323573e-01 9.16841686e-01 -2.94429958e-01
-1.15131581e+00 -1.07981718e+00 1.20087147e+00 -9.21384454e-01
9.18380618e-01 -5.26456594e-01 -8.20943654e-01 8.87840748e-01
2.66211271e-01 -2.71623641e-01 6.79777324e-01 6.60098851e-01
-3.49362999e-01 4.10373770e-02 -7.86374927e-01 7.81252563e-01
1.03917491e+00 -3.87737483e-01 -1.32152998e+00 3.75736624e-01
9.85650182e-01 -1.71441227e-01 -2.98405826e-01 4.25046176e-01
3.34586710e-01 -5.18756330e-01 1.07527876e+00 -1.14515436e+00
7.40724981e-01 -2.72915568e-02 -1.29310917e-02 -1.69825208e+00
-3.80269945e-01 -9.02689636e-01 -1.27364054e-01 1.19818580e+00
8.63716602e-01 -5.21385431e-01 7.76688278e-01 5.86365879e-01
-4.65375125e-01 -8.61618221e-01 -3.89812350e-01 -7.93720305e-01
2.68570811e-01 -7.05812395e-01 1.22231141e-01 9.08311963e-01
5.76118380e-02 1.04053676e+00 4.92052324e-02 -7.53947645e-02
5.52385151e-01 -1.47703230e-01 7.26754010e-01 -1.41246533e+00
-3.12587291e-01 -7.71331668e-01 1.31422207e-01 -1.20119929e+00
4.51756299e-01 -1.21464086e+00 1.06156267e-01 -1.46859753e+00
4.79425676e-02 -3.75917196e-01 -5.36265373e-01 7.63277829e-01
-4.90930825e-01 -9.89450049e-03 1.38601005e-01 8.75672046e-03
-6.72979116e-01 5.47311366e-01 1.02367711e+00 -2.31493130e-01
-5.25803149e-01 2.74461925e-01 -1.55056560e+00 8.36465657e-01
9.02116537e-01 -5.68093538e-01 -7.28910923e-01 -7.68475831e-01
4.43113506e-01 -4.19701964e-01 9.94302332e-02 -6.85739219e-01
2.41181552e-02 -9.36179757e-02 1.49129838e-01 -5.21335483e-01
1.64021507e-01 -7.26074219e-01 -6.11553907e-01 3.62495422e-01
-9.25990999e-01 9.41020921e-02 5.10894537e-01 4.32792991e-01
-2.22713977e-01 -4.73092914e-01 6.29834473e-01 -2.06122309e-01
-4.53449875e-01 2.03228906e-01 -9.09422413e-02 5.22397578e-01
6.48398936e-01 3.40905748e-02 1.63846998e-03 -2.53003597e-01
-8.34438503e-01 4.43403840e-01 1.28505984e-02 5.31509221e-01
3.79246294e-01 -9.88884509e-01 -9.91626143e-01 2.41203257e-03
1.48707747e-01 2.68522501e-02 -4.36189830e-01 6.83385909e-01
2.81235538e-02 8.04530442e-01 -1.85546309e-01 -5.18926919e-01
-1.14609540e+00 6.12867296e-01 1.74735729e-02 -7.81744659e-01
-3.66778702e-01 1.01452100e+00 3.55760545e-01 -5.61957657e-01
5.12569487e-01 -3.93032998e-01 -2.36372322e-01 2.22578704e-01
4.28128034e-01 8.56139511e-02 2.01562107e-01 -2.27717027e-01
-5.74471988e-02 2.16701105e-01 -2.75541872e-01 -2.56982625e-01
1.67073190e+00 -1.03697747e-01 1.15861103e-01 3.81824374e-01
1.16441917e+00 -1.19240843e-01 -1.17325485e+00 -4.81640130e-01
3.13792259e-01 -1.61346316e-01 -3.39751132e-02 -8.36978734e-01
-5.43980122e-01 1.05561316e+00 1.56698693e-02 1.57139331e-01
1.16094458e+00 1.14073731e-01 3.60173553e-01 1.02801931e+00
4.78281155e-02 -1.01885140e+00 3.79904360e-01 9.68830049e-01
1.02742422e+00 -1.46384180e+00 -7.51686916e-02 -1.69207588e-01
-6.71023965e-01 1.02262628e+00 5.91383100e-01 5.19950017e-02
4.82553124e-01 -4.11889255e-02 -7.05472678e-02 -2.47009590e-01
-1.01122546e+00 -4.15043622e-01 4.43029702e-01 7.43762255e-01
6.62896395e-01 -3.86650443e-01 -3.74322623e-01 7.23765671e-01
-4.20261621e-01 -2.30103865e-01 1.99234784e-01 7.45950043e-01
-5.29874027e-01 -1.23323917e+00 -1.84898511e-01 6.17760956e-01
-7.39342690e-01 -5.57108700e-01 -4.90950018e-01 8.94217789e-01
-8.58205333e-02 7.51330078e-01 -9.32276845e-02 9.48226973e-02
2.85841286e-01 4.00608987e-01 3.92423064e-01 -1.06591606e+00
-4.86321419e-01 -2.78751999e-01 2.22117603e-01 -3.64814311e-01
-4.41136867e-01 -5.78455567e-01 -7.69533932e-01 2.73792505e-01
-1.27654254e-01 2.86786050e-01 4.83332813e-01 1.20444250e+00
2.64901131e-01 5.03552496e-01 3.48184228e-01 -7.23213911e-01
-8.14162135e-01 -9.59441185e-01 6.76793605e-02 5.13374627e-01
4.26881135e-01 -5.01019716e-01 -4.05101895e-01 2.28099227e-01] | [10.577397346496582, 8.296758651733398] |
3a5eea4d-4c6e-45ad-875b-8d3d3b6af9ed | charformer-fast-character-transformers-via | 2106.12672 | null | https://arxiv.org/abs/2106.12672v3 | https://arxiv.org/pdf/2106.12672v3.pdf | Charformer: Fast Character Transformers via Gradient-based Subword Tokenization | State-of-the-art models in natural language processing rely on separate rigid subword tokenization algorithms, which limit their generalization ability and adaptation to new settings. In this paper, we propose a new model inductive bias that learns a subword tokenization end-to-end as part of the model. To this end, we introduce a soft gradient-based subword tokenization module (GBST) that automatically learns latent subword representations from characters in a data-driven fashion. Concretely, GBST enumerates candidate subword blocks and learns to score them in a position-wise fashion using a block scoring network. We additionally introduce Charformer, a deep Transformer model that integrates GBST and operates on the byte level. Via extensive experiments on English GLUE, multilingual, and noisy text datasets, we show that Charformer outperforms a series of competitive byte-level baselines while generally performing on par and sometimes outperforming subword-based models. Additionally, Charformer is fast, improving the speed of both vanilla byte-level and subword-level Transformers by 28%-100% while maintaining competitive quality. We believe this work paves the way for highly performant token-free models that are trained completely end-to-end. | ['Donald Metzler', 'Cong Yu', 'Simon Baumgartner', 'Zhen Qin', 'Dara Bahri', 'Hyung Won Chung', 'Jai Gupta', 'Sebastian Ruder', 'Vinh Q. Tran', 'Yi Tay'] | 2021-06-23 | charformer-fast-character-transformers-via-1 | https://openreview.net/forum?id=JtBRnrlOEFN | https://openreview.net/pdf?id=JtBRnrlOEFN | iclr-2022-4 | ['paraphrase-identification', 'linguistic-acceptability'] | ['natural-language-processing', 'natural-language-processing'] | [ 1.70257509e-01 -9.58077908e-02 -5.18327296e-01 -5.04442036e-01
-1.42762744e+00 -6.76634431e-01 6.62122071e-01 4.54343319e-01
-1.02646029e+00 4.31296378e-01 5.53222537e-01 -6.55259490e-01
5.99976718e-01 -6.39359951e-01 -9.62897897e-01 -5.68754792e-01
-1.94520541e-02 3.86536807e-01 1.43268734e-01 -1.99624956e-01
1.56370685e-01 -1.16569148e-02 -1.09302235e+00 7.44165778e-01
7.05641329e-01 4.97456998e-01 3.53000224e-01 6.91997230e-01
-4.73826438e-01 5.71839869e-01 -2.41687223e-01 -5.86926401e-01
2.42771938e-01 -2.00815941e-03 -9.37283874e-01 -3.64437580e-01
8.84926975e-01 -2.37426892e-01 1.22392483e-01 7.28703976e-01
4.05792356e-01 2.22842302e-02 4.17887419e-01 -6.14932537e-01
-6.27727985e-01 1.44512153e+00 -4.97655749e-01 -1.30387405e-02
2.05363646e-01 1.92035452e-01 1.65548706e+00 -1.26680326e+00
4.41506296e-01 1.35298598e+00 8.54478359e-01 6.09308243e-01
-1.44399226e+00 -6.56079531e-01 7.41224289e-01 -3.61904642e-03
-1.11533725e+00 -4.11558986e-01 4.98418659e-01 -1.00504801e-01
1.59104526e+00 -6.20967969e-02 4.41504121e-01 1.09275126e+00
2.26750240e-01 1.31185567e+00 9.26329374e-01 -7.07757056e-01
1.47097632e-01 -2.28549361e-01 4.90203977e-01 8.62994730e-01
2.82240331e-01 -1.97273880e-01 -6.94809258e-01 -1.84520140e-01
3.36361647e-01 -1.53730243e-01 2.00100169e-01 -2.24708632e-01
-1.38166356e+00 7.53999352e-01 3.72643471e-01 4.14586365e-01
-4.17866617e-01 5.80228031e-01 8.32296550e-01 1.74024910e-01
6.52904272e-01 3.74833941e-01 -7.80205846e-01 -4.49167043e-01
-1.23265493e+00 3.61157864e-01 6.18269384e-01 8.44314456e-01
9.81029153e-01 2.27838278e-01 -4.58478063e-01 8.96636128e-01
1.10607207e-01 4.39611584e-01 7.04587877e-01 -3.74746174e-01
7.31157601e-01 2.00914815e-01 -1.92210302e-01 -1.62008375e-01
-2.55017906e-01 -4.46506590e-01 -4.63290989e-01 -8.56984556e-02
2.53578782e-01 1.36191794e-03 -1.25364101e+00 1.91439950e+00
1.58362284e-01 3.01557153e-01 -1.11902870e-01 5.44490337e-01
4.25086260e-01 7.15218186e-01 6.38834476e-01 1.67221755e-01
1.68717992e+00 -1.06451678e+00 -3.24773639e-01 -4.51491535e-01
1.02887738e+00 -8.56368542e-01 1.68997562e+00 5.72526515e-01
-1.26466572e+00 -5.62570632e-01 -9.64899302e-01 -4.95886952e-01
-3.45435679e-01 3.30868699e-02 6.74179018e-01 7.13549197e-01
-1.07682431e+00 4.10020590e-01 -1.23091590e+00 -2.54344702e-01
1.96230993e-01 2.32483461e-01 -3.14684898e-01 -2.19216734e-01
-1.19119251e+00 8.11995089e-01 4.63972092e-01 -1.58821970e-01
-9.01198208e-01 -1.04757154e+00 -1.22172785e+00 2.84890950e-01
1.52621567e-01 -6.92780316e-01 1.64346278e+00 -8.66112590e-01
-1.53841364e+00 8.39190066e-01 -5.87639153e-01 -7.64233649e-01
2.32014716e-01 -5.26965320e-01 -7.28900209e-02 -1.22419573e-01
2.22661898e-01 6.64981663e-01 8.12342286e-01 -1.08023739e+00
-9.42616522e-01 9.58543718e-02 -8.16077739e-02 4.80165109e-02
-6.26104832e-01 8.79756436e-02 -4.89118457e-01 -9.29685354e-01
-2.58437932e-01 -6.15384996e-01 -2.64981866e-01 -4.64270860e-01
-3.77106786e-01 -5.82970738e-01 3.31068009e-01 -5.48488498e-01
1.28433871e+00 -2.09420562e+00 1.91815440e-02 1.16193809e-01
2.10376978e-01 4.22550410e-01 -6.08006537e-01 6.52773559e-01
-1.66636094e-01 1.45321906e-01 -2.72653908e-01 -1.08229017e+00
4.14403707e-01 2.21140966e-01 -5.64686835e-01 2.60762423e-01
4.31654274e-01 1.02991748e+00 -1.14621937e+00 -3.46606910e-01
1.37827471e-01 4.09781367e-01 -8.22564721e-01 -2.20813463e-03
-3.78920078e-01 -3.84483367e-01 -1.44262061e-01 6.12378895e-01
5.82895994e-01 1.31138995e-01 2.49553427e-01 -1.55356437e-01
-2.50270218e-01 1.13793135e+00 -9.79702592e-01 2.06403279e+00
-1.00806093e+00 4.57730114e-01 -4.82821986e-02 -9.80752409e-01
5.99739492e-01 2.70966887e-01 2.53889471e-01 -5.45060515e-01
-1.73925072e-01 2.85361856e-01 -2.40283802e-01 -1.37550563e-01
1.18671000e+00 -3.14048409e-01 -4.78451133e-01 6.17830276e-01
2.70401508e-01 -1.27842175e-02 4.61654186e-01 5.03537834e-01
1.19773018e+00 2.00659245e-01 2.57580042e-01 -2.89902270e-01
3.14446300e-01 -1.36646762e-01 4.68959123e-01 9.79281366e-01
8.95583332e-02 4.88479912e-01 3.54940563e-01 -4.06411618e-01
-1.18980503e+00 -1.35326529e+00 1.56408437e-02 1.85790515e+00
-3.82480353e-01 -9.14662659e-01 -7.37071216e-01 -7.61545837e-01
2.26235226e-01 7.39516914e-01 -6.10860407e-01 -7.89554343e-02
-1.06047440e+00 -5.94292343e-01 8.47216189e-01 7.28940606e-01
9.68240723e-02 -1.04391682e+00 -1.56214580e-01 6.34695888e-01
-1.57675579e-01 -1.23008764e+00 -8.38758051e-01 6.40011847e-01
-6.36773705e-01 -4.68725502e-01 -5.59611499e-01 -1.07469416e+00
5.22608638e-01 2.27924675e-01 1.40524113e+00 3.55747715e-02
-1.01952501e-01 4.49168645e-02 -5.72662413e-01 -3.54474902e-01
-3.98275822e-01 8.73958230e-01 -3.50180850e-03 -1.92369804e-01
6.11450434e-01 -3.58236104e-01 -4.36364174e-01 -2.48640820e-01
-9.93570626e-01 -1.41121552e-03 8.87065709e-01 9.87417102e-01
6.41973019e-01 -5.04040480e-01 1.96475253e-01 -1.06455219e+00
7.23106861e-01 -3.35828185e-01 -4.47289884e-01 1.08044386e-01
-6.88521087e-01 4.63390023e-01 9.87045288e-01 -4.44140047e-01
-7.89268672e-01 1.41895935e-01 -4.14264649e-01 3.33424024e-02
4.32177596e-02 6.98338270e-01 -5.20591289e-02 3.53371829e-01
5.36762834e-01 4.81069535e-01 -3.94439548e-01 -6.69173896e-01
7.96411216e-01 5.43124497e-01 7.28005290e-01 -1.02904212e+00
9.91354644e-01 4.40710962e-01 -5.94303668e-01 -6.18760526e-01
-9.13558006e-01 -6.86247051e-01 -4.58899260e-01 3.00733566e-01
7.20022857e-01 -1.08725727e+00 -4.46599931e-01 5.00126839e-01
-1.16047645e+00 -7.32364655e-01 -4.50785369e-01 3.82469505e-01
-1.74226403e-01 5.03260851e-01 -9.05498207e-01 -4.49786603e-01
-7.16216981e-01 -1.06322312e+00 1.39597857e+00 -3.10259908e-01
-3.26716959e-01 -9.47971225e-01 2.19166562e-01 5.96671663e-02
5.25424957e-01 -3.27074170e-01 1.03805912e+00 -5.61286867e-01
-4.43520039e-01 -1.56150341e-01 -9.47608128e-02 4.69962746e-01
-8.16005319e-02 -2.55258381e-02 -9.09262955e-01 -5.05369663e-01
-5.61367691e-01 -5.80637515e-01 1.47021174e+00 3.10167521e-01
1.01717746e+00 -2.26653352e-01 -1.96317121e-01 8.19468319e-01
1.34013438e+00 -5.93227148e-01 4.36034411e-01 5.35512567e-01
8.47861111e-01 1.00416534e-01 4.18346733e-01 4.01631624e-01
7.60845482e-01 4.87586826e-01 3.00756603e-01 -4.16131645e-01
-2.43304953e-01 -4.79576707e-01 8.62853885e-01 1.08867967e+00
3.56969029e-01 -2.35852584e-01 -9.27223265e-01 9.75894749e-01
-1.76158106e+00 -1.05680060e+00 4.95806225e-02 2.06124663e+00
1.34201598e+00 5.80430806e-01 2.01187760e-01 1.32015750e-01
3.99913758e-01 3.67063433e-01 -3.11125100e-01 -9.26509798e-01
-2.46374998e-02 8.30348313e-01 8.44735384e-01 8.07642162e-01
-1.10742068e+00 1.71677482e+00 5.94557333e+00 1.13642573e+00
-1.26286423e+00 1.49425119e-01 4.15771991e-01 -8.03510547e-02
-6.66365564e-01 1.25850692e-01 -1.24392056e+00 2.44987980e-01
8.99090052e-01 -3.04060020e-02 3.92075777e-01 8.61279190e-01
2.03509554e-01 1.08348772e-01 -1.12462974e+00 5.95531881e-01
3.33252251e-02 -1.22951519e+00 3.37650508e-01 -2.12522209e-01
5.70932448e-01 4.94034648e-01 1.23469926e-01 6.45027041e-01
8.11673045e-01 -9.29783523e-01 1.15121424e+00 -5.67609333e-02
8.91277015e-01 -7.96304405e-01 4.70127314e-01 1.83487460e-01
-1.55361044e+00 2.00382635e-01 -4.03135538e-01 -1.49695605e-01
2.20626697e-01 6.40763819e-01 -1.04520953e+00 3.76611561e-01
4.14251059e-01 6.52544022e-01 -4.22828704e-01 7.86696970e-01
-6.43960297e-01 1.20842636e+00 -3.36342514e-01 -1.98082387e-01
8.77023876e-01 2.51237810e-01 2.41475061e-01 2.09128594e+00
-8.99197906e-02 -2.86343992e-01 3.25561315e-01 5.17892897e-01
-4.09075707e-01 2.05855832e-01 -5.33794053e-02 6.76793605e-02
6.57798231e-01 1.38798225e+00 -5.51038027e-01 -6.72968149e-01
-5.02167463e-01 1.04333138e+00 7.03729630e-01 2.47489125e-01
-7.78255939e-01 -6.39691412e-01 9.48410630e-01 -2.07363721e-02
7.32712209e-01 -4.93394673e-01 -3.85682702e-01 -1.38233232e+00
1.61203057e-01 -9.62208092e-01 4.04062420e-01 -2.43654266e-01
-1.36260927e+00 6.45950913e-01 -4.60778847e-02 -9.12888587e-01
-3.99358183e-01 -7.18957663e-01 -6.41590297e-01 9.03159559e-01
-1.98052704e+00 -1.37043989e+00 6.87795281e-02 5.42717934e-01
7.49596477e-01 9.93076339e-02 9.11190152e-01 2.19940931e-01
-4.42655802e-01 1.20853245e+00 3.03170264e-01 5.58900893e-01
1.03221333e+00 -1.59602737e+00 1.25224030e+00 1.15323651e+00
4.96672422e-01 9.26407576e-01 5.04549146e-01 -3.59200329e-01
-1.38033855e+00 -1.20847332e+00 1.39807403e+00 -3.63638908e-01
8.83941174e-01 -1.05599821e+00 -8.18134129e-01 7.54188120e-01
3.41001213e-01 2.38460917e-02 5.39073288e-01 5.80431104e-01
-9.14163649e-01 -1.70285836e-01 -6.78420603e-01 7.35191584e-01
9.24182355e-01 -7.65290797e-01 -6.87275469e-01 2.13315234e-01
8.33980322e-01 -3.30235690e-01 -4.92995650e-01 1.91380375e-03
6.58072591e-01 -5.39968729e-01 9.14553583e-01 -7.41709292e-01
4.36048627e-01 -3.90767455e-02 -7.78691545e-02 -1.39867163e+00
-5.58763921e-01 -7.89747596e-01 1.22134097e-01 1.27451873e+00
6.66815758e-01 -5.32626033e-01 8.95331204e-01 1.65288344e-01
-5.10483921e-01 -6.59606934e-01 -9.89128292e-01 -9.86592472e-01
6.39138639e-01 -9.48868692e-01 6.42988324e-01 8.19304228e-01
2.52588481e-01 3.11950266e-01 -2.82188088e-01 -1.09270506e-01
5.75716317e-01 -9.63177755e-02 7.03523993e-01 -4.87269044e-01
-4.10482407e-01 -6.76031530e-01 -1.38821140e-01 -1.52683151e+00
4.25862253e-01 -1.29380548e+00 2.98610300e-01 -1.47694457e+00
1.42976373e-01 -4.89215642e-01 -6.09274626e-01 1.02430296e+00
-5.61198771e-01 2.86770105e-01 1.66457385e-01 -7.24234581e-02
-5.79806089e-01 5.00272453e-01 5.93262076e-01 -3.60314369e-01
2.05452815e-02 -4.72764879e-01 -7.52751827e-01 5.23472071e-01
7.58248866e-01 -6.53143883e-01 -2.06556804e-02 -1.01940238e+00
1.80920005e-01 -7.02770114e-01 7.37618702e-03 -7.01809883e-01
1.08837150e-01 5.42547517e-02 4.13390473e-02 -4.74923104e-01
2.12334245e-01 -3.01481783e-01 -8.22933376e-01 5.04777074e-01
-6.21959388e-01 4.17033464e-01 2.39575580e-01 1.61587700e-01
-1.25710443e-01 -7.45927095e-02 6.21501625e-01 -1.45589665e-01
-6.60478055e-01 3.53598028e-01 -6.23549640e-01 3.87074590e-01
5.32088697e-01 -1.75574005e-01 -5.24777174e-02 4.46613058e-02
-1.80864066e-01 1.67046458e-01 2.76027709e-01 4.13579851e-01
4.01777446e-01 -1.12185705e+00 -1.05571592e+00 3.02152157e-01
3.09000790e-01 -8.34982917e-02 -2.55490929e-01 4.38581288e-01
-3.89470249e-01 4.75499004e-01 2.13831916e-01 -4.25120682e-01
-1.22979188e+00 4.08144593e-01 2.58363150e-02 -7.83591688e-01
-4.80456769e-01 1.14252472e+00 -3.25223953e-02 -4.85290408e-01
3.44898790e-01 -1.04083681e+00 4.25705791e-01 -1.23095863e-01
5.49140573e-01 -7.02021420e-02 4.14024264e-01 -3.91759247e-01
-4.25655276e-01 3.60463142e-01 -6.90019667e-01 -3.19384187e-01
1.35078692e+00 1.49655983e-01 6.79313997e-03 2.67830014e-01
1.21090794e+00 3.30051839e-01 -1.11370909e+00 -6.77035868e-01
1.80149898e-01 -9.48288068e-02 -1.05600525e-02 -7.24206388e-01
-7.92930901e-01 8.71320426e-01 7.22908974e-02 -1.25903338e-01
8.93969893e-01 -1.21902995e-01 1.32536840e+00 5.62276602e-01
1.85226977e-01 -1.13343740e+00 -8.94021522e-03 9.79699254e-01
3.43681127e-01 -1.00654733e+00 -1.80533871e-01 -1.16440348e-01
-3.42228711e-01 1.07510340e+00 3.91591400e-01 -4.13203448e-01
4.91360664e-01 6.58795059e-01 1.90125078e-01 2.84182549e-01
-1.06581652e+00 -2.81496853e-01 1.42218605e-01 3.56443644e-01
8.94451320e-01 2.02396646e-01 -4.07911420e-01 5.31492293e-01
-4.18064982e-01 -2.35112518e-01 1.21385336e-01 9.94368434e-01
-5.86230934e-01 -1.84407723e+00 -2.82419205e-01 2.97284782e-01
-5.63442051e-01 -1.04215670e+00 -1.25546511e-02 6.92395806e-01
1.04864337e-01 6.35415435e-01 8.07733759e-02 -2.39565030e-01
2.68394321e-01 2.44586810e-01 3.73519570e-01 -8.66434693e-01
-1.14365339e+00 -3.13072503e-02 2.04096645e-01 -5.12499213e-01
6.38163239e-02 -7.36704707e-01 -1.43683505e+00 -4.68899131e-01
-1.25898167e-01 2.43891656e-01 6.49672270e-01 1.01147389e+00
2.89927870e-01 3.76788557e-01 4.18273926e-01 -9.29627776e-01
-9.17552590e-01 -1.01542032e+00 -2.75367171e-01 4.92718399e-01
3.31878781e-01 -8.33842382e-02 -5.30000590e-02 1.45982787e-01] | [10.787893295288086, 8.645148277282715] |
b80eff6f-7098-4702-a769-14b4a12e52cd | a-series-of-unfortunate-counterfactual-events | 2010.04687 | null | https://arxiv.org/abs/2010.04687v2 | https://arxiv.org/pdf/2010.04687v2.pdf | A Series of Unfortunate Counterfactual Events: the Role of Time in Counterfactual Explanations | Counterfactual explanations are a prominent example of post-hoc interpretability methods in the explainable Artificial Intelligence research domain. They provide individuals with alternative scenarios and a set of recommendations to achieve a sought-after machine learning model outcome. Recently, the literature has identified desiderata of counterfactual explanations, such as feasibility, actionability and sparsity that should support their applicability in real-world contexts. However, we show that the literature has neglected the problem of the time dependency of counterfactual explanations. We argue that, due to their time dependency and because of the provision of recommendations, even feasible, actionable and sparse counterfactual explanations may not be appropriate in real-world applications. This is due to the possible emergence of what we call "unfortunate counterfactual events." These events may occur due to the retraining of machine learning models whose outcomes have to be explained via counterfactual explanation. Series of unfortunate counterfactual events frustrate the efforts of those individuals who successfully implemented the recommendations of counterfactual explanations. This negatively affects people's trust in the ability of institutions to provide machine learning-supported decisions consistently. We introduce an approach to address the problem of the emergence of unfortunate counterfactual events that makes use of histories of counterfactual explanations. In the final part of the paper we propose an ethical analysis of two distinct strategies to cope with the challenge of unfortunate counterfactual events. We show that they respond to an ethically responsible imperative to preserve the trustworthiness of credit lending organizations, the decision models they employ, and the social-economic function of credit lending. | ['Michele Loi', 'Andrea Ferrario'] | 2020-10-09 | null | null | null | null | ['counterfactual-explanation'] | ['miscellaneous'] | [ 2.23435000e-01 8.61961484e-01 -3.62823784e-01 -2.13522673e-01
-2.35565193e-03 -3.86230797e-01 7.22360373e-01 1.51231363e-01
-3.42202455e-01 1.30353463e+00 6.28038108e-01 -7.77492344e-01
-5.42086363e-01 -4.71921831e-01 -7.07439721e-01 -2.44346127e-01
-6.54325634e-02 3.99760097e-01 -6.96122527e-01 1.36186361e-01
5.98704994e-01 3.28282803e-01 -1.48562181e+00 2.47500628e-01
8.89378786e-01 3.73718917e-01 -3.83946240e-01 3.10468644e-01
9.24321041e-02 8.66538644e-01 -3.78438771e-01 -8.64910245e-01
4.90523040e-01 -4.16734278e-01 -8.31353247e-01 1.41959310e-01
8.53470992e-03 -5.82908094e-01 1.25861943e-01 8.24816465e-01
2.04039618e-01 6.97537661e-02 6.75477028e-01 -1.72422886e+00
-7.50170171e-01 8.19725394e-01 -2.40003303e-01 -3.26451212e-02
4.90207613e-01 3.72251958e-01 1.02391756e+00 -3.57141972e-01
4.51610625e-01 1.25825036e+00 6.12704515e-01 8.67708325e-01
-1.31607878e+00 -6.40517056e-01 3.10126275e-01 3.06058563e-02
-6.47538424e-01 -5.26461244e-01 7.45754063e-01 -5.61001360e-01
8.84372711e-01 7.05330193e-01 8.15968931e-01 1.15756130e+00
3.15494388e-01 1.80728853e-01 1.27217233e+00 -7.14396954e-01
6.03421986e-01 3.84803325e-01 3.93333510e-02 2.77610958e-01
1.03328776e+00 7.52588153e-01 -5.06506026e-01 -8.39776933e-01
4.75894272e-01 2.06204832e-01 -4.00046945e-01 -2.44218364e-01
-9.94507492e-01 1.01334047e+00 2.66117573e-01 2.67354578e-01
-7.35721231e-01 4.73744571e-02 2.73906082e-01 3.30838829e-01
4.27337289e-01 9.61730003e-01 -6.25439465e-01 2.64513176e-02
-6.06931984e-01 3.98849130e-01 8.77377391e-01 1.30173534e-01
3.22360009e-01 2.79665608e-02 2.37249553e-01 1.62850872e-01
4.59257126e-01 1.97608873e-01 6.86108828e-01 -1.21326971e+00
3.70003611e-01 4.88317341e-01 5.97039819e-01 -9.30601358e-01
-1.80205896e-01 -4.86037850e-01 -6.27055585e-01 5.65688550e-01
6.63335085e-01 -2.75417805e-01 -2.75415361e-01 1.54707122e+00
2.35475779e-01 2.51593590e-01 1.62028715e-01 9.16561007e-01
-1.25168875e-01 1.91055816e-02 3.18260044e-01 -7.21824586e-01
8.95501018e-01 -3.36711794e-01 -7.52906203e-01 -2.67056525e-01
8.59100699e-01 -3.32075775e-01 1.02221942e+00 2.74140418e-01
-9.33643997e-01 4.63776514e-02 -9.26475704e-01 5.87237775e-01
9.41603631e-02 -4.30387735e-01 1.18182397e+00 7.79640317e-01
-4.38802391e-01 8.89672458e-01 -3.41268063e-01 -1.60413370e-01
4.65386271e-01 3.06466311e-01 -3.96187901e-01 2.34724790e-01
-1.02210629e+00 9.81331408e-01 1.49761721e-01 1.56361207e-01
-3.44200492e-01 -7.52661705e-01 -5.04101157e-01 3.93345624e-01
4.30786610e-01 -9.68633413e-01 1.04933107e+00 -1.68560934e+00
-8.58521461e-01 6.48324847e-01 5.01711480e-02 -8.73596728e-01
1.01450753e+00 -1.14977211e-01 -2.95157373e-01 -3.60979587e-01
3.99587542e-01 1.20236473e-02 6.21892452e-01 -1.16499662e+00
-5.62499583e-01 -4.85393703e-01 1.20234214e-01 9.57133323e-02
-8.11070502e-02 -1.07079260e-01 1.00847387e+00 -5.96753359e-01
-7.56166503e-02 -8.05462241e-01 -5.98761678e-01 -2.29805976e-01
-4.20247287e-01 -6.33108392e-02 2.52499729e-01 -1.98349774e-01
1.00045788e+00 -1.78305614e+00 -6.30913317e-01 8.71176049e-02
1.58973530e-01 -7.58673474e-02 2.37879664e-01 2.21360177e-01
-5.73989749e-01 7.36910701e-01 2.10085548e-02 -1.01739287e-01
1.89501315e-01 4.19665836e-02 -8.24832797e-01 4.81687456e-01
-4.43954505e-02 5.34587204e-01 -5.91463387e-01 -6.69222772e-02
1.18139662e-01 -8.73410627e-02 -5.40949583e-01 1.35295466e-01
1.48688424e-02 5.99391758e-02 -4.70687598e-01 1.81574732e-01
4.59734023e-01 3.83607335e-02 4.84212816e-01 2.42858306e-01
-1.71356171e-01 4.82457966e-01 -1.16486883e+00 7.45624661e-01
-1.89138919e-01 4.47553486e-01 -3.60658914e-01 -1.06918669e+00
4.53681290e-01 7.64279366e-01 9.89880934e-02 -4.38219130e-01
9.72202122e-02 5.48222542e-01 5.55313341e-02 -6.31481826e-01
2.15320095e-01 -8.99004281e-01 1.42992944e-01 1.27299500e+00
-5.95981956e-01 1.32085979e-01 -3.69100034e-01 9.91083831e-02
7.95239210e-01 -3.63403279e-03 9.34487700e-01 -3.84012043e-01
1.27282977e-01 7.76149286e-03 8.53437126e-01 9.22236741e-01
-2.84230649e-01 5.79504728e-01 5.24839103e-01 -1.12999523e+00
-1.01216888e+00 -7.19399214e-01 1.24147557e-01 4.44463819e-01
-3.99879634e-01 3.11039299e-01 -4.59420413e-01 -1.11989892e+00
3.65739167e-01 1.42099655e+00 -1.02464485e+00 -2.83690602e-01
-3.61527085e-01 -8.36809099e-01 2.00419873e-01 3.43768805e-01
3.20212573e-01 -8.98972392e-01 -1.23605752e+00 2.20361337e-01
-6.45901039e-02 -5.04962564e-01 -1.35565758e-01 -1.67003885e-01
-9.50050175e-01 -1.38342774e+00 -2.77523398e-01 3.28335285e-01
7.40854979e-01 1.75217330e-01 9.73801613e-01 7.10896492e-01
2.43686229e-01 2.75389403e-01 -1.48638487e-01 -8.94763649e-01
-6.13284767e-01 -6.11332238e-01 4.22372133e-01 -6.49298951e-02
4.82649505e-01 -7.85719156e-01 -5.82830131e-01 -2.00936031e-02
-7.52927721e-01 1.36668578e-01 3.41893256e-01 8.73619735e-01
-8.65484178e-02 -5.31894527e-02 1.21245825e+00 -1.17100179e+00
8.62538874e-01 -8.21632028e-01 -3.49622399e-01 4.98911589e-01
-1.31874347e+00 1.53211594e-01 5.96374571e-01 -5.87686121e-01
-1.34922338e+00 -2.54018933e-01 5.48575997e-01 1.60240918e-01
-3.54249388e-01 5.04197598e-01 -1.29938945e-01 3.73162210e-01
9.94252443e-01 -3.78196687e-01 3.89827900e-02 -2.53882676e-01
2.18409941e-01 6.60582483e-01 2.15139478e-01 -5.70249856e-01
7.13730812e-01 6.65517867e-01 -7.81425983e-02 -2.20343798e-01
-8.71889532e-01 1.89095318e-01 -1.01309150e-01 -1.50863722e-01
3.93229753e-01 -4.83120233e-01 -8.07395637e-01 -1.45829707e-01
-1.17449379e+00 -3.96782272e-02 -5.42663515e-01 7.70411909e-01
-6.97907507e-01 3.23757023e-01 1.33551002e-01 -1.35748744e+00
-1.70575008e-01 -6.39566600e-01 5.91097809e-02 2.27562770e-01
-9.44089174e-01 -9.93205369e-01 -8.60890746e-02 5.58003068e-01
2.04539046e-01 5.22679746e-01 9.05614555e-01 -9.87758994e-01
-3.52266580e-01 -2.72341460e-01 1.54669628e-01 -6.57810792e-02
2.53799438e-01 -3.00219618e-02 -9.73572731e-01 -4.43128198e-02
3.59746128e-01 4.67903446e-03 4.78474915e-01 3.68477672e-01
4.59038526e-01 -1.31391132e+00 -3.18143696e-01 1.99188162e-02
1.22015202e+00 3.06866378e-01 3.57120812e-01 4.82546449e-01
4.02624346e-02 1.18269598e+00 6.42279625e-01 6.77102625e-01
3.30888003e-01 4.25820321e-01 4.29542810e-01 6.76728040e-02
5.47249317e-01 -4.16909099e-01 2.02310517e-01 -3.93054307e-01
-3.06275070e-01 1.24295257e-01 -7.80542314e-01 8.06785226e-01
-2.11991978e+00 -1.34672725e+00 -1.72146395e-01 2.47538376e+00
3.38553071e-01 2.02750698e-01 1.49222136e-01 2.70730883e-01
6.40461266e-01 -3.37133557e-01 -4.83615696e-01 -1.02383876e+00
9.80259404e-02 -3.68141353e-01 3.37821692e-01 4.48392779e-01
-4.65976208e-01 1.36119425e-01 6.10920143e+00 1.13346718e-01
-8.59854043e-01 -6.21932326e-03 1.07224417e+00 -3.23401481e-01
-1.00732279e+00 4.83349085e-01 2.50403155e-02 5.87741494e-01
9.23348963e-01 -8.13457370e-01 2.03436852e-01 8.52711380e-01
7.25831687e-01 -1.94761232e-01 -1.48766625e+00 2.63064742e-01
-2.39664733e-01 -1.52485144e+00 6.34355322e-02 3.23870659e-01
7.09792554e-01 -5.91407061e-01 1.07471280e-01 -1.66407362e-01
5.23268998e-01 -1.19885397e+00 1.26808262e+00 4.46087450e-01
3.36479098e-01 -7.09382176e-01 9.13744867e-01 6.32253826e-01
-2.32134119e-01 -6.60027504e-01 -1.84693709e-01 -1.10986364e+00
1.70269574e-03 6.84693694e-01 -1.14499104e+00 3.59155238e-01
2.82713205e-01 9.26576853e-02 -1.88698284e-02 6.81599796e-01
-4.46209967e-01 5.78808427e-01 -5.71263693e-02 4.73428816e-02
6.28250837e-02 6.99380189e-02 4.96656835e-01 5.23567200e-01
3.57581258e-01 4.61059660e-01 -6.45020127e-01 1.06656575e+00
2.08362117e-01 -9.14728045e-02 -1.09642720e+00 5.54552898e-02
6.46275580e-01 5.60010493e-01 -5.95487714e-01 -1.52104750e-01
-2.84406513e-01 5.00388145e-01 1.20297000e-01 4.36884433e-01
-3.27144653e-01 4.50343281e-01 5.52825511e-01 5.20158887e-01
-3.95958036e-01 3.98608059e-01 -8.52869153e-01 -1.28783679e+00
2.17202395e-01 -1.17287707e+00 5.79460263e-01 -5.44189990e-01
-1.24301136e+00 1.10613346e-01 -4.53434661e-02 -8.35724175e-01
-5.23644328e-01 -4.67691310e-02 -1.06550157e+00 7.22102642e-01
-1.17416322e+00 -9.78996098e-01 5.34430325e-01 -2.63036862e-02
3.13528299e-01 -1.53381854e-01 8.29931319e-01 -4.88430530e-01
-2.88160920e-01 4.42526519e-01 -1.83502972e-01 -4.51537400e-01
3.55595648e-01 -1.10014415e+00 1.80682853e-01 1.04331017e+00
1.41033098e-01 8.47639203e-01 1.08763945e+00 -8.58730972e-01
-6.58165276e-01 -6.79202735e-01 1.68905640e+00 -5.78284323e-01
3.77549797e-01 1.76492468e-01 -7.45739520e-01 9.77721035e-01
-1.74078792e-02 -3.41470718e-01 7.61655033e-01 3.14573228e-01
-2.75447994e-01 1.77204490e-01 -1.54187191e+00 8.55593443e-01
7.82311201e-01 -3.07879955e-01 -1.07286680e+00 1.36517227e-01
5.53722799e-01 4.79648739e-01 -1.48299351e-01 8.21648315e-02
8.06943774e-01 -1.35083878e+00 7.02039003e-01 -1.17776799e+00
4.69632834e-01 8.71189833e-02 9.83222723e-02 -1.07955205e+00
-1.75883502e-01 -8.65009785e-01 3.25698376e-01 1.13683295e+00
4.75397438e-01 -1.09094834e+00 6.86993301e-01 1.67949641e+00
4.17923063e-01 -6.04091942e-01 -1.24976015e+00 -4.12845701e-01
7.15152547e-02 -4.14640367e-01 1.18741846e+00 1.58226871e+00
5.25941610e-01 -1.26021564e-01 -4.33587343e-01 -5.76776080e-02
7.57235825e-01 3.69637012e-01 6.18447959e-01 -1.33579397e+00
-3.12881112e-01 -2.85094142e-01 -5.16331196e-02 6.90017734e-03
3.39888841e-01 -3.78345013e-01 -3.56135786e-01 -9.86761987e-01
4.55682993e-01 -4.08760875e-01 -1.52702466e-01 3.69641155e-01
-3.80742699e-01 -4.03140873e-01 5.09999573e-01 2.37801373e-01
-2.30181869e-02 3.09251517e-01 5.71106970e-01 2.37070337e-01
-1.38248578e-01 2.36436471e-01 -1.31691480e+00 1.24786997e+00
8.06329668e-01 -6.19019687e-01 -4.38435704e-01 -4.71045151e-02
5.75977027e-01 2.24395052e-01 8.41401815e-01 -3.36936861e-01
-3.25302780e-02 -9.07774866e-01 1.11886702e-01 3.31598371e-01
-2.23133266e-01 -1.02834761e+00 7.91990399e-01 9.66291726e-01
-6.55392230e-01 -1.25808999e-01 -2.62588859e-01 5.17723024e-01
2.23731250e-01 -3.69092941e-01 3.60785276e-01 -3.23983312e-01
-7.35760434e-04 -2.66702592e-01 -5.34537077e-01 -3.36075187e-01
1.18020713e+00 -5.19991100e-01 -3.77444118e-01 -6.05450213e-01
-6.21397436e-01 -2.67433822e-02 7.14367926e-01 1.92677096e-01
6.08835936e-01 -1.07233572e+00 -8.46867740e-01 -9.07965600e-02
-3.10334593e-01 -6.40028536e-01 1.45551771e-01 6.12756193e-01
-8.70975032e-02 5.52614927e-01 -3.38752210e-01 5.32789528e-01
-9.28432345e-01 6.96926117e-01 5.10135949e-01 -2.68822819e-01
-3.58069688e-01 5.04730880e-01 4.59298193e-01 -5.90514205e-02
-1.19968578e-01 -3.22974175e-02 -7.83820376e-02 -7.10799545e-02
3.94530892e-01 5.74764609e-01 -2.30261400e-01 -2.86791503e-01
-3.52694333e-01 -2.07792565e-01 -3.71788442e-02 -2.01596588e-01
1.49418211e+00 -2.91668862e-01 -3.07703316e-02 1.95040181e-01
4.34916884e-01 2.27316190e-02 -1.17347586e+00 2.22142711e-01
2.25543901e-01 -9.24527168e-01 -3.71461809e-01 -1.06789851e+00
-5.44540107e-01 5.76206326e-01 1.72349006e-01 4.69079554e-01
9.55041289e-01 -2.80523986e-01 1.50213286e-01 6.09212853e-02
3.74795914e-01 -1.04705477e+00 -3.73522371e-01 -4.63185370e-01
1.17086029e+00 -1.19113111e+00 2.54208207e-01 -2.39174172e-01
-5.99646032e-01 8.94010842e-01 3.93517882e-01 2.37076834e-01
3.33179563e-01 -2.24933952e-01 5.65658733e-02 -7.50585422e-02
-1.08378971e+00 3.66819769e-01 4.31672018e-03 6.95318162e-01
4.20940667e-01 5.61480522e-01 -9.82061803e-01 1.12496364e+00
-3.83422554e-01 1.42246380e-01 9.75504220e-01 4.68338370e-01
-1.23153172e-01 -8.52612197e-01 -6.11801803e-01 5.97453952e-01
-8.02905560e-01 7.89058656e-02 -6.12093091e-01 8.53441656e-01
3.12256366e-01 1.13013685e+00 -3.37799117e-02 5.32726608e-02
1.76422521e-01 -1.01324366e-02 -1.55578122e-01 -4.79653627e-01
-6.65650547e-01 -3.97684038e-01 5.09352207e-01 -4.62043345e-01
-3.96561831e-01 -9.48582530e-01 -1.15799880e+00 -5.54797888e-01
-4.49091107e-01 6.07153654e-01 1.84253141e-01 1.38415003e+00
4.80793655e-01 -1.97728932e-01 6.76652968e-01 -2.89973855e-01
-1.03830254e+00 -5.70618629e-01 -5.38512647e-01 6.94122434e-01
5.14914870e-01 -5.36724806e-01 -7.71214366e-01 -1.17258161e-01] | [8.739544868469238, 5.679496765136719] |
9d21acdf-e337-480c-9138-618e4ce00941 | improving-eeg-based-emotion-recognition-by | 2303.11421 | null | https://arxiv.org/abs/2303.11421v1 | https://arxiv.org/pdf/2303.11421v1.pdf | Improving EEG-based Emotion Recognition by Fusing Time-frequency And Spatial Representations | Using deep learning methods to classify EEG signals can accurately identify people's emotions. However, existing studies have rarely considered the application of the information in another domain's representations to feature selection in the time-frequency domain. We propose a classification network of EEG signals based on the cross-domain feature fusion method, which makes the network more focused on the features most related to brain activities and thinking changes by using the multi-domain attention mechanism. In addition, we propose a two-step fusion method and apply these methods to the EEG emotion recognition network. Experimental results show that our proposed network, which combines multiple representations in the time-frequency domain and spatial domain, outperforms previous methods on public datasets and achieves state-of-the-art at present. | ['Jing Xiao', 'Ning Cheng', 'Jianzong Wang', 'xulong Zhang', 'Kexin Zhu'] | 2023-03-14 | null | null | null | null | ['eeg-emotion-recognition'] | ['miscellaneous'] | [-8.33322108e-02 -8.11995506e-01 3.08805048e-01 -6.01903796e-01
-3.94316733e-01 3.25198434e-02 1.95168525e-01 1.36268297e-02
-4.86076057e-01 9.96607006e-01 3.19138914e-01 3.72492880e-01
-6.70520782e-01 -7.39893436e-01 -9.62791070e-02 -6.71581686e-01
-2.18642130e-01 -3.49294782e-01 -1.77753493e-01 -2.34121323e-01
5.06628752e-01 3.82597208e-01 -1.49650419e+00 5.93909204e-01
1.10618818e+00 1.66783798e+00 -1.74827293e-01 4.42555137e-02
4.74207699e-02 5.26732266e-01 -9.39739823e-01 -1.03627190e-01
-1.44306153e-01 -5.47926009e-01 -5.57114065e-01 -4.32762504e-01
-3.02033365e-01 9.65749547e-02 -3.25692266e-01 1.18387604e+00
9.55152988e-01 3.92115355e-01 6.93438351e-01 -1.41754830e+00
-9.41158175e-01 3.35071564e-01 -5.95371366e-01 8.12621653e-01
5.52181721e-01 -3.04457217e-01 4.63492453e-01 -8.68407607e-01
-1.48259765e-02 9.26538169e-01 6.86499774e-01 1.77826568e-01
-7.84559727e-01 -1.31224322e+00 3.61613184e-01 8.30169082e-01
-1.66219735e+00 -1.86072454e-01 1.22652805e+00 -3.55492413e-01
1.17652559e+00 8.66164565e-02 8.61734092e-01 1.30582738e+00
8.74232769e-01 5.86604774e-01 1.46654582e+00 -8.40808749e-02
1.20127656e-01 1.29521087e-01 5.43622255e-01 2.70170182e-01
-6.05563745e-02 -1.34428024e-01 -8.89983118e-01 -2.80594975e-01
4.08461004e-01 3.00013483e-01 -4.22225356e-01 4.95411009e-01
-1.18019140e+00 6.59703791e-01 5.04186034e-01 8.15007806e-01
-1.03981483e+00 -1.03774779e-01 5.13463080e-01 4.61901516e-01
9.29334939e-01 5.91655552e-01 -6.78898990e-01 -1.51409179e-01
-8.81754816e-01 -4.91787344e-02 5.06210744e-01 3.51142526e-01
5.98604977e-01 5.83815575e-02 -3.92723948e-01 9.15577710e-01
1.98085740e-01 2.90558934e-01 9.72891986e-01 -3.81210476e-01
5.11074625e-03 7.72806942e-01 -3.80267173e-01 -1.36092389e+00
-7.66915560e-01 -5.81637681e-01 -1.13784313e+00 -1.03837229e-01
-4.35295045e-01 -6.80191338e-01 -5.37157178e-01 1.69693899e+00
-1.05013009e-02 6.03783727e-01 1.96624741e-01 9.37906861e-01
1.17663860e+00 5.72538853e-01 1.24261327e-01 -2.86172301e-01
1.63541996e+00 -7.08480716e-01 -1.00984418e+00 -2.41568666e-02
3.13038118e-02 -2.77386963e-01 4.31000471e-01 7.94329464e-01
-8.53340149e-01 -7.30941474e-01 -1.16520846e+00 4.29569185e-01
-8.42995465e-01 2.13314816e-01 8.65407646e-01 5.74786425e-01
-1.02954125e+00 5.63143849e-01 -4.33123946e-01 -2.47696832e-01
6.52103007e-01 6.68295801e-01 -4.77319807e-01 2.97679037e-01
-1.88607180e+00 1.03233302e+00 3.56483996e-01 1.84497144e-02
-1.85127765e-01 -5.11188030e-01 -6.71522737e-01 3.27978194e-01
-2.69421965e-01 -5.46077132e-01 5.89939833e-01 -1.45610952e+00
-1.58501315e+00 1.36155799e-01 -2.48461902e-01 -1.41954139e-01
-2.42420748e-01 -6.32594451e-02 -1.19917452e+00 4.46443558e-01
-7.10593760e-02 3.98545831e-01 7.04081178e-01 -4.59503084e-01
-4.85056370e-01 -6.66978538e-01 -3.61065865e-01 1.89510494e-01
-8.31706703e-01 3.19062799e-01 1.26476288e-01 -6.64777458e-01
7.60984793e-02 -2.73223966e-01 1.02208748e-01 -3.43395233e-01
-4.92661148e-02 -6.56103134e-01 7.51890898e-01 -6.63248837e-01
1.23076260e+00 -2.25959849e+00 1.19084485e-01 5.72627246e-01
3.30118746e-01 -1.13222547e-01 -9.12866890e-02 8.43302235e-02
-4.66479361e-01 -7.46100545e-02 8.25237557e-02 9.71793383e-02
-1.18326480e-02 -3.78801703e-01 -8.17065239e-02 5.65804243e-01
4.46118504e-01 8.18368733e-01 -6.69391215e-01 -2.92999417e-01
1.69220105e-01 7.77574718e-01 -2.01680586e-01 3.68598700e-02
8.60576034e-01 4.15205598e-01 -6.65290058e-01 6.36994898e-01
8.66912246e-01 -1.36575883e-03 -2.18124494e-01 -5.22511721e-01
-9.14220586e-02 3.08407485e-01 -1.08780205e+00 1.72936249e+00
-2.38123134e-01 7.13339031e-01 -3.67953539e-01 -1.30228424e+00
1.14385033e+00 5.01364350e-01 8.27006340e-01 -9.95308518e-01
6.28551841e-01 -3.70665230e-02 1.55750871e-01 -6.82130694e-01
-6.17343262e-02 -6.61807135e-02 -1.30873471e-01 3.24045509e-01
6.01178229e-01 4.07414198e-01 -2.18687862e-01 -2.33797118e-01
1.04836357e+00 -2.34371141e-01 4.25395072e-01 -5.23962677e-01
5.92792034e-01 -6.27573311e-01 6.97970748e-01 4.90822285e-01
-4.59176421e-01 7.20645264e-02 2.78115869e-01 -5.16748905e-01
-1.25892758e-01 -6.94379389e-01 -4.88016725e-01 1.10561085e+00
1.73514649e-01 -2.70829290e-01 -6.20251238e-01 -5.50523937e-01
-3.09076495e-02 5.87616146e-01 -9.03596759e-01 -8.66360605e-01
1.66348130e-01 -1.19708538e+00 5.38508654e-01 7.63174891e-01
9.48876858e-01 -1.22971904e+00 -7.11711347e-01 2.92514652e-01
-2.35252976e-01 -6.11075521e-01 -1.20134883e-01 4.02947336e-01
-4.54584211e-01 -7.83194959e-01 -7.52598107e-01 -5.36823392e-01
3.99671823e-01 5.75645454e-02 7.83327937e-01 -9.93769318e-02
-1.32709160e-01 6.05809033e-01 -6.35876417e-01 -7.11096287e-01
7.80446529e-01 -5.80455698e-02 2.92286485e-01 4.24154043e-01
1.08459663e+00 -8.88412893e-01 -8.38899672e-01 7.75305107e-02
-5.73095441e-01 -4.31067854e-01 6.21706486e-01 8.45802307e-01
1.87759608e-01 4.58517164e-01 1.29073715e+00 -1.17299415e-01
1.32833505e+00 -8.31678629e-01 2.42250487e-01 2.90424526e-01
-4.45020020e-01 -2.82077521e-01 4.95880336e-01 -7.43518353e-01
-1.18610644e+00 -1.98398739e-01 -4.56388295e-02 -2.83166915e-01
-3.71364534e-01 6.79476202e-01 -5.37486263e-02 -4.53381389e-01
4.29265261e-01 4.66058344e-01 -3.43951136e-01 -1.69441879e-01
-8.81454349e-02 8.78857255e-01 2.82208055e-01 -4.25661087e-01
-1.05699159e-01 1.50401339e-01 -3.75925779e-01 -5.56708097e-01
-4.04249281e-01 -3.38770449e-01 -5.07943511e-01 -3.54970545e-01
9.79990304e-01 -9.87533987e-01 -9.96300340e-01 5.78122973e-01
-1.17922175e+00 5.00242412e-01 1.20917976e-01 1.04349029e+00
-3.13911974e-01 6.13250621e-02 -5.24093151e-01 -9.62984383e-01
-7.15363681e-01 -1.02629757e+00 8.93943191e-01 5.44955790e-01
-3.20436180e-01 -7.08589613e-01 1.84828360e-02 -2.66382843e-01
7.10140109e-01 5.57407327e-02 6.94501162e-01 -9.84390080e-01
2.55759567e-01 -8.40583220e-02 -2.89551169e-01 2.23753780e-01
2.17596158e-01 -4.78756189e-01 -1.18280375e+00 2.03761850e-02
5.78007162e-01 -2.78010935e-01 1.05521810e+00 4.84839588e-01
1.67135084e+00 -5.17565175e-04 -4.01110947e-01 5.97543299e-01
1.20346773e+00 7.36775160e-01 8.41486394e-01 2.08792761e-01
1.78533673e-01 4.38552499e-01 2.83409115e-02 7.35381365e-01
4.44889963e-01 3.31642449e-01 -2.41204165e-03 -1.79416791e-01
4.33539331e-01 6.30218506e-01 2.71991491e-01 8.52208972e-01
-3.32725406e-01 -9.39987600e-02 -5.89346111e-01 4.11421090e-01
-1.77717268e+00 -1.15607548e+00 2.23507032e-01 1.59477615e+00
5.51080108e-01 -2.06574783e-01 -6.34342134e-02 2.28073478e-01
7.32814014e-01 -1.93019897e-01 -7.16897547e-01 -4.20789510e-01
-3.10161859e-01 7.28491843e-01 -4.96020801e-02 -3.42072099e-01
-1.28855276e+00 3.19210708e-01 6.75448132e+00 7.22919285e-01
-1.33372724e+00 3.84444326e-01 6.24323487e-01 -1.87092885e-01
7.64700472e-02 -7.27241755e-01 -2.82233238e-01 8.02870691e-01
1.11113501e+00 -3.17775846e-01 6.43796802e-01 5.46697319e-01
2.91095152e-02 -5.83759099e-02 -8.10094059e-01 1.54703116e+00
5.57657003e-01 -8.61245632e-01 -3.21820647e-01 -2.03820333e-01
3.94433111e-01 -3.44717279e-02 6.24702089e-02 5.54375827e-01
-1.20988384e-01 -1.05188191e+00 3.90417069e-01 1.20775390e+00
4.06338304e-01 -1.04585004e+00 9.69526768e-01 6.02825396e-02
-1.29747045e+00 -4.08041805e-01 -2.83506364e-01 -3.07988435e-01
-1.64096892e-01 7.28577793e-01 -4.29044366e-02 7.70034790e-01
1.20753944e+00 1.09508908e+00 -5.33662021e-01 1.09319603e+00
9.71356630e-02 4.68447775e-01 -1.05907314e-01 -3.40090662e-01
-6.27231821e-02 -1.26137525e-01 6.13233484e-02 1.30397236e+00
8.47139657e-01 5.52511394e-01 -5.17872442e-03 9.56452847e-01
-5.01240380e-02 2.62520403e-01 -5.31730175e-01 9.71109718e-02
3.60759079e-01 1.41304493e+00 -7.23595440e-01 -5.45595646e-01
-4.79440302e-01 1.14132166e+00 1.52554080e-01 3.49633127e-01
-1.02111149e+00 -1.12294865e+00 7.59116352e-01 -8.10230434e-01
1.31732021e-02 1.66444927e-01 -3.08794022e-01 -1.22154653e+00
-5.20567223e-03 -6.69033825e-01 3.71111155e-01 -1.13546193e+00
-1.92111492e+00 1.09343338e+00 -4.93931659e-02 -1.06250572e+00
2.11684376e-01 -6.60301387e-01 -8.71105492e-01 1.06961119e+00
-1.51478267e+00 -6.63461089e-01 -3.20321947e-01 1.21524453e+00
2.73084283e-01 -3.40956450e-01 1.07361221e+00 6.27501309e-01
-5.26396453e-01 4.73033190e-01 -7.46964710e-03 1.20262951e-01
8.03953171e-01 -8.56489003e-01 -3.62249285e-01 4.62887913e-01
-4.06276286e-02 6.58328831e-01 1.55914547e-02 -4.13845122e-01
-1.05532956e+00 -9.11402524e-01 6.23365223e-01 -2.98077092e-02
3.15359563e-01 -2.40152806e-01 -1.00176525e+00 2.65960187e-01
9.12081242e-01 -1.64312392e-01 1.22580624e+00 2.66135067e-01
-3.41395214e-02 -3.83915305e-01 -1.41970789e+00 8.29919428e-02
6.29240036e-01 -5.56074440e-01 -1.01138568e+00 1.99837551e-01
2.80263007e-01 4.49428447e-02 -1.13499808e+00 4.91507083e-01
6.25535786e-01 -7.04860091e-01 8.69719684e-01 -6.05183542e-01
1.62378788e-01 -1.08950615e-01 -2.73518056e-01 -1.97732508e+00
-9.14074361e-01 -4.49999459e-02 9.56426486e-02 1.10361516e+00
1.02526277e-01 -1.00217748e+00 -8.96949396e-02 2.58277655e-01
-1.01548187e-01 -8.49462807e-01 -1.07616472e+00 -4.89797652e-01
-2.59699315e-01 -4.48828340e-01 1.02529740e+00 1.30933297e+00
6.77159131e-01 3.76916379e-01 -2.87915736e-01 1.04798973e-02
4.55303229e-02 3.47822011e-02 -1.66279301e-01 -1.73706460e+00
2.60054499e-01 -6.97809696e-01 -7.65412331e-01 -2.98505396e-01
5.75989783e-01 -8.53374004e-01 5.11567900e-03 -1.43942356e+00
4.67233062e-01 2.30619133e-01 -1.40116894e+00 7.00416028e-01
-1.86988562e-01 3.15410644e-01 -8.19174722e-02 -2.38430902e-01
-7.77215660e-01 9.76036549e-01 8.56629610e-01 -2.15174198e-01
-2.28703871e-01 -4.09292638e-01 -1.21715105e+00 7.55482197e-01
8.95261884e-01 -3.69065285e-01 -4.44180697e-01 -1.90875813e-01
5.89474244e-03 -2.68511593e-01 1.73358321e-01 -1.53286517e+00
4.39094841e-01 -5.97968996e-02 1.40127707e+00 -4.60266173e-01
5.98082066e-01 -8.03869128e-01 -4.83680516e-02 1.79398343e-01
-2.23466471e-01 3.63282651e-01 3.73511404e-01 3.98266643e-01
-4.10066545e-01 2.85452574e-01 4.94132608e-01 -4.58900183e-02
-7.79063046e-01 2.32311949e-01 -6.52589381e-01 -3.16742033e-01
1.15570867e+00 -2.05315471e-01 -4.79117036e-01 -2.91663051e-01
-7.40325511e-01 1.73740715e-01 -4.73695368e-01 5.85355759e-01
1.00828552e+00 -1.75833750e+00 -5.27224779e-01 5.98613322e-01
1.29195815e-03 -1.00391209e+00 5.32290399e-01 1.17397165e+00
4.00863051e-01 2.85911262e-01 -9.45761025e-01 -2.96969175e-01
-9.38186765e-01 3.71011347e-01 5.55597007e-01 1.15130790e-01
-3.05804968e-01 8.08309436e-01 1.34812608e-01 -1.65468782e-01
-1.91586521e-02 -2.52066612e-01 -9.34671998e-01 4.72424001e-01
9.14753556e-01 3.40134323e-01 3.58236253e-01 -7.55039990e-01
-9.10956562e-01 5.53848743e-01 1.86084554e-01 3.97525914e-02
1.67243600e+00 8.96857381e-02 -4.13105160e-01 4.91541654e-01
1.30217016e+00 -5.40802896e-01 -6.33875608e-01 -1.21831968e-01
-3.13537300e-01 -2.30564713e-01 4.84796226e-01 -1.20431101e+00
-1.48071778e+00 9.25420642e-01 1.14562190e+00 9.11656246e-02
1.85426939e+00 -2.67652512e-01 5.84859133e-01 5.28425515e-01
3.88973147e-01 -1.34381247e+00 1.71899036e-01 5.02088130e-01
9.13036346e-01 -9.76690173e-01 -1.56742111e-02 1.81839764e-01
-7.72957385e-01 1.35048449e+00 8.15961123e-01 -3.63398045e-01
1.27011323e+00 1.07151896e-01 -2.09228992e-01 -4.17370766e-01
-8.62043977e-01 2.17562784e-02 7.35910773e-01 5.09278119e-01
5.23120880e-01 7.38972127e-02 -6.12355113e-01 1.71352267e+00
-7.54105672e-02 2.88869470e-01 -1.02254981e-02 7.34558403e-01
-4.75211293e-01 -7.76721358e-01 -5.00522256e-01 9.26932275e-01
-6.94310546e-01 -2.04484582e-01 -4.86604095e-01 3.40531498e-01
6.37066007e-01 1.15428627e+00 2.02282593e-01 -8.81910443e-01
3.74014080e-01 6.02063715e-01 4.22989547e-01 -1.45051360e-01
-8.63333046e-01 -7.01783516e-04 -3.07151884e-01 -6.53111756e-01
-7.91489184e-01 -5.13831317e-01 -1.10658264e+00 4.61510830e-02
-4.06913519e-01 2.66833067e-01 5.97339332e-01 1.12986219e+00
9.01655376e-01 1.13670218e+00 6.36633158e-01 -1.00733340e+00
6.80160522e-02 -1.33531833e+00 -8.75556529e-01 3.78327996e-01
1.91219077e-01 -1.19843376e+00 -2.11876541e-01 -3.64503056e-01] | [13.16024398803711, 3.4683995246887207] |
92440b6a-d83e-4462-b5f9-5775bb6a90ff | transductive-matrix-completion-with | 2302.09834 | null | https://arxiv.org/abs/2302.09834v1 | https://arxiv.org/pdf/2302.09834v1.pdf | Transductive Matrix Completion with Calibration for Multi-Task Learning | Multi-task learning has attracted much attention due to growing multi-purpose research with multiple related data sources. Moreover, transduction with matrix completion is a useful method in multi-label learning. In this paper, we propose a transductive matrix completion algorithm that incorporates a calibration constraint for the features under the multi-task learning framework. The proposed algorithm recovers the incomplete feature matrix and target matrix simultaneously. Fortunately, the calibration information improves the completion results. In particular, we provide a statistical guarantee for the proposed algorithm, and the theoretical improvement induced by calibration information is also studied. Moreover, the proposed algorithm enjoys a sub-linear convergence rate. Several synthetic data experiments are conducted, which show the proposed algorithm out-performs other existing methods, especially when the target matrix is associated with the feature matrix in a nonlinear way. | ['Zhonglei Wang', 'Xiaojun Mao', 'Yasi Zhang', 'Hengfang Wang'] | 2023-02-20 | null | null | null | null | ['matrix-completion', 'multi-label-learning'] | ['methodology', 'methodology'] | [ 2.41224855e-01 -3.25940043e-01 -2.26246595e-01 -3.50595504e-01
-1.25443304e+00 -3.17684054e-01 1.67973995e-01 6.21551499e-02
-3.82835120e-01 6.21234715e-01 2.17039645e-01 3.06350231e-01
-3.38934451e-01 -2.29856521e-01 -6.23792827e-01 -9.85842228e-01
4.04665291e-01 2.13363051e-01 -3.71076643e-01 4.88044918e-02
-6.26440123e-02 -1.78572923e-01 -9.24222052e-01 1.78233936e-01
9.04253066e-01 7.91099787e-01 2.99015671e-01 1.99705094e-01
1.66063800e-01 6.56840444e-01 -6.15762547e-02 -3.44747424e-01
2.98694760e-01 -8.08579661e-03 -5.81510782e-01 5.51366210e-01
2.81282872e-01 -3.09132896e-02 -2.38003537e-01 1.17818463e+00
4.19321090e-01 1.78528264e-01 6.39999926e-01 -1.44860113e+00
-5.46317816e-01 3.26471925e-01 -1.14873838e+00 -2.57224619e-01
9.66219679e-02 -4.95301902e-01 1.17768919e+00 -1.44511724e+00
2.01770082e-01 1.31576860e+00 5.11984468e-01 2.72775412e-01
-1.28173566e+00 -8.70020092e-01 1.17609188e-01 1.55719668e-01
-1.59042633e+00 -3.71044457e-01 9.19672072e-01 -5.03633976e-01
3.72144394e-02 -2.62277946e-03 1.35301605e-01 7.60004520e-01
9.47214216e-02 1.14727056e+00 1.24983156e+00 -4.28529352e-01
-7.47756734e-02 3.33938330e-01 1.55363142e-01 8.68090272e-01
3.67940873e-01 -2.55983561e-01 -4.87208784e-01 -4.20749366e-01
6.01097882e-01 5.73627114e-01 -1.64732710e-01 -5.30193508e-01
-1.38810444e+00 9.05167818e-01 4.98465635e-02 3.33973289e-01
-4.43278491e-01 8.26945081e-02 4.65076327e-01 4.33800161e-01
7.47403622e-01 -3.74965847e-01 -3.50345194e-01 2.20327854e-01
-7.01528251e-01 5.22347987e-02 5.57568252e-01 1.23618054e+00
1.19837666e+00 -1.15984986e-02 -1.10578112e-01 1.15145910e+00
5.60252964e-01 7.13422179e-01 3.55373174e-01 -7.54060626e-01
7.59991646e-01 4.68113393e-01 8.04872662e-02 -1.26243913e+00
-6.36969388e-01 -5.54147184e-01 -1.23127270e+00 -5.06862819e-01
4.45824653e-01 -4.32267994e-01 -5.65438457e-02 1.88649750e+00
5.84248185e-01 4.88857001e-01 2.23038062e-01 8.39903831e-01
4.08306420e-01 6.57749414e-01 -7.63603225e-02 -7.71048963e-01
1.48311388e+00 -1.05283046e+00 -1.04686272e+00 -3.73036981e-01
7.36712635e-01 -9.42775130e-01 9.43892598e-01 4.19917852e-01
-5.30258834e-01 -5.56074560e-01 -8.75627756e-01 5.12993783e-02
3.62847358e-01 5.84616184e-01 8.09455276e-01 6.14227772e-01
-5.35299599e-01 1.36090025e-01 -5.07721066e-01 -1.52681023e-01
2.61966050e-01 4.34071451e-01 -6.43122017e-01 -6.41948104e-01
-9.16647077e-01 4.08406019e-01 3.90464544e-01 1.84856772e-01
-7.24222064e-01 -4.60126698e-01 -7.55352914e-01 -2.54878372e-01
6.19367957e-01 -5.56566119e-01 1.06721163e+00 -6.66308880e-01
-1.20029163e+00 6.14539385e-01 -3.29446733e-01 2.63571441e-01
4.66786653e-01 -3.95367742e-01 -3.37900549e-01 1.67188421e-01
3.25852931e-01 1.08647250e-01 1.26094818e+00 -1.19085777e+00
-7.29246318e-01 -6.09583676e-01 -1.60240039e-01 4.16790485e-01
-8.18255067e-01 -1.04312912e-01 -5.34487545e-01 -7.34566271e-01
3.16441655e-01 -1.13525891e+00 -2.66986966e-01 -8.44716504e-02
-2.87820190e-01 -4.13897008e-01 6.22158945e-01 -4.75436389e-01
1.01579702e+00 -2.38727665e+00 3.27232510e-01 1.50977656e-01
5.06086826e-01 -1.19195648e-01 -2.59421706e-01 4.94607687e-01
-1.25102907e-01 -2.70775199e-01 -3.21005493e-01 -6.56480849e-01
-7.74476975e-02 2.42098887e-02 -1.57826975e-01 8.89406860e-01
-1.73159882e-01 6.60446465e-01 -9.38039899e-01 -6.84950471e-01
1.01869144e-02 4.15971875e-01 -2.10439458e-01 1.72812477e-01
2.88568079e-01 5.76082051e-01 -8.16117465e-01 6.14005744e-01
9.67631876e-01 -5.01070738e-01 2.80617118e-01 -5.37462473e-01
2.03761220e-01 -5.49913526e-01 -1.50699389e+00 2.09857059e+00
-4.86210078e-01 1.23930022e-01 3.13282609e-01 -1.21016407e+00
8.68684828e-01 5.85289538e-01 8.74801278e-01 -3.02992135e-01
1.44338325e-01 3.94179136e-01 -2.77188301e-01 -4.80650932e-01
3.21764588e-01 -5.12340665e-01 -1.21703885e-01 6.23898327e-01
-8.25498030e-02 3.55581135e-01 1.77972689e-01 3.16775233e-01
7.58149385e-01 2.70303953e-02 4.62445170e-01 -3.46578419e-01
8.15334976e-01 -4.63712841e-01 9.78623569e-01 4.02055323e-01
-1.69492796e-01 2.23376185e-01 1.19774789e-01 -1.27171993e-01
-8.50576878e-01 -6.38804376e-01 -3.43273468e-02 1.21561813e+00
2.07673326e-01 -4.34419662e-01 -4.71571684e-01 -6.99473798e-01
-1.76708512e-02 1.76060483e-01 -5.03302813e-01 -8.99109393e-02
-3.33998561e-01 -1.18795347e+00 3.60644370e-01 2.16662556e-01
4.17644769e-01 -3.32648873e-01 2.61630982e-01 3.02833378e-01
-6.57784462e-01 -1.16600633e+00 -8.66229594e-01 -1.16765685e-01
-9.85308826e-01 -1.18168700e+00 -9.12265658e-01 -1.00433803e+00
8.89232397e-01 1.02192485e+00 3.15630466e-01 -1.37138143e-01
2.19031200e-01 4.78015959e-01 -3.18158656e-01 -8.48794430e-02
-1.53941974e-01 3.73533294e-02 3.15443754e-01 7.89253175e-01
3.22854131e-01 -5.97141504e-01 -2.87483215e-01 4.41140890e-01
-1.25174499e+00 3.45190227e-01 9.68553305e-01 1.00232744e+00
7.58471549e-01 8.41249973e-02 9.57944632e-01 -1.10447466e+00
6.79742336e-01 -6.57790959e-01 -4.18932676e-01 5.40788949e-01
-8.31578195e-01 2.36394390e-01 5.95096588e-01 -5.14454842e-01
-1.15923965e+00 4.80666846e-01 3.82418245e-01 -5.77519000e-01
2.71532506e-01 6.92205667e-01 -5.35442472e-01 -1.10323012e-01
3.55415404e-01 3.36798966e-01 5.65951839e-02 -6.96931303e-01
4.17774230e-01 6.69240475e-01 1.59921184e-01 -8.32924426e-01
1.17200673e+00 6.41754866e-01 2.18337670e-01 -4.87739801e-01
-1.11115074e+00 -8.84165287e-01 -7.22550035e-01 -2.66004264e-01
3.59413028e-01 -1.29917717e+00 -9.04263973e-01 5.32608688e-01
-1.07349741e+00 3.55221272e-01 3.20428073e-01 8.97789240e-01
-4.11981374e-01 7.91916549e-01 -7.43195295e-01 -8.62565279e-01
-2.09532171e-01 -1.09955573e+00 8.88172686e-01 -1.09616250e-01
4.30472046e-01 -1.15658379e+00 2.41626665e-01 7.18732595e-01
-3.39656062e-02 5.99072017e-02 6.89905345e-01 -5.93876719e-01
-3.97254676e-01 -2.46018961e-01 -5.27544379e-01 1.56266183e-01
4.80863363e-01 -5.92600584e-01 -8.91228616e-01 -6.80187106e-01
3.62271309e-01 -5.30940711e-01 8.03891361e-01 1.76379159e-01
8.47193599e-01 -3.31814200e-01 -2.91291386e-01 3.90806526e-01
1.49557102e+00 -1.91709325e-01 8.89433101e-02 9.42898020e-02
1.11131430e+00 7.11693227e-01 9.50112462e-01 8.75413954e-01
5.67059517e-01 5.33727407e-01 2.08704859e-01 -7.55712315e-02
3.25687289e-01 -1.02885619e-01 3.37571263e-01 1.53794384e+00
2.03508943e-01 1.38143823e-01 -5.57690263e-01 3.19587678e-01
-2.37389517e+00 -7.19573021e-01 -5.06589234e-01 2.17201066e+00
1.05654681e+00 -4.04927760e-01 2.96797138e-02 1.63018286e-01
9.90723670e-01 9.06330571e-02 -7.09424198e-01 3.42165887e-01
-2.53464401e-01 -2.90526867e-01 4.30556685e-01 4.76637036e-01
-1.22965693e+00 6.95335507e-01 5.74574423e+00 1.14878201e+00
-6.58073604e-01 4.52493191e-01 2.58821577e-01 4.20159958e-02
-3.49066764e-01 1.83839321e-01 -6.74229741e-01 2.29817867e-01
4.38726813e-01 -4.34592277e-01 4.77785885e-01 7.99840748e-01
2.21494034e-01 1.19168684e-01 -9.86472189e-01 1.44985664e+00
3.79422933e-01 -7.88109779e-01 -3.88811305e-02 1.94227874e-01
6.86538279e-01 -3.55411828e-01 5.16223572e-02 2.29978025e-01
1.21675685e-01 -4.90431130e-01 4.75822508e-01 4.64979887e-01
9.40968812e-01 -8.11001480e-01 5.99133790e-01 5.63374341e-01
-1.41998315e+00 -1.03704765e-01 -6.21632338e-01 -1.16242401e-01
1.60171896e-01 1.06915152e+00 -6.14333987e-01 1.06206620e+00
1.74752006e-03 1.08709562e+00 -6.56261444e-01 7.84047604e-01
-1.14549184e-02 5.64239860e-01 -3.12011428e-02 2.78068632e-01
-2.94627845e-02 -8.16494584e-01 3.74928653e-01 9.74874020e-01
1.15867823e-01 -1.11398548e-01 6.98961556e-01 3.16699415e-01
-2.58049071e-01 8.21245015e-01 -4.91318733e-01 3.35428976e-02
3.49295616e-01 1.72153473e+00 -4.13024843e-01 -2.14165986e-01
-8.03018272e-01 1.01264441e+00 5.03657937e-01 3.85060906e-01
-7.24238992e-01 -2.36309811e-01 2.36428678e-01 -5.38431346e-01
-1.33790940e-01 -1.86146587e-01 -5.37361093e-02 -1.49053681e+00
2.17688411e-01 -8.30047309e-01 4.65809673e-01 -3.79500747e-01
-1.50504220e+00 1.44385338e-01 -2.27235526e-01 -1.42422962e+00
8.11922848e-02 -1.86479047e-01 -2.28009801e-02 5.53013682e-01
-1.53072166e+00 -1.52066469e+00 -2.43981421e-01 9.48617160e-01
3.65493238e-01 -2.71525174e-01 6.57575309e-01 9.18802321e-01
-8.96836162e-01 7.18379557e-01 6.53788090e-01 4.04748060e-02
1.19259536e+00 -9.52946305e-01 -4.83213842e-01 7.55902588e-01
2.13673607e-01 5.82950115e-01 4.33671623e-01 -5.84733427e-01
-1.63594818e+00 -1.26586068e+00 6.12015009e-01 6.24209680e-02
7.24691033e-01 -2.67680585e-01 -7.83599854e-01 8.29687417e-01
-4.35854271e-02 3.09974551e-02 1.22683096e+00 3.50750834e-01
-7.02224791e-01 -4.64840472e-01 -9.28702176e-01 1.68045551e-01
6.67578101e-01 -7.10371375e-01 -2.05262274e-01 7.81258345e-01
7.28832841e-01 -1.98319793e-01 -9.68473494e-01 2.04736978e-01
5.44496357e-01 -4.22351301e-01 7.27273643e-01 -5.00477791e-01
1.61710560e-01 -3.97073328e-01 -5.22561073e-01 -1.22328484e+00
-6.59209907e-01 -5.31481922e-01 -5.05256951e-02 1.40549386e+00
2.28077784e-01 -5.53240240e-01 5.97038805e-01 5.24986506e-01
-5.58707714e-02 -4.51904863e-01 -9.18176830e-01 -7.86365926e-01
-1.23439305e-01 -3.98428619e-01 1.19313896e-01 1.17128360e+00
1.15285814e-02 7.93484986e-01 -1.19699359e+00 7.12802708e-02
1.14154708e+00 2.80268252e-01 7.86009431e-01 -1.36701369e+00
-4.26644444e-01 2.75423497e-01 5.89625500e-02 -1.15723693e+00
4.64208901e-01 -1.20121169e+00 -6.67352825e-02 -1.24314690e+00
8.09658706e-01 -5.76713979e-01 -5.27899623e-01 5.05297124e-01
-5.53631127e-01 6.56507760e-02 2.17272326e-01 7.46765375e-01
-9.63613451e-01 8.05803537e-01 1.34236097e+00 -1.67524725e-01
6.06945716e-03 1.28629193e-01 -9.18275058e-01 5.69870591e-01
7.13139594e-01 -8.38609993e-01 -4.78563339e-01 -4.28583413e-01
4.22206253e-01 3.32140595e-01 -5.11876941e-02 -6.86378360e-01
3.08059633e-01 -1.98097318e-01 3.16219474e-03 -3.94140899e-01
5.45894206e-01 -9.25387740e-01 1.87738240e-01 3.31531256e-01
-3.94443810e-01 -9.73523706e-02 -2.69970745e-01 1.23080707e+00
-1.18605964e-01 -2.97413737e-01 6.62520111e-01 1.87525764e-01
-2.66689748e-01 6.63502932e-01 -7.51591921e-02 1.56734571e-01
9.04182196e-01 2.66578019e-01 3.20126154e-02 -2.57259816e-01
-4.88566041e-01 3.92058760e-01 1.07975721e-01 3.72129261e-01
6.04441226e-01 -1.82759607e+00 -9.91051674e-01 -6.98489249e-02
4.27823901e-01 -9.93642882e-02 4.49105650e-01 1.24692941e+00
2.86279291e-01 3.43824625e-01 2.08049282e-01 -4.95893270e-01
-1.25125802e+00 7.22454250e-01 -2.83190697e-01 -3.81751329e-01
-2.89466679e-01 3.93118292e-01 3.69158834e-01 -5.21724403e-01
-3.94353420e-02 2.05963179e-01 -2.04692334e-01 3.18155736e-01
6.12203121e-01 5.53662956e-01 -1.94093212e-01 -7.50478566e-01
-2.38107353e-01 7.08293438e-01 -3.41102540e-01 -1.13669463e-01
1.33253908e+00 -5.26086032e-01 -3.21422637e-01 8.49135220e-01
1.55494869e+00 2.28475854e-01 -1.09311450e+00 -1.10834336e+00
1.02094710e-01 -6.46171212e-01 1.45546729e-02 -4.82758880e-02
-1.19772530e+00 6.88360989e-01 3.15921903e-01 -1.97733536e-01
1.07127285e+00 -3.15225691e-01 8.23250353e-01 6.44422948e-01
4.79141891e-01 -1.11184835e+00 3.85791838e-01 1.54895067e-01
6.29987061e-01 -1.60122907e+00 2.39373386e-01 -6.92689955e-01
-6.72002912e-01 9.27787006e-01 3.96390617e-01 5.00915311e-02
6.99336052e-01 -2.76633650e-01 -1.37268439e-01 -4.89450730e-02
-6.87709570e-01 -8.40364322e-02 9.83414054e-02 3.38563561e-01
3.62392932e-01 1.93727732e-01 -6.07388616e-01 6.88750029e-01
4.33650345e-01 5.33972010e-02 4.89807695e-01 6.71145380e-01
-3.66149902e-01 -1.41994512e+00 -6.06054306e-01 3.95164400e-01
-5.12855947e-01 -1.01403043e-01 1.71482675e-02 3.11190575e-01
-2.68228263e-01 1.37103200e+00 -6.79686606e-01 -4.90687042e-01
8.62726495e-02 2.11576396e-03 4.32551861e-01 -6.25975251e-01
-1.74137950e-01 6.23399079e-01 -1.49610922e-01 -2.38470927e-01
-7.87116706e-01 -8.23573053e-01 -1.29713786e+00 -1.67974144e-01
-7.05986261e-01 4.15367633e-01 5.00343919e-01 1.14445806e+00
3.15282822e-01 1.17947415e-01 1.21070087e+00 -3.22918653e-01
-7.44105875e-01 -1.07805729e+00 -1.01835012e+00 6.05707169e-01
1.64068446e-01 -7.16895103e-01 -3.83818954e-01 1.51778921e-01] | [8.037503242492676, 4.512552261352539] |
6a1776a7-6e60-4f25-bcd9-ca61078f1299 | joint-noise-tolerant-learning-and-meta-camera | 2103.04618 | null | https://arxiv.org/abs/2103.04618v1 | https://arxiv.org/pdf/2103.04618v1.pdf | Joint Noise-Tolerant Learning and Meta Camera Shift Adaptation for Unsupervised Person Re-Identification | This paper considers the problem of unsupervised person re-identification (re-ID), which aims to learn discriminative models with unlabeled data. One popular method is to obtain pseudo-label by clustering and use them to optimize the model. Although this kind of approach has shown promising accuracy, it is hampered by 1) noisy labels produced by clustering and 2) feature variations caused by camera shift. The former will lead to incorrect optimization and thus hinders the model accuracy. The latter will result in assigning the intra-class samples of different cameras to different pseudo-label, making the model sensitive to camera variations. In this paper, we propose a unified framework to solve both problems. Concretely, we propose a Dynamic and Symmetric Cross-Entropy loss (DSCE) to deal with noisy samples and a camera-aware meta-learning algorithm (MetaCam) to adapt camera shift. DSCE can alleviate the negative effects of noisy samples and accommodate the change of clusters after each clustering step. MetaCam simulates cross-camera constraint by splitting the training data into meta-train and meta-test based on camera IDs. With the interacted gradient from meta-train and meta-test, the model is enforced to learn camera-invariant features. Extensive experiments on three re-ID benchmarks show the effectiveness and the complementary of the proposed DSCE and MetaCam. Our method outperforms the state-of-the-art methods on both fully unsupervised re-ID and unsupervised domain adaptive re-ID. | ['Nicu Sebe', 'Shaozi Li', 'Yaojin Lin', 'Yuanzheng Cai', 'Zhiming Luo', 'Zhun Zhong', 'Fengxiang Yang'] | 2021-03-08 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Yang_Joint_Noise-Tolerant_Learning_and_Meta_Camera_Shift_Adaptation_for_Unsupervised_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Yang_Joint_Noise-Tolerant_Learning_and_Meta_Camera_Shift_Adaptation_for_Unsupervised_CVPR_2021_paper.pdf | cvpr-2021-1 | ['unsupervised-person-re-identification'] | ['computer-vision'] | [ 1.58304930e-01 -3.65072876e-01 -8.13860819e-02 -5.96466780e-01
-6.02402091e-01 -3.86786401e-01 6.21823788e-01 -4.10956740e-02
-6.66501641e-01 5.60111105e-01 1.21454515e-01 4.16369289e-01
2.94434391e-02 -3.94853741e-01 -3.82569313e-01 -9.06581700e-01
4.52424020e-01 5.65289438e-01 9.68145281e-02 3.51467818e-01
1.34916291e-01 2.45756209e-01 -1.67310226e+00 9.62532535e-02
1.06861448e+00 6.28331184e-01 2.15645939e-01 3.43582064e-01
-6.92294762e-02 4.60610747e-01 -5.79335093e-01 -6.11775756e-01
4.37215954e-01 -5.45444548e-01 -5.45963764e-01 5.02884567e-01
3.22632849e-01 -1.10165313e-01 -1.24266349e-01 1.32660234e+00
6.94224298e-01 5.19342303e-01 7.93095469e-01 -1.33643353e+00
-4.86982852e-01 2.40879968e-01 -7.90560782e-01 1.95887331e-02
2.23113865e-01 -1.45479292e-02 5.17466664e-01 -8.14818919e-01
3.10468882e-01 1.21642160e+00 7.36962199e-01 1.00863683e+00
-1.26256490e+00 -7.04223752e-01 3.92032444e-01 3.96687508e-01
-1.61101365e+00 -2.86329418e-01 8.90172899e-01 -5.26274741e-01
4.19344306e-01 1.54837295e-01 5.60411692e-01 1.04588592e+00
-3.77918899e-01 6.55480385e-01 1.26655924e+00 -5.49293280e-01
3.72466862e-01 5.18152595e-01 2.92668968e-01 3.56602460e-01
1.33121476e-01 3.29699814e-02 -2.96804160e-01 -3.89423445e-02
3.45812649e-01 2.62324870e-01 -2.54374892e-01 -4.25118357e-01
-1.02802205e+00 5.34069121e-01 1.91968441e-01 2.54408717e-01
-5.30677252e-02 -3.91143918e-01 3.68685663e-01 7.04410374e-02
2.99045056e-01 1.39023036e-01 -2.94123590e-01 3.18064131e-02
-9.35011923e-01 2.46087313e-02 4.59988385e-01 8.24306965e-01
8.83235097e-01 -4.00407463e-01 -1.84472591e-01 1.23618913e+00
2.18382046e-01 4.94463891e-01 9.45537925e-01 -6.90887988e-01
5.10089219e-01 7.19359815e-01 1.74551710e-01 -1.04111493e+00
-3.65956426e-01 -4.29888546e-01 -9.25661445e-01 -8.46822560e-02
5.74700236e-01 -3.86752337e-02 -9.44685698e-01 1.72129059e+00
3.96932930e-01 6.88403428e-01 -1.05251476e-01 9.97405469e-01
5.43870389e-01 4.87154454e-01 2.19081298e-01 -3.58959049e-01
9.64966238e-01 -1.15305459e+00 -6.10350192e-01 -2.88672715e-01
6.26013160e-01 -4.78807449e-01 7.95979202e-01 3.39127451e-01
-6.39215767e-01 -1.03043723e+00 -9.89235222e-01 3.55592400e-01
-3.94757986e-01 2.76523620e-01 -9.79383141e-02 9.37465906e-01
-8.77683878e-01 3.56176674e-01 -5.34379721e-01 -4.03311074e-01
1.60452813e-01 5.08471489e-01 -3.58120888e-01 -2.11825714e-01
-8.33270013e-01 5.92315376e-01 6.12993658e-01 1.16967984e-01
-6.22184575e-01 -4.92893010e-01 -5.33395767e-01 -1.00926094e-01
3.57927322e-01 -4.60160702e-01 9.32138026e-01 -1.33857954e+00
-1.60942221e+00 9.82988477e-01 -2.85062909e-01 -1.85627416e-01
6.22889817e-01 -4.45616879e-02 -7.67946362e-01 -1.14698380e-01
2.09498167e-01 5.43755591e-01 9.81851161e-01 -1.54144549e+00
-8.46402168e-01 -5.09877086e-01 -4.05649811e-01 4.70453650e-01
-5.70646644e-01 -2.80941218e-01 -8.99567604e-01 -6.66882336e-01
-7.84159638e-03 -1.25639200e+00 -1.72188982e-01 -6.48266673e-01
-5.25652826e-01 -1.92981377e-01 8.46236467e-01 -5.46834230e-01
1.26802647e+00 -2.20804358e+00 2.07466215e-01 3.57400298e-01
2.81605516e-02 5.40281951e-01 -1.03988521e-01 -4.00217474e-02
-1.62078425e-01 -1.00063230e-03 -2.34053642e-01 -7.50470519e-01
-1.72581479e-01 1.49863556e-01 1.03294656e-01 5.51157176e-01
-2.66618311e-01 5.03277302e-01 -8.15530300e-01 -6.06196761e-01
5.41190505e-01 3.76902312e-01 -4.56120253e-01 4.05586123e-01
1.71513215e-01 8.00079465e-01 -2.33191386e-01 5.50812662e-01
9.85298514e-01 -1.31572053e-01 2.66737729e-01 -3.37779641e-01
5.20619787e-02 -2.79217690e-01 -1.52713871e+00 1.38788188e+00
-2.20582962e-01 1.28579810e-01 -1.54843941e-01 -1.16790366e+00
9.13031995e-01 7.89065287e-02 5.25596976e-01 -6.11796260e-01
8.39204639e-02 9.45746899e-02 -4.04623866e-01 -4.53960538e-01
3.45739543e-01 -5.02269045e-02 -1.08880335e-02 3.84400129e-01
-1.94871537e-02 4.60062921e-01 8.91952496e-03 -1.00518294e-01
4.55924898e-01 7.25632608e-02 1.53946713e-01 -1.79749534e-01
1.03903973e+00 -3.32482964e-01 9.15208995e-01 7.48484612e-01
-5.70601344e-01 6.46202803e-01 -1.54700249e-01 -3.38061243e-01
-9.07339394e-01 -9.19120431e-01 -5.34519702e-02 9.58738208e-01
6.31644189e-01 -2.10029766e-01 -1.00820851e+00 -1.05656195e+00
-1.35110825e-01 5.27109087e-01 -4.76166248e-01 -3.26003969e-01
-5.25558949e-01 -1.14285207e+00 2.17849016e-01 5.00464976e-01
9.78411019e-01 -5.88225484e-01 -1.17577679e-01 1.70290068e-01
-5.65116823e-01 -1.14991522e+00 -7.66870797e-01 -2.19419878e-02
-7.14275122e-01 -1.04025376e+00 -8.71560156e-01 -8.84226918e-01
1.03469193e+00 4.74790215e-01 7.09500015e-01 7.84539878e-02
-1.00117885e-01 5.76332986e-01 -4.63475645e-01 7.20695630e-02
-3.67941409e-01 7.35213864e-04 3.62559170e-01 5.81195951e-01
7.73522496e-01 -3.52272779e-01 -5.75043917e-01 7.46875048e-01
-7.49679804e-01 -1.06778525e-01 1.89280093e-01 1.01902997e+00
7.53819764e-01 3.97507548e-01 4.62469846e-01 -8.90875697e-01
2.84785718e-01 -3.85708183e-01 -6.15127802e-01 5.51598430e-01
-9.65690613e-01 3.99704687e-02 7.44863868e-01 -6.92997694e-01
-1.25443089e+00 4.34968382e-01 1.85308829e-01 -3.89523715e-01
-3.40463519e-01 1.11532293e-01 -4.97819781e-01 -5.09185158e-02
2.94453055e-01 3.01366687e-01 -2.20551133e-01 -5.88991761e-01
1.57544911e-01 9.78876889e-01 7.76717067e-01 -4.90623921e-01
9.33983207e-01 5.95886230e-01 -2.79112965e-01 -6.77174330e-01
-7.89338589e-01 -8.83607209e-01 -1.02098525e+00 -4.28044409e-01
1.03010118e+00 -1.04958212e+00 -3.78796548e-01 8.15867007e-01
-7.63417542e-01 -1.56648368e-01 -7.97201991e-02 5.04268467e-01
-2.27063373e-01 8.43584895e-01 -1.25710517e-01 -7.87469089e-01
-4.83168401e-02 -1.28198802e+00 7.66515136e-01 6.44408226e-01
1.27415329e-01 -1.06522393e+00 1.72747776e-01 5.21449625e-01
-1.66099332e-02 7.70750493e-02 4.48087245e-01 -6.55385792e-01
-1.87687293e-01 -2.54732490e-01 -9.42764059e-02 5.39352596e-01
3.35026264e-01 -2.86652833e-01 -1.22662473e+00 -5.76194942e-01
-7.87400454e-02 -1.70018896e-01 8.10938835e-01 2.93113500e-01
1.23172820e+00 -1.86239362e-01 -5.50521791e-01 8.68892789e-01
1.54408455e+00 1.98884353e-01 4.97147679e-01 5.63081384e-01
9.51792061e-01 5.70675850e-01 6.52061760e-01 5.47280550e-01
6.23439193e-01 1.00483084e+00 1.79385304e-01 -1.95794050e-02
-3.07478732e-03 -3.05990309e-01 3.86389673e-01 7.10821092e-01
-1.99493706e-01 -1.35342255e-01 -6.37272298e-01 5.30871153e-01
-1.99826801e+00 -9.88899231e-01 -6.50554448e-02 2.53388047e+00
5.24749756e-01 -1.64965048e-01 4.78596479e-01 2.17551827e-01
1.29483831e+00 -1.54494330e-01 -5.91678143e-01 -1.73607673e-02
-1.20175332e-01 -2.87215114e-01 6.18002713e-01 4.11751956e-01
-1.35946810e+00 7.63752639e-01 5.11230612e+00 7.51047134e-01
-9.97805536e-01 2.63193935e-01 6.54054463e-01 8.09446499e-02
1.71092004e-01 -1.45486861e-01 -1.06243563e+00 9.64001179e-01
6.93262815e-01 7.12257698e-02 5.86521685e-01 9.12464917e-01
1.63661912e-02 -6.47150129e-02 -1.13707364e+00 1.55045736e+00
4.17497933e-01 -6.37937546e-01 -9.98564959e-02 1.89794190e-02
1.03177464e+00 -2.41457209e-01 6.37033880e-02 2.48634934e-01
1.59824893e-01 -6.36571586e-01 7.19842494e-01 3.77612293e-01
7.21068203e-01 -8.47095370e-01 8.68459880e-01 3.28531414e-01
-1.34037638e+00 -5.11212468e-01 -5.53526163e-01 3.52783471e-01
-6.33774372e-03 4.46642339e-01 -5.64045012e-01 5.29774189e-01
9.44166303e-01 8.78902972e-01 -9.14141834e-01 1.06492209e+00
-7.08203539e-02 3.36610526e-01 -1.64927810e-01 2.70531178e-01
-1.12148076e-01 -4.42038596e-01 5.19435167e-01 1.32740605e+00
2.14522272e-01 -4.42330651e-02 3.57672632e-01 5.41503489e-01
4.72401977e-02 -2.24769115e-02 -1.74191132e-01 4.09255564e-01
7.07992792e-01 1.12904656e+00 -6.71964169e-01 -4.15218502e-01
-4.65712309e-01 1.54488587e+00 3.18485707e-01 5.57585120e-01
-8.35132062e-01 3.67878452e-02 4.08265859e-01 8.91407356e-02
3.03733945e-01 4.35946994e-02 -1.72209635e-01 -1.40529633e+00
6.42071851e-03 -7.91378438e-01 7.83668160e-01 -2.65314639e-01
-1.54847860e+00 3.51762921e-01 7.98333436e-02 -1.48835742e+00
-1.86770365e-01 -4.30930376e-01 -4.07351345e-01 5.52802563e-01
-1.48770928e+00 -1.07404125e+00 -6.10212505e-01 9.00707543e-01
5.05915403e-01 -3.37876976e-01 5.40228307e-01 6.31104350e-01
-9.65724468e-01 1.09762609e+00 3.98495525e-01 2.17790902e-01
1.07629085e+00 -1.24460363e+00 -1.21469475e-01 1.03057659e+00
-2.53846422e-02 3.64897460e-01 4.47737306e-01 -5.86751938e-01
-9.47683632e-01 -1.35416508e+00 7.51972318e-01 -4.29219037e-01
4.26380597e-02 -2.74668366e-01 -7.47735798e-01 6.47472680e-01
-3.06615770e-01 -4.76670824e-02 8.02367866e-01 1.28460897e-03
-3.70384812e-01 -4.89255697e-01 -1.31673288e+00 4.14140701e-01
9.67739642e-01 -4.23964024e-01 -4.11975324e-01 1.79383814e-01
1.86599493e-01 -8.85151848e-02 -7.90753484e-01 2.78315991e-01
4.17988122e-01 -1.14871573e+00 1.02122486e+00 -1.64724559e-01
-1.45550400e-01 -5.75744450e-01 1.40239233e-02 -1.40898371e+00
-6.05186641e-01 -2.77255654e-01 1.84710011e-01 1.70943630e+00
5.16878553e-02 -6.05782986e-01 6.27889037e-01 9.44040298e-01
1.30460560e-01 -1.32841498e-01 -9.72038925e-01 -9.45143461e-01
-1.55507088e-01 -1.71044961e-01 7.41130888e-01 1.12962615e+00
-1.03352040e-01 8.05154219e-02 -7.14603543e-01 3.32400292e-01
8.86410713e-01 -1.20393962e-01 8.51980269e-01 -1.22562563e+00
-3.00467640e-01 -3.22321057e-01 -4.51794803e-01 -9.72387254e-01
2.95705944e-01 -7.97679365e-01 4.46736403e-02 -1.05382371e+00
5.86635828e-01 -5.90876281e-01 -4.16555405e-01 2.71168411e-01
-5.03099740e-01 2.94555634e-01 3.34455669e-01 5.94362378e-01
-9.05226767e-01 5.18523932e-01 7.00566828e-01 -2.70626634e-01
-5.50062299e-01 1.30974799e-01 -5.14706433e-01 7.72864938e-01
7.29672253e-01 -5.49628973e-01 -3.27461064e-01 -3.04073870e-01
-2.31195912e-01 -4.64832038e-01 4.50944990e-01 -1.26897299e+00
4.19537812e-01 3.51731367e-02 6.40627265e-01 -4.89478678e-01
1.90485343e-01 -1.12480462e+00 1.74873829e-01 3.24221462e-01
-2.35672355e-01 -7.89090917e-02 -1.81351244e-01 8.33459437e-01
-1.70847073e-01 -3.10123086e-01 1.13518178e+00 -1.09675527e-01
-8.37983012e-01 2.90397108e-01 -1.90968856e-01 3.04597355e-02
1.08792830e+00 -5.67114711e-01 -1.24992775e-02 -2.75542706e-01
-7.33239412e-01 3.29611897e-01 7.65791893e-01 4.35155809e-01
3.96662712e-01 -1.36638868e+00 -5.97303391e-01 2.74218768e-01
3.07330340e-01 -1.00206681e-01 4.80470330e-01 6.44924998e-01
-2.00006850e-02 8.01497400e-02 6.91951439e-02 -7.60884047e-01
-1.36710715e+00 7.83697724e-01 5.08090496e-01 -3.88950735e-01
-4.17836666e-01 6.79478586e-01 1.63196430e-01 -6.61393106e-01
4.92239177e-01 1.33256570e-01 -3.82770061e-01 2.80253794e-02
6.72422945e-01 6.74420655e-01 -1.17611945e-01 -9.88581717e-01
-4.28688943e-01 9.63527918e-01 -1.76765829e-01 2.25501396e-02
9.79337215e-01 -6.05090618e-01 7.20568150e-02 3.43690544e-01
1.16979945e+00 -1.64798036e-01 -1.36693096e+00 -3.53005439e-01
5.36540225e-02 -4.89429623e-01 -1.62933350e-01 -7.65179694e-01
-1.00798357e+00 5.71470976e-01 1.19372749e+00 -1.65087208e-01
1.32698607e+00 -1.62436023e-01 8.01259577e-01 2.16591597e-01
2.22950682e-01 -1.75152194e+00 2.38944769e-01 1.05294585e-01
2.08510801e-01 -1.46528518e+00 -5.02443910e-02 -3.20453376e-01
-8.63263547e-01 9.40598428e-01 7.86200047e-01 1.22351646e-01
6.46802902e-01 -6.78682104e-02 8.20915923e-02 3.41431141e-01
3.25783938e-02 -9.28010046e-02 2.66329050e-01 8.80771518e-01
-4.08973694e-02 1.57085374e-01 -8.74907374e-02 6.74420297e-01
1.27557695e-01 7.29708672e-02 1.82849377e-01 6.02074206e-01
-1.53534308e-01 -1.30779028e+00 -7.02812850e-01 2.19110459e-01
-2.23059610e-01 3.15282196e-01 -3.70700717e-01 6.01087570e-01
5.68670690e-01 1.22892869e+00 1.23648373e-02 -7.52296805e-01
2.01291338e-01 1.17634818e-01 3.58122230e-01 -3.98571014e-01
-4.92833704e-01 1.59312919e-01 -3.09158623e-01 -2.73491681e-01
-8.48468840e-01 -8.31456542e-01 -1.01267135e+00 -1.37274712e-01
-4.48059648e-01 1.19312249e-01 4.06204760e-01 9.61005092e-01
3.46459717e-01 1.85575366e-01 1.04055893e+00 -7.60490596e-01
-5.62284887e-01 -8.41078222e-01 -6.67693496e-01 8.94950271e-01
2.46336967e-01 -7.19623089e-01 -5.87027371e-01 3.67175341e-01] | [14.806422233581543, 1.0976905822753906] |
f70977f0-8118-42df-8342-1f9d07d0cc8e | fully-automated-2d-and-3d-convolutional | 2103.14734 | null | https://arxiv.org/abs/2103.14734v2 | https://arxiv.org/pdf/2103.14734v2.pdf | Fully Automated 2D and 3D Convolutional Neural Networks Pipeline for Video Segmentation and Myocardial Infarction Detection in Echocardiography | Cardiac imaging known as echocardiography is a non-invasive tool utilized to produce data including images and videos, which cardiologists use to diagnose cardiac abnormalities in general and myocardial infarction (MI) in particular. Echocardiography machines can deliver abundant amounts of data that need to be quickly analyzed by cardiologists to help them make a diagnosis and treat cardiac conditions. However, the acquired data quality varies depending on the acquisition conditions and the patient's responsiveness to the setup instructions. These constraints are challenging to doctors especially when patients are facing MI and their lives are at stake. In this paper, we propose an innovative real-time end-to-end fully automated model based on convolutional neural networks (CNN) to detect MI depending on regional wall motion abnormalities (RWMA) of the left ventricle (LV) from videos produced by echocardiography. Our model is implemented as a pipeline consisting of a 2D CNN that performs data preprocessing by segmenting the LV chamber from the apical four-chamber (A4C) view, followed by a 3D CNN that performs a binary classification to detect if the segmented echocardiography shows signs of MI. We trained both CNNs on a dataset composed of 165 echocardiography videos each acquired from a distinct patient. The 2D CNN achieved an accuracy of 97.18% on data segmentation while the 3D CNN achieved 90.9% of accuracy, 100% of precision and 95% of recall on MI detection. Our results demonstrate that creating a fully automated system for MI detection is feasible and propitious. | ['Tahir Hamid', 'Rashid Mazhar', 'Ridha Hamila', 'Serkan Kiranyaz', 'Christopher J. Henry', 'Sheela Ramanna', 'Oumaima Hamila'] | 2021-03-26 | null | null | null | null | ['myocardial-infarction-detection'] | ['medical'] | [ 3.11702788e-01 -1.59317926e-01 6.92690760e-02 -3.46369624e-01
-4.43816394e-01 -8.15842688e-01 -2.37743124e-01 1.08353205e-01
-3.15419704e-01 3.98669183e-01 -2.71444172e-01 -7.65102625e-01
1.34368762e-01 -6.06891930e-01 -4.39309955e-01 -5.20737052e-01
-4.63281333e-01 6.04792118e-01 -7.50617832e-02 3.85085255e-01
-1.52547462e-02 7.90979505e-01 -1.06887507e+00 2.15529978e-01
6.41441941e-01 1.04922831e+00 4.14712578e-02 1.54626346e+00
3.62427115e-01 5.26771963e-01 -7.36839175e-01 3.23668689e-01
4.12070811e-01 -6.21343911e-01 -6.30503833e-01 2.34793112e-01
5.61710298e-01 -8.62490356e-01 1.38418109e-03 5.27374446e-01
9.95591342e-01 -5.18229961e-01 3.02794933e-01 -7.73446083e-01
-5.88799678e-02 3.69012535e-01 -1.39635578e-01 7.87038326e-01
-1.07958578e-01 4.63931531e-01 3.82768452e-01 -7.19711244e-01
7.59114265e-01 5.63258290e-01 7.50615358e-01 4.32061762e-01
-9.79824245e-01 -3.98279130e-01 -5.57558835e-01 -1.37138471e-01
-1.06930315e+00 -1.38173699e-01 5.78222036e-01 -8.65040183e-01
5.99548459e-01 3.47266972e-01 9.88353848e-01 4.45783108e-01
2.85999328e-01 3.10762972e-01 8.05583000e-01 -2.55704254e-01
5.67487143e-02 -1.84968576e-01 -3.90573107e-02 7.54491389e-01
2.06411198e-01 9.00527239e-02 -3.39111127e-02 -6.52464181e-02
1.14377689e+00 2.16338664e-01 -4.46309000e-01 -1.67462811e-01
-1.57593644e+00 4.59024876e-01 7.24277198e-02 4.43910271e-01
-3.95129055e-01 -2.36665569e-02 6.42218411e-01 3.71457905e-01
1.70424759e-01 3.46313566e-01 -7.00708389e-01 -2.52403468e-01
-1.14140081e+00 1.76500246e-01 6.22569740e-01 4.97102916e-01
4.68548387e-02 8.24375451e-02 -2.94360936e-01 6.71682179e-01
2.33724102e-01 6.19194150e-01 4.75359619e-01 -1.09292471e+00
4.03594762e-01 6.20234787e-01 1.50269195e-02 -9.39194381e-01
-7.59825170e-01 -8.24840069e-01 -1.20197201e+00 4.17763323e-01
5.96403658e-01 -5.61115682e-01 -1.07315075e+00 1.19559431e+00
2.54390031e-01 1.34544790e-01 -1.48598999e-01 1.38237405e+00
1.17547917e+00 3.37195694e-01 -6.98277876e-02 -2.35494465e-01
1.43652999e+00 -5.41024923e-01 -5.10187685e-01 4.37758453e-02
1.06876659e+00 -6.73812866e-01 7.32140422e-01 3.19310546e-01
-1.23111880e+00 -8.26503992e-01 -1.11207378e+00 2.24287957e-01
4.98441085e-02 6.21721506e-01 6.19801022e-02 7.79427052e-01
-1.05992162e+00 9.55054522e-01 -9.11025882e-01 -1.36856347e-01
7.39453197e-01 5.01215577e-01 -4.77394074e-01 1.85915768e-01
-1.02375960e+00 5.80056489e-01 3.17069888e-01 5.34369171e-01
-7.25195348e-01 -8.80310833e-01 -7.41205394e-01 7.75779039e-02
2.17017263e-01 -9.32423234e-01 9.45374727e-01 -9.07905579e-01
-1.27885711e+00 1.39483261e+00 2.37788379e-01 -4.16597247e-01
7.53249288e-01 -1.73161179e-01 -1.73413575e-01 5.74115276e-01
-3.08735557e-02 4.42982584e-01 6.02340400e-01 -5.96701503e-01
-5.93774796e-01 -4.25596505e-01 -1.29987434e-01 -7.15655535e-02
2.97500253e-01 1.36642784e-01 -3.16202402e-01 -6.93000197e-01
3.12881708e-01 -1.09057593e+00 -1.83457658e-01 1.62485436e-01
-3.95129919e-01 4.25995022e-01 8.62514853e-01 -1.24036694e+00
1.10851943e+00 -1.98924434e+00 -3.30645479e-02 1.55690432e-01
8.18123341e-01 9.99222934e-01 2.52674490e-01 2.72996929e-02
-2.72493392e-01 5.71801662e-01 -3.25755358e-01 6.23742752e-02
-6.05921268e-01 -2.47556325e-02 2.73102462e-01 3.86711001e-01
1.79762244e-01 1.06537473e+00 -6.25427723e-01 -5.75016916e-01
5.79443455e-01 4.71432924e-01 -3.22509766e-01 5.24978995e-01
3.15691620e-01 1.12525296e+00 -2.29530379e-01 7.49315083e-01
6.72625422e-01 -5.01164913e-01 5.40855110e-01 -7.64013156e-02
1.70201771e-02 -1.82634637e-01 -9.70005095e-01 1.45765686e+00
-1.39052734e-01 6.96749985e-01 1.39882118e-01 -1.18087232e+00
7.59156406e-01 8.60325277e-01 6.01400018e-01 -3.40494901e-01
4.26969469e-01 2.58116812e-01 5.39172173e-01 -1.02567923e+00
-3.77988935e-01 -4.82462421e-02 2.38866955e-01 3.43150944e-01
-1.78007081e-01 4.48102169e-02 2.99431801e-01 -1.63213477e-01
1.17878103e+00 8.68940204e-02 2.99255401e-01 -1.46049276e-01
5.95283568e-01 -3.89388856e-03 8.70085180e-01 9.02926147e-01
-5.01061916e-01 1.27173281e+00 6.57009065e-01 -1.13968098e+00
-1.08074296e+00 -9.30168092e-01 -9.63630676e-02 1.18489064e-01
-2.85015225e-01 -4.14900854e-02 -7.26791739e-01 -7.56548285e-01
-2.35508025e-01 -2.37962797e-01 -3.05209547e-01 1.59514338e-01
-1.07307827e+00 -6.10997081e-01 4.97528762e-01 8.11237216e-01
5.91861308e-01 -1.21486807e+00 -1.47079027e+00 4.32089984e-01
-2.57831305e-01 -1.23151553e+00 -2.31895179e-01 -1.62930578e-01
-1.31816947e+00 -1.36036348e+00 -1.01120710e+00 -7.86569893e-01
6.35740876e-01 -1.92068845e-01 1.26008260e+00 4.53437001e-01
-8.60817790e-01 -1.30015761e-01 -2.61252642e-01 -5.10587156e-01
-6.36276484e-01 1.14209577e-01 -1.95128709e-01 -1.74490497e-01
-2.97480941e-01 -5.65658510e-01 -1.08115602e+00 9.97469723e-02
-6.57576859e-01 2.31513605e-01 4.82187301e-01 6.90302074e-01
5.83022058e-01 -4.11085099e-01 5.09151101e-01 -1.15629971e+00
2.73349851e-01 -2.89077550e-01 -6.34245038e-01 -5.98063245e-02
-4.71517324e-01 -7.14029431e-01 6.69287443e-01 -8.64611492e-02
-3.64303797e-01 2.92211980e-01 -2.45988175e-01 -6.72145069e-01
-5.82362115e-01 5.96204638e-01 1.41252518e-01 1.65437177e-01
4.73363996e-01 -9.20430273e-02 2.70408005e-01 -1.33915886e-01
-2.12433949e-01 7.45974243e-01 7.73278177e-01 6.78661419e-03
4.80737984e-01 3.26380640e-01 3.16121966e-01 -6.22607350e-01
-5.71924627e-01 -3.69777739e-01 -1.10888171e+00 -6.55380368e-01
1.20690000e+00 -5.99120975e-01 -6.39894366e-01 5.86703122e-01
-9.89087343e-01 -2.10575640e-01 -1.75317764e-01 5.87958992e-01
-4.07423943e-01 4.53906745e-01 -7.01018870e-01 -4.85907257e-01
-8.91582012e-01 -1.16666675e+00 8.45967233e-01 2.20479876e-01
-2.15704143e-01 -8.79573107e-01 -1.53960019e-01 6.25728011e-01
5.20068467e-01 8.66606772e-01 8.93522561e-01 -5.52592337e-01
-5.72492957e-01 -4.37518805e-01 -1.89824700e-01 6.93729520e-01
1.83585182e-01 1.09327763e-01 -7.87900388e-01 -1.82954997e-01
-2.41675422e-01 7.92794451e-02 4.24922913e-01 1.02526772e+00
1.22818124e+00 2.02871904e-01 -5.35067357e-02 7.52643526e-01
1.14619052e+00 4.72408742e-01 5.38922668e-01 -5.72926924e-02
7.98069119e-01 3.04729313e-01 5.70610642e-01 3.57157588e-01
6.07606582e-02 2.78231293e-01 4.26580787e-01 -7.42953598e-01
-1.40861809e-01 4.72665131e-01 -3.14399689e-01 7.55560696e-01
-4.14858341e-01 -9.93234962e-02 -1.32895982e+00 4.58161622e-01
-1.64000010e+00 -7.00567842e-01 -4.88720149e-01 2.32133484e+00
5.96694291e-01 1.45483181e-01 1.66432947e-01 4.41645950e-01
7.80518353e-01 -2.74240911e-01 -3.95174026e-01 -4.40667421e-01
1.00518763e-01 3.27971518e-01 3.14220399e-01 2.24305466e-02
-1.40313268e+00 2.98431098e-01 6.16194010e+00 -2.37039655e-01
-1.67128897e+00 -9.97439548e-02 1.17618060e+00 1.75612450e-01
4.77757215e-01 -2.25904822e-01 -2.24396348e-01 4.73180592e-01
9.88824368e-01 3.25229079e-01 -2.07249239e-01 4.95392233e-01
5.54668128e-01 -1.37929916e-01 -9.34110463e-01 1.04812908e+00
-1.27746612e-01 -1.70665443e+00 -4.74252641e-01 -8.69526565e-02
3.81806791e-01 -6.61564842e-02 -3.10350686e-01 4.75874229e-04
-7.45866120e-01 -1.05658627e+00 2.91710883e-01 6.47150338e-01
1.13785172e+00 -4.50979978e-01 1.19122541e+00 4.36155736e-01
-8.80648792e-01 1.26747772e-01 2.43565306e-01 -1.08101077e-01
2.13021874e-01 7.71165967e-01 -1.25203514e+00 3.05055499e-01
7.67618835e-01 5.82701921e-01 -3.16832811e-01 1.04871655e+00
1.04518138e-01 9.79912519e-01 -1.41984895e-02 4.95139420e-01
-1.26197144e-01 -1.40776828e-01 7.36082911e-01 1.12165642e+00
3.94484520e-01 8.93091112e-02 4.60406154e-01 8.68820131e-01
-6.79072440e-02 2.18224660e-01 -4.91882622e-01 -8.89721960e-02
9.68309641e-02 1.43554747e+00 -1.18012023e+00 -6.54214680e-01
-2.87299305e-01 6.57163680e-01 -2.79239595e-01 8.70229527e-02
-7.60600626e-01 -3.98026139e-01 1.35893106e-01 5.57561576e-01
1.75689295e-01 -3.09930816e-02 -5.53044617e-01 -9.70740736e-01
2.76718199e-01 -9.44599926e-01 4.03877646e-01 -6.96630776e-01
-6.85338914e-01 5.81401646e-01 -2.87459105e-01 -1.44948459e+00
-4.95673329e-01 -6.30679607e-01 -6.77153409e-01 1.05254948e+00
-1.26351106e+00 -9.87673759e-01 -5.92325628e-01 -1.10069498e-01
4.82869714e-01 -7.57582933e-02 9.53920126e-01 5.68224609e-01
-6.31278574e-01 1.07565008e-01 -5.40799856e-01 6.65796220e-01
6.05804145e-01 -1.46581495e+00 2.32563645e-01 8.30422401e-01
-3.55421096e-01 5.13543904e-01 3.78595948e-01 -6.11054420e-01
-1.00674307e+00 -1.15373063e+00 1.01256847e+00 -2.20205739e-01
-1.05172180e-01 1.13110714e-01 -6.59493387e-01 5.14383256e-01
-1.80790842e-01 7.71192312e-01 8.13946187e-01 -4.19261158e-01
3.25616658e-01 2.21817419e-02 -1.15805316e+00 2.14199424e-01
6.53932810e-01 -2.09556311e-01 -2.76571959e-01 2.30497345e-01
1.67906567e-01 -9.98904943e-01 -1.26209879e+00 8.19500685e-01
8.41844261e-01 -1.07205760e+00 9.24008250e-01 -5.69249511e-01
8.27359021e-01 -3.53645653e-01 3.95779401e-01 -8.07267487e-01
-1.30138323e-01 -4.17192668e-01 -5.00479713e-02 7.64038384e-01
3.62136930e-01 -4.24882859e-01 7.11282492e-01 4.03111190e-01
-2.39464641e-01 -1.07804561e+00 -6.96637571e-01 -2.35112995e-01
-1.14945054e-01 -4.64185089e-01 1.80308446e-01 9.39133704e-01
-5.09240866e-01 6.09410321e-03 -1.74064279e-01 2.79351503e-01
2.84769773e-01 2.45384142e-01 6.82768166e-01 -1.44998419e+00
-8.97046551e-02 -2.38405123e-01 -6.66553736e-01 -3.18204612e-01
-4.46402401e-01 -8.42163861e-01 -2.69023329e-01 -1.53924763e+00
1.19423196e-01 -4.02867377e-01 -3.66650790e-01 2.75615841e-01
-2.90011615e-01 6.40316784e-01 1.99828103e-01 2.02959642e-01
-9.98401791e-02 -5.25085211e-01 1.48379195e+00 7.11366907e-02
-4.46152091e-01 2.72336334e-01 -1.98910594e-01 8.57291520e-01
9.71558988e-01 -3.55688065e-01 -2.86797494e-01 -2.45540529e-01
9.46186557e-02 8.11900496e-01 6.82501554e-01 -1.25920165e+00
-2.14831501e-01 3.59613419e-01 7.82392383e-01 -8.86973381e-01
-1.44709855e-01 -5.70534885e-01 9.91844311e-02 9.38181460e-01
-2.84422129e-01 6.42231926e-02 4.68019396e-02 -5.77515736e-02
-2.34257221e-01 -4.52289283e-02 8.87276769e-01 -4.19959813e-01
-3.02498370e-01 5.26402652e-01 -7.09440827e-01 3.66713673e-01
1.00303423e+00 -3.60121012e-01 1.47595137e-01 -2.74669617e-01
-1.27022195e+00 8.07590559e-02 -4.95324805e-02 1.86690882e-01
7.24831700e-01 -8.72698426e-01 -1.08529019e+00 3.89011323e-01
-8.59023929e-02 2.26891056e-01 5.60058296e-01 1.34356081e+00
-1.35727477e+00 3.02662671e-01 -4.12498504e-01 -1.24220860e+00
-1.71161342e+00 1.42231017e-01 7.46735156e-01 -3.51175934e-01
-1.13614964e+00 5.46891689e-01 -9.19170678e-02 -4.16415751e-01
6.43721223e-02 -7.91351140e-01 -2.97363132e-01 -1.46555036e-01
4.78650630e-01 2.70217210e-01 2.21487194e-01 -4.07142699e-01
-2.79332072e-01 5.84596395e-01 1.80147782e-01 2.07423478e-01
1.04873884e+00 -1.14998974e-01 3.61439399e-02 3.77616704e-01
9.38504815e-01 -3.39560539e-01 -9.02761102e-01 1.11200079e-01
-4.56286371e-01 -4.22080725e-01 1.77370623e-01 -1.05702949e+00
-1.42169642e+00 1.27364469e+00 1.20825803e+00 2.04326913e-01
1.20865953e+00 -2.94145256e-01 1.00942218e+00 -5.92722371e-02
-1.00322038e-01 -9.19741333e-01 -3.28220218e-01 1.06237985e-01
6.42853260e-01 -1.41160309e+00 -2.40217090e-01 -4.32121783e-01
-5.13893187e-01 1.50415277e+00 3.48691881e-01 -6.62943125e-02
7.87484229e-01 3.18044782e-01 8.64847541e-01 -2.18537241e-01
-2.90611446e-01 9.61149205e-03 3.77558500e-01 5.35296619e-01
6.91078246e-01 1.73609719e-01 -3.78362030e-01 5.18872499e-01
1.11439951e-01 4.24812496e-01 5.73052704e-01 1.20675397e+00
-1.87211990e-01 -8.15585554e-01 -2.04903945e-01 7.05807269e-01
-1.03269625e+00 8.97834823e-02 -1.13105133e-01 6.19658113e-01
6.88132644e-01 7.65054941e-01 1.68019510e-03 -9.41930339e-02
2.52177328e-01 1.82931334e-01 2.45885804e-01 -7.34630704e-01
-9.49742019e-01 2.30168566e-01 6.27677068e-02 -3.90561014e-01
-2.74712533e-01 -6.07784212e-01 -1.43273711e+00 1.34014308e-01
9.90141034e-02 -3.72882009e-01 6.94927573e-01 1.08529842e+00
5.86045802e-01 1.02921224e+00 6.09196186e-01 -7.04654276e-01
-1.62093997e-01 -8.99052799e-01 -5.35018384e-01 4.55774188e-01
4.69160646e-01 -9.75072756e-02 -3.19412015e-02 6.66518807e-01] | [14.203413009643555, -2.4207427501678467] |
247ac899-0a25-4093-b894-c390e9719b3c | contrastive-learning-of-global-and-local-2 | null | null | http://proceedings.neurips.cc/paper/2021/hash/38ef4b66cb25e92abe4d594acb841471-Abstract.html | http://proceedings.neurips.cc/paper/2021/file/38ef4b66cb25e92abe4d594acb841471-Paper.pdf | Contrastive Learning of Global and Local Video Representations | Contrastive learning has delivered impressive results for various tasks in the self-supervised regime. However, existing approaches optimize for learning representations specific to downstream scenarios, i.e., global representations suitable for tasks such as classification or local representations for tasks such as detection and localization. While they produce satisfactory results in the intended downstream scenarios, they often fail to generalize to tasks that they were not originally designed for. In this work, we propose to learn video representations that generalize to both the tasks which require global semantic information (e.g., classification) and the tasks that require local fine-grained spatio-temporal information (e.g., localization). We achieve this by optimizing two contrastive objectives that together encourage our model to learn global-local visual information given audio signals. We show that the two objectives mutually improve the generalizability of the learned global-local representations, significantly outperforming their disjointly learned counterparts. We demonstrate our approach on various tasks including action/sound classification, lip reading, deepfake detection, event and sound localization. | ['Yale Song', 'Daniel McDuff', 'Zhaoyang Zeng', 'Shuang Ma'] | 2021-12-01 | null | https://openreview.net/forum?id=txWfwhc6gi | https://openreview.net/pdf?id=txWfwhc6gi | neurips-2021-12 | ['sound-classification'] | ['audio'] | [ 4.17223066e-01 -2.31452912e-01 -4.01041538e-01 -2.97606438e-01
-1.18477404e+00 -5.58999419e-01 5.92831135e-01 2.01505259e-01
-2.11285338e-01 5.11912644e-01 4.52967107e-01 -9.79702473e-02
-8.79831165e-02 -4.44487959e-01 -8.18557799e-01 -6.64933681e-01
-2.05210343e-01 -2.41859425e-02 1.80879384e-01 1.83319803e-02
2.23375127e-01 4.90796000e-01 -2.01645947e+00 5.08382320e-01
4.58737731e-01 1.25768614e+00 2.67554522e-01 8.66701424e-01
1.21374570e-01 7.53576517e-01 -5.08081496e-01 3.49637449e-01
-2.97889262e-02 -4.37373281e-01 -8.59407365e-01 1.56840727e-01
7.57923245e-01 -1.34093121e-01 -3.36278230e-01 8.49756718e-01
5.24180353e-01 3.97767961e-01 7.66122878e-01 -1.46193206e+00
-7.09578335e-01 2.89196342e-01 -5.09200335e-01 3.79924119e-01
3.90514523e-01 1.52157873e-01 1.27831769e+00 -8.67270172e-01
1.06954038e-01 1.25341368e+00 6.34242475e-01 6.79379165e-01
-1.17976868e+00 -4.98193949e-01 5.04433632e-01 3.38124365e-01
-1.14883804e+00 -9.46043372e-01 6.20026410e-01 -4.96483535e-01
8.41351032e-01 1.56452730e-01 1.72235087e-01 1.36327422e+00
-9.42621529e-02 1.03778052e+00 9.63583767e-01 -3.28282416e-01
2.43581846e-01 -9.55306590e-02 2.55336855e-02 3.17087531e-01
-1.70432836e-01 1.80472229e-02 -9.65644181e-01 9.97437686e-02
6.89074218e-01 -9.01029631e-02 -3.96656781e-01 -3.48340243e-01
-1.15920365e+00 6.71764612e-01 3.09932500e-01 4.93755966e-01
-3.73627424e-01 4.20312643e-01 5.40764689e-01 3.17898661e-01
6.14395618e-01 2.79011786e-01 -7.19963014e-01 -1.71042502e-01
-1.02200091e+00 -1.09417543e-01 3.49123091e-01 8.09073985e-01
6.43611014e-01 2.81090230e-01 -3.90631318e-01 9.82277691e-01
2.21767262e-01 2.83833772e-01 7.44806111e-01 -7.87956476e-01
4.20762867e-01 9.13958475e-02 3.64566445e-02 -7.92866528e-01
-5.78885555e-01 -4.65988725e-01 -5.92931926e-01 1.25044972e-01
3.80446374e-01 -1.12898439e-01 -1.00314689e+00 2.26010060e+00
7.67193809e-02 4.31880981e-01 9.98836756e-02 8.83520842e-01
1.06303859e+00 7.41821706e-01 2.73422569e-01 -2.27183700e-01
1.14716065e+00 -1.13268447e+00 -6.47807240e-01 -4.19085532e-01
4.62552160e-01 -5.73364139e-01 1.17613196e+00 2.10061416e-01
-1.02621925e+00 -7.65061438e-01 -8.52836072e-01 -3.13651040e-02
-2.40941092e-01 2.51117051e-01 5.14970958e-01 2.62519419e-01
-1.39427984e+00 3.31441253e-01 -6.50730610e-01 -6.54386282e-01
3.12134087e-01 1.98292285e-01 -3.72851819e-01 2.39754468e-01
-1.06287014e+00 5.63035607e-01 1.90680772e-01 -9.54647660e-02
-1.16340554e+00 -4.91058946e-01 -9.14330006e-01 2.32543781e-01
4.01077241e-01 -4.91412431e-01 1.46734250e+00 -1.22688282e+00
-1.40051925e+00 8.73631179e-01 -1.84410930e-01 -4.32927758e-01
2.14906096e-01 -2.96403587e-01 -2.38616124e-01 3.25018108e-01
2.35584959e-01 7.69676507e-01 1.18590772e+00 -1.06085408e+00
-7.26985157e-01 -2.70342678e-01 2.64799446e-01 3.38500500e-01
-5.68189859e-01 3.52064781e-02 -2.90387958e-01 -8.74889076e-01
-1.32272218e-03 -6.31842196e-01 1.34816796e-01 2.30928078e-01
-4.02159601e-01 -4.92619544e-01 1.02043402e+00 -6.74506545e-01
8.24441850e-01 -2.35024548e+00 1.41955778e-01 -9.17663425e-02
-4.42395955e-02 2.17321545e-01 -3.96162987e-01 3.82912785e-01
-2.58978963e-01 -2.41994504e-02 -2.90541612e-02 -5.63981831e-01
-4.25439142e-03 -4.24028933e-03 -2.63795078e-01 5.30033231e-01
3.20224375e-01 8.97814810e-01 -1.11103976e+00 -3.25935304e-01
6.82798475e-02 4.08374578e-01 -5.58401167e-01 3.49038243e-01
-2.29226872e-01 5.68890095e-01 -4.15250838e-01 5.55623174e-01
2.40698949e-01 -3.44887882e-01 -9.54826828e-03 -1.08338334e-01
1.94178410e-02 4.12797749e-01 -1.03798163e+00 1.91867208e+00
-8.71164262e-01 9.73627746e-01 4.98142928e-01 -1.50847900e+00
6.36887968e-01 5.11950433e-01 5.60296714e-01 -6.11240268e-01
-2.60119200e-01 -1.97949391e-02 -3.05101126e-01 -5.86486697e-01
2.41276816e-01 -3.04018408e-01 -5.92874624e-02 4.89295423e-01
3.23276430e-01 2.09077224e-01 -2.53678232e-01 3.88754643e-02
1.10842073e+00 3.15232545e-01 4.12030041e-01 -2.44977072e-01
4.01827365e-01 -3.59344602e-01 3.49085927e-01 8.00326467e-01
-2.67532796e-01 7.96749473e-01 3.24021310e-01 5.17554767e-02
-3.84542406e-01 -1.08792830e+00 -4.84189056e-02 1.80383325e+00
2.03263506e-01 -5.17740965e-01 -6.01870894e-01 -8.07295918e-01
-1.00795835e-01 3.31569821e-01 -5.05726039e-01 -3.90101701e-01
-2.85308778e-01 -2.76734293e-01 5.48757255e-01 9.20416534e-01
2.97219396e-01 -1.19573855e+00 -6.45654976e-01 2.08644494e-01
-2.05077350e-01 -1.17825735e+00 -5.67195117e-01 3.49361926e-01
-6.81133270e-01 -9.51081455e-01 -7.73962617e-01 -1.00608003e+00
3.00035298e-01 3.85799915e-01 9.94300067e-01 -8.29003453e-02
-3.67555678e-01 1.05183256e+00 -3.39961052e-01 -2.26294115e-01
2.13860236e-02 -9.65541452e-02 1.26104742e-01 4.06383812e-01
1.33347586e-01 -6.41833425e-01 -6.71690881e-01 2.42755413e-01
-8.65277290e-01 -1.51989445e-01 3.91052753e-01 9.84708667e-01
3.82613838e-01 -2.53550142e-01 9.62629139e-01 -2.98064590e-01
5.53134084e-01 -5.28697670e-01 -4.92097922e-02 2.83698857e-01
-4.99355048e-02 -9.13081318e-02 4.72042948e-01 -7.16485977e-01
-8.24869156e-01 -1.47081437e-02 -2.39381194e-01 -5.44654548e-01
-5.68459690e-01 3.19316804e-01 -2.03005880e-01 7.78918862e-02
6.38011813e-01 3.01394284e-01 -8.29927027e-02 -5.18054008e-01
3.32527637e-01 7.98236609e-01 5.74892044e-01 -5.35102904e-01
4.55760539e-01 2.62321115e-01 -1.63240388e-01 -1.01745903e+00
-9.42307293e-01 -7.28731692e-01 -4.38060910e-01 -1.12745270e-01
1.03145802e+00 -1.05151427e+00 -6.22113466e-01 4.67598259e-01
-1.09015954e+00 -4.44507927e-01 -4.76934671e-01 4.94683534e-01
-1.06507897e+00 2.19078600e-01 -3.41089666e-01 -9.46570396e-01
7.45812207e-02 -1.08438122e+00 1.56252003e+00 1.56475212e-02
-2.61361837e-01 -1.21433568e+00 -3.44042368e-02 4.66081314e-02
4.97263581e-01 7.71662369e-02 7.49389410e-01 -6.52925253e-01
-3.96761119e-01 6.59487695e-02 -1.87923685e-01 5.38671970e-01
3.26699197e-01 -2.85426676e-01 -1.49922776e+00 -2.32595146e-01
-2.80702740e-01 -7.33648598e-01 1.29435790e+00 5.84699392e-01
1.60424519e+00 -3.55419606e-01 -3.98051202e-01 6.21122241e-01
9.29837883e-01 7.45663866e-02 3.28715473e-01 4.12435122e-02
5.21910489e-01 8.28609824e-01 6.55464828e-01 3.99559140e-01
2.79727459e-01 1.13612413e+00 5.68752170e-01 -3.20984393e-01
-4.73806620e-01 -3.77072871e-01 4.83573884e-01 4.08448279e-01
5.51611856e-02 -3.55452895e-01 -7.74657071e-01 7.86514163e-01
-2.19760227e+00 -9.52022433e-01 3.87499750e-01 1.94700778e+00
6.99702740e-01 -2.65682429e-01 2.45578319e-01 1.81936294e-01
8.77637744e-01 3.91460359e-01 -5.93121648e-01 -3.65406871e-01
2.50226501e-02 2.81719387e-01 -6.76849559e-02 3.30377221e-01
-1.46564901e+00 1.07038140e+00 6.38005781e+00 7.63932467e-01
-1.29991019e+00 4.53590065e-01 4.97183293e-01 -1.19730569e-01
-1.34697571e-01 -2.97321290e-01 -4.65720981e-01 2.41632000e-01
8.97825658e-01 2.08512351e-01 2.98220545e-01 6.45436943e-01
2.45462641e-01 1.58383667e-01 -1.39668512e+00 1.27810693e+00
1.45076632e-01 -1.12536156e+00 -7.51358643e-02 -1.88151330e-01
6.10743105e-01 -1.88564047e-01 3.71495813e-01 3.96962404e-01
-1.12538837e-01 -1.21222305e+00 8.79893839e-01 4.70411569e-01
9.49624419e-01 -4.19105381e-01 2.29041010e-01 4.20200109e-01
-1.33007622e+00 -2.39434019e-01 3.41295227e-02 -7.30782077e-02
1.28154354e-02 2.11441502e-01 -5.16730964e-01 2.79082566e-01
8.64213407e-01 9.58153784e-01 -3.61390889e-01 1.01998329e+00
-2.77571380e-01 6.19749844e-01 -1.74269095e-01 2.07984567e-01
3.08698237e-01 2.22075671e-01 6.82698846e-01 1.23466301e+00
3.05244237e-01 -1.93193421e-01 2.56110907e-01 6.21620238e-01
-9.97535959e-02 1.94960654e-01 -8.43728781e-01 -8.49350691e-02
3.49036485e-01 1.07184887e+00 -4.12961155e-01 -9.91266668e-02
-4.81525630e-01 9.36792791e-01 3.96709144e-01 5.88920355e-01
-8.01631808e-01 -3.89908522e-01 1.05434442e+00 9.05762464e-02
4.25126612e-01 -1.56435415e-01 -1.08309247e-01 -1.11446571e+00
1.82946041e-01 -7.09557712e-01 5.33605754e-01 -7.97987223e-01
-1.25165486e+00 3.93661916e-01 3.37059125e-02 -1.27828705e+00
-5.64064026e-01 -7.38671720e-01 -8.51106465e-01 5.83521068e-01
-1.60857236e+00 -1.33507574e+00 -1.75033569e-01 1.04255152e+00
6.81478858e-01 -1.55940473e-01 7.57977843e-01 1.24455161e-01
-3.19177181e-01 8.13055754e-01 -7.09688962e-02 4.83210152e-03
1.00092173e+00 -1.24968076e+00 -1.22824557e-01 7.38511980e-01
5.28995752e-01 4.47335482e-01 4.45639312e-01 -1.55466869e-01
-1.18387187e+00 -1.18620384e+00 7.03630626e-01 -1.08186014e-01
6.56283498e-01 -4.52525139e-01 -5.75216234e-01 7.84271657e-01
-4.56884224e-03 3.95805389e-01 6.06789291e-01 4.37616199e-01
-4.98891562e-01 -1.98741555e-01 -9.89182293e-01 3.26704741e-01
1.35959876e+00 -1.05579174e+00 -4.27419662e-01 6.01901412e-01
6.32782400e-01 -2.08478630e-01 -5.29619932e-01 3.08091640e-01
4.46807623e-01 -8.79204750e-01 1.09196103e+00 -9.88587439e-01
2.07142800e-01 -1.82693109e-01 -4.52182174e-01 -1.29698443e+00
-1.53782591e-01 -7.26301968e-01 -3.46140474e-01 1.40990984e+00
3.68910670e-01 -5.83157361e-01 3.50259185e-01 -2.09458753e-01
-4.03068513e-01 -5.90273142e-01 -1.11213338e+00 -8.68534923e-01
2.83914525e-02 -6.70003057e-01 3.78029764e-01 7.01304555e-01
4.27909382e-02 4.63286102e-01 -5.46081662e-01 2.79489040e-01
3.01255018e-01 2.95924932e-01 5.54152727e-01 -1.19268262e+00
-5.19700587e-01 -5.08062065e-01 -7.60311604e-01 -1.54516542e+00
5.14301836e-01 -8.16212714e-01 3.83340597e-01 -1.52923167e+00
2.11835757e-01 -3.38136256e-01 -4.59174544e-01 7.64652312e-01
-1.03271879e-01 4.30760205e-01 4.40331846e-02 9.14458632e-02
-8.45766425e-01 5.20148277e-01 1.02789652e+00 -2.59252161e-01
-1.80297524e-01 2.33661950e-01 -9.37523484e-01 6.48716450e-01
8.30278993e-01 -5.33857383e-02 -5.31774282e-01 -4.47149515e-01
-2.36400709e-01 1.06404424e-01 7.87848532e-01 -1.03231955e+00
2.74559319e-01 -1.43994451e-01 2.69076318e-01 -2.05386087e-01
6.60303414e-01 -5.36187589e-01 -6.15111291e-01 -7.19017759e-02
-7.03286111e-01 -2.66176283e-01 2.82140493e-01 8.19825888e-01
-6.17427349e-01 1.43100340e-02 8.87428463e-01 1.12573303e-01
-1.02009189e+00 4.11183864e-01 -3.26111495e-01 2.00002745e-01
1.09146249e+00 -1.75718635e-01 -2.32227758e-01 -9.17486429e-01
-1.22365463e+00 8.86248518e-03 8.56352895e-02 6.90123737e-01
6.92903340e-01 -1.36814427e+00 -6.23213768e-01 2.05576077e-01
3.30085933e-01 -2.18859106e-01 1.18369743e-01 1.04959905e+00
1.91746369e-01 5.37981987e-01 -4.45518196e-02 -1.03569865e+00
-1.29666829e+00 4.96894240e-01 3.05391341e-01 3.44127677e-02
-6.78472519e-01 1.04001427e+00 8.73324513e-01 -2.53725141e-01
7.42161989e-01 -4.32060361e-01 -3.80174577e-01 1.26256734e-01
5.60165763e-01 1.32145062e-01 1.66229486e-01 -6.70616925e-01
-5.31525314e-01 7.48169541e-01 2.60004044e-01 2.63942573e-02
1.26046169e+00 -2.74666578e-01 3.84679377e-01 6.12902164e-01
1.36363888e+00 -5.62872775e-02 -1.52873695e+00 -2.48304561e-01
-4.81522866e-02 -2.48584956e-01 2.49549776e-01 -7.04666793e-01
-1.10468197e+00 1.42469835e+00 6.97639644e-01 3.15787017e-01
1.37088573e+00 5.22866964e-01 4.63545799e-01 1.26777798e-01
3.25221360e-01 -9.11720574e-01 5.20798862e-01 6.31317377e-01
9.29758191e-01 -1.14404595e+00 -5.26441813e-01 -1.03287622e-01
-6.35466576e-01 1.14372718e+00 4.98031259e-01 1.68329582e-01
5.46853125e-01 2.09494933e-01 -3.77513953e-02 -2.34553162e-02
-8.38237941e-01 -7.26083934e-01 6.85219944e-01 9.54022467e-01
4.95161057e-01 -7.69613963e-03 2.71782279e-01 5.53882718e-01
2.05532342e-01 -3.36044699e-01 3.72643247e-02 9.59433377e-01
-5.25196731e-01 -7.26107121e-01 -2.55787432e-01 3.12387139e-01
-3.51111352e-01 4.65952512e-03 -5.43464959e-01 7.77165949e-01
1.63218483e-01 1.30009007e+00 2.89778918e-01 -3.16742301e-01
1.68191835e-01 1.39507487e-01 5.13529360e-01 -8.67152333e-01
-3.04092705e-01 2.65832573e-01 1.43401444e-01 -7.06121325e-01
-6.03452444e-01 -7.69974828e-01 -9.73678648e-01 2.03712314e-01
-2.19728857e-01 -5.56224361e-02 4.88501102e-01 9.95490551e-01
2.75468588e-01 5.70198298e-01 5.35567522e-01 -1.01675940e+00
-6.25362039e-01 -8.94066155e-01 -6.60457969e-01 4.04169947e-01
8.16069722e-01 -8.47184002e-01 -4.19214308e-01 3.82708684e-02] | [14.647201538085938, 4.938182353973389] |
c17bf0a0-8d8b-4337-b62d-6f5ad05d0f1e | lion-latent-point-diffusion-models-for-3d | 2210.06978 | null | https://arxiv.org/abs/2210.06978v1 | https://arxiv.org/pdf/2210.06978v1.pdf | LION: Latent Point Diffusion Models for 3D Shape Generation | Denoising diffusion models (DDMs) have shown promising results in 3D point cloud synthesis. To advance 3D DDMs and make them useful for digital artists, we require (i) high generation quality, (ii) flexibility for manipulation and applications such as conditional synthesis and shape interpolation, and (iii) the ability to output smooth surfaces or meshes. To this end, we introduce the hierarchical Latent Point Diffusion Model (LION) for 3D shape generation. LION is set up as a variational autoencoder (VAE) with a hierarchical latent space that combines a global shape latent representation with a point-structured latent space. For generation, we train two hierarchical DDMs in these latent spaces. The hierarchical VAE approach boosts performance compared to DDMs that operate on point clouds directly, while the point-structured latents are still ideally suited for DDM-based modeling. Experimentally, LION achieves state-of-the-art generation performance on multiple ShapeNet benchmarks. Furthermore, our VAE framework allows us to easily use LION for different relevant tasks: LION excels at multimodal shape denoising and voxel-conditioned synthesis, and it can be adapted for text- and image-driven 3D generation. We also demonstrate shape autoencoding and latent shape interpolation, and we augment LION with modern surface reconstruction techniques to generate smooth 3D meshes. We hope that LION provides a powerful tool for artists working with 3D shapes due to its high-quality generation, flexibility, and surface reconstruction. Project page and code: https://nv-tlabs.github.io/LION. | ['Karsten Kreis', 'Sanja Fidler', 'Or Litany', 'Zan Gojcic', 'Francis Williams', 'Arash Vahdat', 'Xiaohui Zeng'] | 2022-10-12 | null | null | null | null | ['3d-shape-generation', 'point-cloud-generation'] | ['computer-vision', 'computer-vision'] | [-7.21040368e-02 -3.91775183e-03 2.68338650e-01 4.54411171e-02
-8.50118279e-01 -5.94331980e-01 7.67835736e-01 -2.70084888e-01
3.20168912e-01 3.75271529e-01 1.86622486e-01 -2.99000442e-01
2.41558895e-01 -1.37447333e+00 -9.55846429e-01 -6.79041564e-01
1.46056488e-01 9.48096871e-01 2.75987182e-02 -2.50358433e-01
-3.95133719e-02 9.90180135e-01 -1.53247178e+00 3.71455252e-01
8.68169248e-01 8.12102258e-01 3.44119787e-01 6.31595373e-01
-3.49408716e-01 1.29736692e-01 -2.40128025e-01 -3.17353904e-01
3.95238400e-01 -1.68037638e-01 -4.91106242e-01 3.67062353e-02
4.49044675e-01 -3.50204170e-01 -7.52828121e-02 7.11005449e-01
5.41293561e-01 9.10059456e-03 1.06436467e+00 -1.24284363e+00
-9.25390422e-01 1.07201546e-01 -3.77295464e-01 -5.95288813e-01
1.22831181e-01 6.07511699e-01 7.48133004e-01 -1.32266009e+00
9.97794986e-01 1.49985456e+00 7.35998869e-01 7.60286272e-01
-1.64436138e+00 -5.53452194e-01 -1.43773213e-01 -5.52417219e-01
-1.21658611e+00 -3.96595925e-01 9.93522167e-01 -7.26957142e-01
1.04593849e+00 3.80423456e-01 8.94585252e-01 1.30832791e+00
1.86650410e-01 8.74471903e-01 8.98500860e-01 -4.20504436e-03
3.48814487e-01 -1.48290247e-01 -5.39104462e-01 6.15146339e-01
-2.45504051e-01 2.20560402e-01 -2.94170558e-01 -4.20455277e-01
1.63643944e+00 -1.02132536e-01 -1.57915682e-01 -4.80319113e-01
-1.48600531e+00 9.08728182e-01 5.20591497e-01 1.90912411e-01
-6.45390332e-01 6.21573508e-01 -1.88927483e-02 1.99433804e-01
8.51892769e-01 3.79479349e-01 -2.96837509e-01 -1.54741526e-01
-1.17718053e+00 7.19757199e-01 6.70750022e-01 1.15636432e+00
6.47800148e-01 5.86188853e-01 -2.83175677e-01 9.48729157e-01
5.27662039e-01 7.47390211e-01 5.79792727e-03 -1.49835336e+00
8.35836604e-02 4.48210657e-01 4.50971089e-02 -8.72003198e-01
5.29883578e-02 -2.91095525e-01 -1.07315242e+00 8.36431146e-01
4.74283472e-03 8.72490853e-02 -1.25104737e+00 1.53420985e+00
3.77074808e-01 2.94114321e-01 -1.66171953e-01 7.27110267e-01
1.07737613e+00 1.05792952e+00 -1.04694832e-02 1.79016545e-01
9.09196675e-01 -6.36573553e-01 -2.83373117e-01 2.02765331e-01
2.98227251e-01 -9.57493722e-01 1.16995955e+00 3.72478396e-01
-1.62918210e+00 -5.56026101e-01 -7.58143067e-01 -3.85226756e-01
-1.12649217e-01 -1.91946894e-01 7.26200938e-01 2.06629887e-01
-1.44805527e+00 9.61656749e-01 -1.19440222e+00 2.70262663e-03
6.44101799e-01 1.88657045e-01 -2.05453828e-01 7.69788101e-02
-7.76186764e-01 4.29419905e-01 -1.58850983e-01 -1.56783611e-01
-9.20071065e-01 -1.03285849e+00 -8.94512713e-01 -8.64723474e-02
-3.71404171e-01 -1.45045805e+00 1.06133699e+00 -5.84556580e-01
-1.64885640e+00 8.88763130e-01 -1.54666200e-01 -1.10551998e-01
7.24911630e-01 2.30574220e-01 1.16446525e-01 3.16561684e-02
7.43718371e-02 1.32487428e+00 1.14489079e+00 -1.65914392e+00
4.73718494e-02 -1.18208408e-01 -4.02354449e-01 9.09451768e-02
3.12792808e-02 -4.16289836e-01 -4.45856184e-01 -1.06918859e+00
2.51391143e-01 -8.48106563e-01 -2.65921086e-01 5.45315742e-01
-2.46269345e-01 -2.52191037e-01 1.08557725e+00 -6.29667342e-01
5.78072965e-01 -2.03764534e+00 4.22856450e-01 1.80648595e-01
3.36210489e-01 7.73085561e-03 -2.75178730e-01 4.36223000e-01
-1.03326961e-01 4.09811795e-01 -6.43108845e-01 -1.00731456e+00
3.38730693e-01 3.33265632e-01 -4.01658952e-01 9.51222032e-02
4.10262048e-01 1.26252341e+00 -6.88653708e-01 -2.92077273e-01
4.15587604e-01 9.81334865e-01 -9.24761832e-01 1.44294947e-01
-7.38319397e-01 8.10555458e-01 -3.09641361e-01 8.37497056e-01
8.19337964e-01 -3.05390507e-01 -4.80389416e-01 -3.03126305e-01
-1.30398378e-01 2.76055522e-02 -1.08231485e+00 1.91678345e+00
-6.82510436e-01 3.77503633e-01 4.43825215e-01 -1.71341866e-01
1.10187614e+00 4.42960888e-01 5.84112465e-01 -2.72594571e-01
-1.19677544e-01 4.16350216e-01 -5.48241258e-01 1.40246615e-01
5.30890703e-01 -3.29740345e-01 2.57886291e-01 4.75107372e-01
8.12271088e-02 -1.11503279e+00 -1.67011529e-01 3.08682114e-01
7.93356538e-01 5.27140200e-01 -5.64770341e-01 -2.17144191e-01
-1.41613083e-02 -1.30454004e-01 1.79160267e-01 2.95518070e-01
5.55077136e-01 1.01368749e+00 2.60234803e-01 -3.46928596e-01
-1.58799648e+00 -1.61609256e+00 -2.55974948e-01 4.75326627e-01
-2.89627880e-01 -2.80812204e-01 -5.92779934e-01 -2.21453920e-01
2.80176759e-01 1.13120127e+00 -4.71602142e-01 1.59286335e-01
-5.08384109e-01 -2.04528019e-01 4.19237226e-01 5.70795476e-01
1.71094671e-01 -1.27111542e+00 5.10245338e-02 2.59479523e-01
5.81842735e-02 -6.61022365e-01 -5.21370173e-01 -2.39373446e-01
-1.21843231e+00 -4.71922725e-01 -1.08072698e+00 -7.33859539e-01
7.17119455e-01 -7.38672465e-02 1.41282821e+00 1.84486941e-01
-2.49731746e-02 6.56308353e-01 6.63714781e-02 -3.06316495e-01
-9.09893692e-01 -1.00379623e-01 -5.51614612e-02 -1.74952462e-01
-3.07381392e-01 -1.24215031e+00 -6.94802940e-01 2.03603134e-01
-1.23884749e+00 5.28105199e-01 2.46979132e-01 6.99613094e-01
1.10747576e+00 -2.98523277e-01 3.41357380e-01 -4.46032107e-01
6.93499804e-01 -4.00350809e-01 -6.35431826e-01 -2.18304887e-01
-3.60160470e-01 7.97043964e-02 2.87100077e-01 -2.12269768e-01
-9.57769275e-01 5.27783148e-02 -8.43052983e-01 -9.57817316e-01
-1.45058215e-01 2.38123313e-01 -9.35944021e-02 2.18128990e-02
5.91495395e-01 3.15773666e-01 3.26079398e-01 -7.18739569e-01
6.39436126e-01 3.79059970e-01 3.39861184e-01 -9.54050958e-01
8.26692522e-01 6.59335434e-01 1.18783221e-01 -9.32765424e-01
-1.45439571e-02 1.58757836e-01 -5.98772705e-01 -1.40337959e-01
9.81555939e-01 -8.94441426e-01 -5.05571067e-01 6.04811311e-01
-1.36519086e+00 -8.98125529e-01 -6.32991076e-01 -9.08066928e-02
-9.82373774e-01 2.11761981e-01 -8.05668592e-01 -5.49797595e-01
-6.13392711e-01 -1.35854948e+00 1.66465926e+00 -2.19772667e-01
-3.51374209e-01 -1.19309258e+00 -5.08568883e-02 5.38239554e-02
5.33778131e-01 6.11218572e-01 1.04496443e+00 1.85369700e-01
-9.91793633e-01 9.71062481e-03 -8.22024420e-02 4.31140453e-01
6.59582391e-02 3.14142674e-01 -9.91498113e-01 -1.82609424e-01
-1.51312947e-01 -2.25675732e-01 8.43486607e-01 7.06797719e-01
1.23565876e+00 -2.12053046e-01 -3.05840701e-01 1.10713565e+00
1.19447041e+00 -2.73516178e-01 6.96474254e-01 -1.75132036e-01
9.79938686e-01 3.08136225e-01 -6.13701344e-02 3.39334667e-01
3.51447970e-01 6.65436685e-01 4.84828740e-01 -2.83467352e-01
-4.82809007e-01 -4.97816920e-01 2.17646092e-01 1.16860199e+00
-2.84926891e-01 -9.04927626e-02 -9.26212728e-01 5.03659427e-01
-1.66744602e+00 -8.38572323e-01 -4.73228753e-01 1.98148584e+00
9.42698598e-01 1.93540342e-02 -3.67199555e-02 -2.30279073e-01
3.53629410e-01 1.82880118e-01 -5.56949139e-01 -4.46771175e-01
-1.90095320e-01 4.99360651e-01 1.87702045e-01 6.78222358e-01
-7.77682960e-01 1.01896787e+00 5.68813992e+00 1.15093756e+00
-1.17662668e+00 2.19271079e-01 6.74031675e-01 -3.22073698e-02
-1.14836931e+00 -1.88864842e-01 -5.32574177e-01 3.52544039e-01
5.30894458e-01 -3.00853439e-02 5.38664281e-01 8.96732867e-01
3.12495589e-01 3.89139324e-01 -9.72047925e-01 1.17332649e+00
-4.51298177e-01 -1.90635693e+00 4.19748127e-01 3.13725233e-01
1.00869572e+00 2.23438784e-01 2.67092228e-01 1.87290668e-01
5.09688914e-01 -1.27361786e+00 8.73450339e-01 7.50704229e-01
1.24106979e+00 -6.62882328e-01 8.04162174e-02 4.44614917e-01
-1.10210264e+00 5.98944426e-01 -3.81542206e-01 3.05017561e-01
4.34891164e-01 7.22117186e-01 -4.96166945e-01 3.83695126e-01
4.49532360e-01 7.69206166e-01 -1.32527128e-01 7.82383502e-01
-1.70847699e-01 5.05063236e-01 -6.99399352e-01 3.45109046e-01
1.78822741e-01 -4.50717121e-01 9.16088283e-01 9.78606522e-01
6.20011926e-01 6.68077022e-02 1.13660485e-01 1.77511811e+00
-1.38487667e-01 -1.32236093e-01 -8.17031145e-01 -3.09114158e-03
4.03705150e-01 1.10821831e+00 -6.89351618e-01 -3.14514875e-01
3.83517183e-02 1.13561749e+00 7.52120884e-03 3.84185523e-01
-6.64568126e-01 5.06786220e-02 7.67311931e-01 4.99790758e-01
3.30468118e-01 -7.82037735e-01 -7.83386052e-01 -1.07169092e+00
-2.14592278e-01 -6.28929496e-01 -2.23379105e-01 -1.21383607e+00
-1.52531719e+00 8.05907488e-01 -6.41413331e-02 -1.26486290e+00
-2.57931978e-01 -4.51565802e-01 -5.96743584e-01 1.17573857e+00
-1.11003518e+00 -1.68453228e+00 -2.19287485e-01 5.58812439e-01
6.33771241e-01 1.50450263e-02 9.87489462e-01 2.06136048e-01
4.40577716e-02 1.98814005e-01 -3.25868949e-02 -1.05458468e-01
2.98191726e-01 -1.33598697e+00 1.20385242e+00 3.72340143e-01
2.06501797e-01 6.46124005e-01 4.81508195e-01 -1.10134161e+00
-1.59741318e+00 -1.20802951e+00 3.66797030e-01 -8.23577881e-01
2.55027711e-01 -3.32136214e-01 -9.08495307e-01 6.61502361e-01
4.51366119e-02 -8.24221224e-02 3.22572887e-01 -2.53249168e-01
-3.04906182e-02 3.90777767e-01 -1.12689888e+00 7.13814318e-01
1.11922002e+00 -3.85246754e-01 -3.67067575e-01 4.52391565e-01
8.59917939e-01 -7.21854746e-01 -1.19982815e+00 4.39751834e-01
4.51907814e-01 -8.22931826e-01 1.28006244e+00 -1.71544656e-01
8.68454099e-01 -2.76010990e-01 -2.26214886e-01 -1.44273257e+00
-5.44126213e-01 -6.93986893e-01 -3.29421848e-01 1.16167712e+00
4.02479500e-01 -4.33418125e-01 9.08836603e-01 7.49469876e-01
-5.19735575e-01 -9.90157485e-01 -7.70647109e-01 -7.16747701e-01
5.10746181e-01 -7.65459239e-01 8.88074219e-01 9.03366208e-01
-8.88753533e-01 -6.60576764e-03 -1.81212902e-01 -4.07563746e-02
7.68991172e-01 3.09502900e-01 9.03017998e-01 -1.22023499e+00
-3.40764046e-01 -7.07621455e-01 2.27995906e-02 -1.37790239e+00
-1.82150118e-02 -1.24114847e+00 -2.14362845e-01 -2.00452757e+00
-4.40335959e-01 -7.85349309e-01 3.64590734e-01 5.14261365e-01
1.83662951e-01 4.50361818e-01 3.04512382e-01 3.51801634e-01
1.63643554e-01 1.06520045e+00 1.72255170e+00 -1.73994333e-01
-5.25632918e-01 8.36030953e-03 -3.60394418e-01 7.61814058e-01
5.37175715e-01 -3.69562030e-01 -3.02314311e-01 -6.64211214e-01
1.44583806e-01 2.85408676e-01 7.44735837e-01 -8.67064714e-01
-4.02170606e-03 -1.64840728e-01 6.24385476e-01 -8.88532639e-01
7.92023718e-01 -6.22987986e-01 6.81923330e-01 1.44409612e-01
1.83660194e-01 -9.94649231e-02 3.67078900e-01 2.24928752e-01
-4.46371324e-02 1.55501738e-01 8.48481238e-01 -3.31582904e-01
-2.84485012e-01 7.50089407e-01 -7.60784298e-02 -2.77191937e-01
5.76603889e-01 -4.83846098e-01 1.63151309e-01 -5.32918870e-01
-9.25720930e-01 -1.99343842e-02 1.09488690e+00 3.73480767e-01
9.48686004e-01 -1.79615664e+00 -8.79311442e-01 5.56947112e-01
-1.72514528e-01 5.20077646e-01 3.43596965e-01 4.91882771e-01
-8.10389042e-01 -9.09244716e-02 -6.64562657e-02 -8.51946950e-01
-8.64067256e-01 2.48638347e-01 1.62180215e-01 -2.87177935e-02
-8.98315907e-01 1.05229390e+00 3.29722255e-01 -7.98291266e-01
-1.47642136e-01 -5.09338200e-01 4.76025552e-01 -1.35102794e-01
1.56586483e-01 1.92099303e-01 8.05988982e-02 -5.36230743e-01
7.76928663e-02 7.10932136e-01 4.89326686e-01 -3.08827311e-01
1.76114118e+00 2.98339039e-01 -2.50312358e-01 2.93575913e-01
1.05833316e+00 2.29247242e-01 -1.64479160e+00 1.68227956e-01
-5.83639622e-01 -3.49578351e-01 3.59308481e-01 -6.10586703e-01
-1.15200901e+00 9.87503469e-01 2.37618074e-01 4.63367403e-02
8.20450246e-01 5.57230366e-03 1.24611270e+00 -2.57309765e-01
5.21275342e-01 -6.08428478e-01 6.39929995e-02 5.22721410e-01
1.62847292e+00 -6.94165111e-01 -6.26783669e-02 -4.20900524e-01
-5.03743351e-01 1.09006858e+00 1.68618202e-01 -4.31613386e-01
8.90437484e-01 4.81243312e-01 -1.11034244e-01 -4.42377001e-01
-8.34377348e-01 4.97176573e-02 5.35375416e-01 6.86347485e-01
3.52778345e-01 1.63362667e-01 2.52326399e-01 4.09576416e-01
-4.16338354e-01 2.23787967e-02 1.73246637e-01 6.46062851e-01
-8.45352933e-02 -1.45586848e+00 -4.94076073e-01 5.14323950e-01
5.15740104e-02 -2.37438887e-01 -2.42223158e-01 5.13178706e-01
3.00631970e-02 3.44768733e-01 1.74582124e-01 -1.94313854e-01
3.02029371e-01 5.04774414e-02 5.59525907e-01 -6.98189318e-01
-4.59899664e-01 3.85482728e-01 -1.62461996e-01 -6.45341694e-01
6.18460439e-02 -6.38842285e-01 -1.14261127e+00 -5.60768068e-01
6.24912865e-02 -1.42219931e-01 9.05603111e-01 3.22765708e-01
7.64045835e-01 4.43676174e-01 2.93650597e-01 -1.61841834e+00
-1.81815520e-01 -7.84406245e-01 -4.40552205e-01 4.79489654e-01
1.32500380e-01 -6.60346150e-01 -1.76074296e-01 1.92114025e-01] | [8.939461708068848, -3.6220602989196777] |
d5cb9e8b-7159-4d17-b704-a6470fd4842b | transformerfusion-monocular-rgb-scene | 2107.02191 | null | https://arxiv.org/abs/2107.02191v1 | https://arxiv.org/pdf/2107.02191v1.pdf | TransformerFusion: Monocular RGB Scene Reconstruction using Transformers | We introduce TransformerFusion, a transformer-based 3D scene reconstruction approach. From an input monocular RGB video, the video frames are processed by a transformer network that fuses the observations into a volumetric feature grid representing the scene; this feature grid is then decoded into an implicit 3D scene representation. Key to our approach is the transformer architecture that enables the network to learn to attend to the most relevant image frames for each 3D location in the scene, supervised only by the scene reconstruction task. Features are fused in a coarse-to-fine fashion, storing fine-level features only where needed, requiring lower memory storage and enabling fusion at interactive rates. The feature grid is then decoded to a higher-resolution scene reconstruction, using an MLP-based surface occupancy prediction from interpolated coarse-to-fine 3D features. Our approach results in an accurate surface reconstruction, outperforming state-of-the-art multi-view stereo depth estimation methods, fully-convolutional 3D reconstruction approaches, and approaches using LSTM- or GRU-based recurrent networks for video sequence fusion. | ['Matthias Nießner', 'Angela Dai', 'Justus Thies', 'Pablo Palafox', 'Aljaž Božič'] | 2021-07-05 | null | http://proceedings.neurips.cc/paper/2021/hash/0a87257e5308197df43230edf4ad1dae-Abstract.html | http://proceedings.neurips.cc/paper/2021/file/0a87257e5308197df43230edf4ad1dae-Paper.pdf | neurips-2021-12 | ['stereo-depth-estimation', '3d-scene-reconstruction'] | ['computer-vision', 'computer-vision'] | [ 6.16693616e-01 7.22326338e-02 1.47223592e-01 -2.70532519e-01
-1.11123610e+00 -3.42891455e-01 5.90216517e-01 -1.34013832e-01
-2.87482798e-01 4.40592438e-01 4.47938740e-01 1.14446782e-01
8.66500661e-02 -1.08648896e+00 -1.09415615e+00 -6.83978558e-01
1.60128430e-01 6.03900194e-01 4.39376980e-01 -3.19368183e-03
3.80670875e-01 6.32536590e-01 -2.23063397e+00 8.73146355e-01
1.12213567e-01 1.68100882e+00 6.74575925e-01 9.06240046e-01
-1.79435924e-01 1.25988519e+00 1.70493990e-01 1.51425719e-01
3.25307757e-01 -1.51818857e-01 -8.03757668e-01 2.59117752e-01
7.56660819e-01 -7.48434603e-01 -6.59450710e-01 5.97498596e-01
2.54530519e-01 9.78713408e-02 4.41704988e-01 -4.34660047e-01
-3.66939157e-02 -3.46558876e-02 -2.57018715e-01 5.41366860e-02
9.79696035e-01 -1.97806612e-01 7.28460073e-01 -1.30622470e+00
1.00921392e+00 1.35360634e+00 5.78461707e-01 3.21332157e-01
-1.23734152e+00 -1.03717938e-01 -3.23242173e-02 2.75403917e-01
-1.09082127e+00 -5.06407201e-01 9.24154103e-01 -4.41138238e-01
1.70789158e+00 1.62180990e-01 1.05844748e+00 8.70780170e-01
5.96296132e-01 5.08191586e-01 1.25938487e+00 -2.32604310e-01
3.90097886e-01 -3.97233188e-01 -5.14453292e-01 9.51219261e-01
-5.28599620e-01 3.06642681e-01 -1.11366677e+00 1.67439450e-02
1.39479923e+00 4.10726190e-01 -4.28312451e-01 -5.52858055e-01
-1.38111186e+00 6.61010742e-01 5.46193540e-01 2.28312947e-02
-8.67419720e-01 2.81765014e-01 4.03913945e-01 3.18358511e-01
9.35563028e-01 4.86335577e-03 -4.47700351e-01 -3.91547270e-02
-1.20303881e+00 3.63359481e-01 4.83902395e-01 4.44899380e-01
1.21441007e+00 8.01092461e-02 9.29170698e-02 4.24667627e-01
5.46358943e-01 4.64576662e-01 4.36932236e-01 -1.60256743e+00
3.54374528e-01 5.97525239e-01 -3.01302392e-02 -6.44003928e-01
-2.18858495e-01 9.64058787e-02 -7.11313725e-01 5.66397965e-01
-1.53541356e-01 5.43581724e-01 -9.71432805e-01 1.18076849e+00
4.10600007e-01 4.69711900e-01 -3.37367393e-02 9.43110168e-01
8.81336570e-01 6.70103431e-01 -3.88680249e-01 -1.82009354e-01
1.21416533e+00 -5.88423073e-01 -2.16068268e-01 -8.91560838e-02
3.25555831e-01 -5.70736885e-01 4.86304969e-01 4.88781512e-01
-1.45358133e+00 -6.13311350e-01 -8.26535106e-01 -5.70803881e-01
-1.54141247e-01 -3.74111116e-01 3.20309699e-01 -7.14395475e-03
-1.63476896e+00 8.41876686e-01 -9.71254945e-01 -2.10033819e-01
3.78913194e-01 4.10286158e-01 -8.60373735e-01 -2.51611590e-01
-8.23805988e-01 1.01130092e+00 3.50021929e-01 -4.21964526e-02
-1.18351901e+00 -7.14890480e-01 -1.47098053e+00 -1.44174099e-01
1.52104646e-02 -1.30953789e+00 1.24102759e+00 -7.74022102e-01
-1.58092034e+00 1.08575153e+00 -6.97236359e-01 -5.78546107e-01
3.06029141e-01 -1.89141095e-01 3.60438138e-01 5.38063824e-01
8.84821638e-02 8.75179350e-01 1.02510262e+00 -1.44587171e+00
-7.93320000e-01 -7.66910076e-01 5.90361580e-02 6.74991131e-01
4.23489034e-01 -4.03605640e-01 -3.48849893e-01 -3.16307724e-01
6.42041564e-01 -3.43674690e-01 -3.01715106e-01 1.12031274e-01
-7.56824911e-02 9.13150012e-02 1.01110816e+00 -8.59604657e-01
5.66863120e-01 -1.91494691e+00 7.52636671e-01 -1.43013284e-01
4.55932379e-01 -3.57128143e-01 3.77549268e-02 2.77960956e-01
-2.78396346e-03 -4.77633685e-01 -2.30429977e-01 -1.05109143e+00
-4.18297619e-01 2.95718700e-01 -5.09654522e-01 5.78074515e-01
-3.89194973e-02 8.75742018e-01 -9.11222219e-01 -3.26400310e-01
1.18045676e+00 9.57677305e-01 -6.40706062e-01 5.93462944e-01
-3.85818273e-01 6.80840433e-01 -3.77905726e-01 6.75713718e-01
4.40196067e-01 -3.03802550e-01 -2.66337126e-01 -4.50699627e-01
-5.09031057e-01 5.73257506e-01 -7.52680302e-01 2.43845129e+00
-7.84715772e-01 6.29431009e-01 1.27741799e-01 -9.74278450e-01
7.90249407e-01 4.01561260e-01 6.90459311e-01 -8.00491571e-01
1.62183225e-01 2.42979273e-01 -1.10625803e+00 -1.67125925e-01
7.43676603e-01 -1.76121280e-01 3.62231880e-02 2.14568496e-01
2.83789158e-01 -6.37785733e-01 -3.49362016e-01 3.69765772e-03
1.09372759e+00 7.95359612e-01 2.19639003e-01 7.67108649e-02
4.94346052e-01 -1.38802275e-01 8.81466568e-02 4.57730561e-01
3.12103957e-01 1.03972137e+00 -6.96184207e-03 -8.31654727e-01
-1.40260947e+00 -1.13584435e+00 5.84501736e-02 6.82769239e-01
6.21611662e-02 -2.32646212e-01 -5.89075565e-01 -1.66143686e-01
-8.74215588e-02 4.13371921e-01 -8.55602562e-01 7.20487311e-02
-5.33246934e-01 2.47361958e-01 -4.71122228e-02 2.47881755e-01
4.82378632e-01 -1.15975952e+00 -1.38328099e+00 5.09945452e-01
-3.00689369e-01 -1.18802381e+00 -2.89488677e-02 5.53839564e-01
-1.22163391e+00 -9.14407670e-01 -7.43162513e-01 -5.60018718e-01
3.54605168e-01 4.08341438e-01 1.19916797e+00 -3.17798138e-01
-4.72270995e-02 5.55357695e-01 -1.21833652e-01 2.42278561e-01
-3.75287026e-01 -4.56588745e-01 -1.61028989e-02 -1.83025876e-03
-1.50327794e-02 -9.44893003e-01 -7.27451682e-01 -5.78325838e-02
-8.23082745e-01 6.16821706e-01 1.06369182e-01 8.97664607e-01
1.10199678e+00 -3.41587424e-01 -1.91304132e-01 -4.79655147e-01
-2.88964987e-01 -2.59081960e-01 -6.67752624e-01 -8.48762617e-02
7.48129338e-02 1.20333716e-01 5.93738079e-01 2.14725524e-01
-9.84874189e-01 4.95475709e-01 -3.66880387e-01 -1.24864578e+00
-3.16825390e-01 3.39442760e-01 8.14073756e-02 -2.65075654e-01
4.21346217e-01 7.20590293e-01 -1.31390831e-02 -3.88834983e-01
3.22683692e-01 3.52346510e-01 7.60040641e-01 -6.03137836e-02
3.20691824e-01 7.94027328e-01 1.52211100e-01 -8.35354567e-01
-1.01115155e+00 -4.92348552e-01 -1.11559951e+00 -4.97070044e-01
1.18281245e+00 -1.42350364e+00 -6.82905257e-01 6.55701339e-01
-1.48455894e+00 -4.31966126e-01 -5.98993301e-01 3.06706131e-01
-1.33161449e+00 3.76284689e-01 -7.80295014e-01 -6.97299123e-01
-2.99729943e-01 -1.30818582e+00 2.07008362e+00 -2.51142472e-01
-1.25742272e-01 -9.57050562e-01 1.87896848e-01 5.23886561e-01
1.50430962e-01 5.44688702e-01 5.92809677e-01 2.48888046e-01
-1.11180234e+00 1.77090779e-01 -1.32350281e-01 1.12962984e-01
2.73609143e-02 -3.53028685e-01 -1.46066773e+00 -1.20992452e-01
3.35934699e-01 -5.22150457e-01 1.12508285e+00 7.55288243e-01
1.15435100e+00 -6.50651753e-02 -3.09174269e-01 1.00862575e+00
1.52976525e+00 -1.71709359e-01 6.96864247e-01 2.56103903e-01
9.68209505e-01 3.12211901e-01 3.69279832e-01 6.62258446e-01
7.67885149e-01 6.62343085e-01 8.67326379e-01 -1.18090816e-01
-2.19522983e-01 -5.09261131e-01 4.21107829e-01 6.01433873e-01
-1.83516413e-01 1.44070769e-02 -6.82524502e-01 3.43740761e-01
-1.74958289e+00 -1.06162643e+00 2.56366342e-01 2.35856771e+00
4.55500960e-01 -5.40132970e-02 -2.17154160e-01 3.95607978e-01
2.10410833e-01 4.77923185e-01 -5.96094310e-01 -2.80835241e-01
-1.98939890e-01 2.68393099e-01 5.78395963e-01 7.59028316e-01
-9.56392765e-01 1.06558108e+00 6.23821402e+00 5.51838398e-01
-1.38912153e+00 1.37300089e-01 7.17782259e-01 -2.37961814e-01
-6.58413887e-01 -1.60511687e-01 -3.72677386e-01 1.30218878e-01
1.01802838e+00 3.50823998e-01 7.15463638e-01 5.63067019e-01
2.13502914e-01 -4.30119902e-01 -1.19941401e+00 1.50652242e+00
1.74153671e-01 -1.83810103e+00 3.11348736e-01 2.20295087e-01
5.77075064e-01 6.84802830e-01 -2.44641855e-01 -9.71651003e-02
2.56412923e-01 -1.03974664e+00 1.17210639e+00 8.95826697e-01
1.09299564e+00 -6.39076889e-01 3.40331912e-01 4.40142006e-01
-1.50099277e+00 2.78370231e-02 -3.63770455e-01 -2.56102979e-01
6.14069104e-01 5.47677994e-01 -4.89676744e-01 7.74652898e-01
1.06142235e+00 1.17254472e+00 -1.56116590e-01 5.34456193e-01
2.03493521e-01 -1.75700143e-01 -3.72610122e-01 5.98434925e-01
1.76632315e-01 -3.83174531e-02 4.60292786e-01 8.64819467e-01
7.18191564e-01 2.23168686e-01 2.13731885e-01 7.71041334e-01
2.61837337e-02 -4.18901116e-01 -1.16947484e+00 4.56860244e-01
2.30837330e-01 1.02866232e+00 -5.30953705e-01 -5.79407394e-01
-3.65616411e-01 1.52708900e+00 5.37461162e-01 2.11884946e-01
-2.46102124e-01 3.82298708e-01 5.63255787e-01 2.22229108e-01
6.12202942e-01 -2.30497792e-01 -2.33968869e-01 -1.19739962e+00
-8.25711247e-03 -3.79949510e-01 1.71687678e-01 -1.37709379e+00
-9.21420515e-01 8.56237710e-01 -2.62718827e-01 -1.32000661e+00
-7.36858308e-01 -5.62193215e-01 -1.39260396e-01 1.01169658e+00
-1.60399532e+00 -1.22915256e+00 -4.46332604e-01 8.72178257e-01
6.68634772e-01 5.93919680e-02 1.06202281e+00 -1.44945368e-01
3.69275510e-01 -2.88886189e-01 -1.05873510e-01 -4.88485366e-01
1.34810656e-01 -1.10585177e+00 6.82847738e-01 3.79764527e-01
1.03996716e-01 -1.21691585e-01 3.98371100e-01 -4.86213893e-01
-1.81456709e+00 -1.13109410e+00 9.83797252e-01 -7.07462907e-01
2.24342033e-01 -4.83762771e-01 -7.83434272e-01 6.47787511e-01
4.57483083e-02 3.82851630e-01 2.26581082e-01 -4.56891119e-01
-3.37796390e-01 1.33458361e-01 -1.30073929e+00 1.93082228e-01
1.22862673e+00 -1.17423868e+00 -5.70358157e-01 4.73661609e-02
7.15352476e-01 -9.15862560e-01 -1.09892738e+00 3.82774949e-01
6.13416970e-01 -1.61311412e+00 1.09913552e+00 8.26653093e-02
8.09151173e-01 -3.51119846e-01 -7.91921854e-01 -1.16635931e+00
-2.59878844e-01 -2.31265321e-01 -3.77285212e-01 3.19141924e-01
-8.36605877e-02 -3.61662865e-01 9.42182899e-01 2.25158110e-01
-4.13016647e-01 -8.42855752e-01 -1.53974676e+00 2.39700973e-02
-4.12386060e-01 -7.11839318e-01 2.67684817e-01 5.22171021e-01
-2.78463125e-01 1.89978153e-01 -4.00677323e-01 9.72897410e-02
8.45832169e-01 3.60701829e-01 5.36546648e-01 -1.17519093e+00
-2.52437025e-01 -2.31709823e-01 -7.61698365e-01 -1.44033349e+00
3.13135266e-01 -7.12633193e-01 9.68917981e-02 -1.80184996e+00
6.73709586e-02 1.13149630e-02 -1.83353350e-02 3.43821824e-01
4.06349391e-01 6.88376307e-01 4.15165946e-02 1.69649869e-01
-5.57108939e-01 8.37589204e-01 1.21975708e+00 5.22465408e-02
2.44635921e-02 -4.95959252e-01 -1.55385286e-02 6.22703195e-01
1.08421817e-01 -7.81280026e-02 -2.01320544e-01 -5.68019748e-01
1.00257054e-01 8.15829754e-01 8.05780411e-01 -1.09071362e+00
1.98186681e-01 2.69761920e-01 7.12922633e-01 -1.15790904e+00
1.18165076e+00 -9.17913854e-01 3.64014834e-01 2.61307716e-01
3.42857726e-02 3.63768861e-02 1.26475349e-01 4.68156099e-01
-3.74779731e-01 3.85091454e-01 6.18930876e-01 -6.14746869e-01
-9.16026175e-01 5.97070813e-01 -4.55341458e-01 -5.28594553e-01
7.19339788e-01 -7.91167676e-01 3.77088845e-01 -4.25795466e-01
-7.54633784e-01 -2.29581699e-01 9.57297623e-01 3.24771643e-01
1.13707829e+00 -1.28147352e+00 -5.84565163e-01 6.30403757e-01
-7.49479756e-02 5.93953669e-01 6.32287383e-01 4.78712499e-01
-4.91826177e-01 5.58758080e-01 -2.96252728e-01 -1.19693947e+00
-1.05331075e+00 3.63478571e-01 7.03043103e-01 -2.15402663e-01
-1.02331269e+00 8.41397047e-01 5.34045100e-01 -4.94424880e-01
-3.29218321e-02 -5.93273997e-01 -5.24088815e-02 -9.01121050e-02
5.06114721e-01 9.48352143e-02 3.00655812e-01 -1.05025434e+00
-1.64741024e-01 1.00906599e+00 3.48378271e-01 -4.09152955e-01
1.48247278e+00 -3.99167389e-01 -1.74209759e-01 8.33905935e-01
1.34766519e+00 -5.93705058e-01 -1.90727246e+00 -3.04596543e-01
-5.07006705e-01 -5.43706357e-01 5.35771966e-01 -3.24437201e-01
-9.49957609e-01 8.19681585e-01 5.94721317e-01 -1.62076488e-01
1.16086066e+00 1.61508381e-01 7.10006177e-01 1.35079071e-01
1.02509046e+00 -4.68760252e-01 -9.98554677e-02 8.57105553e-01
9.13631260e-01 -1.11220062e+00 -3.83087955e-02 -1.66475594e-01
-2.07924008e-01 1.21903837e+00 1.51219741e-01 -3.07533056e-01
6.62258267e-01 2.72544205e-01 -1.48682371e-01 -4.88958865e-01
-9.51115608e-01 -1.72738899e-02 2.79239416e-01 5.02690434e-01
3.80797327e-01 -2.22491056e-01 8.40079904e-01 -2.37705112e-01
-3.17518800e-01 2.30137601e-01 2.46459454e-01 7.87316740e-01
-5.45366347e-01 -7.48468935e-01 -3.11679244e-01 4.93329167e-01
5.29892035e-02 -2.00701252e-01 -3.48732173e-02 2.85682052e-01
5.85479662e-02 5.26779234e-01 5.83970547e-01 -3.93999785e-01
3.09972733e-01 -5.34545165e-03 9.79602933e-01 -5.85233271e-01
-4.86213893e-01 2.72948056e-01 -1.26768291e-01 -1.36133873e+00
-6.35751367e-01 -8.54444563e-01 -1.25033462e+00 -2.08423689e-01
2.02995762e-01 -2.17180267e-01 5.84833384e-01 1.11942589e+00
3.70532393e-01 4.66271400e-01 7.12184846e-01 -2.01671839e+00
1.84728086e-01 -7.19213247e-01 -4.99487281e-01 3.80805321e-02
8.44471037e-01 -6.86212540e-01 -1.40433088e-01 2.38206431e-01] | [8.655566215515137, -2.8701303005218506] |
19a16e18-322e-4c34-8cee-6d24ec3c75b0 | task-aware-multi-task-learning-for-speech-to | null | null | https://ieeexplore.ieee.org/document/9414703 | https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9414703 | TASK AWARE MULTI-TASK LEARNING FOR SPEECH TO TEXT TASKS | In general, the direct Speech-to-text translation (ST) is jointly trained with Automatic Speech Recognition (ASR), and Machine Translation (MT) tasks. However, the issues with the current joint learning strategies inhibit the knowledge transfer across these tasks. We propose a task modulation network which allows the model to learn task specific features, while learning the shared features simultaneously. This proposed approach removes the need for separate finetuning step resulting in a single model which performs all these tasks. This single model achieves a performance of 28.64 BLEU score on ST MuST-C English-German, WER of 11.61% on ASR TEDLium v3, 23.35 BLEU score on MT WMT’15 English-German task. This sets a new state-of-the-art performance (SOTA) on the ST task while outperforming the existing end-to-end ASR systems. | ['Inchul Hwang', 'Chanwoo Kim', 'Sangha Kim', 'Seokchan Ahn', 'Hyojung Han', 'Beomseok Lee', 'Nikhil Kumar Lakumarapu', 'Mohd Abbas Zaidi', 'Sathish Indurthi'] | 2021-06-10 | null | null | null | icassp-2021-6 | ['speech-to-text-translation'] | ['natural-language-processing'] | [ 5.03800631e-01 3.34754825e-01 -3.72154146e-01 -4.40956533e-01
-1.66767001e+00 -5.21057069e-01 1.06475353e+00 -5.23700655e-01
-6.85771227e-01 9.18413103e-01 2.64061570e-01 -7.85721660e-01
4.17488068e-01 5.48173711e-02 -6.71723008e-01 -6.08658075e-01
5.35373688e-01 6.61867499e-01 -6.70133252e-03 -2.39108592e-01
-1.73758164e-01 1.25213325e-01 -7.95224547e-01 6.13156438e-01
1.07291234e+00 8.07230055e-01 7.26028144e-01 9.34974670e-01
3.75373773e-02 5.58971763e-01 -8.07587624e-01 -5.25160909e-01
2.90245451e-02 -4.12104756e-01 -8.58492911e-01 -7.52392784e-02
5.01737237e-01 -1.84260234e-02 -5.89845419e-01 7.17963278e-01
7.88542211e-01 9.68508497e-02 7.45155632e-01 -6.82318330e-01
-1.05695403e+00 8.53117824e-01 -2.46506482e-01 2.94988722e-01
4.56671603e-02 7.72503093e-02 1.02052367e+00 -1.39424360e+00
3.63681883e-01 1.30369449e+00 2.44360372e-01 9.23978925e-01
-1.04886281e+00 -5.93211949e-01 -3.03039290e-02 6.97659329e-02
-1.30195653e+00 -1.19254422e+00 2.72609651e-01 -1.45475104e-01
1.70812237e+00 4.12344754e-01 2.93984916e-02 1.51008868e+00
1.80753544e-01 1.01876915e+00 1.20533049e+00 -6.80657387e-01
-1.00621104e-01 7.81970024e-02 -3.24006915e-01 4.67706382e-01
-1.74412429e-01 1.23098604e-01 -1.00553012e+00 2.86316901e-01
5.59161067e-01 -5.37430942e-01 -3.26964736e-01 2.63599694e-01
-1.61975420e+00 5.12416601e-01 -6.73937285e-03 6.11141026e-01
-2.09651679e-01 9.26752388e-02 4.34352726e-01 7.80704439e-01
6.99532866e-01 3.12919766e-01 -9.61282134e-01 -4.74529356e-01
-9.27858174e-01 -3.38313580e-01 6.28952205e-01 1.08642876e+00
5.00414729e-01 5.00243545e-01 -7.17849076e-01 1.20029950e+00
3.41552734e-01 1.09092855e+00 6.94880307e-01 -5.27670741e-01
1.20508981e+00 1.39210433e-01 -1.37372753e-02 -4.30427976e-02
6.28190786e-02 -7.30032742e-01 -7.66187727e-01 -3.59390587e-01
2.21383810e-01 -4.31064844e-01 -1.19906580e+00 1.84084558e+00
-2.27769285e-01 -1.51474401e-01 4.76063520e-01 7.13798285e-01
7.20993459e-01 1.11017346e+00 1.73067790e-03 -4.70601171e-01
1.07033432e+00 -1.49519801e+00 -1.22839212e+00 -6.53121233e-01
9.08735394e-01 -1.31859314e+00 1.16408646e+00 1.19449422e-01
-1.35350215e+00 -6.70362055e-01 -9.79161263e-01 -2.36135527e-01
-2.06923991e-01 7.88731694e-01 8.20277557e-02 6.79226995e-01
-1.35337949e+00 3.25117111e-01 -8.31357241e-01 -4.65195358e-01
-6.37274683e-02 5.02296627e-01 -3.27628613e-01 -9.55605228e-03
-1.42050898e+00 1.54711902e+00 2.91716486e-01 1.20335951e-01
-1.04872680e+00 -3.78413349e-01 -7.27131128e-01 -6.62161782e-02
1.76533088e-01 -6.54757798e-01 1.69214714e+00 -1.05055606e+00
-2.25536609e+00 9.33373213e-01 -5.03939092e-01 -4.78991240e-01
5.08847713e-01 -4.98721868e-01 -7.29521751e-01 -2.27547869e-01
-1.91358268e-01 6.10595644e-01 8.94699395e-01 -7.38833964e-01
-5.37041724e-01 -1.46031782e-01 -5.40664196e-01 6.48174763e-01
-2.67401159e-01 4.53353405e-01 -2.62749642e-01 -7.39022613e-01
-3.95211056e-02 -9.11270797e-01 2.89507974e-02 -7.42588341e-01
-2.75622964e-01 -3.82550329e-01 7.22371995e-01 -1.13392758e+00
1.45754385e+00 -1.95025980e+00 4.00360912e-01 -4.40441459e-01
-4.91783172e-01 8.30510974e-01 -4.80628401e-01 6.05034530e-01
1.24187795e-02 1.15624122e-01 -2.23849416e-01 -7.90620863e-01
2.55752751e-03 1.16528004e-01 -1.61174178e-01 2.57557780e-01
2.74429411e-01 1.30136764e+00 -5.94156802e-01 -1.56067520e-01
2.94306964e-01 3.27945381e-01 -7.03549013e-02 5.27900875e-01
-1.02147110e-01 6.36900485e-01 -2.11505935e-01 3.97675127e-01
3.32910001e-01 7.71894306e-02 2.34080464e-01 2.74356276e-01
-2.07032636e-01 1.13684070e+00 -4.87860739e-01 2.06585598e+00
-8.57417226e-01 7.56461918e-01 5.01445308e-02 -9.10446227e-01
1.13741422e+00 9.20182407e-01 -8.56588632e-02 -8.21060777e-01
1.30064203e-03 7.72332072e-01 2.51887739e-01 -2.00202554e-01
2.67062992e-01 -5.59267700e-02 -1.45747112e-02 2.69528300e-01
4.19660121e-01 -1.69318169e-01 -2.96644121e-01 -1.38211489e-01
1.04056013e+00 5.58350943e-02 3.69720876e-01 -3.53914678e-01
6.03801668e-01 -1.95134178e-01 4.70062852e-01 6.28442705e-01
-1.64746359e-01 5.47214568e-01 -2.28259772e-01 -1.72081396e-01
-1.08942175e+00 -1.23626268e+00 2.03777269e-01 1.24519682e+00
-4.73960012e-01 -3.55556339e-01 -8.77435565e-01 -7.93496728e-01
-4.43479240e-01 1.07205462e+00 -1.65288553e-01 -2.74347156e-01
-9.39555705e-01 -3.95246357e-01 8.23595464e-01 2.16709033e-01
5.10868669e-01 -1.10115278e+00 1.83894679e-01 2.82418281e-01
-7.43749917e-01 -1.43559039e+00 -1.04259086e+00 2.36499995e-01
-9.61786449e-01 -2.28092566e-01 -9.75599706e-01 -1.03326476e+00
8.98538679e-02 2.90409237e-01 1.07648468e+00 -5.31555355e-01
3.92438740e-01 -1.13992486e-03 -3.46567929e-01 -1.46889612e-01
-9.04672682e-01 6.05810583e-01 2.43401915e-01 3.79238948e-02
3.81726682e-01 -3.64586532e-01 -1.57251686e-01 3.86574149e-01
-3.62530291e-01 2.95108229e-01 1.17224121e+00 9.73803580e-01
4.24960613e-01 -8.26909244e-01 9.27735865e-01 -3.67897332e-01
4.76429790e-01 -9.95260999e-02 -3.15155596e-01 6.35069609e-01
-6.41520083e-01 7.40917306e-03 4.93613660e-01 -5.81210375e-01
-1.27063382e+00 3.00030690e-03 -1.52228624e-01 -2.05333978e-01
-3.39336544e-02 4.15137500e-01 -4.47282761e-01 2.76548594e-01
6.60945296e-01 7.42995620e-01 -9.71661434e-02 -6.84160113e-01
4.69707966e-01 1.52043545e+00 4.00974602e-01 -3.32591712e-01
7.06374943e-01 -3.71065378e-01 -5.39987981e-01 -8.63057375e-01
-8.02252531e-01 -3.85591537e-01 -6.40999258e-01 1.69586167e-01
9.51956570e-01 -1.31049669e+00 -1.30229279e-01 4.04754072e-01
-1.55138159e+00 -6.46618068e-01 6.27425238e-02 7.86209702e-01
-8.63137126e-01 1.74301639e-01 -7.07153320e-01 -6.74331903e-01
-7.30503321e-01 -1.28382850e+00 1.17510819e+00 -3.94542366e-01
-2.74801314e-01 -7.24791348e-01 3.70435752e-02 8.32092583e-01
6.19854569e-01 -7.33678102e-01 7.46520281e-01 -8.85341942e-01
-5.04025877e-01 2.67737091e-01 -1.20893359e-01 8.56800973e-01
4.11624461e-01 -3.70449871e-01 -1.09232283e+00 -5.73297322e-01
-3.12137343e-02 -3.52512360e-01 7.19538271e-01 3.34508538e-01
5.36126256e-01 -4.49882299e-01 -4.17827554e-02 4.75961357e-01
9.68090653e-01 3.20356488e-01 5.45805991e-01 1.99737288e-02
6.68702602e-01 3.52219105e-01 6.24679387e-01 -1.72841981e-01
3.15471560e-01 1.10069525e+00 -1.15907930e-01 1.11208279e-02
-8.51873457e-01 -2.95488328e-01 1.22793639e+00 1.75111032e+00
1.77237019e-01 -4.80277896e-01 -9.44006264e-01 5.94504476e-01
-1.82214880e+00 -5.45456231e-01 -5.88373132e-02 2.21349502e+00
1.15120709e+00 4.01045121e-02 -2.40209535e-01 -2.60096997e-01
8.95839930e-01 2.45395988e-01 -1.72353953e-01 -8.70389223e-01
-2.60315627e-01 2.87706792e-01 5.43009102e-01 8.84544671e-01
-8.10056806e-01 1.65924823e+00 6.44389677e+00 1.26058829e+00
-1.09605849e+00 6.69913769e-01 4.56083953e-01 5.80299571e-02
-1.65590331e-01 -1.37575299e-01 -9.13260579e-01 2.05658108e-01
1.64594710e+00 -1.85810477e-01 7.52378762e-01 4.60818946e-01
4.36715037e-01 2.86084622e-01 -1.22955632e+00 1.21226037e+00
3.93300876e-02 -1.09866500e+00 2.69290507e-01 7.78010190e-02
9.22302246e-01 5.57445467e-01 2.40308657e-01 6.57392919e-01
2.61087716e-01 -1.22433507e+00 7.00761616e-01 1.26309618e-01
1.27457774e+00 -5.88129342e-01 5.93754888e-01 5.76439202e-01
-8.79869044e-01 1.81710407e-01 -2.03621432e-01 5.66170104e-02
1.31016061e-01 4.53959197e-01 -1.21446931e+00 6.27396762e-01
1.00706041e-01 4.01094228e-01 -2.83987373e-01 4.86720949e-01
-4.61448818e-01 1.00287318e+00 -1.33066326e-01 -1.24285161e-01
3.44679564e-01 1.01001188e-01 8.02647710e-01 1.65270185e+00
5.55173814e-01 -5.22451997e-01 5.50351106e-02 3.58856201e-01
-3.90350580e-01 2.37674057e-01 -5.78148544e-01 -2.82285571e-01
5.82965553e-01 8.52049172e-01 1.16039567e-01 -4.20126170e-01
-4.36641455e-01 1.37891209e+00 5.79208314e-01 6.05023503e-01
-5.54028690e-01 -9.74062458e-02 7.37436175e-01 -2.65050918e-01
2.51122773e-01 -7.07147539e-01 -4.15813446e-01 -1.26063955e+00
2.20189810e-01 -1.19665217e+00 -1.83323368e-01 -5.13874233e-01
-1.03330266e+00 8.49155486e-01 -4.14272547e-01 -1.12582421e+00
-5.42197049e-01 -5.10372460e-01 -3.09763163e-01 1.31319058e+00
-1.67807233e+00 -1.42450142e+00 5.08273363e-01 5.03916204e-01
1.28665781e+00 -6.32782638e-01 9.46067810e-01 2.66764998e-01
-4.84929025e-01 9.63240683e-01 3.76296073e-01 1.31199155e-02
1.12525344e+00 -1.19913340e+00 9.82311845e-01 9.95994329e-01
3.16834867e-01 4.15505618e-01 3.49490643e-01 -6.12787545e-01
-1.30845737e+00 -1.12962699e+00 1.78629577e+00 -5.79184592e-01
7.13649571e-01 -7.00954676e-01 -6.32206082e-01 7.40810156e-01
6.42832160e-01 -1.93382204e-01 4.62680966e-01 1.54540360e-01
-2.89432675e-01 -8.58526528e-02 -8.74375582e-01 7.18025029e-01
1.22359383e+00 -9.19973671e-01 -7.21442699e-01 3.70449871e-01
1.14768398e+00 -2.58177161e-01 -8.38061869e-01 3.69177669e-01
2.94425637e-01 -2.73865670e-01 6.35827601e-01 -7.22461402e-01
9.66566056e-02 -5.80681376e-02 -6.45538330e-01 -1.76528180e+00
-2.88758516e-01 -1.28432167e+00 -3.33860308e-01 1.14677942e+00
1.05327642e+00 -6.47545576e-01 1.74192354e-01 1.44495338e-01
-7.02535272e-01 -4.49927688e-01 -1.49940407e+00 -1.18449998e+00
3.15127879e-01 -2.18822017e-01 3.32944959e-01 9.40408587e-01
-1.27510443e-01 9.66650546e-01 -4.92851108e-01 3.16957943e-02
1.65835425e-01 -4.60229337e-01 5.92157602e-01 -7.82456160e-01
-3.62310439e-01 -3.17076862e-01 1.86945081e-01 -1.33109581e+00
2.21562728e-01 -1.21795917e+00 2.11704910e-01 -1.51918876e+00
4.11819741e-02 -1.09122902e-01 -3.49150717e-01 4.82308656e-01
-2.01888412e-01 -1.73470423e-01 3.20022702e-01 2.78754979e-01
-4.38295931e-01 8.17554235e-01 1.53786588e+00 -5.64962290e-02
-1.47846967e-01 4.40748148e-02 -3.71548116e-01 2.38608420e-01
9.98427868e-01 -6.00512087e-01 -3.32080185e-01 -1.12750518e+00
-3.75647396e-01 2.61213362e-01 -2.82267481e-01 -8.20910037e-01
1.60694003e-01 -2.02848092e-01 2.40387134e-02 -4.75628316e-01
5.20437837e-01 -5.03407598e-01 -1.75022691e-01 2.63219804e-01
-5.14630198e-01 1.23806469e-01 7.50923529e-02 1.99728534e-01
-3.04203779e-01 8.13061222e-02 9.00256872e-01 3.52827720e-02
-2.54524022e-01 1.13937430e-01 -5.55879891e-01 8.57244655e-02
5.65218151e-01 2.21828697e-03 -3.59959632e-01 -4.04657483e-01
-7.11125076e-01 -2.23829653e-02 2.67712609e-03 8.04608107e-01
4.37396258e-01 -1.39725578e+00 -1.30720878e+00 1.91230372e-01
6.25502318e-02 -3.34446311e-01 -3.04917730e-02 9.37594175e-01
5.96499257e-02 9.26783860e-01 1.66107923e-01 -4.68403906e-01
-1.09981215e+00 7.93549642e-02 3.21729124e-01 -5.37122130e-01
-2.34637931e-01 7.77236521e-01 8.46147463e-02 -7.34483302e-01
1.82316050e-01 -1.82163045e-02 9.39539671e-02 -2.84079283e-01
3.42434257e-01 3.82572323e-01 5.53291023e-01 -7.94375241e-01
-3.74757320e-01 2.37699524e-01 -3.68312091e-01 -7.77239561e-01
9.46383059e-01 -5.90771556e-01 1.48349762e-01 5.34720898e-01
1.30251718e+00 -1.40357167e-02 -9.63903010e-01 -6.38799787e-01
1.72467247e-01 -1.69907406e-01 2.76943922e-01 -1.52603734e+00
-5.38875341e-01 1.18496585e+00 5.35612404e-01 -3.42610449e-01
8.11622798e-01 -6.65767714e-02 1.05834818e+00 7.46816278e-01
1.63137361e-01 -1.40134645e+00 -2.15709396e-02 1.15944910e+00
1.28788984e+00 -1.22808492e+00 -6.35385573e-01 -3.12987417e-01
-7.77661204e-01 9.41190541e-01 4.48510766e-01 2.24955782e-01
2.08575144e-01 1.47551820e-01 3.96566033e-01 5.34879744e-01
-1.13924646e+00 -2.98069358e-01 6.15232587e-01 6.07819676e-01
8.16186190e-01 4.90052044e-01 -4.62718934e-01 4.07321930e-01
-2.31212229e-01 -1.73041493e-01 4.89604920e-02 5.75554729e-01
-5.56387126e-01 -1.36287963e+00 -8.13823640e-02 -1.32554872e-02
-6.76481009e-01 -5.70700347e-01 -6.12057328e-01 4.24588233e-01
-5.04528284e-01 1.41972184e+00 -2.81059593e-01 -4.97270137e-01
2.57073820e-01 5.31941593e-01 4.11689430e-01 -8.78285587e-01
-7.84312487e-01 3.10197920e-01 6.36084199e-01 -2.90756881e-01
-5.06340042e-02 -5.24008453e-01 -7.95979440e-01 4.86631040e-03
-3.33522022e-01 8.77125934e-02 1.01705492e+00 1.07803249e+00
5.30140996e-01 4.56281632e-01 8.59274507e-01 -4.77581799e-01
-9.10303473e-01 -1.69315708e+00 -8.55360776e-02 -1.02629013e-01
4.94390190e-01 -2.46508896e-01 -2.61945754e-01 -8.78088698e-02] | [14.491800308227539, 7.230256080627441] |
15759296-399d-4098-8f3d-7dc3dfb1f7dc | multilingual-coreference-resolution-in | 2208.01307 | null | https://arxiv.org/abs/2208.01307v2 | https://arxiv.org/pdf/2208.01307v2.pdf | Multilingual Coreference Resolution in Multiparty Dialogue | Existing multiparty dialogue datasets for entity coreference resolution are nascent, and many challenges are still unaddressed. We create a large-scale dataset, Multilingual Multiparty Coref (MMC), for this task based on TV transcripts. Due to the availability of gold-quality subtitles in multiple languages, we propose reusing the annotations to create silver coreference resolution data in other languages (Chinese and Farsi) via annotation projection. On the gold (English) data, off-the-shelf models perform relatively poorly on MMC, suggesting that MMC has broader coverage of multiparty coreference than prior datasets. On the silver data, we find success both using it for data augmentation and training from scratch, which effectively simulates the zero-shot cross-lingual setting. | ['Benjamin Van Durme', 'Mahsa Yarmohammadi', 'Patrick Xia', 'Boyuan Zheng'] | 2022-08-02 | null | null | null | null | ['coreference-resolution'] | ['natural-language-processing'] | [-1.15794204e-02 4.63828325e-01 -3.71150464e-01 -3.58238041e-01
-1.76910233e+00 -1.02884901e+00 7.49626577e-01 -2.93127567e-01
-6.29422188e-01 9.37659979e-01 1.24446416e+00 -7.25627616e-02
2.77014792e-01 -9.91474912e-02 -4.57073241e-01 -1.64017767e-01
3.31159741e-01 1.26128805e+00 2.29562055e-02 -6.92033827e-01
-3.77575070e-01 -2.14562789e-01 -1.05733371e+00 8.15500200e-01
7.91016698e-01 1.11115813e-01 4.25982401e-02 4.22386229e-01
-2.50834674e-01 5.61531007e-01 -2.82791227e-01 -9.96655405e-01
1.58309594e-01 -4.17050391e-01 -1.49972701e+00 -6.48906529e-01
5.99009275e-01 5.11454903e-02 -2.86025912e-01 8.56052399e-01
6.72185600e-01 1.00968853e-01 1.80500954e-01 -1.02827549e+00
-4.19537932e-01 1.36318552e+00 -4.39203441e-01 -5.86796738e-02
6.54092848e-01 -2.43154258e-01 1.54706204e+00 -8.74567568e-01
1.36962140e+00 1.38274896e+00 9.30920720e-01 8.12561989e-01
-1.36080587e+00 -9.14434612e-01 -7.81410858e-02 3.78288291e-02
-1.18373740e+00 -1.00873017e+00 5.30768454e-01 -1.03699066e-01
1.14915812e+00 3.96916300e-01 1.55742586e-01 1.72162414e+00
-5.75385034e-01 1.01063621e+00 9.85628664e-01 -6.52967274e-01
-5.12984633e-01 3.36866006e-02 2.09877968e-01 1.56014994e-01
-3.24456632e-01 -5.16531765e-02 -7.05811560e-01 -4.21149015e-01
1.48187265e-01 -6.62214220e-01 -7.19726443e-01 -4.75291125e-02
-1.09220636e+00 9.05876994e-01 4.33395356e-02 3.29729855e-01
1.12451343e-02 -3.77728939e-01 5.63753784e-01 4.30997580e-01
3.69793981e-01 5.96973836e-01 -8.69774580e-01 -4.96344000e-01
-6.42661691e-01 2.97793001e-01 1.24803555e+00 1.30045462e+00
5.17607093e-01 -6.61877632e-01 -8.62452909e-02 1.38707399e+00
2.22538337e-02 5.32586575e-01 1.11500084e-01 -1.42837453e+00
1.32655501e+00 3.62516284e-01 1.90447003e-01 -4.50357795e-01
-6.28024161e-01 -1.23183737e-02 -4.97044146e-01 -5.84384978e-01
6.42514288e-01 -4.55434382e-01 -2.91456908e-01 2.21113539e+00
2.64328331e-01 -2.32900962e-01 6.09137893e-01 9.12096977e-01
1.17425656e+00 5.39165616e-01 1.56687692e-01 -4.35521394e-01
1.44005120e+00 -1.07650137e+00 -9.42914903e-01 -3.91638100e-01
8.55188429e-01 -9.26172674e-01 1.02939880e+00 8.68453830e-02
-9.87426400e-01 5.76493889e-03 -4.54136401e-01 -4.24467385e-01
-1.01527542e-01 -1.79718703e-01 8.52242053e-01 2.70854115e-01
-7.60372341e-01 3.14175904e-01 -5.59933186e-01 -5.72920263e-01
-1.53964937e-01 1.08938798e-01 -8.16667080e-01 -1.72527432e-01
-1.71619844e+00 1.01588416e+00 3.78668487e-01 -1.46828622e-01
-4.77857649e-01 -8.21693063e-01 -9.42643225e-01 -1.40042871e-01
4.51754659e-01 -1.68831393e-01 1.48486137e+00 -6.18801415e-01
-1.36215448e+00 1.21219957e+00 -1.77063912e-01 -3.05007458e-01
4.25570339e-01 -4.57672775e-01 -7.17704654e-01 -8.79747644e-02
3.76267344e-01 8.65565062e-01 -1.12531759e-01 -1.27864742e+00
-8.01937342e-01 -3.34411561e-01 1.04333580e-01 4.65786755e-01
-5.33204414e-02 5.08686543e-01 -9.95354891e-01 -3.21842402e-01
-1.38141751e-01 -1.25746810e+00 1.04345441e-01 -9.46370959e-01
-4.76422131e-01 -2.81684250e-01 6.49331987e-01 -7.07826138e-01
1.18579841e+00 -1.93334997e+00 2.50633329e-01 -1.92568406e-01
-2.75095582e-01 2.48931766e-01 -4.98899907e-01 6.52425170e-01
-1.38295561e-01 5.36389686e-02 -1.25380814e-01 -6.72342539e-01
-1.31540611e-01 2.78240681e-01 -2.19661519e-01 2.11137906e-01
-1.24217086e-01 8.37863147e-01 -1.00281250e+00 -7.76311338e-01
-3.62873495e-01 1.94366708e-01 -9.98848200e-01 1.27295032e-01
-4.66332063e-02 6.93722844e-01 -2.16204911e-01 5.21551490e-01
5.02767324e-01 9.76868272e-02 9.07259941e-01 -4.16844636e-01
-4.40580100e-01 6.62757277e-01 -9.69459593e-01 2.37704420e+00
-6.66559517e-01 5.86499035e-01 5.42604506e-01 -4.05318916e-01
7.02542782e-01 8.06807220e-01 3.23599458e-01 -7.27074862e-01
6.90473535e-05 4.43101406e-01 3.34013291e-02 -2.92106211e-01
8.53601396e-01 9.94204171e-03 -7.23896265e-01 6.67448640e-01
4.96210337e-01 4.61850539e-02 6.04711138e-02 3.79858971e-01
9.57739949e-01 1.18535355e-01 -1.43076675e-02 -2.56844193e-01
3.23344290e-01 4.67051834e-01 1.22279108e+00 5.56187510e-01
-1.48143619e-01 7.71860898e-01 2.61887461e-01 -4.57126535e-02
-7.52856433e-01 -7.56444395e-01 -3.16050351e-01 1.38376439e+00
-1.09371260e-01 -7.67457008e-01 -7.38358796e-01 -8.40718150e-01
-3.57612938e-01 6.40012801e-01 -3.38385522e-01 4.78393704e-01
-9.06979859e-01 -6.54419959e-01 9.75412905e-01 2.77419865e-01
3.38548124e-01 -1.07666445e+00 7.74248242e-02 4.85806942e-01
-1.28221095e+00 -1.62036228e+00 -1.01424253e+00 2.09071428e-01
-3.53575617e-01 -1.27828598e+00 -3.37051272e-01 -7.99549341e-01
-1.49544915e-02 1.23417102e-01 1.56265259e+00 -3.65943044e-01
3.30100022e-02 4.07855868e-01 -3.93626183e-01 -1.44158423e-01
-5.55486798e-01 4.38987106e-01 1.76194534e-01 -4.36557353e-01
7.05714703e-01 -4.71147090e-01 -1.10313430e-01 7.97073394e-02
-1.45212725e-01 8.90827030e-02 2.95617670e-01 9.40017819e-01
4.50202614e-01 -8.14088225e-01 5.79035759e-01 -1.54116523e+00
4.79612082e-01 -3.92402023e-01 -3.54795933e-01 3.57340962e-01
-2.45734155e-01 1.99731085e-02 2.69268215e-01 -2.35776722e-01
-1.58014643e+00 -7.45510086e-02 -2.45263040e-01 -2.95132697e-01
1.60935316e-02 4.75490123e-01 -4.27633196e-01 3.91306430e-01
4.46884334e-01 -4.26524520e-01 -3.01002800e-01 -9.82570648e-01
6.73524439e-01 9.44016159e-01 1.29099572e+00 -1.02971983e+00
2.81980008e-01 1.14317566e-01 -6.48655832e-01 -5.39728284e-01
-1.13582182e+00 -6.08589292e-01 -8.77154410e-01 8.18674341e-02
9.83619809e-01 -1.32510448e+00 -7.22617269e-01 6.35999441e-02
-1.54308069e+00 -3.38795632e-01 8.84127170e-02 5.54440141e-01
-3.85075241e-01 1.25964493e-01 -9.63129640e-01 -5.66355169e-01
-7.88330376e-01 -9.54461873e-01 9.42779422e-01 -4.89496684e-04
-7.21020341e-01 -1.00581884e+00 6.99554741e-01 8.45654428e-01
-5.11720739e-02 1.12930462e-01 7.81966269e-01 -1.03058350e+00
-1.51583135e-01 2.51789931e-02 -3.30598354e-02 -2.13755682e-01
-1.19215228e-01 -1.85543448e-01 -9.91170943e-01 -3.39091033e-01
-5.53835213e-01 -8.52315784e-01 6.61034346e-01 -6.31195679e-02
2.24269822e-01 -2.23447770e-01 -4.23884094e-01 6.64996207e-01
1.09857988e+00 -7.68914819e-02 3.49149019e-01 5.80941439e-01
4.98164117e-01 6.90403700e-01 7.75041282e-01 1.88429192e-01
1.01494741e+00 1.06478083e+00 -1.23781621e-01 -1.40068692e-03
-1.82334408e-01 -4.63419855e-01 2.50096023e-01 1.34866822e+00
-1.10137202e-01 -1.29391909e-01 -1.17126262e+00 8.47392678e-01
-2.02957845e+00 -1.18055868e+00 -4.00561601e-01 1.90328264e+00
1.54584467e+00 -1.60022408e-01 -1.93168327e-01 -6.71148598e-01
8.41690540e-01 1.31696582e-01 2.19611712e-02 -2.58679688e-01
-6.42974019e-01 1.10601768e-01 3.18274498e-01 8.54722500e-01
-1.30297542e+00 1.29955506e+00 6.49769640e+00 3.12842906e-01
-5.14324784e-01 4.31563586e-01 1.78318948e-01 -1.67690888e-01
-4.27809685e-01 4.32871282e-01 -1.26494372e+00 2.16437221e-01
1.05070722e+00 9.12825912e-02 4.54852462e-01 4.03101236e-01
-1.06778190e-01 1.81767032e-01 -1.20755696e+00 8.60799730e-01
-1.01755923e-02 -1.40281665e+00 -3.30468565e-01 -1.40324011e-01
8.18357646e-01 6.61871612e-01 -4.85740483e-01 8.81889760e-01
1.17925787e+00 -5.86537421e-01 4.56368119e-01 9.12460387e-02
1.02289534e+00 -6.92776501e-01 8.99608374e-01 3.57977778e-01
-9.53530550e-01 1.43024027e-01 -1.27662584e-01 3.51605892e-01
5.56520164e-01 -1.37698606e-01 -5.09850383e-01 6.74216270e-01
8.43239069e-01 4.46808040e-01 -2.48141795e-01 3.08095753e-01
-2.54290521e-01 3.94747585e-01 -4.64181215e-01 6.53796673e-01
2.59209812e-01 -1.41420797e-01 6.88689530e-01 1.58451664e+00
9.37583297e-03 4.02377158e-01 3.06800276e-01 5.09980381e-01
-5.86119413e-01 2.19654083e-01 -4.08290267e-01 2.22993135e-01
1.21219289e+00 1.50347388e+00 1.85271308e-01 -8.82706642e-02
-8.77253532e-01 9.33947444e-01 8.45273435e-01 2.64762789e-01
-5.18480957e-01 -6.44852817e-02 8.26380372e-01 -5.17895699e-01
-1.81445882e-01 2.32554585e-01 2.60831743e-01 -1.45912564e+00
-4.28354770e-01 -1.41044199e+00 1.05166769e+00 -3.79116416e-01
-1.20184469e+00 6.78251922e-01 -7.02816099e-02 -1.00885975e+00
-6.22287512e-01 -8.08026418e-02 -5.83010852e-01 9.35447335e-01
-1.33168209e+00 -1.45233655e+00 1.70661174e-02 8.58430445e-01
5.41183114e-01 -3.11215997e-01 1.35038829e+00 5.16667366e-01
-6.87759042e-01 9.83605444e-01 4.06667627e-02 6.19906366e-01
1.55303514e+00 -1.14472187e+00 3.85495633e-01 6.15205407e-01
3.12783718e-01 7.65526533e-01 7.44294167e-01 -5.24650872e-01
-1.35870802e+00 -8.59414697e-01 1.32060552e+00 -9.41512227e-01
8.12199473e-01 -5.52493632e-01 -8.94578278e-01 1.12514710e+00
6.30825460e-01 -2.64312685e-01 1.03765297e+00 1.13213503e+00
-6.23172462e-01 2.80103683e-01 -9.66799736e-01 5.77754259e-01
1.22211361e+00 -9.17786598e-01 -1.08786309e+00 5.62978759e-02
7.82484412e-01 -6.55365169e-01 -1.31654859e+00 4.62048113e-01
4.79851007e-01 -5.12406826e-01 7.37832844e-01 -8.79675627e-01
2.63366997e-01 2.76038647e-01 -5.25184274e-01 -1.44555974e+00
-2.05085680e-01 -9.97780919e-01 7.26465732e-02 2.12797260e+00
8.20910990e-01 -2.00587630e-01 6.10319376e-01 8.82873833e-01
-3.40377688e-01 1.80593431e-01 -1.08249485e+00 -4.09551382e-01
3.56370956e-01 -4.12232816e-01 4.54127908e-01 1.60095632e+00
5.96414626e-01 1.09305036e+00 -5.49163938e-01 1.13184899e-01
6.43846095e-01 4.24479902e-01 9.16008055e-01 -1.41445291e+00
-2.97937155e-01 3.54636721e-02 5.97347200e-01 -7.89156199e-01
7.65022397e-01 -1.12020493e+00 1.51415139e-01 -1.05043197e+00
5.44132471e-01 -6.43514991e-01 9.64256451e-02 8.31788719e-01
-3.40163410e-01 3.11436683e-01 3.92646402e-01 5.97605348e-01
-7.95758069e-01 4.15850163e-01 8.73926938e-01 -9.93463695e-02
-4.53429312e-01 -3.97734255e-01 -7.92019784e-01 7.97712624e-01
2.44882524e-01 -4.18753028e-01 1.16978996e-01 -8.39609206e-01
-9.63202491e-02 5.93109906e-01 -4.17975187e-01 -3.77332926e-01
3.92438322e-01 -1.30087696e-03 -1.29315421e-01 -6.98632002e-01
4.05068427e-01 -5.64358652e-01 2.18425542e-01 -2.26787046e-01
-5.84806144e-01 -5.02561480e-02 2.96916842e-01 2.34647900e-01
-4.17602837e-01 -1.82944015e-01 5.88959575e-01 -2.30442494e-01
-8.41305375e-01 9.39100161e-02 1.66605721e-04 9.78063822e-01
2.15431213e-01 5.58421850e-01 -7.18787909e-01 -4.73101348e-01
-8.19641888e-01 7.02711761e-01 3.54751527e-01 7.93541968e-01
-1.64154261e-01 -1.44399071e+00 -1.08625114e+00 -1.45238817e-01
4.70690578e-01 -2.32768759e-01 2.41373152e-01 7.80205190e-01
1.75640106e-01 6.60747528e-01 4.63468768e-02 -2.93901294e-01
-1.49016047e+00 1.88104570e-01 1.47665396e-01 -6.08020067e-01
-6.51172996e-01 7.47387826e-01 6.52707443e-02 -1.27802694e+00
1.66636005e-01 4.72398758e-01 -3.46614867e-01 4.70121950e-01
4.33033407e-01 1.45062543e-02 -6.15201667e-02 -8.72894645e-01
-5.39047658e-01 1.88278839e-01 -3.32018077e-01 -4.83027965e-01
1.55347669e+00 -3.51916015e-01 -7.25186393e-02 3.64306569e-01
1.24081647e+00 5.15421093e-01 -9.56663609e-01 -5.32539904e-01
4.28882718e-01 -4.48510163e-02 -1.24069877e-01 -1.03489375e+00
-7.84937978e-01 4.53389198e-01 1.74706399e-01 -4.41149354e-01
6.43988550e-01 3.60737145e-01 7.48016536e-01 5.97998500e-01
7.15725541e-01 -1.44702184e+00 -4.21376109e-01 1.29142022e+00
9.81238306e-01 -1.41524172e+00 -2.68694103e-01 -3.35233539e-01
-1.08238876e+00 7.45682776e-01 6.74182475e-01 5.18528938e-01
2.25024447e-01 5.53240180e-01 5.64412534e-01 -1.38370872e-01
-1.06408370e+00 -2.88157076e-01 -9.24475789e-02 5.80589354e-01
9.13420975e-01 8.26238394e-02 -3.52232456e-01 1.25411856e+00
-4.79055375e-01 -3.88052315e-01 5.07679462e-01 3.99267435e-01
1.36542007e-01 -1.49695539e+00 -2.12897360e-01 -1.23461992e-01
-8.18375111e-01 -5.49927056e-01 -6.58322155e-01 1.18199658e+00
-1.64749950e-01 1.09882677e+00 2.03329504e-01 -2.19974428e-01
6.38497114e-01 2.71135390e-01 2.52092242e-01 -6.73854768e-01
-6.82271004e-01 9.45827514e-02 1.14521253e+00 -6.43087745e-01
-5.81364214e-01 -1.04582965e+00 -1.18039453e+00 -3.01437795e-01
-3.52685750e-01 6.94078147e-01 2.12302640e-01 8.86948109e-01
3.55975360e-01 -4.95434627e-02 4.53375429e-01 -7.17378795e-01
-2.30663374e-01 -1.23272502e+00 -1.92183539e-01 8.51508737e-01
-7.72256330e-02 -4.69426841e-01 -1.41406864e-01 -2.23608226e-01] | [9.300653457641602, 9.556510925292969] |
19ab74e9-6e9f-429b-b6f0-59c2eb99ae22 | the-muse-2023-multimodal-sentiment-analysis | 2305.03369 | null | https://arxiv.org/abs/2305.03369v1 | https://arxiv.org/pdf/2305.03369v1.pdf | The MuSe 2023 Multimodal Sentiment Analysis Challenge: Mimicked Emotions, Cross-Cultural Humour, and Personalisation | The MuSe 2023 is a set of shared tasks addressing three different contemporary multimodal affect and sentiment analysis problems: In the Mimicked Emotions Sub-Challenge (MuSe-Mimic), participants predict three continuous emotion targets. This sub-challenge utilises the Hume-Vidmimic dataset comprising of user-generated videos. For the Cross-Cultural Humour Detection Sub-Challenge (MuSe-Humour), an extension of the Passau Spontaneous Football Coach Humour (Passau-SFCH) dataset is provided. Participants predict the presence of spontaneous humour in a cross-cultural setting. The Personalisation Sub-Challenge (MuSe-Personalisation) is based on the Ulm-Trier Social Stress Test (Ulm-TSST) dataset, featuring recordings of subjects in a stressed situation. Here, arousal and valence signals are to be predicted, whereas parts of the test labels are made available in order to facilitate personalisation. MuSe 2023 seeks to bring together a broad audience from different research communities such as audio-visual emotion recognition, natural language processing, signal processing, and health informatics. In this baseline paper, we introduce the datasets, sub-challenges, and provided feature sets. As a competitive baseline system, a Gated Recurrent Unit (GRU)-Recurrent Neural Network (RNN) is employed. On the respective sub-challenges' test datasets, it achieves a mean (across three continuous intensity targets) Pearson's Correlation Coefficient of .4727 for MuSe-Mimic, an Area Under the Curve (AUC) value of .8310 for MuSe-Humor and Concordance Correlation Coefficient (CCC) values of .7482 for arousal and .7827 for valence in the MuSe-Personalisation sub-challenge. | ['Björn W. Schuller', 'Erik Cambria', 'Alan Cowen', 'Andreas König', 'Eva-Maria Meßner', 'Panagiotis Tzirakis', 'Chris Gagne', 'Steffen Klug', 'Niklas Müller', 'Alexander Kathan', 'Alice Baird', 'Shahin Amiriparian', 'Lukas Christ'] | 2023-05-05 | null | null | null | null | ['multimodal-sentiment-analysis', 'multimodal-sentiment-analysis'] | ['computer-vision', 'natural-language-processing'] | [ 4.80360277e-02 -6.50385991e-02 3.94576579e-01 -4.00756240e-01
-8.95541966e-01 -2.88080812e-01 4.05820966e-01 1.61254272e-01
-3.05333048e-01 4.28026587e-01 5.08739114e-01 6.00825548e-01
2.24399850e-01 -7.55158961e-02 -1.23414341e-02 -5.91144323e-01
-4.23742890e-01 -1.09004825e-01 -6.16440892e-01 -6.29232049e-01
-2.54841615e-02 2.03929786e-02 -1.66264117e+00 9.16432142e-01
1.98401555e-01 1.22527874e+00 -4.97497380e-01 1.04290533e+00
6.06623113e-01 1.02264643e+00 -6.41119897e-01 -6.59087718e-01
-8.84806812e-02 -8.78298163e-01 -8.42979074e-01 -1.16352990e-01
1.29413038e-01 3.71252537e-01 -2.82039326e-02 6.28615737e-01
1.11788511e+00 5.65643847e-01 5.07676721e-01 -1.45976496e+00
-3.02238196e-01 4.07325059e-01 -3.99904817e-01 3.50639485e-02
9.50397491e-01 1.87814429e-01 9.78240967e-01 -9.48401511e-01
7.48911440e-01 1.03813660e+00 9.87931490e-01 8.15607250e-01
-1.11260045e+00 -6.36604309e-01 -3.75917524e-01 4.46742833e-01
-1.14454663e+00 -3.50136399e-01 9.69230235e-01 -3.16298753e-01
9.89036918e-01 5.28831303e-01 7.26823270e-01 1.91528237e+00
1.76075220e-01 9.19965029e-01 1.29996061e+00 -1.29924715e-01
2.78467923e-01 4.25786227e-01 -1.53692439e-01 2.53724068e-01
-9.73649621e-01 -1.38006747e-01 -1.10192633e+00 -2.17113987e-01
1.59012586e-01 -4.75894570e-01 -4.68973756e-01 1.94517180e-01
-1.13347423e+00 7.61636555e-01 3.51140618e-01 5.58818519e-01
-6.30897343e-01 -9.44400281e-02 1.27669859e+00 7.19568729e-01
4.99014258e-01 6.62958682e-01 -1.83911279e-01 -5.88184416e-01
-7.75942385e-01 2.98719406e-01 8.71251166e-01 2.90057391e-01
2.11397037e-01 1.24125652e-01 -4.59615439e-01 1.29590213e+00
-1.33698374e-01 2.33035654e-01 9.73881960e-01 -8.85152459e-01
-7.39376247e-02 4.95650798e-01 -1.45860434e-01 -1.30157232e+00
-1.04859436e+00 -3.50624800e-01 -1.09208298e+00 2.04408653e-02
-1.30199909e-01 -4.65569228e-01 -2.88549602e-01 2.06316924e+00
1.82071194e-01 1.63378105e-01 2.63229162e-01 1.24091208e+00
1.31840885e+00 7.60281503e-01 1.64480746e-01 -6.95410550e-01
1.40915668e+00 -6.96623445e-01 -6.56140685e-01 -1.34176239e-01
5.30262232e-01 -7.67825186e-01 1.15453613e+00 9.42525864e-01
-1.17359269e+00 -5.87982178e-01 -9.54635978e-01 1.09686725e-01
-4.25607294e-01 -1.26964048e-01 2.15570759e-02 6.23571932e-01
-1.03233612e+00 5.64063609e-01 -3.51956151e-02 -5.13046324e-01
-7.44529739e-02 1.50132075e-01 -7.28666902e-01 3.80208105e-01
-1.72574341e+00 9.19503629e-01 -5.48295863e-02 2.62814730e-01
-7.25532770e-01 -6.19096994e-01 -9.53747511e-01 -2.90663511e-01
-3.38890433e-01 -4.06087101e-01 1.17056513e+00 -1.65971828e+00
-1.76370347e+00 1.45924449e+00 2.12931007e-01 -4.03337121e-01
3.30469608e-01 -2.66772777e-01 -7.74627090e-01 3.29997271e-01
-1.20795935e-01 7.33322740e-01 8.43977988e-01 -9.99914348e-01
-7.46761337e-02 -1.81286231e-01 -3.95586699e-01 5.00676751e-01
-4.22006428e-01 4.02272224e-01 -6.44779671e-03 -5.69409370e-01
-5.64847708e-01 -8.77758265e-01 1.94611233e-02 -9.04261351e-01
-3.44951659e-01 -1.46568790e-01 6.02688253e-01 -6.98382735e-01
1.30959213e+00 -2.50725651e+00 5.78445137e-01 2.94722855e-01
1.13046885e-01 -2.25803386e-02 -4.35057729e-01 5.51711679e-01
-5.18842459e-01 -2.35776097e-01 -6.37805983e-02 -6.84454679e-01
2.02166736e-01 -1.68795124e-01 -9.77500603e-02 4.70136285e-01
6.36263564e-02 8.08778763e-01 -8.54352891e-01 -2.21646503e-01
-6.32218197e-02 6.92749798e-01 -3.08178604e-01 4.21435356e-01
1.80453435e-01 3.98328990e-01 2.37429857e-01 4.81099814e-01
1.43004641e-01 2.10192576e-01 1.48035869e-01 -2.78125644e-01
1.11017257e-01 -1.00268081e-01 -8.81667078e-01 1.57018769e+00
-4.94632363e-01 8.35223377e-01 1.65286973e-01 -4.85836118e-01
1.38519335e+00 9.14492846e-01 7.12954760e-01 -8.13406527e-01
5.19358158e-01 2.94508152e-02 -2.66767740e-01 -8.00233245e-01
6.58732593e-01 -7.27415085e-01 -7.67067671e-01 3.51181149e-01
2.43907630e-01 -1.16768211e-01 7.28217736e-02 1.63905770e-01
1.22569668e+00 -4.96989954e-03 5.14485657e-01 -1.58501801e-03
6.17365241e-01 -5.13755560e-01 5.30490100e-01 2.32044831e-01
-8.14652205e-01 8.53568673e-01 7.98958600e-01 -3.32025617e-01
-7.08761752e-01 -7.26429880e-01 8.48011449e-02 1.44490421e+00
-4.01057810e-01 -6.02637589e-01 -6.42408371e-01 -2.34903678e-01
-4.61816430e-01 6.45401061e-01 -1.13413632e+00 -4.11829025e-01
-1.12780660e-01 -7.51282454e-01 8.33670378e-01 2.62438655e-01
2.30698958e-01 -1.76491213e+00 -1.03165686e+00 1.52130857e-01
-6.61672354e-01 -1.00991869e+00 -2.36855641e-01 3.00997198e-01
-2.04684585e-02 -8.25898767e-01 -6.08666301e-01 -6.52460992e-01
-2.29679674e-01 -4.15721714e-01 1.26883936e+00 -3.27181846e-01
-3.98638129e-01 9.22932982e-01 -6.91517889e-01 -3.72606635e-01
-2.85904557e-01 -2.02033430e-01 2.28599653e-01 4.50928390e-01
4.93337631e-01 -6.85025275e-01 -5.04580498e-01 2.33475342e-01
-6.97213650e-01 -1.10284574e-01 7.67502561e-02 1.02386272e+00
2.36970097e-01 -5.70589066e-01 1.01457584e+00 -3.87742251e-01
9.92416561e-01 -8.21761549e-01 4.35863942e-01 -1.02717713e-01
-2.63901241e-02 -8.95039499e-01 5.24388731e-01 -6.67652190e-01
-9.53979850e-01 -5.61292619e-02 -7.65912011e-02 -7.31456935e-01
-2.88965821e-01 5.56830406e-01 1.18605517e-01 3.79219413e-01
1.07049537e+00 -7.19787553e-02 -5.74931242e-02 1.17437549e-01
1.12589233e-01 8.73260736e-01 1.05910063e+00 -3.85404497e-01
4.33576927e-02 -1.08680889e-01 -8.59602764e-02 -8.70336354e-01
-8.25856447e-01 -6.65877283e-01 -3.74212354e-01 -8.29335451e-01
9.37514722e-01 -1.01782596e+00 -1.06624424e+00 4.78160083e-01
-9.85066175e-01 -4.42290574e-01 -1.56878084e-01 3.16993833e-01
-9.19194996e-01 1.46943867e-01 -9.32824910e-01 -1.21076465e+00
-8.99834454e-01 -6.41466498e-01 9.66196477e-01 1.95830733e-01
-1.15198028e+00 -8.53331387e-01 6.28826916e-01 5.64961672e-01
1.38332546e-01 1.03133619e+00 5.44411421e-01 -8.91441882e-01
9.03685331e-01 -2.96536237e-01 2.31098637e-01 4.82753307e-01
-1.38954177e-01 -3.25965315e-01 -1.35225010e+00 -3.41556631e-02
1.47932336e-01 -1.24107540e+00 5.46712816e-01 1.00650743e-01
7.25382566e-01 -1.12852290e-01 5.71439028e-01 1.06244706e-01
7.49477625e-01 7.28849173e-02 8.28251600e-01 2.87441283e-01
2.74691641e-01 1.00749528e+00 5.24899423e-01 8.56656253e-01
2.96111584e-01 5.84495604e-01 4.15948600e-01 9.29294303e-02
3.47416878e-01 1.53107226e-01 9.28551912e-01 8.37521315e-01
-6.61474019e-02 2.54650526e-02 -8.58067691e-01 6.11653864e-01
-1.91201591e+00 -1.27849209e+00 -4.05457884e-01 1.96046305e+00
9.09740865e-01 -3.87932271e-01 7.29951322e-01 4.60877776e-01
4.68238175e-01 2.23741293e-01 -2.43442819e-01 -1.22985470e+00
-4.75745022e-01 4.69452366e-02 -7.17436731e-01 3.32034290e-01
-1.13173199e+00 5.26921690e-01 5.19025373e+00 5.24580419e-01
-1.17477715e+00 1.48549795e-01 8.32448006e-01 -7.74191916e-01
1.96517944e-01 -8.92349958e-01 8.15290119e-03 3.65299821e-01
1.60748720e+00 -4.22166139e-02 4.18135166e-01 6.51032448e-01
2.81666577e-01 -4.18487052e-03 -9.84045088e-01 1.49912524e+00
7.57075250e-01 -4.89226073e-01 -5.82355738e-01 -5.15783250e-01
5.89916408e-01 -5.90382814e-02 3.53345156e-01 8.49076033e-01
-4.76331413e-01 -1.30488133e+00 4.68293786e-01 7.53697991e-01
9.05894339e-01 -1.18364668e+00 1.14768255e+00 -6.01106249e-02
-7.87652373e-01 -1.29417315e-01 1.49872079e-01 -2.05104887e-01
3.11693102e-01 4.91126716e-01 -6.13358617e-01 4.14502174e-01
9.60096776e-01 7.28322923e-01 -3.50566417e-01 5.61280191e-01
-4.19470221e-02 5.96331835e-01 -9.90538672e-02 -1.55892655e-01
2.64850259e-01 1.17215417e-01 6.52528942e-01 1.92113066e+00
1.30667701e-01 1.79963365e-01 -1.03168435e-01 4.39522386e-01
-2.33942315e-01 5.82453907e-01 -4.36494827e-01 9.04057324e-02
-4.52810675e-02 1.90526342e+00 -2.20959052e-01 -6.22292124e-02
1.04382537e-01 1.28650808e+00 1.29626676e-01 1.34859234e-01
-8.06474328e-01 -4.06958669e-01 3.98804575e-01 -5.04571736e-01
-1.95776954e-01 5.06564081e-01 -2.95059234e-01 -9.06918585e-01
-3.87731463e-01 -1.32482195e+00 6.47218943e-01 -1.20164990e+00
-1.53849590e+00 9.06255126e-01 -1.96774006e-01 -1.08476770e+00
-6.16171122e-01 -3.95302862e-01 -8.39606822e-01 6.78163946e-01
-6.46176577e-01 -1.22253692e+00 -4.28348452e-01 7.64683664e-01
4.27139461e-01 -1.63280472e-01 1.09075177e+00 3.02527696e-01
-6.95947230e-01 6.91426873e-01 -5.88648796e-01 -1.31865526e-02
1.17891693e+00 -1.18507314e+00 -3.64963710e-01 2.27497771e-01
-1.05070181e-01 2.07541123e-01 1.05159473e+00 -3.91648293e-01
-1.00197411e+00 -1.04219067e+00 1.03833234e+00 -3.85461450e-01
7.42397666e-01 -3.09813291e-01 -9.02206481e-01 3.59435081e-01
5.46282709e-01 -2.70983875e-01 1.33180285e+00 3.42329592e-01
-4.74884301e-01 1.65235296e-01 -1.15762854e+00 5.11702120e-01
5.39159596e-01 -6.53713048e-01 -4.08335805e-01 1.84752658e-01
1.57534346e-01 -3.39232594e-01 -1.22841883e+00 4.72692668e-01
8.74697864e-01 -1.37592196e+00 5.04118979e-01 -6.41945601e-01
9.74360228e-01 1.51254103e-01 -4.83509064e-01 -1.45915020e+00
-2.37247288e-01 -9.68445182e-01 1.83524508e-02 1.28723109e+00
1.74836934e-01 3.43145244e-02 4.24247235e-01 4.18294609e-01
-7.37833008e-02 -9.69696999e-01 -1.14587235e+00 -1.73496023e-01
-1.61657318e-01 -8.17274094e-01 -3.41804475e-02 1.26295233e+00
8.62029135e-01 8.74377847e-01 -9.89474893e-01 -3.96910638e-01
2.26096958e-01 -2.07423866e-01 7.76535988e-01 -9.62354839e-01
-3.26323122e-01 -5.29916584e-01 -6.62890971e-01 6.10196032e-02
3.25010687e-01 -8.29942107e-01 3.34703587e-02 -8.86465013e-01
3.38683397e-01 5.88276327e-01 -7.12379098e-01 6.60175562e-01
5.97220883e-02 7.69092143e-01 4.07633960e-01 -2.53447652e-01
-9.01750207e-01 8.27979803e-01 5.73316157e-01 9.07555819e-02
-5.39779723e-01 -3.14137459e-01 -6.13017440e-01 6.93039358e-01
8.15910757e-01 -6.97996542e-02 -1.70425028e-01 5.58108389e-01
7.19770491e-01 2.72616476e-01 5.97043931e-01 -9.82125342e-01
-1.14713311e-01 1.72084510e-01 4.68739629e-01 -3.71035457e-01
9.16117609e-01 -3.39772612e-01 1.11639738e-01 2.25714684e-01
-5.19048274e-01 4.14327979e-02 4.34260011e-01 1.99972093e-01
-2.79585093e-01 -5.28643504e-02 8.85419250e-01 6.19962327e-02
-4.87038165e-01 -3.06942970e-01 -5.92180789e-01 -5.57111204e-03
9.65236843e-01 6.08227141e-02 -1.70030028e-01 -1.01826358e+00
-9.88576651e-01 2.77023494e-01 -2.61726491e-02 6.42752290e-01
7.75799513e-01 -1.48956954e+00 -9.48831201e-01 -1.30315959e-01
2.63602912e-01 -7.41342187e-01 8.58089387e-01 1.31983960e+00
2.29125127e-01 -8.60941187e-02 -4.87639695e-01 -3.80714774e-01
-1.59613025e+00 2.35704422e-01 4.73841161e-01 -2.48060957e-01
-3.54050934e-01 8.34323943e-01 -6.02431074e-02 -5.67405581e-01
-4.94743465e-03 4.68327522e-01 -4.81967151e-01 7.25850046e-01
7.72861421e-01 6.32086098e-01 -4.06398214e-02 -1.16337860e+00
-2.60073870e-01 1.41056836e-01 3.41977954e-01 -3.50323528e-01
1.39468515e+00 -1.08505510e-01 -1.91149935e-01 1.22033572e+00
1.29942608e+00 -1.31751016e-01 -6.22566760e-01 6.25086129e-02
-1.84514821e-01 2.78463781e-01 -3.38170752e-02 -1.26918340e+00
-6.88711107e-01 7.43139446e-01 8.24206769e-01 2.08573535e-01
1.50250244e+00 -2.92483330e-01 6.70969427e-01 2.06464574e-01
-5.56451408e-03 -1.30159330e+00 3.96946043e-01 6.32592320e-01
1.50213003e+00 -9.85029280e-01 -3.12774807e-01 2.99742490e-01
-1.51553261e+00 1.09661865e+00 7.56046534e-01 9.22636986e-02
1.81616962e-01 -1.89295143e-01 4.01439637e-01 -3.86031628e-01
-1.21407700e+00 4.85665798e-02 5.05414724e-01 4.45287973e-01
7.09256887e-01 1.43663779e-01 -1.63415179e-01 1.32718611e+00
-6.63965285e-01 -1.41003370e-01 5.54350436e-01 4.29573834e-01
-6.57279566e-02 -3.72552842e-01 -5.72960794e-01 1.61279157e-01
-4.72747117e-01 1.98452771e-01 -9.26069200e-01 5.20092249e-01
1.79322660e-01 1.12930536e+00 -1.39857084e-01 -8.59625459e-01
5.04420698e-01 5.29606879e-01 2.35879675e-01 -2.97659993e-01
-1.42707491e+00 1.55792356e-01 5.09211302e-01 -8.16575825e-01
-6.21388972e-01 -7.49043882e-01 -1.13054836e+00 -1.29810423e-01
3.89514953e-01 2.38123506e-01 5.14057636e-01 6.53856635e-01
2.64220357e-01 5.22111714e-01 8.58697832e-01 -1.24607682e+00
5.14087118e-02 -1.30172205e+00 -8.29006553e-01 9.08574522e-01
1.52530476e-01 -3.69060159e-01 -5.46569049e-01 2.47476231e-02] | [13.37167739868164, 5.072886943817139] |
7c75211c-2273-4c3d-a330-0f7956c10f59 | sparsity-by-redundancy-solving-l-1-with-a | 2210.01212 | null | https://arxiv.org/abs/2210.01212v4 | https://arxiv.org/pdf/2210.01212v4.pdf | spred: Solving $L_1$ Penalty with SGD | We propose to minimize a generic differentiable objective with $L_1$ constraint using a simple reparametrization and straightforward stochastic gradient descent. Our proposal is the direct generalization of previous ideas that the $L_1$ penalty may be equivalent to a differentiable reparametrization with weight decay. We prove that the proposed method, \textit{spred}, is an exact differentiable solver of $L_1$ and that the reparametrization trick is completely ``benign" for a generic nonconvex function. Practically, we demonstrate the usefulness of the method in (1) training sparse neural networks to perform gene selection tasks, which involves finding relevant features in a very high dimensional space, and (2) neural network compression task, to which previous attempts at applying the $L_1$-penalty have been unsuccessful. Conceptually, our result bridges the gap between the sparsity in deep learning and conventional statistical learning. | ['ZiHao Wang', 'Liu Ziyin'] | 2022-10-03 | null | null | null | null | ['neural-network-compression', 'neural-network-compression'] | ['methodology', 'miscellaneous'] | [ 3.37744743e-01 2.24745691e-01 -1.57998353e-01 -6.19056046e-01
-1.04100251e+00 -1.16884194e-01 6.56420663e-02 -1.36212468e-01
-4.69040811e-01 1.14034855e+00 -6.14888221e-02 -2.08666012e-01
-5.03488660e-01 -5.76591253e-01 -1.04595959e+00 -1.15017736e+00
-2.93780982e-01 2.47510672e-01 -5.46413362e-01 -2.25499496e-01
2.84222752e-01 5.78672111e-01 -1.40131104e+00 -2.24089608e-01
8.72570097e-01 1.15773559e+00 1.12308986e-01 4.46617782e-01
1.04854055e-01 3.19214553e-01 -3.61290336e-01 -3.73309106e-01
4.42207813e-01 -4.14079547e-01 -7.03429759e-01 -3.51405554e-02
6.90153658e-01 1.80740073e-01 -2.00326130e-01 1.11438572e+00
5.55005610e-01 4.62506235e-01 6.29267335e-01 -8.86008978e-01
-7.31066942e-01 5.81169665e-01 -7.74722099e-01 1.67682186e-01
-1.72552556e-01 -2.18010128e-01 1.20912337e+00 -1.11529660e+00
3.49489540e-01 1.12736309e+00 1.09474015e+00 3.98001611e-01
-1.40253234e+00 -5.00790119e-01 7.89447427e-02 -1.29926845e-01
-1.58967757e+00 -4.47349548e-01 8.23830605e-01 -3.33946586e-01
8.32336903e-01 3.29449475e-01 2.85515487e-01 7.65053928e-01
1.05949186e-01 8.89010072e-01 6.75551057e-01 -3.78595144e-01
3.61800730e-01 -6.85370341e-02 1.75784886e-01 1.18255663e+00
2.36387908e-01 -2.96711773e-02 -6.22076213e-01 -2.51306683e-01
6.71322346e-01 -1.98405478e-02 -2.18122363e-01 -3.65781665e-01
-8.85474741e-01 1.16537619e+00 1.27216995e-01 2.86552131e-01
-1.62790596e-01 5.32240927e-01 4.51713353e-01 4.60104585e-01
8.13914180e-01 7.66034245e-01 -6.58874571e-01 -1.83714405e-01
-1.23156655e+00 1.77939355e-01 7.01358855e-01 1.04038537e+00
6.99234009e-01 6.53265178e-01 2.88572715e-04 1.10966575e+00
7.28366151e-02 3.68898720e-01 5.56341708e-01 -1.14704430e+00
3.10075253e-01 1.54950038e-01 -1.35742590e-01 -9.66976106e-01
-5.22474170e-01 -7.04647124e-01 -1.30961478e+00 -4.94043417e-02
3.75922382e-01 -4.68252152e-01 -5.55725396e-01 2.08958507e+00
-2.48251110e-02 4.20166254e-01 -1.16258234e-01 5.40197968e-01
6.09412491e-01 5.31842232e-01 -1.62440151e-01 -4.01621222e-01
8.63806725e-01 -5.75359762e-01 -5.85164309e-01 8.44478011e-02
8.47190559e-01 -5.43976724e-01 1.07159376e+00 4.66949880e-01
-1.48061323e+00 -3.17537904e-01 -9.61109340e-01 -2.43745729e-01
-1.60277754e-01 4.77742255e-01 1.04868937e+00 7.30956137e-01
-1.14474285e+00 9.46804345e-01 -7.27330148e-01 -1.25148622e-02
6.52155936e-01 7.17948437e-01 -3.83540213e-01 4.06043157e-02
-9.17858839e-01 5.47720373e-01 2.02100024e-01 2.48796389e-01
-7.27047920e-01 -9.61483240e-01 -9.66183662e-01 2.31505245e-01
3.28169882e-01 -7.78238177e-01 5.74634433e-01 -6.11068249e-01
-1.54904723e+00 1.05658805e+00 -4.50836629e-01 -6.79666936e-01
2.16157958e-01 -2.15408385e-01 1.95804715e-01 -3.07035208e-01
-1.45228520e-01 3.93913716e-01 1.12191629e+00 -8.02757442e-01
-4.51699197e-01 -4.77240503e-01 -2.93686718e-01 6.85128802e-03
-4.08710718e-01 -1.80700958e-01 -9.80206430e-02 -1.13856745e+00
1.74849495e-01 -9.97734368e-01 -4.41251308e-01 -1.27533332e-01
-6.40741706e-01 -2.30643496e-01 5.10499537e-01 -5.47226429e-01
1.20731091e+00 -2.20272303e+00 5.50102413e-01 4.34392810e-01
1.87766537e-01 1.51304737e-01 -4.12695259e-01 1.75862968e-01
-4.13540363e-01 1.99113607e-01 -7.09728003e-01 -5.89635432e-01
-6.10085996e-03 1.60732418e-01 -2.84297347e-01 9.25077736e-01
2.73032397e-01 7.74369478e-01 -7.39764929e-01 -1.40095949e-01
-2.28657708e-01 5.12873888e-01 -8.81455302e-01 -1.95120320e-01
-1.07903399e-01 9.84261110e-02 -4.93226886e-01 7.46226311e-01
7.61717916e-01 -1.35385215e-01 -8.06504041e-02 -1.05664134e-01
-5.12821302e-02 -9.17875618e-02 -1.08264565e+00 2.02652884e+00
-3.38605136e-01 7.86400437e-01 3.74926597e-01 -1.87998247e+00
1.02941549e+00 7.77588785e-02 8.55739117e-01 -1.75309405e-02
2.42247313e-01 2.64186651e-01 -3.89935017e-01 -3.34307820e-01
3.17356259e-01 -4.31749761e-01 -5.57238609e-02 2.73341209e-01
2.79480904e-01 -1.66460723e-01 1.52197003e-01 -2.64743090e-01
1.09644783e+00 6.57830015e-02 1.32131517e-01 -6.76465452e-01
5.12627542e-01 -3.32504362e-01 6.77991331e-01 8.27494681e-01
2.95607746e-02 6.73420012e-01 6.80690110e-01 -3.91515762e-01
-9.93417799e-01 -8.34085047e-01 -5.89964449e-01 1.42066240e+00
-3.27980697e-01 -4.49598394e-02 -7.00959444e-01 -5.30636489e-01
1.93330720e-01 6.23892903e-01 -6.71396673e-01 -1.80688933e-01
-8.37550819e-01 -1.37919402e+00 6.95616186e-01 3.09197217e-01
3.91215056e-01 -6.24853492e-01 1.09616844e-02 5.68001978e-02
7.24277571e-02 -6.05420828e-01 -6.23676300e-01 5.78831911e-01
-1.06434548e+00 -8.05600762e-01 -9.22127903e-01 -1.20783973e+00
7.99561024e-01 7.27994666e-02 1.01416731e+00 -3.94833349e-02
-4.87248778e-01 1.89700976e-01 -1.65058430e-02 -3.94424528e-01
-6.82750018e-03 2.37361252e-01 1.54397920e-01 6.07516728e-02
3.29560131e-01 -8.52029026e-01 -4.92870480e-01 6.22169040e-02
-9.89033639e-01 -3.22119802e-01 3.77624094e-01 1.24060905e+00
8.99934471e-01 2.67684869e-02 8.57286751e-01 -1.04175580e+00
8.61460567e-01 -5.79970121e-01 -6.09998524e-01 1.35752067e-01
-5.94731987e-01 3.86443287e-01 7.53417134e-01 -2.05521658e-01
-6.68131292e-01 2.00779527e-01 -4.74357724e-01 -4.38219190e-01
2.70184159e-01 8.26835692e-01 2.30161265e-01 -3.86107624e-01
7.92480230e-01 4.04423237e-01 1.70337602e-01 -6.25710607e-01
4.59664762e-01 1.99104860e-01 3.09494913e-01 -7.49973536e-01
7.67484307e-01 5.82728565e-01 6.27741039e-01 -1.00011683e+00
-8.72687340e-01 -2.14923009e-01 -4.81453091e-01 3.37820619e-01
4.60750431e-01 -7.65625477e-01 -9.14546311e-01 8.39497149e-02
-8.36768270e-01 -1.86212420e-01 -7.48832524e-01 5.66996336e-01
-1.06574440e+00 4.10011142e-01 -5.13902545e-01 -6.06911302e-01
-4.73894656e-01 -1.12993228e+00 1.04058242e+00 -1.52963743e-01
9.00511220e-02 -1.00553536e+00 1.44485816e-01 1.34131104e-01
4.49405879e-01 3.53070766e-01 1.21075094e+00 -5.92299700e-01
-1.24837600e-01 -3.96291167e-01 -1.18884638e-01 6.14910543e-01
3.82816978e-02 -1.51659563e-01 -8.44759047e-01 -6.06699049e-01
4.61810470e-01 -5.06893992e-01 1.14652419e+00 9.21897709e-01
1.73539412e+00 -3.91863883e-01 5.19429985e-03 1.30324972e+00
1.38044691e+00 -3.18071432e-02 4.57597762e-01 -8.20992216e-02
5.21794915e-01 4.19332206e-01 1.97308779e-01 6.63798869e-01
-1.53885171e-01 4.12498683e-01 3.21356595e-01 -2.31751680e-01
5.89112304e-02 1.50115371e-01 1.58887759e-01 9.23299491e-01
-2.58349568e-01 1.54286817e-01 -5.05219579e-01 5.34779370e-01
-1.85760152e+00 -8.49415541e-01 1.96936950e-01 2.30081820e+00
9.40625787e-01 -2.99470156e-01 -6.22848943e-02 -1.51656061e-01
6.26211584e-01 8.51945281e-02 -7.91815877e-01 -4.90223229e-01
-4.04651314e-01 7.35188603e-01 6.35544062e-01 4.27197635e-01
-1.18584657e+00 8.21064293e-01 7.23158216e+00 1.33584249e+00
-1.05153120e+00 -6.39887601e-02 9.13187861e-01 -3.48581672e-01
-3.84475052e-01 -3.12565178e-01 -8.45614791e-01 3.66867155e-01
6.48527026e-01 -3.18407267e-01 5.28156936e-01 1.02705812e+00
3.72844279e-01 1.74168572e-01 -1.26584804e+00 1.14376771e+00
1.66787952e-01 -1.59778643e+00 -8.96439850e-02 -5.82808480e-02
8.49685907e-01 7.46450573e-02 5.91567576e-01 3.95599455e-01
2.79758424e-01 -1.38941741e+00 4.54982594e-02 3.77020329e-01
8.75356078e-01 -9.59377587e-01 3.84718508e-01 1.95463955e-01
-9.49661434e-01 -1.21352285e-01 -7.61298299e-01 3.43164980e-01
4.08433937e-02 9.00917351e-01 -6.37536168e-01 1.65820241e-01
3.16676795e-01 7.83369899e-01 -1.84369102e-01 1.00086093e+00
1.47076145e-01 6.34159386e-01 -4.11043257e-01 5.79870269e-02
1.95810616e-01 -6.95768714e-01 8.75590682e-01 1.47089696e+00
6.39602602e-01 -1.78418737e-02 -1.09544955e-01 8.48238051e-01
-5.70851028e-01 2.56439000e-01 -6.78143680e-01 -1.77115276e-02
2.36344319e-02 8.19120228e-01 -2.68256813e-01 7.89496824e-02
-2.28127852e-01 6.91491663e-01 5.13650537e-01 4.84664440e-01
-7.41227269e-01 -7.31278419e-01 8.32047880e-01 -1.54552951e-01
4.55326319e-01 -4.13885087e-01 -5.38222313e-01 -1.29043484e+00
6.65908158e-02 -7.18823075e-01 4.61532176e-01 -3.36069137e-01
-1.32798994e+00 3.60233456e-01 -2.92264819e-01 -9.89271581e-01
-2.22645357e-01 -7.20905602e-01 -3.15605730e-01 8.62521052e-01
-1.62584949e+00 -7.96597362e-01 2.81780779e-01 6.22412443e-01
5.79928100e-01 -5.39479792e-01 8.56871605e-01 3.07560563e-01
-8.60788405e-01 9.68861043e-01 7.16898322e-01 -1.16479628e-01
3.78348172e-01 -1.26384592e+00 -7.76106631e-03 6.69523180e-01
8.18829145e-03 8.55564058e-01 7.75551558e-01 -2.73549616e-01
-1.64206862e+00 -1.14270389e+00 7.47863531e-01 1.35612816e-01
5.84775150e-01 -2.61206269e-01 -9.35656130e-01 7.04447329e-01
-1.95889190e-01 9.17103421e-03 1.12319934e+00 3.38126987e-01
-7.90217891e-02 -1.95080131e-01 -1.39807117e+00 4.93943751e-01
1.09288669e+00 -4.09762919e-01 -3.22981834e-01 8.84409130e-01
6.53565824e-01 -8.62900168e-02 -1.12376869e+00 4.92938221e-01
4.77395475e-01 -4.60773319e-01 1.28553593e+00 -1.10907793e+00
4.71965373e-01 2.11550936e-01 -3.98702979e-01 -1.27467847e+00
-3.65906805e-01 -1.05892098e+00 -1.07151695e-01 6.19446635e-01
5.58935046e-01 -5.20686388e-01 1.25499284e+00 4.86672014e-01
-4.45570201e-01 -1.19701409e+00 -1.45112979e+00 -7.21363723e-01
5.31801820e-01 -4.04427797e-01 1.80874839e-01 1.11146331e+00
7.80919492e-02 -7.47360140e-02 -6.74482703e-01 -1.06675364e-01
6.28920019e-01 4.48646024e-02 3.91626805e-01 -1.17381704e+00
-5.44437110e-01 -6.39074445e-01 -2.43600518e-01 -1.40093684e+00
5.74614525e-01 -1.17751002e+00 3.93192321e-02 -9.24120724e-01
1.07575223e-01 -6.95877910e-01 -4.34911698e-01 3.94598693e-01
-1.31201111e-02 1.55527696e-01 -1.43360406e-01 7.80221447e-02
-2.63725013e-01 8.38569045e-01 1.09821868e+00 -1.18432723e-01
-3.40215117e-01 2.12691247e-01 -9.83953357e-01 8.88869286e-01
6.63238168e-01 -4.51395005e-01 -2.65496105e-01 -5.76599717e-01
4.54644859e-01 -2.25125141e-02 -1.71627998e-02 -6.41901314e-01
1.32689461e-01 -2.72017986e-01 8.69586691e-02 -1.56107754e-01
6.06794119e-01 -4.56588507e-01 -3.16367686e-01 3.10471535e-01
-6.99463606e-01 -6.33998886e-02 1.18694142e-01 5.05055666e-01
-2.21117929e-01 -6.32868409e-01 9.46245790e-01 -1.54033586e-01
-3.10076177e-01 6.48995757e-01 -1.94871947e-01 4.14114475e-01
7.97546983e-01 -1.26236126e-01 7.17892721e-02 -2.22907811e-01
-1.03833652e+00 -5.96223436e-02 5.36247268e-02 -2.96106875e-01
7.85185039e-01 -1.22490907e+00 -9.60878432e-01 3.54562163e-01
-3.91576231e-01 1.59242779e-01 7.72333071e-02 8.55791807e-01
-5.00427604e-01 4.63175029e-01 1.36220157e-01 -4.38331127e-01
-9.22289908e-01 4.20557052e-01 4.42218214e-01 -2.40106806e-01
-4.39418316e-01 1.48869872e+00 1.50355518e-01 -1.90383583e-01
4.83047247e-01 -3.24456632e-01 -8.19216892e-02 6.89191837e-03
2.33385816e-01 4.46210533e-01 2.26484880e-01 -4.06571090e-01
-9.02140588e-02 6.49396479e-01 -1.50433118e-02 1.19901516e-01
1.75450814e+00 1.99231207e-01 -3.03402185e-01 1.91186622e-01
1.77388906e+00 -1.47852391e-01 -1.04820669e+00 -4.45888758e-01
-1.99991897e-01 -4.33818281e-01 1.68757930e-01 -7.22045153e-02
-1.29673088e+00 8.53414178e-01 4.12924975e-01 1.35347480e-02
1.13539302e+00 -1.50885597e-01 7.68493176e-01 9.69665289e-01
3.05903137e-01 -1.17641199e+00 -6.32990301e-02 6.92916274e-01
1.07441461e+00 -1.06556129e+00 1.50028244e-01 -3.88639569e-01
-2.64466077e-01 1.04713809e+00 1.12640880e-01 -4.58633095e-01
9.69919384e-01 3.00309598e-01 -6.85099900e-01 -1.72160774e-01
-5.38286984e-01 6.06091060e-02 2.73634881e-01 4.26210970e-01
6.43969715e-01 -9.19735581e-02 -6.11948609e-01 4.12336200e-01
-3.22911918e-01 5.77998757e-02 1.98718578e-01 8.59044850e-01
-4.83203411e-01 -9.54404712e-01 -1.49238825e-01 7.48612344e-01
-6.29069924e-01 -3.63958925e-01 5.40586077e-02 4.71026570e-01
8.08956698e-02 4.14029866e-01 -9.39633697e-02 -9.54375640e-02
1.55923724e-01 7.49903768e-02 3.85819048e-01 -5.75906575e-01
-3.14959168e-01 -7.16381241e-03 -1.88699663e-01 -4.59206104e-01
-2.82292068e-01 -6.54297769e-01 -1.07056665e+00 -3.23735446e-01
-2.22599968e-01 9.70342085e-02 7.99315095e-01 9.68535244e-01
4.23469961e-01 2.98880994e-01 6.79518044e-01 -6.62768722e-01
-1.00978935e+00 -6.73860490e-01 -9.83057082e-01 3.58927757e-01
3.68370801e-01 -6.05781972e-01 -5.62759936e-01 8.59925225e-02] | [7.955105304718018, 3.8680319786071777] |
dea24688-3ef1-4160-b13b-f4c38fb95146 | modeling-task-interactions-in-document-level | 2205.01909 | null | https://arxiv.org/abs/2205.01909v1 | https://arxiv.org/pdf/2205.01909v1.pdf | Modeling Task Interactions in Document-Level Joint Entity and Relation Extraction | We target on the document-level relation extraction in an end-to-end setting, where the model needs to jointly perform mention extraction, coreference resolution (COREF) and relation extraction (RE) at once, and gets evaluated in an entity-centric way. Especially, we address the two-way interaction between COREF and RE that has not been the focus by previous work, and propose to introduce explicit interaction namely Graph Compatibility (GC) that is specifically designed to leverage task characteristics, bridging decisions of two tasks for direct task interference. Our experiments are conducted on DocRED and DWIE; in addition to GC, we implement and compare different multi-task settings commonly adopted in previous work, including pipeline, shared encoders, graph propagation, to examine the effectiveness of different interactions. The result shows that GC achieves the best performance by up to 2.3/5.1 F1 improvement over the baseline. | ['Jinho D. Choi', 'Liyan Xu'] | 2022-05-04 | null | https://aclanthology.org/2022.naacl-main.395 | https://aclanthology.org/2022.naacl-main.395.pdf | naacl-2022-7 | ['document-level-relation-extraction', 'joint-entity-and-relation-extraction'] | ['natural-language-processing', 'natural-language-processing'] | [ 2.29256332e-01 7.09818304e-01 -1.66039586e-01 -2.09678933e-01
-9.77055013e-01 -5.64336538e-01 8.02653730e-01 2.52077192e-01
-6.58905864e-01 5.87545395e-01 6.26624227e-01 -4.91098493e-01
-1.77602887e-01 -3.54519129e-01 -7.93837428e-01 -2.76356250e-01
-3.34418058e-01 6.02881908e-01 2.49668077e-01 -1.19497858e-01
-2.21628755e-01 1.13756336e-01 -8.13954711e-01 3.50021392e-01
7.39461482e-01 4.58387494e-01 2.30082899e-01 6.00872099e-01
3.33736986e-02 9.17757928e-01 -4.18468446e-01 -8.81881893e-01
-1.28744729e-02 8.86657238e-02 -1.30157375e+00 -1.72530532e-01
3.44450325e-01 -4.84642088e-02 -4.43033636e-01 8.89375985e-01
6.78780615e-01 1.53147252e-02 4.19018477e-01 -1.07304418e+00
-4.61296618e-01 1.35541141e+00 -7.71524012e-01 2.41774336e-01
3.00025821e-01 -1.07949547e-01 1.34902370e+00 -7.53060758e-01
9.17541564e-01 1.35658765e+00 6.10056639e-01 3.14963371e-01
-1.22440052e+00 -6.29330933e-01 3.15410435e-01 1.40115261e-01
-1.17449188e+00 -5.84556043e-01 5.58072805e-01 -1.91129863e-01
1.56460690e+00 6.46341071e-02 -6.22077053e-03 1.26603937e+00
-2.26773452e-02 8.80428076e-01 7.24864960e-01 -5.91720164e-01
-2.83650786e-01 -1.03897646e-01 5.72988749e-01 4.08601642e-01
3.61654282e-01 -6.55860901e-02 -5.80406070e-01 1.46783456e-01
1.32016703e-01 -4.92327005e-01 -7.67916977e-01 -2.18259782e-01
-1.19172096e+00 4.80661154e-01 5.63547611e-01 4.71317917e-01
-3.25715303e-01 -1.23187120e-03 5.86834490e-01 2.80636132e-01
4.66048717e-01 5.27309418e-01 -7.38709986e-01 -1.05833389e-01
-6.51456475e-01 2.09841967e-01 1.05921876e+00 1.22369432e+00
4.42219734e-01 -6.80120885e-01 -7.83859670e-01 8.22640419e-01
3.38319063e-01 1.16019078e-01 6.77152351e-02 -6.23583734e-01
9.88888979e-01 5.78752279e-01 -1.88647509e-01 -8.13193023e-01
-7.11899877e-01 -8.36409807e-01 -6.32592201e-01 -5.75713634e-01
3.82264316e-01 -4.65085983e-01 -7.13609040e-01 2.12197948e+00
2.54410595e-01 2.14394107e-01 2.19484955e-01 7.96396017e-01
1.09007895e+00 1.80728346e-01 4.69587445e-01 -1.96139619e-01
1.79550707e+00 -1.30950475e+00 -1.04218864e+00 -4.98997480e-01
1.09066939e+00 -6.90027654e-01 8.45207155e-01 1.57601796e-02
-1.09437430e+00 -2.72183418e-01 -9.77657378e-01 -4.60486799e-01
-2.67525822e-01 2.11042270e-01 6.43294394e-01 2.53804654e-01
-9.43382978e-01 5.36822736e-01 -8.29663098e-01 -4.80600238e-01
1.40295178e-01 3.90663773e-01 -5.96631765e-01 6.29001122e-04
-1.60049033e+00 1.27896893e+00 5.28031886e-01 2.21179649e-01
-4.81860697e-01 -7.79007196e-01 -9.85933840e-01 3.31382453e-01
8.23989034e-01 -8.61915290e-01 1.26315176e+00 -2.02006415e-01
-1.12949371e+00 8.96911025e-01 -2.51682520e-01 -5.64924777e-01
2.95448214e-01 -6.18263543e-01 -3.54330301e-01 -1.78234816e-01
-1.24271207e-01 5.43624043e-01 1.37961969e-01 -1.12830567e+00
-4.08474892e-01 -3.77085268e-01 4.15817112e-01 3.39780867e-01
-1.31995499e-01 4.71428841e-01 -9.56892550e-01 -3.93757582e-01
-2.88093746e-01 -9.40903962e-01 1.34357437e-01 -9.11671042e-01
-9.67663229e-01 -4.48338866e-01 5.66789091e-01 -9.45047796e-01
1.61183989e+00 -2.04460049e+00 3.85177135e-01 -1.52310589e-03
3.96097273e-01 4.40038025e-01 -2.17460632e-01 6.42005742e-01
-1.93612471e-01 2.31039196e-01 -3.37880701e-01 -8.33041847e-01
9.88666341e-02 -8.96458551e-02 2.28546068e-01 1.88747853e-01
3.68626922e-01 1.16271675e+00 -8.29109073e-01 -3.92918676e-01
-3.50381047e-01 5.00238478e-01 -5.63627422e-01 2.44844005e-01
-9.83542800e-02 1.97938651e-01 -3.64714026e-01 3.30913514e-01
7.15469301e-01 -4.25758392e-01 7.97373533e-01 -5.35433233e-01
-5.12913316e-02 9.60120082e-01 -1.13510966e+00 1.80309319e+00
-4.87070739e-01 4.42141563e-01 2.61879593e-01 -6.86552346e-01
5.61727405e-01 4.95320559e-01 5.36471903e-02 -6.36088312e-01
6.36811852e-02 -9.63521004e-02 4.53050226e-01 -3.28677177e-01
4.67308164e-01 3.04656178e-01 -1.41935021e-01 3.71494979e-01
3.94856334e-01 6.32048488e-01 3.00255388e-01 7.87055612e-01
1.60880053e+00 2.61995107e-01 3.33872616e-01 -2.64421314e-01
4.31212068e-01 -1.74098253e-01 6.01751089e-01 6.24816298e-01
2.03024030e-01 4.02799159e-01 9.31562483e-01 3.25160116e-01
-5.47547758e-01 -5.42321086e-01 1.10401399e-01 1.13276112e+00
1.21160842e-01 -8.64387572e-01 -7.41795659e-01 -1.00856304e+00
-8.20897892e-02 7.43829548e-01 -5.51151693e-01 -2.89962322e-01
-9.36895788e-01 -8.12005997e-01 7.43537188e-01 4.62133020e-01
5.08965492e-01 -1.10851085e+00 -2.93991864e-01 3.31220180e-01
-4.44099516e-01 -1.76604557e+00 -6.84979796e-01 3.66106331e-01
-4.40180808e-01 -1.23016024e+00 -3.54207814e-01 -4.41794485e-01
1.46501720e-01 1.46296456e-01 1.52618098e+00 5.83585352e-02
7.69730806e-02 8.18645731e-02 -4.97787595e-01 -5.00903912e-02
3.77777964e-02 7.13850200e-01 -3.94960374e-01 -1.44280955e-01
4.87804115e-01 -5.06896734e-01 -3.74864757e-01 -7.46053159e-02
-4.73526776e-01 2.08242655e-01 8.61256719e-01 7.93457031e-01
1.67898953e-01 -2.53783613e-01 5.64311087e-01 -1.60356879e+00
7.39541113e-01 -5.34153581e-01 -2.03308806e-01 5.63048899e-01
-7.15380967e-01 2.11930305e-01 2.30889380e-01 -1.33428946e-01
-1.28550529e+00 -1.86537579e-01 -9.49478894e-03 -1.59518868e-01
-5.50704487e-02 8.32090676e-01 -5.71721673e-01 3.74676436e-01
3.47857833e-01 -3.34488720e-01 -3.96826357e-01 -7.41819501e-01
4.61078048e-01 6.28700852e-01 6.84254229e-01 -6.66050851e-01
5.21052897e-01 -1.29034013e-01 -2.95363456e-01 -3.03536326e-01
-9.49601114e-01 -5.66804707e-01 -6.63349152e-01 3.39884430e-01
1.00475883e+00 -1.12943304e+00 -9.21382129e-01 1.36012614e-01
-1.60282624e+00 -3.51475269e-01 1.05283350e-01 5.09390473e-01
-6.89876452e-03 2.11276174e-01 -1.00198781e+00 -6.10275388e-01
-6.34594977e-01 -1.26987123e+00 1.12488580e+00 4.92523760e-02
-2.48380199e-01 -9.99265432e-01 1.06704287e-01 5.19372582e-01
2.97186375e-01 1.70989737e-01 1.12409997e+00 -1.07674778e+00
-4.83777940e-01 2.77374625e-01 -6.36931002e-01 -1.55205563e-01
-1.10020354e-01 -2.65722781e-01 -1.11031651e+00 -3.94863725e-01
-5.10342300e-01 -2.76245326e-01 1.02794671e+00 -2.01718695e-03
6.09465241e-01 -2.31704310e-01 -7.81647027e-01 4.89210814e-01
1.21592355e+00 1.24735339e-02 7.06124067e-01 2.53367066e-01
9.24944043e-01 5.83620012e-01 4.94722426e-01 -1.00433566e-01
1.03671050e+00 1.08295846e+00 2.40682289e-01 -2.94487983e-01
-4.96378928e-01 -2.50537753e-01 3.19590390e-01 7.31825292e-01
-7.51202032e-02 -5.22722840e-01 -9.19737816e-01 5.30455410e-01
-2.09017968e+00 -7.48127639e-01 -3.69294614e-01 1.96675050e+00
1.04486883e+00 2.65397161e-01 7.46233910e-02 5.25713265e-02
6.60534680e-01 1.67399138e-01 -9.93703529e-02 -3.03578049e-01
2.23367047e-02 3.71277422e-01 6.16242826e-01 6.05779469e-01
-1.22724104e+00 1.33795643e+00 5.44724703e+00 5.35865664e-01
-1.00792849e+00 2.58341610e-01 1.74044847e-01 8.54499042e-02
-2.31192812e-01 3.68219316e-01 -9.19097185e-01 9.42396000e-02
8.95226717e-01 1.24156348e-01 2.27608949e-01 2.61026084e-01
-1.86265305e-01 7.85335302e-02 -1.38584495e+00 4.98627037e-01
-2.67659485e-01 -9.26371098e-01 -2.46760994e-01 8.24287012e-02
1.98106453e-01 2.25223079e-01 -4.44845527e-01 7.02692032e-01
5.57164729e-01 -8.92016888e-01 4.06135201e-01 1.72025159e-01
5.86868942e-01 -5.60947120e-01 9.52166855e-01 2.42783964e-01
-1.28237545e+00 1.94003075e-01 3.77215743e-02 1.69847026e-01
4.62079555e-01 7.64348149e-01 -8.03355098e-01 1.14444101e+00
5.34065545e-01 6.30413115e-01 -5.45386314e-01 7.00956941e-01
-5.09942830e-01 6.28494322e-01 -6.27297089e-02 3.02824587e-01
1.63877055e-01 2.10839808e-01 5.99714279e-01 1.72085667e+00
-6.22726083e-02 1.04351938e-01 1.60705850e-01 7.27985263e-01
-5.41726112e-01 7.94379637e-02 -3.13207000e-01 2.02156585e-02
8.87043655e-01 1.55765522e+00 -2.87423551e-01 -1.42473787e-01
-5.99287868e-01 8.77824128e-01 1.02996838e+00 3.02823275e-01
-9.43271399e-01 -4.70768034e-01 4.53960478e-01 -3.06035555e-03
3.72707516e-01 -5.70417382e-02 2.66307481e-02 -1.19977963e+00
8.23066011e-02 -8.55771661e-01 6.16186559e-01 -3.07396591e-01
-1.06127524e+00 7.07809389e-01 1.14455760e-01 -5.42862117e-01
-2.59460568e-01 -3.36505800e-01 -7.54148424e-01 9.47875738e-01
-1.65834427e+00 -1.40303051e+00 -3.83842364e-02 4.52108115e-01
1.08103469e-01 2.23843828e-01 7.06878066e-01 7.22386956e-01
-1.05369389e+00 1.07026577e+00 -6.65911317e-01 3.48262936e-01
9.87264037e-01 -1.45216215e+00 6.92390084e-01 1.04434741e+00
1.74370527e-01 7.95773566e-01 5.10071337e-01 -7.03447700e-01
-1.47148848e+00 -1.08049798e+00 1.66623020e+00 -2.83142835e-01
5.18929243e-01 -5.86924076e-01 -8.85681808e-01 1.11168647e+00
4.69469070e-01 -9.80564579e-02 4.25992638e-01 9.07695115e-01
-5.34222662e-01 1.46201104e-01 -9.36714232e-01 4.32249874e-01
1.59092987e+00 -3.70512813e-01 -5.98527312e-01 1.36072472e-01
1.08252084e+00 -6.75945282e-01 -1.15444160e+00 7.32935131e-01
1.31450191e-01 -7.19041288e-01 8.66794527e-01 -6.91832006e-01
2.78748691e-01 -1.01990856e-01 -5.37481308e-02 -1.38082278e+00
-7.01804399e-01 -7.90011227e-01 -3.94779444e-01 1.80888915e+00
9.08782005e-01 -5.56458831e-01 2.25859448e-01 5.96974015e-01
-3.27206492e-01 -6.97385073e-01 -5.97288728e-01 -5.24890840e-01
-1.15539856e-01 -1.33157805e-01 5.91577053e-01 1.11834681e+00
2.76273131e-01 1.05550575e+00 -3.08753192e-01 3.52512538e-01
4.09110487e-01 1.36027694e-01 7.91110277e-01 -1.08678007e+00
-6.34396970e-01 -3.73197824e-01 2.37204850e-01 -9.70875323e-01
4.48247701e-01 -1.13661075e+00 -5.72685413e-02 -1.61149180e+00
3.49216223e-01 -5.31518161e-01 -3.84707719e-01 7.43558347e-01
-5.86698174e-01 -1.28550529e-01 4.03665811e-01 8.79128501e-02
-7.27717519e-01 3.40212792e-01 1.18774641e+00 4.30658422e-02
-1.88777372e-01 -2.93862194e-01 -1.27835751e+00 3.38906825e-01
3.76036674e-01 -4.46568131e-01 -3.15391064e-01 -8.52214813e-01
1.00392379e-01 2.03451797e-01 2.00656634e-02 -5.69177210e-01
5.45724988e-01 3.92369956e-01 -1.58271298e-01 -2.72860289e-01
1.06918499e-01 -6.43137634e-01 9.11439806e-02 6.37566745e-02
-4.71734345e-01 5.91124315e-03 1.18387505e-01 3.25741410e-01
-2.30982006e-01 6.19505607e-02 3.54814738e-01 1.90233603e-01
-4.17693973e-01 2.00321078e-01 4.10774767e-01 3.56735379e-01
6.80168033e-01 4.09631997e-01 -6.86130285e-01 -7.56262988e-02
-8.20906043e-01 5.63951790e-01 -6.59813210e-02 4.66903001e-01
3.95947620e-02 -1.01959562e+00 -9.94803190e-01 -2.77690291e-01
1.80752009e-01 6.94396570e-02 9.96884927e-02 1.24281192e+00
1.56185776e-01 5.33650160e-01 4.05698061e-01 -2.17955068e-01
-1.26878881e+00 5.02136946e-01 2.81762987e-01 -1.17350245e+00
-8.18472743e-01 8.77162278e-01 1.09147832e-01 -4.80184019e-01
4.33909297e-01 -1.97940066e-01 -4.49553639e-01 6.27168491e-02
3.05858493e-01 2.58880585e-01 5.24458885e-01 -4.99752581e-01
-7.27071524e-01 2.19581112e-01 -5.30517578e-01 5.06091490e-02
1.26637650e+00 -7.06434175e-02 -1.05247281e-01 -1.20763499e-02
1.15419924e+00 1.71481147e-01 -9.49419320e-01 -5.62940776e-01
6.36474252e-01 1.28410548e-01 2.13740215e-01 -1.20054841e+00
-1.20485270e+00 6.78855062e-01 1.85838640e-02 1.32080046e-02
9.62441862e-01 2.58467287e-01 9.15749788e-01 1.76376656e-01
2.57516414e-01 -6.60665393e-01 -5.81571877e-01 7.84550905e-01
8.18403602e-01 -1.03797853e+00 -6.24601990e-02 -8.63266945e-01
-6.56716585e-01 5.43357790e-01 7.32456326e-01 9.17054042e-02
5.99856675e-01 7.19105303e-01 -2.96676219e-01 -3.34745735e-01
-1.14295328e+00 -6.05487943e-01 4.99010593e-01 2.55813420e-01
1.06560004e+00 1.12073377e-01 -4.95925426e-01 9.97516811e-01
-8.80784530e-04 -1.55349299e-02 9.27248225e-02 7.50152528e-01
1.47036448e-01 -1.45195484e+00 2.98522323e-01 1.82098106e-01
-8.05721164e-01 -5.30173957e-01 -4.76899624e-01 1.00131500e+00
-8.51929095e-03 9.77958798e-01 -3.75155848e-03 -4.74441677e-01
6.48359478e-01 2.16102198e-01 5.32241821e-01 -6.81363106e-01
-1.12128270e+00 -2.74945367e-02 8.40829849e-01 -5.99836767e-01
-3.20437431e-01 -5.04688501e-01 -1.37454665e+00 -1.42561004e-01
-4.55420643e-01 1.14306383e-01 2.60184050e-01 1.15243268e+00
7.60987163e-01 9.20792043e-01 1.91762447e-01 -5.26450932e-01
-2.83625364e-01 -1.40455902e+00 -1.21900000e-01 5.55022478e-01
1.42199755e-01 -7.82468438e-01 -2.38915831e-01 -2.98273534e-01] | [9.275753021240234, 8.854002952575684] |
8e55cbab-ff1e-4e0d-a5c5-77b05df97eac | zero-shot-keyword-spotting-for-visual-speech | 1807.08469 | null | http://arxiv.org/abs/1807.08469v2 | http://arxiv.org/pdf/1807.08469v2.pdf | Zero-shot keyword spotting for visual speech recognition in-the-wild | Visual keyword spotting (KWS) is the problem of estimating whether a text
query occurs in a given recording using only video information. This paper
focuses on visual KWS for words unseen during training, a real-world, practical
setting which so far has received no attention by the community. To this end,
we devise an end-to-end architecture comprising (a) a state-of-the-art visual
feature extractor based on spatiotemporal Residual Networks, (b) a
grapheme-to-phoneme model based on sequence-to-sequence neural networks, and
(c) a stack of recurrent neural networks which learn how to correlate visual
features with the keyword representation. Different to prior works on KWS,
which try to learn word representations merely from sequences of graphemes
(i.e. letters), we propose the use of a grapheme-to-phoneme encoder-decoder
model which learns how to map words to their pronunciation. We demonstrate that
our system obtains very promising visual-only KWS results on the challenging
LRS2 database, for keywords unseen during training. We also show that our
system outperforms a baseline which addresses KWS via automatic speech
recognition (ASR), while it drastically improves over other recently proposed
ASR-free KWS methods. | ['Georgios Tzimiropoulos', 'Themos Stafylakis'] | 2018-07-23 | zero-shot-keyword-spotting-for-visual-speech-1 | http://openaccess.thecvf.com/content_ECCV_2018/html/Themos_Stafylakis_Zero-shot_keyword_search_ECCV_2018_paper.html | http://openaccess.thecvf.com/content_ECCV_2018/papers/Themos_Stafylakis_Zero-shot_keyword_search_ECCV_2018_paper.pdf | eccv-2018-9 | ['visual-keyword-spotting'] | ['computer-vision'] | [ 3.74246627e-01 -1.13825127e-01 -4.29164842e-02 -6.26457781e-02
-1.04982483e+00 -5.54363668e-01 8.35201502e-01 -1.49417192e-01
-6.50817335e-01 2.35698253e-01 2.74814934e-01 -6.47772491e-01
5.89423299e-01 -8.21415931e-02 -1.14982629e+00 -7.21537054e-01
1.31723598e-01 1.65354922e-01 1.02090023e-01 -1.10475242e-01
3.68356436e-01 2.80482501e-01 -1.49090397e+00 5.05828977e-01
7.49818310e-02 8.79816592e-01 6.43476129e-01 1.25519288e+00
-3.68846022e-02 1.09236097e+00 -5.96342981e-01 -2.57280450e-02
-7.00349510e-02 -5.76791704e-01 -6.06968045e-01 6.34203330e-02
7.50521064e-01 -3.41577500e-01 -9.37962711e-01 8.34821045e-01
5.58932006e-01 1.62708133e-01 7.49598920e-01 -9.42459166e-01
-1.37963450e+00 2.89086699e-01 -3.93151164e-01 3.74705940e-01
4.11148012e-01 2.57087857e-01 1.18521249e+00 -1.36850858e+00
8.46223772e-01 1.11589396e+00 1.49211526e-01 6.33986056e-01
-1.18732798e+00 -3.13140839e-01 3.30909342e-01 7.64928639e-01
-1.61363935e+00 -7.63590455e-01 6.92294121e-01 -2.95330435e-01
1.70140696e+00 3.19003493e-01 7.04358578e-01 1.66697717e+00
1.63753271e-01 1.18339908e+00 7.95223475e-01 -7.56920576e-01
9.19169858e-02 1.91294938e-01 3.26890424e-02 8.30540359e-01
-3.83271873e-01 -3.81437242e-02 -1.03269958e+00 1.37957707e-01
5.44123232e-01 -1.69664890e-01 -7.51761138e-01 -4.85867143e-01
-1.36989927e+00 6.89519584e-01 2.92886823e-01 3.96064043e-01
-3.37455183e-01 5.74608862e-01 3.76956761e-01 5.18928349e-01
4.75028545e-01 1.23073071e-01 -3.74581039e-01 -2.61663586e-01
-1.24018049e+00 -1.66144565e-01 5.82235336e-01 9.93859768e-01
4.83374476e-01 4.21669602e-01 -4.42364037e-01 6.82259381e-01
2.48327747e-01 6.45797193e-01 8.16802382e-01 -1.41095266e-01
5.30791759e-01 4.88542728e-02 8.38247128e-03 -7.92547882e-01
-1.35461047e-01 -1.43223852e-01 -4.94771481e-01 -5.07490560e-02
-2.30382159e-02 3.70403260e-01 -1.34879756e+00 1.61770689e+00
-1.31931871e-01 5.47301650e-01 1.88029602e-01 1.01904154e+00
9.97790992e-01 1.15719473e+00 -2.37151548e-01 -2.43052244e-01
1.36736882e+00 -1.16326272e+00 -8.89389455e-01 -3.21977109e-01
5.80291510e-01 -4.87503469e-01 1.37410796e+00 3.88077289e-01
-1.10741639e+00 -4.14528370e-01 -1.08318853e+00 -4.27402407e-01
-7.55726576e-01 3.55530560e-01 -2.85584535e-02 2.11918667e-01
-1.69264472e+00 1.23665243e-01 -7.31639326e-01 -4.45692778e-01
-2.33339462e-02 1.14009477e-01 -5.56887209e-01 9.16997865e-02
-1.21557188e+00 8.99763525e-01 3.97439837e-01 3.36464435e-01
-1.12528670e+00 -3.16585809e-01 -1.23233449e+00 1.52555212e-01
7.63940096e-01 -6.22216821e-01 1.15433884e+00 -1.08721614e+00
-1.56781328e+00 9.40096259e-01 -6.00633264e-01 -5.68478465e-01
1.04519740e-01 -2.18458876e-01 -5.97984493e-01 5.73928893e-01
-2.90367723e-01 6.88908041e-01 1.50480461e+00 -1.19519091e+00
-4.26419348e-01 -1.75816342e-01 -2.74542481e-01 4.26228464e-01
-4.49460804e-01 1.28718421e-01 -1.11674809e+00 -1.03066707e+00
-2.71176666e-01 -8.20527732e-01 3.00665766e-01 -8.71578082e-02
-5.91289222e-01 -3.33474517e-01 9.41043973e-01 -1.16132200e+00
1.19366574e+00 -2.23975492e+00 5.84629238e-01 3.51680666e-02
2.09630772e-01 4.18105543e-01 -3.38718623e-01 5.91986537e-01
-2.24634334e-01 -8.81107002e-02 6.01189323e-02 -6.30561590e-01
-3.03252414e-02 8.02180693e-02 -6.49289429e-01 7.55195141e-01
1.88062161e-01 1.44004250e+00 -7.56267369e-01 -2.88896084e-01
3.57107133e-01 6.94793105e-01 -1.24375567e-01 4.96034086e-01
-2.27590606e-01 -9.71436203e-02 5.57641014e-02 6.25449538e-01
1.08239613e-01 -3.98383379e-01 -1.13430023e-02 -1.04453221e-01
2.49544699e-02 3.28508258e-01 -7.11215377e-01 1.96531498e+00
-5.52185118e-01 1.06392431e+00 -7.23476931e-02 -1.30940187e+00
5.65284610e-01 5.97103238e-01 -1.70719355e-01 -1.01558638e+00
-1.55033663e-01 9.45676267e-02 -6.00985289e-01 -5.64359367e-01
7.49961197e-01 2.15949789e-01 -7.77294412e-02 1.64156199e-01
5.31916440e-01 2.57262319e-01 -1.11455306e-01 5.50226510e-01
1.16421294e+00 1.46828339e-01 3.72861713e-01 3.29670310e-02
4.31610286e-01 -2.63425648e-01 -1.97574750e-01 1.03112435e+00
9.05973464e-02 9.02434886e-01 2.96003163e-01 -1.43294558e-01
-9.92646396e-01 -1.10268605e+00 4.29938853e-01 1.18487716e+00
1.36262625e-01 -4.70626771e-01 -5.72328448e-01 -6.27583206e-01
-1.88589334e-01 8.04714441e-01 -8.43666077e-01 -2.66563088e-01
-5.77581763e-01 2.41090097e-02 5.86224437e-01 5.77476442e-01
-1.13612786e-01 -1.40227532e+00 -5.55480540e-01 3.55689414e-02
-1.68647975e-01 -1.41686225e+00 -1.06192696e+00 4.20337886e-01
-2.34678954e-01 -6.65248513e-01 -1.23239183e+00 -1.07663155e+00
5.23411453e-01 5.13687313e-01 1.01853800e+00 -1.42248645e-01
-4.80621636e-01 1.02751601e+00 -5.70357740e-01 -7.25975186e-02
-2.31799364e-01 -6.25010654e-02 -6.56754151e-02 2.90198743e-01
3.72907996e-01 -2.95379132e-01 -6.87078416e-01 -5.51733561e-02
-8.93273056e-01 2.50778139e-01 7.00500846e-01 8.85517061e-01
6.01563334e-01 -6.20107770e-01 1.39412954e-01 -4.50643182e-01
5.75133562e-01 -2.96120107e-01 -5.68553507e-01 6.35778666e-01
-5.74704051e-01 2.03619242e-01 7.08518803e-01 -6.76180780e-01
-5.50152481e-01 9.45154876e-02 -8.46995134e-03 -1.02539194e+00
-1.32439747e-01 4.53916103e-01 9.60223973e-02 2.37677433e-03
3.47534359e-01 1.14938998e+00 -2.64003187e-01 -4.80537921e-01
7.38930702e-01 8.92150044e-01 1.01757646e+00 5.78716435e-02
8.08530867e-01 3.63945544e-01 -4.50802475e-01 -1.37818789e+00
-3.25301230e-01 -9.93955791e-01 -3.64302754e-01 -2.44623348e-01
9.78054404e-01 -1.19267666e+00 -6.44440889e-01 3.00220072e-01
-1.35099614e+00 -3.76054943e-01 -1.42806664e-01 3.82967234e-01
-7.49561429e-01 5.93586564e-01 -6.25373363e-01 -8.63003731e-01
-4.75475043e-01 -9.71362293e-01 1.57282352e+00 -1.39689967e-01
-7.45123178e-02 -8.29053879e-01 1.48492958e-02 9.60651413e-02
3.40669125e-01 -1.46212712e-01 6.13567770e-01 -6.84746146e-01
-6.31277442e-01 -9.01301280e-02 -2.07242906e-01 2.22440064e-01
-1.42806783e-01 -3.66211295e-01 -1.04244006e+00 -5.71439862e-01
-1.07806583e-04 -4.83336121e-01 1.33065903e+00 2.61703908e-01
1.30178702e+00 -5.39965153e-01 -3.17947328e-01 7.21205890e-01
1.42654395e+00 3.42148066e-01 6.55685544e-01 1.27474517e-01
1.01445127e+00 2.82973111e-01 1.79274365e-01 1.55364648e-01
3.39662671e-01 1.05055571e+00 3.49919975e-01 -1.21547915e-01
-4.14531499e-01 -4.79716182e-01 7.30498433e-01 1.04770064e+00
2.56680995e-01 -8.09441805e-01 -6.25522614e-01 9.26122069e-01
-1.92120850e+00 -8.05969775e-01 3.17446202e-01 2.01864982e+00
6.19220853e-01 -5.14618009e-02 -2.91284006e-02 -7.28259832e-02
6.40897572e-01 6.51405096e-01 -5.98832965e-01 -5.35393476e-01
-2.27158844e-01 2.99250692e-01 4.92950678e-01 5.19368947e-01
-9.37955081e-01 1.27621841e+00 5.65194654e+00 1.04710770e+00
-1.35075212e+00 1.02931663e-01 3.16070497e-01 -2.51839817e-01
-3.23214054e-01 -2.69883811e-01 -6.38476908e-01 3.01130772e-01
1.20040977e+00 2.54120231e-01 7.95832515e-01 7.18030751e-01
1.19500227e-01 7.00515062e-02 -1.06766701e+00 1.37321305e+00
8.05661440e-01 -1.44767988e+00 2.39200503e-01 -8.02551582e-02
1.04845867e-01 1.07286140e-01 2.82402724e-01 2.87976235e-01
-1.45168692e-01 -1.27152741e+00 1.10791683e+00 5.63512862e-01
1.31431806e+00 -4.95400727e-01 1.79347038e-01 2.62017787e-01
-1.47838891e+00 2.86731809e-01 -2.35937491e-01 4.33482617e-01
2.23085314e-01 6.09475784e-02 -1.09188926e+00 4.28701431e-01
6.46569073e-01 7.42542505e-01 -5.12828112e-01 5.91198146e-01
-3.88289779e-01 7.81528592e-01 -5.23262471e-02 -2.88979590e-01
5.11164725e-01 2.26187229e-01 5.90252161e-01 1.79659832e+00
3.34200978e-01 -1.40843481e-01 -1.17914595e-01 6.28946185e-01
-2.52064109e-01 2.77233452e-01 -9.11543667e-01 -4.32043642e-01
4.19532834e-03 1.04025638e+00 -7.13643789e-01 -4.18883055e-01
-7.36666799e-01 1.90766883e+00 5.68375468e-01 8.92185867e-01
-6.61332846e-01 -6.71192944e-01 3.93611312e-01 -1.10399254e-01
9.87129629e-01 -3.59662592e-01 4.91443515e-01 -1.48394942e+00
3.33012253e-01 -7.05364823e-01 5.86314015e-02 -1.13527620e+00
-8.84039044e-01 6.64959550e-01 -2.16263726e-01 -1.05621505e+00
-6.44513845e-01 -8.28779340e-01 -3.23897183e-01 9.57425177e-01
-1.59783494e+00 -1.09862649e+00 6.08451627e-02 8.34568262e-01
9.52470899e-01 -2.17289269e-01 7.13259697e-01 -9.07024145e-02
-1.60981685e-01 7.02239990e-01 2.20637083e-01 2.06954673e-01
6.77141309e-01 -1.29201615e+00 9.86594498e-01 9.37548041e-01
1.04522860e+00 4.70745087e-01 7.07228184e-01 -5.59590459e-01
-1.90987182e+00 -9.95339930e-01 1.11897576e+00 -3.91609162e-01
7.35222220e-01 -1.16614962e+00 -1.05920386e+00 8.32727492e-01
5.65091252e-01 3.12171161e-01 3.04317772e-01 -3.31230342e-01
-5.67094564e-01 3.36295992e-01 -3.49235058e-01 7.83320487e-01
1.05964291e+00 -1.29577053e+00 -6.80673182e-01 3.63540530e-01
1.02138710e+00 -4.28737640e-01 -1.17066391e-01 -5.27902059e-02
5.88192344e-01 -6.33921385e-01 1.10126972e+00 -7.74408400e-01
1.23042010e-01 -2.94892341e-01 -1.84855565e-01 -1.35718930e+00
-5.97531311e-02 -7.08049953e-01 -5.51018953e-01 8.41172040e-01
4.10231858e-01 -2.67039210e-01 5.04479885e-01 -9.39818099e-02
-8.11838731e-02 -6.70253992e-01 -1.15316832e+00 -8.71828496e-01
-4.80890095e-01 -6.06585979e-01 -1.85479283e-01 7.47111738e-01
2.17365026e-02 5.19939363e-01 -1.02350640e+00 2.28082538e-01
2.21839681e-01 9.65086967e-02 4.09621149e-01 -4.48266476e-01
-5.29118180e-01 -2.69447774e-01 -5.50831795e-01 -1.66235840e+00
3.82127762e-01 -8.57747436e-01 5.08216143e-01 -1.57028937e+00
1.52063191e-01 7.60094285e-01 -2.66852915e-01 4.67558831e-01
-1.74761221e-01 3.21673930e-01 1.70076489e-01 7.90710002e-02
-9.13909018e-01 7.19178498e-01 7.81886339e-01 -3.15173656e-01
-3.24112289e-02 -5.10365665e-01 -3.55124354e-01 3.17537934e-01
3.76235932e-01 -3.14056098e-01 -3.41092616e-01 -3.00237626e-01
1.51304677e-01 3.54089469e-01 4.67874795e-01 -5.16232729e-01
5.47191620e-01 1.77243650e-01 2.48151064e-01 -7.37187266e-01
6.77735388e-01 -6.69180155e-01 -3.37523907e-01 1.12027653e-01
-3.96993250e-01 3.16359133e-01 1.64979562e-01 1.01210845e+00
-3.07259023e-01 -2.24430710e-02 3.15743417e-01 -1.72195539e-01
-1.16740608e+00 1.33364782e-01 -7.12004066e-01 -8.12392458e-02
8.64252269e-01 -2.97760993e-01 -3.16202998e-01 -7.96758711e-01
-6.81590855e-01 1.12540290e-01 2.44276658e-01 6.78078949e-01
1.22640800e+00 -1.25895178e+00 -6.16561711e-01 1.44489333e-01
5.05142868e-01 -5.38492560e-01 1.78866848e-01 5.66526592e-01
-2.23783493e-01 9.15038764e-01 3.36836666e-01 -6.07578754e-01
-1.45781386e+00 1.00888145e+00 1.02585271e-01 -2.22210102e-02
-8.97698283e-01 1.10254502e+00 4.51234370e-01 6.71053678e-02
7.01234221e-01 -4.15322542e-01 -2.45035082e-01 4.81385849e-02
6.23766303e-01 -1.58990428e-01 1.97615027e-01 -9.39142048e-01
-5.42824686e-01 5.93721807e-01 -2.89532542e-01 -4.71197039e-01
1.06934559e+00 -2.75613070e-01 3.86261791e-01 9.09287453e-01
1.65773320e+00 -2.87257612e-01 -1.12333322e+00 -4.27799255e-01
-6.66690618e-02 -1.58464283e-01 2.38655955e-01 -5.87984920e-01
-8.70573223e-01 1.06308734e+00 6.03853106e-01 9.51566175e-02
1.07875192e+00 3.00681204e-01 8.00781667e-01 6.52953744e-01
-3.93226817e-02 -1.14373267e+00 3.04875642e-01 4.33224589e-01
1.10397661e+00 -1.18425012e+00 -5.26449382e-01 6.26712665e-02
-7.45488822e-01 1.19397295e+00 1.82420146e-02 -1.03254728e-01
3.97930235e-01 1.38542488e-01 -5.11247944e-03 -2.40388915e-01
-1.13167453e+00 -4.38515514e-01 8.55859756e-01 4.21303749e-01
8.74082297e-02 -2.86329597e-01 2.19746113e-01 9.56282318e-02
6.04010522e-02 -1.03129849e-01 2.86233872e-01 8.51870298e-01
-2.53705472e-01 -6.09393179e-01 -1.81086883e-01 2.75597692e-01
-3.90281081e-01 -6.35454416e-01 -5.57680428e-01 6.50245428e-01
-4.86524820e-01 7.63459265e-01 9.26516280e-02 -4.04996812e-01
2.94896126e-01 4.49066103e-01 4.82427180e-01 -6.24215961e-01
-4.05570567e-01 2.80526966e-01 -1.29863247e-01 -7.39124835e-01
-2.61197984e-01 -3.98458540e-01 -8.86206985e-01 2.19461232e-01
-2.32231796e-01 3.50161418e-02 7.55286694e-01 1.02447271e+00
3.43569487e-01 5.62492073e-01 3.95674556e-01 -9.51112866e-01
-3.67170215e-01 -8.58746231e-01 -6.95010781e-01 3.73058677e-01
8.92487168e-01 -3.83864850e-01 -4.71344739e-01 2.41676703e-01] | [10.697278022766113, 1.2371104955673218] |
d31d26a1-4a29-4dae-ac17-db625da5f120 | a-survey-in-adversarial-defences-and | 2203.06414 | null | https://arxiv.org/abs/2203.06414v4 | https://arxiv.org/pdf/2203.06414v4.pdf | A Survey of Adversarial Defences and Robustness in NLP | In the past few years, it has become increasingly evident that deep neural networks are not resilient enough to withstand adversarial perturbations in input data, leaving them vulnerable to attack. Various authors have proposed strong adversarial attacks for computer vision and Natural Language Processing (NLP) tasks. As a response, many defense mechanisms have also been proposed to prevent these networks from failing. The significance of defending neural networks against adversarial attacks lies in ensuring that the model's predictions remain unchanged even if the input data is perturbed. Several methods for adversarial defense in NLP have been proposed, catering to different NLP tasks such as text classification, named entity recognition, and natural language inference. Some of these methods not only defend neural networks against adversarial attacks but also act as a regularization mechanism during training, saving the model from overfitting. This survey aims to review the various methods proposed for adversarial defenses in NLP over the past few years by introducing a novel taxonomy. The survey also highlights the fragility of advanced deep neural networks in NLP and the challenges involved in defending them. | ['Balaraman Ravindran', 'Mitesh M. Khapra', 'Sumanth Doddapaneni', 'Shreya Goyal'] | 2022-03-12 | null | null | null | null | ['adversarial-defense'] | ['adversarial'] | [ 2.39749804e-01 3.41265887e-01 2.59703457e-01 -5.05572140e-01
-1.92083061e-01 -1.26116943e+00 6.98743999e-01 1.18276447e-01
-5.81885397e-01 6.81270063e-01 -2.56099273e-02 -4.49567199e-01
1.70566410e-01 -8.97737324e-01 -8.46731663e-01 -7.34455287e-01
-6.64169863e-02 2.24068224e-01 1.27809629e-01 -4.76740271e-01
9.59072337e-02 1.23397720e+00 -1.00189018e+00 3.21435094e-01
7.04615176e-01 8.66928458e-01 -6.39297426e-01 8.62822235e-01
-2.13146120e-01 1.04680645e+00 -1.17088532e+00 -8.03583860e-01
5.10167956e-01 -1.20715201e-02 -8.65371883e-01 -7.06283391e-01
5.91727436e-01 -1.75131351e-01 -8.79223347e-01 1.39490545e+00
5.58673561e-01 1.43971741e-01 4.96657014e-01 -1.51257455e+00
-9.08869445e-01 6.27085388e-01 1.78682413e-02 2.28887767e-01
5.08045144e-02 2.73201644e-01 6.94207132e-01 -5.40362358e-01
1.90551311e-01 1.50689423e+00 7.12733984e-01 1.18289149e+00
-8.84507954e-01 -9.16804552e-01 3.78754556e-01 -5.22640310e-02
-1.08345270e+00 -4.37009662e-01 7.16409028e-01 -2.94136137e-01
1.11036944e+00 3.86489868e-01 2.19918266e-02 1.68113482e+00
7.07270622e-01 7.71347046e-01 6.02017224e-01 -2.25376695e-01
3.33989143e-01 7.13515803e-02 1.54678538e-01 4.78931636e-01
1.95384488e-01 3.80781531e-01 -1.89061686e-01 -6.69578373e-01
1.49929479e-01 -3.00208807e-01 -1.81686953e-01 2.50472333e-02
-6.37463331e-01 1.04718292e+00 5.08915484e-01 4.07881945e-01
-1.84815869e-01 1.74085245e-01 8.67499530e-01 4.70403016e-01
3.45056981e-01 7.61331379e-01 -7.29576766e-01 4.76800144e-01
-3.96432847e-01 3.51529509e-01 9.46254194e-01 6.03755951e-01
2.56936029e-02 6.41945004e-01 8.91768932e-02 6.28132403e-01
3.55350167e-01 6.05779111e-01 5.35021842e-01 -5.95039666e-01
5.76120257e-01 3.87108088e-01 -2.01811135e-01 -1.47747254e+00
-3.61127138e-01 -2.16014311e-01 -1.38949740e+00 6.27421081e-01
4.27567482e-01 -5.46689034e-01 -1.08481050e+00 1.91224706e+00
1.95719585e-01 1.34844612e-02 4.94348526e-01 4.00408864e-01
7.49893248e-01 7.70080030e-01 3.60492766e-01 1.43815517e-01
9.24176931e-01 -5.09918332e-01 -5.41184604e-01 -4.74324346e-01
2.51566172e-01 -5.12624681e-01 6.55361414e-01 4.77313727e-01
-8.80827606e-01 -5.19589603e-01 -1.11735761e+00 3.72253545e-02
-9.05393064e-01 -7.30892658e-01 3.97878945e-01 1.06462872e+00
-7.25915611e-01 7.62763858e-01 -8.56334627e-01 2.34186817e-02
6.96349978e-01 6.23169482e-01 -4.29636389e-01 3.22149515e-01
-1.92927635e+00 9.87049818e-01 6.01123154e-01 1.93070307e-01
-1.03949308e+00 -6.04895890e-01 -7.76381969e-01 1.77295670e-01
-1.29250139e-01 -5.36955476e-01 9.45984542e-01 -1.19684613e+00
-1.29588521e+00 7.32067287e-01 5.29766083e-01 -1.03487992e+00
6.35676086e-01 -4.16113257e-01 -7.06377089e-01 9.67947207e-03
-4.26965684e-01 3.08368772e-01 9.75047171e-01 -9.46046948e-01
-2.12810218e-01 -4.24510509e-01 1.98615417e-01 -1.18839264e-01
-7.15010405e-01 3.97462457e-01 2.88787127e-01 -9.20896769e-01
-3.96278888e-01 -9.80883121e-01 -4.66477752e-01 1.02677837e-01
-6.97675228e-01 -1.27365768e-01 1.12726986e+00 -2.30202556e-01
8.83102357e-01 -2.03885722e+00 -3.59827504e-02 1.64896905e-01
1.76041305e-01 1.16399217e+00 -2.78604269e-01 4.38760668e-01
-4.59040403e-01 5.79002619e-01 -2.66565561e-01 3.79580120e-03
-6.85719624e-02 4.01949346e-01 -1.15643370e+00 5.66894829e-01
3.78983974e-01 8.47665548e-01 -6.27483130e-01 -4.83415276e-02
6.83096051e-02 6.69889510e-01 -3.06063950e-01 1.97643757e-01
-3.41425359e-01 3.97732109e-01 -6.55726016e-01 4.80718434e-01
8.51727545e-01 2.38344386e-01 -5.79406507e-02 1.97087005e-01
5.29398322e-01 -2.39324868e-02 -8.38301778e-01 7.44547844e-01
-1.32105738e-01 8.30893815e-01 1.26076922e-01 -1.18376040e+00
9.75009203e-01 5.73478758e-01 2.60341823e-01 -2.60302693e-01
4.61542517e-01 3.02315615e-02 2.17200160e-01 -3.42865229e-01
2.40413085e-01 -1.38554797e-01 -4.00423229e-01 2.87102729e-01
-6.15277402e-02 -6.68168887e-02 -3.25111985e-01 3.23554307e-01
1.26921034e+00 -4.41164196e-01 2.23931983e-01 -9.83829424e-02
1.05467248e+00 -1.63110435e-01 6.60339355e-01 9.84850645e-01
-6.41331553e-01 2.71938860e-01 3.70074213e-01 -8.20901394e-01
-9.40368950e-01 -1.16854239e+00 8.78341570e-02 9.97321188e-01
-4.07730311e-01 1.38956279e-01 -7.98776925e-01 -1.10269725e+00
1.00176208e-01 6.42905354e-01 -7.80298233e-01 -8.91148329e-01
-6.77181780e-01 -7.52625346e-01 1.61532629e+00 6.06484056e-01
6.73355937e-01 -1.35142565e+00 -1.84222445e-01 1.20126970e-01
8.25926587e-02 -1.23910022e+00 -1.54927030e-01 3.68424833e-01
-7.89080024e-01 -1.14312279e+00 -3.90125573e-01 -8.29974651e-01
5.97333670e-01 -2.35189825e-01 1.08205831e+00 1.90377370e-01
-1.27121463e-01 1.29668012e-01 -3.18558484e-01 -9.60061789e-01
-1.09551990e+00 -3.84008810e-02 6.58903360e-01 -1.02057673e-01
4.18810517e-01 -5.50160885e-01 -1.21404432e-01 1.53224826e-01
-1.26420450e+00 -8.56791198e-01 1.79153010e-01 8.69471371e-01
1.80636764e-01 3.23109329e-01 7.66565859e-01 -1.13765252e+00
8.83098781e-01 -5.52934229e-01 -4.92021739e-01 2.75953740e-01
-2.90724725e-01 -8.52408484e-02 1.45530748e+00 -8.67954731e-01
-7.85094380e-01 -1.37618082e-02 -7.63358891e-01 -6.09802306e-01
-5.23369253e-01 1.88299552e-01 -4.57184106e-01 -6.04087472e-01
1.21310055e+00 1.48958385e-01 -1.60703331e-01 -2.24237606e-01
2.37723351e-01 4.53403980e-01 4.83058959e-01 -4.19794410e-01
1.26527297e+00 4.09419656e-01 7.12205470e-02 -8.03171217e-01
-1.05894339e+00 2.14850470e-01 -5.20277023e-01 5.98772056e-02
7.09927619e-01 -3.42286348e-01 -8.86198223e-01 1.05304253e+00
-1.35500860e+00 -1.24228470e-01 -6.97650239e-02 3.89921553e-02
-2.51555890e-01 5.78673244e-01 -7.98992932e-01 -6.98342204e-01
-7.32661009e-01 -7.82754779e-01 1.88679710e-01 2.60010302e-01
1.17091471e-02 -1.15246129e+00 1.12335615e-01 2.31263369e-01
5.86619437e-01 8.88537645e-01 9.84158337e-01 -1.45694005e+00
5.20730503e-02 -6.68532431e-01 1.96796715e-01 9.70345736e-01
3.13499570e-02 2.53522456e-01 -1.25756478e+00 -3.94377768e-01
4.02927607e-01 -4.93599534e-01 6.77558780e-01 5.97064085e-02
1.16451812e+00 -8.21546137e-01 -1.51041552e-01 7.13747323e-01
1.09966409e+00 4.04651046e-01 7.84313917e-01 4.45915818e-01
7.24796295e-01 6.67032421e-01 3.70414883e-01 -1.56937152e-01
-5.07560909e-01 9.68979150e-02 1.00987422e+00 -8.48361384e-03
5.96171439e-01 1.89060327e-02 5.10487676e-01 2.07789123e-01
3.97733837e-01 -7.96992838e-01 -1.13780141e+00 2.52038181e-01
-1.49849713e+00 -1.09939730e+00 -3.11480425e-02 1.72563052e+00
6.66483104e-01 4.97115642e-01 -3.29999059e-01 1.82329148e-01
7.09936440e-01 3.80737096e-01 -9.39434052e-01 -9.32639360e-01
-3.41395617e-01 2.66328096e-01 5.56934357e-01 4.00879681e-01
-1.54117131e+00 1.28295696e+00 6.93701982e+00 4.91584808e-01
-1.23926365e+00 -1.83436662e-01 6.72591388e-01 1.96601599e-02
2.10403815e-01 -4.23357636e-01 -6.36313796e-01 3.27707261e-01
1.14610469e+00 -2.40667492e-01 7.85171613e-02 1.13086772e+00
-1.55831978e-01 6.59933329e-01 -8.54597032e-01 4.38701242e-01
-1.73876826e-02 -1.29209149e+00 3.85657042e-01 -8.51735547e-02
5.75883687e-01 4.39189643e-01 2.19552472e-01 4.61814106e-01
8.15433681e-01 -1.32786107e+00 3.04707557e-01 2.03892142e-01
2.75967419e-01 -1.22170210e+00 7.54196525e-01 6.15643442e-01
-7.03501940e-01 -1.97494939e-01 -5.69119394e-01 7.51431435e-02
-4.15411452e-03 3.14304322e-01 -6.41981661e-01 2.49419793e-01
7.41043270e-01 8.28342363e-02 -4.36349571e-01 3.58037800e-01
-2.87569582e-01 6.39388084e-01 -1.77079514e-01 1.14672549e-01
4.93874311e-01 3.13207507e-01 8.62716734e-01 9.92906392e-01
-4.07971233e-01 9.23749804e-02 9.44971517e-02 5.16996920e-01
-2.81614453e-01 -9.83082280e-02 -1.05027223e+00 -5.25691807e-01
4.58620548e-01 8.21178913e-01 -4.55813289e-01 -1.50677308e-01
-7.11009651e-02 9.77434397e-01 3.78133386e-01 2.96591580e-01
-9.28459585e-01 -6.86803699e-01 1.12531507e+00 -3.69772047e-01
-5.54674864e-02 -3.66943091e-01 -2.67809868e-01 -1.06267631e+00
-1.42808139e-01 -1.39568627e+00 5.90323269e-01 -3.76041919e-01
-1.62158453e+00 1.08344376e+00 -4.16901797e-01 -9.09194350e-01
-1.21830933e-01 -9.84608114e-01 -7.38364100e-01 8.25064480e-01
-1.13645065e+00 -8.16209316e-01 2.86237121e-01 9.44462299e-01
3.13386828e-01 -7.39715457e-01 1.11845410e+00 4.88075316e-02
-6.76182926e-01 9.59355950e-01 8.29440728e-02 7.57005930e-01
6.87414587e-01 -9.32083726e-01 9.69031215e-01 1.25002551e+00
1.40132129e-01 9.37091529e-01 8.69406819e-01 -4.96118933e-01
-1.08149529e+00 -1.28384769e+00 7.32209206e-01 -6.66783333e-01
7.95161903e-01 -5.61467290e-01 -1.38817012e+00 7.15445578e-01
7.53816739e-02 2.26535812e-01 8.29636633e-01 -2.24120140e-01
-8.14531326e-01 3.03909369e-02 -1.72343874e+00 7.55266130e-01
5.21326900e-01 -6.54830039e-01 -7.22065151e-01 5.09101927e-01
9.50869024e-01 -4.58001167e-01 -6.51663661e-01 3.43940943e-01
3.25557619e-01 -7.84321666e-01 1.27141380e+00 -1.30964363e+00
2.08367422e-01 1.67518239e-02 -7.09407255e-02 -1.14135993e+00
-2.29331002e-01 -8.11083734e-01 -2.70636529e-01 1.24450517e+00
2.07391813e-01 -1.02413249e+00 9.41993594e-01 8.43171239e-01
8.13493580e-02 -5.67836344e-01 -9.55159843e-01 -8.71301234e-01
7.69146681e-01 -5.09854257e-01 3.80196869e-01 1.22740960e+00
-5.40471494e-01 1.36247426e-01 -4.52182978e-01 7.09250510e-01
4.32154715e-01 -7.16915369e-01 6.32780790e-01 -1.31518960e+00
-6.37394115e-02 -3.63085806e-01 -6.18612945e-01 -3.87451917e-01
6.55946136e-01 -8.00849795e-01 -1.75507404e-02 -9.14355338e-01
-5.96329510e-01 -1.50742188e-01 -4.50119644e-01 7.26109982e-01
-8.68694484e-02 3.08858633e-01 3.09317023e-01 8.50260556e-02
1.38958737e-01 2.93980241e-01 7.31212318e-01 -4.08282369e-01
5.64175583e-02 2.44354814e-01 -9.18073714e-01 1.09186852e+00
1.26483953e+00 -9.80810463e-01 -3.69164169e-01 -4.54488426e-01
3.57851714e-01 -4.91152644e-01 3.28624845e-01 -9.36926067e-01
2.40226835e-01 -3.86134475e-01 5.01364410e-01 -1.57412797e-01
1.47732854e-01 -1.01325357e+00 -2.60825664e-01 8.32532048e-01
-4.82471734e-01 -8.39814171e-03 5.38457692e-01 6.61655843e-01
-2.15406477e-01 -3.25283706e-01 1.43620729e+00 -1.32373661e-01
-4.33941990e-01 5.04120886e-01 -6.71340883e-01 2.48646349e-01
1.20354187e+00 1.07246555e-01 -4.12143886e-01 -2.23766908e-01
-8.41998160e-01 3.17432940e-01 1.54567242e-01 7.91240275e-01
5.07074118e-01 -9.66228008e-01 -7.33921051e-01 2.02073336e-01
-3.29781085e-01 -4.29166183e-02 3.15434098e-01 7.82629773e-02
-6.70664787e-01 4.58709985e-01 -3.81932825e-01 -9.52610560e-03
-1.53053117e+00 1.03730631e+00 7.65975416e-01 -3.82561684e-01
-4.70335275e-01 1.24515212e+00 2.85846442e-01 -5.57552576e-01
5.67445099e-01 3.03862337e-02 -2.24020600e-01 -1.74705520e-01
7.82858253e-01 8.69551003e-02 1.42381698e-01 -5.89284718e-01
-6.31663382e-01 9.49442908e-02 -5.91214597e-01 3.27195644e-01
1.24362767e+00 2.54101843e-01 8.46532453e-03 5.82658574e-02
1.10811269e+00 3.78068015e-02 -8.89592648e-01 -2.32660651e-01
-3.81413139e-02 -1.50248453e-01 -1.72909856e-01 -8.88105273e-01
-1.13849342e+00 1.01849806e+00 3.48502666e-01 5.75054467e-01
1.15336192e+00 -5.85827172e-01 9.09439147e-01 9.46978927e-01
9.98367667e-02 -9.56160426e-01 1.83754683e-01 1.13967288e+00
1.15429771e+00 -1.02313864e+00 -1.92379877e-01 -6.95409998e-02
-6.06229842e-01 1.29655421e+00 6.43524170e-01 -6.33446813e-01
7.21868277e-01 4.81079400e-01 4.61010158e-01 5.41377626e-02
-5.68062425e-01 6.01098597e-01 1.42826408e-01 9.35949743e-01
1.63784340e-01 -3.85525495e-01 1.16025314e-01 3.62645447e-01
-2.87101984e-01 -5.70597529e-01 4.75460500e-01 9.49477673e-01
-3.22042048e-01 -1.19511330e+00 -5.91553986e-01 5.74633554e-02
-1.07199538e+00 -1.56640574e-01 -1.03828764e+00 6.82806909e-01
-3.52485813e-02 9.83790159e-01 -2.63245285e-01 -4.59089398e-01
3.90148729e-01 2.25231782e-01 -1.44679114e-01 -5.52258670e-01
-1.19260228e+00 -8.00638974e-01 -7.77972415e-02 -5.53924024e-01
-6.70678318e-02 -2.98015445e-01 -1.14115942e+00 -5.75433135e-01
-1.11055553e-01 1.25619352e-01 5.35549641e-01 8.52465987e-01
3.65197629e-01 5.48548520e-01 7.95989156e-01 -5.76934040e-01
-1.00864780e+00 -6.84355855e-01 -2.55052894e-01 2.74082601e-01
6.22934937e-01 -2.27931306e-01 -6.96130991e-01 -1.82235867e-01] | [5.708001613616943, 7.837150573730469] |
c1d98c3d-b9e7-4d04-827e-242becfa07fd | gaussian-induced-convolution-for-graphs | 1811.04393 | null | http://arxiv.org/abs/1811.04393v1 | http://arxiv.org/pdf/1811.04393v1.pdf | Gaussian-Induced Convolution for Graphs | Learning representation on graph plays a crucial role in numerous tasks of
pattern recognition. Different from grid-shaped images/videos, on which local
convolution kernels can be lattices, however, graphs are fully coordinate-free
on vertices and edges. In this work, we propose a Gaussian-induced convolution
(GIC) framework to conduct local convolution filtering on irregular graphs.
Specifically, an edge-induced Gaussian mixture model is designed to encode
variations of subgraph region by integrating edge information into weighted
Gaussian models, each of which implicitly characterizes one component of
subgraph variations. In order to coarsen a graph, we derive a vertex-induced
Gaussian mixture model to cluster vertices dynamically according to the
connection of edges, which is approximately equivalent to the weighted graph
cut. We conduct our multi-layer graph convolution network on several public
datasets of graph classification. The extensive experiments demonstrate that
our GIC is effective and can achieve the state-of-the-art results. | ['Jian Yang', 'Zhen Cui', 'Jiatao Jiang', 'Chunyan Xu'] | 2018-11-11 | null | null | null | null | ['learning-representation-on-graph'] | ['methodology'] | [-2.39499807e-01 -1.47295278e-02 -1.68110598e-02 -3.07296336e-01
7.11755008e-02 -3.54827285e-01 4.30299282e-01 -9.62788910e-02
1.98825728e-02 8.83689001e-02 -6.21619401e-04 -4.00482655e-01
-1.23268202e-01 -1.21673203e+00 -6.69929028e-01 -7.76162684e-01
-2.81874567e-01 2.57443130e-01 1.98418394e-01 2.15547923e-02
-1.88648736e-03 6.24511123e-01 -8.95043075e-01 2.90555984e-01
9.17837083e-01 7.22531736e-01 1.70991331e-01 6.87493324e-01
-5.20751357e-01 6.92714214e-01 -9.45144445e-02 -3.86033893e-01
6.98311776e-02 -2.14824736e-01 -6.08409166e-01 7.18534291e-01
1.93351507e-01 6.62068948e-02 -9.17487264e-01 1.49624407e+00
5.11135347e-02 7.74375945e-02 8.53967249e-01 -1.27627337e+00
-1.36151171e+00 6.65299535e-01 -7.65982270e-01 -4.62737642e-02
-5.03595248e-02 1.75931126e-01 8.63675356e-01 -6.44980252e-01
5.29643059e-01 1.46560895e+00 6.10910535e-01 1.21548571e-01
-1.16712356e+00 -5.84182620e-01 8.13100517e-01 2.79925410e-02
-1.63972533e+00 1.50108233e-01 1.08787394e+00 -5.38637340e-01
6.72133446e-01 1.23105966e-01 6.10413551e-01 7.01748371e-01
1.69088647e-01 6.23392940e-01 6.18588567e-01 -1.33899868e-01
-8.77232477e-02 -2.89532572e-01 2.23507613e-01 1.11870253e+00
4.05815929e-01 -2.84153461e-01 1.01004742e-01 -8.47800747e-02
1.22611678e+00 3.51230830e-01 -4.13413912e-01 -4.23647374e-01
-1.03234482e+00 8.61727118e-01 8.44420254e-01 1.73209578e-01
-2.77129680e-01 4.57856447e-01 2.04795316e-01 1.58741891e-01
5.69109917e-01 -1.47072807e-01 -1.31503910e-01 3.15388173e-01
-4.45258647e-01 -6.30714670e-02 7.30262756e-01 1.26039565e+00
1.09060764e+00 -1.51100866e-02 -4.40800369e-01 8.83698165e-01
5.95374048e-01 1.99390247e-01 1.38393968e-01 -5.35635710e-01
2.81197459e-01 1.12207854e+00 -4.68999505e-01 -1.43341887e+00
-3.24581534e-01 -6.53705060e-01 -1.37829292e+00 -1.10989347e-01
1.85304523e-01 -2.16362122e-02 -1.19388080e+00 1.58230042e+00
2.31281161e-01 8.70053470e-01 -3.29854369e-01 6.65756583e-01
1.13379526e+00 4.17824209e-01 6.04340760e-03 1.10293590e-01
1.33069074e+00 -1.05123186e+00 -6.01549268e-01 7.25241052e-03
5.05271733e-01 -5.04709661e-01 8.28441441e-01 6.75887987e-02
-5.20591497e-01 -5.61722100e-01 -8.66401196e-01 6.88362345e-02
-3.59191746e-01 -4.18868065e-02 1.04667592e+00 5.81277251e-01
-1.10286725e+00 6.34280205e-01 -1.11125648e+00 -1.80374637e-01
7.12674737e-01 3.32533643e-02 -4.43276346e-01 -2.69349068e-01
-7.81865716e-01 1.34756193e-01 3.49004179e-01 2.43982762e-01
-6.04002714e-01 -5.09179473e-01 -1.21019220e+00 2.52568305e-01
2.00491071e-01 -6.40406072e-01 6.41509771e-01 -7.56593585e-01
-1.30528784e+00 7.69823909e-01 -4.75507639e-02 -2.37277925e-01
2.61894107e-01 3.74246746e-01 -5.54960847e-01 -5.29685616e-02
-1.53917447e-01 3.02953511e-01 7.69466698e-01 -1.13187373e+00
-2.99653649e-01 -3.26733112e-01 1.07170492e-01 -5.38572036e-02
-1.72370538e-01 -6.80476949e-02 -1.30167472e+00 -7.61184454e-01
2.75624722e-01 -7.44046330e-01 -5.50145924e-01 -6.51946142e-02
-6.15784585e-01 -2.66298145e-01 8.33489835e-01 -5.46139479e-01
1.56784093e+00 -2.13427496e+00 2.14455854e-02 4.84313995e-01
7.99719155e-01 8.70271772e-02 -3.03410262e-01 1.73593238e-01
-1.67403370e-01 2.02028394e-01 -2.87329584e-01 -3.17188531e-01
1.46999568e-01 2.17169136e-01 3.78851630e-02 6.60132945e-01
2.93730140e-01 1.30529046e+00 -8.01294386e-01 -4.36693162e-01
2.52578050e-01 7.06303596e-01 -5.82998812e-01 6.52777925e-02
-1.79735065e-01 2.13510975e-01 -6.48460090e-01 5.20748436e-01
1.23775721e+00 -7.25962281e-01 2.85652399e-01 -2.92554051e-01
2.44418323e-01 -2.10384890e-01 -1.41314280e+00 1.76155722e+00
-2.57757753e-02 2.48472899e-01 1.33155316e-01 -1.22117019e+00
9.08503294e-01 -3.44183967e-02 3.13487649e-01 -6.97180256e-02
3.28864723e-01 -3.12391132e-01 2.31925175e-02 -1.88187018e-01
1.38647571e-01 3.79085124e-01 1.53604876e-02 2.03800350e-01
3.54472458e-01 1.55641899e-01 2.27765754e-01 5.19092619e-01
1.26072299e+00 -8.50404054e-02 7.07348138e-02 -3.88067156e-01
5.04067779e-01 -5.29038787e-01 5.42550862e-01 4.77473617e-01
-5.37815392e-02 7.32042015e-01 7.57233977e-01 -5.17728984e-01
-4.86907154e-01 -1.09537816e+00 3.13711911e-02 5.84259331e-01
1.93945780e-01 -6.16414189e-01 -9.20372009e-01 -7.87023962e-01
1.08947158e-01 1.94217965e-01 -6.40605211e-01 -2.86249638e-01
-2.97361970e-01 -8.97764504e-01 2.42168710e-01 5.07040203e-01
5.15970051e-01 -9.94597554e-01 4.90975887e-01 2.78267056e-01
2.55304366e-01 -1.14993358e+00 -1.05661154e+00 -1.89220592e-01
-6.02563381e-01 -1.42721009e+00 -4.60651070e-01 -1.00596857e+00
1.08894944e+00 4.44660515e-01 1.16875398e+00 2.48375013e-01
-2.95780718e-01 5.12638152e-01 -2.86785573e-01 -4.55290824e-03
6.34939671e-02 -1.11552151e-02 -4.00596172e-01 4.39639360e-01
3.71478230e-01 -8.95149171e-01 -6.03680074e-01 1.40575007e-01
-9.27315235e-01 2.07814932e-01 5.68706155e-01 5.98921537e-01
8.19961250e-01 3.79228890e-01 1.20308392e-01 -1.33971727e+00
9.01745498e-01 -6.64151967e-01 -7.91829228e-01 4.06254172e-01
-5.00652015e-01 2.05253825e-01 6.41210377e-01 -5.61019301e-01
-7.44332433e-01 6.43387670e-03 8.18799138e-02 -8.65153670e-01
-1.47077948e-01 8.36902201e-01 -5.59529722e-01 -4.99949396e-01
2.38105148e-01 2.58475006e-01 2.06393115e-02 -5.17693460e-01
7.49570072e-01 2.67538428e-01 4.80323315e-01 -6.88763857e-01
8.12215745e-01 3.85422677e-01 1.81984246e-01 -6.30050480e-01
-3.63946319e-01 -4.69374180e-01 -5.47726631e-01 -2.07235292e-01
1.00998724e+00 -7.71156907e-01 -7.89507031e-01 9.09843206e-01
-9.98479426e-01 -5.47853947e-01 1.43934086e-01 4.31385636e-01
-2.41269395e-01 7.74571598e-01 -9.34426308e-01 -3.42344642e-01
-1.28331199e-01 -1.31609643e+00 8.74977291e-01 4.33958352e-01
4.69387949e-01 -1.35410094e+00 9.38273966e-03 -2.79569864e-01
3.14977050e-01 5.04606843e-01 8.33195806e-01 -3.77378702e-01
-8.21956396e-01 -2.84204960e-01 -6.72137499e-01 4.23570484e-01
4.08056378e-01 2.42648065e-01 -4.31637228e-01 -3.30450267e-01
-4.68081534e-01 1.97623640e-01 9.96635914e-01 5.68335831e-01
1.69185066e+00 -9.22626033e-02 -5.80084026e-01 1.21930003e+00
1.48063302e+00 -1.29591152e-01 7.91024208e-01 -2.97759801e-01
1.36580670e+00 -2.58075558e-02 -1.59110017e-02 2.78384417e-01
5.45315623e-01 1.59155414e-01 4.61914390e-01 -2.19724283e-01
-2.38637760e-01 -3.94116998e-01 -9.44557562e-02 8.70864451e-01
-2.88387150e-01 -4.03065652e-01 -6.40398622e-01 2.57420242e-01
-1.94201720e+00 -7.09021688e-01 -5.18262446e-01 1.93098414e+00
2.29036778e-01 1.01964362e-01 -1.51299983e-01 -5.00674784e-01
1.24791837e+00 4.80701596e-01 -4.22166407e-01 -3.48006263e-02
-7.10915402e-02 4.70055878e-01 6.36821747e-01 4.84571010e-01
-1.30018783e+00 1.08903778e+00 5.87199354e+00 1.07318735e+00
-9.42701161e-01 -8.92757848e-02 7.27336943e-01 5.57591677e-01
-4.60941106e-01 1.66553888e-03 -3.20042819e-01 7.67490566e-01
3.61316919e-01 -2.09292844e-01 8.10110211e-01 9.17503655e-01
-1.85290322e-01 5.49358368e-01 -7.67688036e-01 1.16360593e+00
-9.86662135e-02 -1.47641122e+00 3.83786500e-01 2.86502838e-01
8.87092412e-01 -5.98262772e-02 -5.55769280e-02 4.72474366e-01
8.14099729e-01 -1.27452230e+00 1.45817503e-01 7.75992930e-01
9.29841399e-01 -8.75636458e-01 3.33362073e-01 1.89349562e-01
-1.97188687e+00 4.63214546e-01 -6.41488314e-01 5.45129962e-02
-6.75982609e-02 5.74539006e-01 -3.56196582e-01 8.26528430e-01
4.88695741e-01 9.11998034e-01 -5.12254417e-01 1.03273618e+00
-3.18935812e-01 6.08764648e-01 -1.08308263e-01 1.13066927e-01
3.95095974e-01 -8.77902806e-01 2.71807253e-01 1.27159929e+00
1.74154773e-01 1.94829002e-01 5.51775753e-01 1.21378243e+00
-5.29433608e-01 8.98914337e-02 -4.19472516e-01 -3.42070073e-01
3.57999355e-01 1.69835067e+00 -9.24775958e-01 -1.79464906e-01
-8.34331810e-01 1.16844845e+00 8.59581828e-01 7.37525702e-01
-9.66055334e-01 -5.33809781e-01 7.99298525e-01 -1.30956650e-01
5.39477646e-01 -3.26395243e-01 2.11345226e-01 -1.33218122e+00
-4.27965745e-02 -6.52286708e-01 4.51838225e-01 -4.59449083e-01
-1.58744359e+00 6.42527461e-01 -3.49642158e-01 -9.89092886e-01
4.45329875e-01 -9.10561323e-01 -1.30342400e+00 1.14692628e+00
-1.21107113e+00 -1.44943547e+00 -7.31191456e-01 8.87250364e-01
-2.21967436e-02 -3.54800560e-02 6.83536291e-01 2.38513991e-01
-6.14666641e-01 6.52752221e-01 -5.98119870e-02 6.92068696e-01
3.49570334e-01 -1.24850583e+00 7.68396080e-01 1.06788218e+00
2.79453129e-01 7.62077808e-01 1.06194638e-01 -8.36539984e-01
-1.59620607e+00 -1.72568917e+00 2.01917872e-01 -1.90644767e-02
7.75303304e-01 -3.46879631e-01 -1.12011099e+00 1.03853726e+00
2.66639646e-02 8.08772326e-01 4.56958115e-01 1.77469373e-01
-6.22062027e-01 4.02362458e-02 -9.76463795e-01 5.20307362e-01
1.44746304e+00 -6.26259565e-01 -2.63242703e-02 4.08282787e-01
1.02348697e+00 -4.08878803e-01 -8.36477697e-01 4.90842968e-01
1.87272415e-01 -6.78243041e-01 8.93481374e-01 -9.02191818e-01
9.38312486e-02 -6.30708754e-01 -6.60488605e-02 -1.45513701e+00
-1.04491544e+00 -5.34531713e-01 -2.44721115e-01 1.24954593e+00
1.53983548e-01 -9.06926572e-01 9.15324569e-01 3.71797979e-01
-3.99614692e-01 -8.66837859e-01 -4.06429887e-01 -6.07054472e-01
1.05128007e-03 -5.71685314e-01 8.67844284e-01 1.03536344e+00
-3.58797103e-01 1.19275033e-01 9.85185988e-03 6.16873145e-01
7.51425982e-01 2.15735540e-01 8.53071630e-01 -1.13635850e+00
-5.60934961e-01 -8.95050347e-01 -9.16945279e-01 -1.10412073e+00
4.18109119e-01 -1.36373949e+00 -3.16305578e-01 -1.74336302e+00
3.54217887e-01 -3.82148772e-01 -5.31040549e-01 1.45163819e-01
-5.60262382e-01 4.86041643e-02 -2.25692749e-01 -2.28832409e-01
-7.35940158e-01 5.18459916e-01 1.44577026e+00 -4.26496536e-01
-6.76573366e-02 -1.19466379e-01 -7.51416147e-01 6.73613012e-01
5.99417150e-01 -2.70982087e-02 -6.20712698e-01 -3.89547884e-01
-5.22221588e-02 -1.98198602e-01 3.86509269e-01 -8.40848863e-01
3.03130537e-01 -2.39380538e-01 2.14621648e-01 -4.61866587e-01
-9.46337953e-02 -5.66054225e-01 5.14580071e-01 2.63684481e-01
7.64055848e-02 -2.16331884e-01 -6.85116872e-02 1.00538599e+00
-2.74607092e-01 2.39989668e-01 7.52508461e-01 -2.39153802e-01
-7.28723943e-01 1.27944648e+00 1.91080477e-02 1.72975603e-02
8.57863903e-01 1.61453545e-01 -2.97580928e-01 -3.18497062e-01
-8.14988971e-01 4.05435205e-01 5.08236885e-01 2.44953096e-01
6.67777121e-01 -1.73383510e+00 -7.40101039e-01 3.30237061e-01
3.20799381e-01 3.09981257e-01 6.28149807e-01 7.51006842e-01
-6.86785698e-01 1.93685908e-02 2.20460519e-01 -5.24225950e-01
-9.54720259e-01 9.98071373e-01 4.74550903e-01 -4.02405441e-01
-8.06765914e-01 1.07114387e+00 7.38476276e-01 -6.06398463e-01
1.83553621e-01 -5.37206590e-01 -1.27601564e-01 -5.59659243e-01
4.16475594e-01 7.65403286e-02 -1.48281097e-01 -5.42918324e-01
-2.06993714e-01 5.25010109e-01 -2.78779957e-02 5.98052680e-01
1.06532347e+00 1.04856998e-01 -5.46192110e-01 6.66938946e-02
1.36303067e+00 5.20915017e-02 -1.24955666e+00 -4.21218097e-01
-4.06294405e-01 -4.05773491e-01 7.42208734e-02 -9.54567343e-02
-1.52667916e+00 6.92837358e-01 1.33164465e-01 2.92202741e-01
9.97424364e-01 1.03457190e-01 6.96533501e-01 7.20887557e-02
2.97657490e-01 -6.26091540e-01 -1.68507755e-01 5.11551976e-01
7.45231807e-01 -1.01290953e+00 -2.74035662e-01 -8.97345662e-01
-3.09699267e-01 1.13078272e+00 6.76160514e-01 -7.68639922e-01
1.26620412e+00 2.89436132e-01 -4.17281657e-01 -5.23261070e-01
-3.95061731e-01 -3.67721885e-01 6.53185606e-01 7.27224767e-01
4.09317166e-01 6.42589271e-01 -1.91131517e-01 9.12711918e-01
1.68280974e-01 -3.73566568e-01 5.49181737e-02 4.03438270e-01
-2.92296857e-01 -1.00269008e+00 -1.00627448e-02 6.75806940e-01
-9.09414068e-02 -2.44535595e-01 -3.26822847e-01 7.94228256e-01
1.18351988e-01 7.21006513e-01 1.89114109e-01 -6.73057973e-01
3.67246121e-02 -2.67719716e-01 4.74619687e-01 -6.46144867e-01
6.92708641e-02 1.54480442e-01 -2.09436789e-01 -6.85164809e-01
-7.82017559e-02 -3.16882163e-01 -1.31523824e+00 -4.67241168e-01
-1.94816917e-01 1.70785740e-01 2.37996563e-01 7.05000281e-01
5.75050950e-01 9.07580376e-01 5.83703935e-01 -7.25070477e-01
-1.14525028e-01 -1.01841497e+00 -1.03630078e+00 5.87700784e-01
-2.05576405e-01 -6.71738923e-01 -4.03180867e-01 -1.08933419e-01] | [7.141427040100098, 6.30488395690918] |
d868e478-2a91-4921-96b5-5608d83bcb0d | slam-for-visually-impaired-people-a-survey | 2212.04745 | null | https://arxiv.org/abs/2212.04745v1 | https://arxiv.org/pdf/2212.04745v1.pdf | SLAM for Visually Impaired People: A Survey | In recent decades, several assistive technologies for visually impaired and blind (VIB) people have been developed to improve their ability to navigate independently and safely. At the same time, simultaneous localization and mapping (SLAM) techniques have become sufficiently robust and efficient to be adopted in the development of assistive technologies. In this paper, we first report the results of an anonymous survey conducted with VIB people to understand their experience and needs; we focus on digital assistive technologies that help them with indoor and outdoor navigation. Then, we present a literature review of assistive technologies based on SLAM. We discuss proposed approaches and indicate their pros and cons. We conclude by presenting future opportunities and challenges in this domain. | ['Alireza Darvishy', 'Davide Scaramuzza', 'Marziyeh Bamdad'] | 2022-12-09 | null | null | null | null | ['simultaneous-localization-and-mapping'] | ['computer-vision'] | [-5.21652400e-01 -1.79430351e-01 1.72141552e-01 -3.98369044e-01
-2.31669530e-01 -5.04670918e-01 4.79387432e-01 3.70026156e-02
-8.18420172e-01 1.14061046e+00 5.12398779e-01 -5.34680784e-01
-1.67741805e-01 -3.57023358e-01 1.86916292e-02 -1.89088970e-01
-1.59051791e-01 3.94922157e-04 1.97159201e-01 -4.12999094e-01
5.09610116e-01 6.53257132e-01 -1.89856994e+00 -4.13864553e-01
1.31771958e+00 4.95169312e-01 7.51819491e-01 5.68492234e-01
-7.40307868e-02 4.28340405e-01 -6.35520875e-01 -1.62084684e-01
2.98878998e-01 -9.59135965e-02 -6.21400595e-01 -3.29213023e-01
3.72135013e-01 -1.02430654e+00 -5.22679389e-01 7.88188815e-01
1.18577659e+00 1.07176207e-01 4.46792066e-01 -1.39600909e+00
-3.71029586e-01 -2.82076448e-01 1.07929558e-01 -1.69832930e-01
1.18169796e+00 6.09582737e-02 2.90984839e-01 -7.35533774e-01
4.05990392e-01 9.87634897e-01 8.57301116e-01 4.84225899e-01
-5.10630488e-01 -3.86172593e-01 2.68463492e-01 5.65306365e-01
-1.49076784e+00 -8.92665863e-01 2.33674660e-01 -4.81340736e-01
1.12568080e+00 6.20569438e-02 1.47184348e+00 6.40662014e-01
-1.40821934e-01 7.78453827e-01 9.69166219e-01 -7.19819546e-01
4.07337457e-01 3.07575583e-01 -5.74653596e-02 5.82615972e-01
1.04463494e+00 -8.29797164e-02 -9.43553388e-01 7.29856789e-02
5.64293444e-01 -2.22529098e-01 -7.81785607e-01 -8.10122967e-01
-9.22073483e-01 5.27292490e-01 6.87036455e-01 6.52637109e-02
-5.43441176e-01 -1.52281374e-01 -1.68170795e-01 1.38694555e-01
-1.97527017e-02 1.24933511e-01 -2.09351942e-01 -5.45998096e-01
-3.82045686e-01 1.37620866e-01 8.41999233e-01 1.33270764e+00
5.21876991e-01 -1.48626208e-01 7.26204216e-02 7.08758891e-01
1.04439306e+00 7.70779610e-01 1.61327913e-01 -9.01782334e-01
2.05026314e-01 4.12810206e-01 9.14774895e-01 -6.55209959e-01
-3.81676137e-01 1.06068216e-01 -1.60059243e-01 8.27051997e-01
4.01107728e-01 -3.86317819e-01 -1.02538300e+00 1.15819538e+00
1.54412277e-02 -3.76816779e-01 -1.11429006e-01 8.74841154e-01
7.99516439e-01 -8.29069614e-02 1.39122218e-01 -7.25884959e-02
9.44286585e-01 -6.81450486e-01 -1.02866864e+00 -6.45884335e-01
4.79400009e-01 -7.83271670e-01 1.04322124e+00 3.50096345e-01
-6.76954389e-01 -1.62031949e-01 -1.01507521e+00 -1.02389425e-01
-5.62460542e-01 1.78110972e-01 6.86943829e-01 1.19974434e+00
-1.68877304e+00 3.41781825e-02 -7.65987635e-01 -1.37542701e+00
3.63714337e-01 3.28755975e-01 -4.48814452e-01 -2.59309173e-01
-6.94611251e-01 1.39193058e+00 -3.41409519e-02 -1.68110095e-02
-4.03045833e-01 -2.07414448e-01 -7.34216452e-01 -8.01505446e-01
-5.36542892e-01 -8.54611695e-01 1.33353853e+00 2.33584386e-03
-1.44368100e+00 9.32463527e-01 -7.57809162e-01 -2.96431422e-01
7.80816317e-01 -8.81649792e-01 -3.82340342e-01 -4.52652611e-02
6.30172193e-01 4.32583123e-01 1.35670811e-01 -1.16785896e+00
-1.36126852e+00 -4.43178773e-01 4.30154279e-02 6.97031558e-01
-3.63461077e-01 6.79491693e-03 -3.94774973e-01 2.16410891e-03
6.43470168e-01 -6.37577355e-01 -1.22820973e-01 8.63657415e-01
-2.69617498e-01 5.42625710e-02 6.42364383e-01 -9.03356969e-01
8.95192683e-01 -1.89513183e+00 -1.56054214e-01 -5.11264168e-02
2.10246697e-01 4.62963223e-01 4.01559770e-01 8.06514323e-01
8.56322229e-01 -1.46023825e-01 1.21346205e-01 -7.81449974e-01
-2.00950518e-01 4.00299877e-01 1.25722736e-01 7.03271270e-01
-7.67330825e-01 5.02043009e-01 -1.03606820e+00 -2.32112944e-01
7.95016527e-01 6.30457044e-01 -7.64024556e-02 3.01688612e-01
7.85196841e-01 7.84774482e-01 -2.58715272e-01 1.11716080e+00
6.37265325e-01 5.46502054e-01 -8.80187675e-02 2.88324744e-01
-9.47160721e-01 4.43845212e-01 -9.85415995e-01 1.73797035e+00
-5.88301897e-01 1.02998328e+00 4.03371215e-01 -4.07779872e-01
7.04232037e-01 2.53969222e-01 1.32556900e-01 -5.91046333e-01
-1.33915797e-01 7.60332286e-01 -4.79879707e-01 -8.36538553e-01
5.44246912e-01 3.53188843e-01 5.39771199e-01 -1.41452014e-01
-3.37071538e-01 -8.57737064e-02 -2.85394937e-01 7.69393221e-02
9.93690908e-01 2.12449849e-01 1.01104641e+00 -3.21290821e-01
6.52703047e-01 -2.46254522e-02 2.48182565e-01 9.27603960e-01
-8.97199214e-01 3.36927533e-01 -5.01326740e-01 -4.73617554e-01
-7.10253716e-01 -1.19499564e+00 -1.85969770e-01 4.83302325e-01
5.16785145e-01 -4.29383308e-01 -2.75797755e-01 -2.90451288e-01
4.50947881e-01 4.57088679e-01 -3.63081545e-02 4.46945548e-01
-9.18182135e-02 -4.34174482e-03 2.49395743e-01 4.18495059e-01
1.26386774e+00 -7.10502565e-01 -9.57213640e-01 2.21904173e-01
-5.51403880e-01 -9.30853546e-01 3.20186019e-01 -2.58525342e-01
-6.31762564e-01 -1.07995081e+00 -1.14348114e+00 -1.10421848e+00
6.69347227e-01 1.19869149e+00 5.15094817e-01 2.60916948e-02
-1.82798937e-01 9.82845843e-01 -6.62413061e-01 -7.90888488e-01
4.05854762e-01 -1.89284369e-01 6.93247974e-01 -5.97515047e-01
6.54410064e-01 -9.04785275e-01 -9.28770065e-01 4.42096144e-01
3.60932887e-01 -2.06413805e-01 4.23387140e-01 1.47912681e-01
1.12092122e-01 -2.34201849e-01 2.27228165e-01 1.91461816e-01
6.98657334e-01 -2.09485114e-01 -5.66097677e-01 2.59611726e-01
-5.59082925e-01 -5.12086630e-01 -1.09091848e-01 7.80357569e-02
-8.13850164e-01 1.64226443e-01 -1.80013582e-01 6.49587333e-01
-3.69753212e-01 1.35080829e-01 -6.01655066e-01 -1.10636437e+00
8.05411577e-01 1.23304725e-01 7.34081566e-02 -7.49705434e-01
3.01163614e-01 1.58289385e+00 5.55867732e-01 -2.04563960e-02
7.49891996e-01 7.54806101e-01 -3.22056979e-01 -1.29874802e+00
-1.88408315e-01 -7.00996399e-01 -7.16743112e-01 -8.26789618e-01
5.48121631e-01 -9.81198967e-01 -6.50155306e-01 1.13888466e+00
-1.13214958e+00 -3.49742770e-01 2.65972316e-03 9.18798566e-01
-4.73869741e-01 3.83120030e-01 1.83013499e-01 -1.47562087e+00
-1.44816265e-01 -8.92052531e-01 6.08534575e-01 6.41480446e-01
-3.24734807e-01 -7.35830843e-01 -6.85604056e-04 2.89001346e-01
7.87916601e-01 2.78012455e-02 5.96756190e-02 2.24917516e-01
-5.67730606e-01 -3.55646312e-01 -3.80651325e-01 -1.26898102e-02
7.36622810e-01 -7.61035562e-01 -9.09035802e-01 -5.93486488e-01
-3.38334084e-01 1.12241849e-01 2.64202893e-01 6.59893155e-01
2.02093627e-02 -2.12191731e-01 -7.59892344e-01 6.63125575e-01
1.10492694e+00 4.35000092e-01 8.66614580e-01 1.00375855e+00
4.42501754e-01 7.94748068e-01 6.79158628e-01 4.99341279e-01
7.94069409e-01 7.27457881e-01 4.01323169e-01 1.22571141e-01
-2.80713350e-01 -4.27859306e-01 1.64591506e-01 1.62079275e-01
-6.95122540e-01 -1.73290431e-01 -1.08267379e+00 7.59098053e-01
-1.90370774e+00 -3.29418540e-01 -4.23270404e-01 2.52344966e+00
1.52246848e-01 -3.99783969e-01 4.06760499e-02 3.38723332e-01
5.01438797e-01 -3.01205903e-01 -4.59987044e-01 2.11312562e-01
-1.71307266e-01 -9.68304724e-02 8.93346131e-01 8.56112361e-01
-9.95761991e-01 1.10101366e+00 7.03521919e+00 -4.14235950e-01
-6.60661817e-01 -9.05474126e-02 -7.13661909e-01 1.25413731e-01
4.84297089e-02 1.12206951e-01 -6.07204139e-01 2.89710522e-01
1.40103355e-01 -6.45349920e-02 6.02431834e-01 9.42047954e-01
5.12495637e-01 -7.46572793e-01 -4.83261943e-01 1.31849718e+00
-1.24582835e-01 -6.98286593e-01 -3.72655600e-01 3.05674791e-01
3.54876965e-01 3.70433092e-01 -1.97167978e-01 -2.18361855e-01
2.02622324e-01 -7.26742029e-01 7.76013136e-01 6.70852125e-01
8.62257004e-01 -4.81644690e-01 8.82069230e-01 1.76460326e-01
-1.37327433e+00 -3.66567075e-01 -4.88169998e-01 -7.89300501e-01
7.84166038e-01 2.59441793e-01 -7.80479312e-01 4.06672239e-01
1.21432769e+00 8.13532829e-01 -5.61039031e-01 2.06109023e+00
-9.72818494e-01 -1.47762701e-01 -2.94890285e-01 -4.91195530e-01
-2.09730044e-01 -3.70439321e-01 8.56618881e-01 5.37963867e-01
9.02343631e-01 1.24951273e-01 -2.24741116e-01 -3.36318240e-02
6.85355723e-01 1.94296017e-01 -1.09912932e+00 5.24397075e-01
1.04402840e+00 6.55965149e-01 -1.13810912e-01 1.10573500e-01
-5.49060047e-01 1.23309481e+00 -8.48166645e-02 6.79216981e-01
2.87178680e-02 -6.91588700e-01 1.25053430e+00 3.75499576e-01
-3.17213356e-01 -9.59334493e-01 -4.68344152e-01 -9.64707077e-01
3.28071117e-01 -2.85874009e-01 -2.99938545e-02 -1.12621367e+00
-7.41732836e-01 3.95636708e-01 -2.51471251e-01 -1.57161856e+00
-6.60544336e-02 -7.10004807e-01 -1.57103851e-01 1.41183472e+00
-1.46207750e+00 -1.29542828e+00 -1.11951029e+00 6.19210124e-01
2.07246378e-01 -5.64167142e-01 1.24235845e+00 2.31171668e-01
1.87096372e-02 2.43831933e-01 3.20102572e-01 -3.78857404e-01
8.49867165e-01 -1.09978056e+00 5.06149292e-01 1.14765763e+00
-4.88574922e-01 1.21359491e+00 8.91600490e-01 -9.10365403e-01
-1.22231984e+00 -4.59528506e-01 1.13736391e+00 -5.15720963e-01
9.33132693e-02 -2.19180152e-01 3.60495970e-02 6.37420535e-01
3.03102899e-02 -3.96314204e-01 6.15352869e-01 2.19384834e-01
1.61205143e-01 3.30390669e-02 -1.73863697e+00 9.05555487e-01
1.82099211e+00 -6.95511103e-01 -4.46949422e-01 2.11828217e-01
1.95706442e-01 -2.49119759e-01 -3.15236039e-02 1.27238199e-01
1.10166848e+00 -1.20291936e+00 9.86881673e-01 6.73249140e-02
-8.06917071e-01 -7.19237685e-01 -4.89876270e-01 -1.14392531e+00
-9.28699151e-02 -7.82685041e-01 -6.11424372e-02 8.96570444e-01
-3.09324145e-01 -1.30326402e+00 6.92898393e-01 8.39158893e-01
-1.43461391e-01 -2.89570596e-02 -1.26166975e+00 -7.66529799e-01
-6.47722781e-01 -5.08186102e-01 3.81934196e-01 2.47985572e-01
3.72072637e-01 -3.63423496e-01 -5.96640348e-01 5.05452633e-01
7.34752774e-01 -8.02891612e-01 9.48525131e-01 -1.70542657e+00
4.27045971e-01 -6.75396770e-02 -1.26531053e+00 -1.31093836e+00
-3.67527634e-01 -1.05334014e-01 6.23206310e-02 -2.62759995e+00
-5.24664581e-01 -4.21298146e-01 2.88381070e-01 4.84533936e-01
1.71680033e-01 1.77221954e-01 -2.61772186e-01 4.34135854e-01
1.13323266e-02 3.66158426e-01 7.00101972e-01 2.22743213e-01
-6.15754008e-01 5.09073496e-01 -9.10097778e-01 7.18697727e-01
1.04863787e+00 3.32661688e-01 -6.43494725e-01 -6.69901431e-01
1.01593740e-01 -3.28914493e-01 6.20715499e-01 -1.42616343e+00
5.77534795e-01 1.07745491e-01 2.24920347e-01 -8.73808801e-01
5.03158271e-01 -8.32204282e-01 -9.40725580e-02 8.32700908e-01
6.66656673e-01 -3.95958841e-01 1.75171614e-01 3.55349392e-01
3.56521189e-01 -1.10734947e-01 4.71849650e-01 9.79034230e-02
-1.20216024e+00 -1.19080491e-01 -8.26332510e-01 -7.32712507e-01
1.07240808e+00 -5.69502294e-01 -4.26761955e-01 -1.05397582e+00
-4.43053097e-01 3.67973596e-01 8.72076273e-01 3.74843478e-01
9.33618009e-01 -1.25560451e+00 -2.64071256e-01 3.86504024e-01
4.42045093e-01 -2.42567942e-01 7.61230662e-02 5.73437750e-01
-1.07736099e+00 7.50921249e-01 -5.82731366e-01 -2.38299981e-01
-1.46968043e+00 -6.75866827e-02 3.73374701e-01 9.31848228e-01
-8.80108178e-01 8.28053951e-01 -6.33739114e-01 -4.20976222e-01
7.71544099e-01 1.03412971e-01 -4.75978583e-01 -8.04739520e-02
8.67664039e-01 7.49878228e-01 -8.04207027e-02 -6.16903603e-01
-9.30428624e-01 7.89495945e-01 4.96125102e-01 -7.01059043e-01
1.12829757e+00 -9.83692050e-01 -1.76819488e-01 -1.48175331e-03
4.36336130e-01 2.34309375e-01 -1.03031754e+00 1.51313484e-01
-4.01212126e-02 -9.21447098e-01 -1.47146238e-02 -1.02076030e+00
-3.04502368e-01 6.00051999e-01 1.17744768e+00 -2.90942714e-02
1.13956535e+00 -2.21419156e-01 3.37634712e-01 7.12149262e-01
1.58188665e+00 -9.25808787e-01 -6.09918594e-01 2.23667786e-01
9.96307909e-01 -1.17442751e+00 8.74252543e-02 -4.18511331e-01
-1.35399371e-01 8.90851319e-01 4.42482769e-01 5.11498809e-01
5.85187554e-01 -4.47895117e-02 5.46278954e-01 2.82814711e-01
6.19465590e-01 -8.31116319e-01 -2.97521986e-02 1.87855422e+00
2.85247117e-01 3.16841863e-02 -7.03085840e-01 2.50012606e-01
-5.06563187e-01 1.03612587e-01 3.79263222e-01 1.59123600e+00
-9.43089068e-01 -1.30404270e+00 -3.67881089e-01 3.09350610e-01
4.69086021e-01 1.56326562e-01 -6.40500784e-01 8.20108831e-01
4.74455617e-02 1.52213156e+00 -5.43343484e-01 -6.67688370e-01
6.28512383e-01 -5.31240962e-02 4.51540798e-01 -1.95391282e-01
3.70127350e-01 -5.84469438e-01 4.92726475e-01 -7.33661115e-01
-4.26529676e-01 -8.15777063e-01 -9.36246872e-01 -3.14121366e-01
9.68492553e-02 -1.19226798e-01 1.09614480e+00 7.56531060e-01
3.05675626e-01 7.99762234e-02 2.55732387e-01 -9.78546679e-01
1.25357836e-01 -9.88523066e-01 -7.64530957e-01 -5.63967466e-01
5.47778428e-01 -8.17082644e-01 -3.63045514e-01 -4.92442340e-01] | [7.5579328536987305, -1.8506495952606201] |
709e440e-9065-4540-a993-ccb831660e3c | representing-and-learning-functions-invariant | 2306.05261 | null | https://arxiv.org/abs/2306.05261v1 | https://arxiv.org/pdf/2306.05261v1.pdf | Representing and Learning Functions Invariant Under Crystallographic Groups | Crystallographic groups describe the symmetries of crystals and other repetitive structures encountered in nature and the sciences. These groups include the wallpaper and space groups. We derive linear and nonlinear representations of functions that are (1) smooth and (2) invariant under such a group. The linear representation generalizes the Fourier basis to crystallographically invariant basis functions. We show that such a basis exists for each crystallographic group, that it is orthonormal in the relevant $L_2$ space, and recover the standard Fourier basis as a special case for pure shift groups. The nonlinear representation embeds the orbit space of the group into a finite-dimensional Euclidean space. We show that such an embedding exists for every crystallographic group, and that it factors functions through a generalization of a manifold called an orbifold. We describe algorithms that, given a standardized description of the group, compute the Fourier basis and an embedding map. As examples, we construct crystallographically invariant neural networks, kernel machines, and Gaussian processes. | ['Peter Orbanz', 'Ryan P. Adams'] | 2023-06-08 | null | null | null | null | ['gaussian-processes'] | ['methodology'] | [ 4.24407959e-01 4.16602433e-01 -5.11357971e-02 -2.29232669e-01
-7.55188391e-02 -6.28013432e-01 1.03192127e+00 -1.09918022e+00
3.85339186e-02 7.64174938e-01 4.15298164e-01 -1.49931997e-01
-3.66129845e-01 -9.57362950e-01 -9.47146297e-01 -1.12115765e+00
-4.80520099e-01 6.90265119e-01 -1.92775995e-01 9.53725912e-03
5.92057943e-01 6.39624834e-01 -1.56745303e+00 6.97207227e-02
5.73385298e-01 6.20269477e-01 1.23268433e-01 7.22884834e-01
-2.74415582e-01 5.20719826e-01 -1.74882915e-02 3.02660644e-01
5.84099948e-01 -3.75632644e-01 -1.07424784e+00 4.76972818e-01
4.87688899e-01 1.65812343e-01 -6.30712330e-01 1.13208461e+00
-1.44949734e-01 1.50282532e-01 1.28634059e+00 -1.20974779e+00
-1.42557287e+00 4.26848531e-01 -3.67331393e-02 -3.84130776e-01
2.97616631e-01 -2.89789617e-01 9.21364665e-01 -8.15816939e-01
1.13773525e+00 1.38557160e+00 4.88305360e-01 5.59573829e-01
-1.64111149e+00 -3.77335161e-01 -2.61409253e-01 6.18044473e-02
-1.32670557e+00 -4.20113593e-01 3.55976015e-01 -6.30279839e-01
9.01146829e-01 4.65778351e-01 5.28115749e-01 7.66782165e-01
6.95835292e-01 7.24265948e-02 1.23992872e+00 -8.02230000e-01
3.63988340e-01 -5.17734230e-01 4.76976693e-01 7.73838401e-01
1.79650187e-01 -6.15211353e-02 -3.60246122e-01 -4.00297999e-01
1.21251404e+00 2.35463947e-01 -5.87421656e-01 -7.35179722e-01
-1.67443633e+00 7.80896366e-01 -6.02421872e-02 5.25521398e-01
-4.18733746e-01 3.81761551e-01 -3.29749376e-01 5.34362435e-01
1.53511763e-01 5.33149004e-01 2.93901898e-02 1.93837509e-01
-3.56100351e-01 3.17513466e-01 1.36729527e+00 1.12033367e+00
7.93304622e-01 -9.12770554e-02 3.62146914e-01 3.20902556e-01
-1.04634501e-02 8.15472782e-01 4.21253443e-01 -1.58109558e+00
-9.47052464e-02 1.37873769e-01 6.43555373e-02 -6.42671943e-01
-4.66849834e-01 1.69660985e-01 -8.13886702e-01 4.45009261e-01
5.09232104e-01 5.85166335e-01 -7.13042080e-01 1.69458342e+00
-1.10934153e-01 1.84904844e-01 2.27343123e-02 3.47791404e-01
2.69064978e-02 7.44426906e-01 -5.31985104e-01 -3.59867066e-01
1.31133938e+00 -6.64124668e-01 -7.84942567e-01 5.22928715e-01
5.51662803e-01 -7.59413779e-01 7.38758683e-01 3.74785244e-01
-1.14037144e+00 -2.23131180e-01 -1.17654526e+00 1.75907522e-01
-1.29569143e-01 -7.18980953e-02 7.03055203e-01 3.73721391e-01
-1.29856801e+00 1.08872855e+00 -9.69449282e-01 -4.22861695e-01
-2.88261443e-01 4.49491948e-01 -8.98926377e-01 5.84122948e-02
-8.15185070e-01 7.24498749e-01 4.60010581e-02 -3.97156984e-01
-2.52095789e-01 -5.83759665e-01 -5.94748795e-01 -1.31956547e-01
-1.48451596e-01 -5.17371595e-01 1.08519733e+00 -6.54271841e-01
-1.48442793e+00 8.31482530e-01 -4.90957171e-01 -2.29763180e-01
1.21067598e-01 2.68592149e-01 -4.52303380e-01 2.46728599e-01
1.48506999e-01 2.04535097e-01 1.30724216e+00 -7.02680349e-01
-2.29044557e-01 -3.38872492e-01 -1.48009285e-01 -7.04983622e-02
-1.04108170e-01 -7.52955228e-02 1.17142409e-01 -4.44299549e-01
7.61085808e-01 -1.41942656e+00 -9.83623043e-02 -3.20532441e-01
-5.59195995e-01 -3.43805581e-01 9.77578044e-01 -5.70714355e-01
7.96655178e-01 -2.06710505e+00 5.57215333e-01 5.87842166e-01
4.14019078e-01 -5.75970948e-01 7.49033466e-02 5.11332691e-01
-6.93205357e-01 1.60850942e-01 -2.02386096e-01 3.65474820e-01
3.93598378e-02 9.51813683e-02 -3.96729201e-01 1.20534134e+00
-5.77963553e-02 6.44033074e-01 -3.55086714e-01 1.35135591e-01
-1.32395597e-02 4.87878948e-01 -8.28543901e-01 -2.37901300e-01
-2.82178503e-02 5.13029575e-01 -5.32378733e-01 3.36925358e-01
7.90539265e-01 -3.35734725e-01 4.24927026e-02 -2.64082730e-01
-4.43895638e-01 2.44810328e-01 -1.55495632e+00 1.06872058e+00
4.56147380e-02 6.19954467e-01 1.22885302e-01 -9.85855639e-01
1.03477943e+00 4.62166697e-01 6.85414732e-01 -1.52698830e-01
-1.62540480e-01 3.33740085e-01 8.59200656e-02 -3.07030767e-01
2.05627710e-01 -1.60494000e-01 1.91285729e-01 9.93384123e-01
1.07554458e-01 4.01849300e-02 -1.88875142e-02 -1.80037230e-01
1.40718937e+00 1.75799981e-01 2.75772363e-01 -8.80185366e-01
6.61053896e-01 -3.99255782e-01 2.20862612e-01 5.67003787e-01
3.25716108e-01 7.48192132e-01 3.75088036e-01 -6.92388833e-01
-1.36532807e+00 -1.19368494e+00 -4.93411988e-01 7.01252878e-01
-2.71086711e-02 -4.65371400e-01 -1.07985008e+00 2.14273646e-01
2.18130574e-01 5.40524662e-01 -6.70698762e-01 -1.91420749e-01
-7.75326550e-01 -5.94366074e-01 6.43185601e-02 1.92494914e-01
4.85730737e-01 -1.18626106e+00 -1.00601958e-02 9.87692475e-02
1.55884415e-01 -6.16055965e-01 -7.73599267e-01 2.55251586e-01
-9.47750568e-01 -1.30552030e+00 -5.32314122e-01 -9.28932786e-01
9.72718894e-01 3.69263530e-01 5.51781714e-01 -3.44358057e-01
-2.64731705e-01 5.73300779e-01 2.66995668e-01 7.32384399e-02
-5.12493789e-01 -1.60173684e-01 7.87395716e-01 5.67119867e-02
5.84017396e-01 -8.69912922e-01 -1.67085946e-01 7.04639792e-01
-7.74328768e-01 -1.14026889e-01 1.14522122e-01 6.72018290e-01
6.96955323e-01 8.26111734e-02 9.49656740e-02 -6.41909480e-01
5.32376647e-01 -2.14098975e-01 -7.42353797e-01 1.42209083e-01
-4.36047047e-01 8.75259221e-01 4.78213578e-01 -3.21259916e-01
-7.17122018e-01 6.51685894e-02 4.86489415e-01 -1.97254837e-01
4.14358154e-02 1.10574968e-01 -2.48117045e-01 -6.21722341e-01
7.67649710e-01 3.21178436e-01 3.77020806e-01 -5.21041930e-01
4.52924699e-01 7.78907537e-01 5.70042133e-01 -9.47589517e-01
1.06156361e+00 9.03169274e-01 5.24987280e-01 -1.34464610e+00
-1.94424301e-01 -3.92270446e-01 -9.76474404e-01 1.22748047e-01
1.13175118e+00 -4.76386249e-01 -9.76136029e-01 2.15750739e-01
-1.17565036e+00 2.69360423e-01 -4.28476036e-01 1.00459015e+00
-1.52045500e+00 3.59294087e-01 -4.98769075e-01 -4.75951761e-01
1.28753362e-02 -1.16186059e+00 7.58939803e-01 -2.51031369e-01
-5.63423693e-01 -7.72235572e-01 2.76505440e-01 -5.46280295e-02
5.07151127e-01 2.45717708e-02 1.14375234e+00 -6.83238804e-01
-9.97574627e-01 -1.36771694e-01 2.36527324e-01 4.99207377e-01
2.16178477e-01 5.51927229e-03 -5.76843739e-01 -1.18767150e-01
6.16819024e-01 2.64418423e-01 6.59430861e-01 6.64513230e-01
1.02609980e+00 -2.80048728e-01 -3.83722216e-01 1.10942078e+00
9.26328719e-01 6.36576772e-01 6.28995657e-01 5.36545992e-01
6.86981678e-01 5.72925627e-01 -4.37544465e-01 -1.89427525e-01
-1.56751573e-01 4.42281008e-01 -1.81924805e-01 4.01607364e-01
1.84563085e-01 1.09736964e-01 4.95732963e-01 1.20498836e+00
-1.12615764e+00 5.31731367e-01 -5.31971037e-01 4.36053388e-02
-1.73281848e+00 -1.26767409e+00 -1.57463297e-01 2.40107965e+00
5.24447083e-01 -2.27276757e-01 -1.84331432e-01 -7.99477175e-02
1.13634729e+00 -1.08715355e-01 -3.74178290e-01 -2.78333217e-01
-3.90882105e-01 6.22298658e-01 1.03192425e+00 8.03394914e-01
-1.07418931e+00 6.38428628e-01 8.11442280e+00 3.97228241e-01
-6.16149366e-01 -9.42673311e-02 -1.24791563e-01 2.49076724e-01
-4.63058054e-01 2.13764042e-01 -7.79832304e-01 1.48844182e-01
1.00296926e+00 -5.46699464e-01 9.19834375e-01 9.89680946e-01
-6.67105168e-02 5.10842085e-01 -1.49447250e+00 9.96955872e-01
1.61904413e-02 -1.70165944e+00 1.10392220e-01 8.92386198e-01
7.61479497e-01 -2.71062613e-01 4.39576469e-02 -6.51488230e-02
2.23333895e-01 -1.27801013e+00 4.56421196e-01 7.72823691e-01
7.36665130e-01 -6.59116507e-01 -5.46867959e-02 5.33262044e-02
-1.02206182e+00 3.05759966e-01 -6.44028306e-01 -8.47878456e-02
9.14480984e-02 -4.92553972e-03 -3.30316037e-01 3.07553913e-02
4.83733058e-01 7.21833825e-01 1.97171271e-01 5.56420028e-01
1.41045704e-01 3.83275419e-01 -3.13135058e-01 2.79611528e-01
1.92752659e-01 -1.27586019e+00 7.56392002e-01 6.54147208e-01
7.86620259e-01 6.00805640e-01 -9.27156284e-02 8.88423920e-01
-1.92488477e-01 -1.44894412e-02 -1.03455055e+00 -3.78947675e-01
2.06839070e-01 9.23519135e-01 -6.02776945e-01 -1.93202749e-01
-5.29685795e-01 8.84833813e-01 -1.78961292e-01 5.77544093e-01
-2.36885026e-01 -6.06938243e-01 1.28332090e+00 4.41544682e-01
5.75568974e-01 -5.44065297e-01 1.89015269e-01 -1.38036668e+00
-3.09863761e-02 -7.20524192e-01 -2.70051777e-01 -7.99767077e-01
-1.16587305e+00 3.89724553e-01 8.71583670e-02 -1.06747746e+00
-5.96949816e-01 -1.26832569e+00 -4.57897186e-01 1.10504246e+00
-6.06214702e-01 -5.97763538e-01 1.57022387e-01 7.29489446e-01
-7.52510056e-02 -6.02081954e-01 1.20916438e+00 -3.47540885e-01
1.51935518e-01 -1.80186659e-01 7.32146740e-01 -2.76462466e-01
4.19844240e-01 -1.50125575e+00 6.00414455e-01 4.46952283e-01
2.67999079e-02 1.07653046e+00 6.83137655e-01 -5.59975028e-01
-1.92543614e+00 -9.05382335e-01 8.68153691e-01 -4.76242661e-01
8.67812753e-01 -3.53937656e-01 -9.45469797e-01 1.31925321e+00
-3.79881412e-02 -1.02320462e-01 7.19940782e-01 1.50748610e-01
-5.78039527e-01 3.03215623e-01 -1.04132318e+00 8.37391794e-01
1.58259523e+00 -9.38774645e-01 -8.15452278e-01 1.00684893e+00
7.75175929e-01 -5.96160367e-02 -1.05178237e+00 -8.70088413e-02
7.34730363e-01 -7.57847190e-01 1.18498278e+00 -8.79071355e-01
6.56523407e-02 -4.86404121e-01 -4.90650624e-01 -1.20887589e+00
-1.00310409e+00 -1.09071815e+00 1.45209566e-01 3.46897155e-01
-1.57472007e-02 -1.05954373e+00 6.30796671e-01 5.06588221e-01
-1.43551737e-01 -2.30936125e-01 -1.05291438e+00 -1.18056703e+00
1.55541271e-01 -2.26928592e-02 7.17818618e-01 9.63352382e-01
2.67553747e-01 3.76259834e-02 -2.32595026e-01 7.74519006e-03
8.74117076e-01 3.22277308e-01 4.59202945e-01 -1.55591738e+00
-2.73407251e-01 -1.97053641e-01 -9.73624706e-01 -7.81981885e-01
5.82596838e-01 -1.42631745e+00 -3.36499780e-01 -1.01821625e+00
2.17002574e-02 -1.75726473e-01 6.49601221e-03 1.20954879e-01
7.24078238e-01 -7.64378011e-02 -5.73698618e-02 6.30230784e-01
-4.73022647e-02 3.06987494e-01 1.24331248e+00 1.30533397e-01
-2.20516682e-01 -1.28021389e-01 -5.45045674e-01 9.91031945e-01
7.14696705e-01 -9.07553583e-02 -5.68500273e-02 1.07257918e-01
4.37476747e-02 -2.15780735e-01 1.93661470e-02 -9.79924798e-01
1.67263970e-01 -3.33245426e-01 2.46591344e-01 -1.25026524e-01
1.17398843e-01 -5.86991549e-01 7.76858985e-01 3.43044132e-01
-1.77145496e-01 1.13476127e-01 -5.61263502e-01 4.06601220e-01
3.76604833e-02 -3.41704875e-01 6.36837006e-01 -3.25800627e-01
-3.93960923e-01 2.76316851e-01 -4.14286315e-01 -2.47734904e-01
9.21062827e-01 -4.38809246e-01 -4.29851830e-01 -4.32110429e-01
-1.03537273e+00 -3.43644291e-01 1.01324451e+00 6.81544095e-02
2.53949434e-01 -1.81579769e+00 -4.29433823e-01 7.45480955e-01
3.36892670e-03 -3.61489415e-01 -1.54108137e-01 6.41257405e-01
-9.71027374e-01 7.13658750e-01 -5.25242031e-01 -5.57409763e-01
-9.98882830e-01 5.04626036e-01 4.58097607e-01 3.90962958e-01
-9.70334649e-01 2.19577909e-01 6.99045777e-01 -6.00040793e-01
-9.73189324e-02 -6.25999093e-01 8.98438171e-02 -3.93959850e-01
9.48571324e-01 4.39862698e-01 -1.73681259e-01 -9.44906592e-01
-2.39863247e-01 1.01147068e+00 1.89378619e-01 -2.30043828e-01
1.44064343e+00 2.73336709e-01 -9.07075524e-01 3.45085949e-01
1.53586376e+00 1.21564120e-01 -8.96394789e-01 -1.71797156e-01
-8.11152384e-02 -1.18085422e-01 -4.06926185e-01 3.61155242e-01
-5.93427420e-01 2.53630340e-01 8.02885219e-02 4.79115903e-01
5.24517357e-01 3.03596407e-01 2.60521173e-01 1.09096336e+00
4.77946132e-01 -9.60240483e-01 -4.84113306e-01 8.44698548e-01
1.11823583e+00 -4.03444678e-01 -2.15330735e-01 -7.21138477e-01
1.30825177e-01 1.65299869e+00 -2.83307582e-01 -6.62600636e-01
1.13145125e+00 1.96442604e-01 -3.82431418e-01 -2.14805529e-01
-6.14139318e-01 2.11845919e-01 2.75737166e-01 7.87975132e-01
4.34555441e-01 3.55408639e-01 -3.55711073e-01 1.60032973e-01
-7.93903649e-01 5.09374915e-03 6.84546828e-01 9.26668704e-01
-9.73210275e-01 -1.13375628e+00 -1.00509918e+00 5.08135200e-01
-2.67501883e-02 1.30340531e-01 -1.18145846e-01 6.51860356e-01
-9.63726714e-02 3.39808643e-01 4.26768035e-01 -3.13718736e-01
2.22822994e-01 7.23445177e-01 8.77558470e-01 -4.10085917e-01
4.13480222e-01 -6.25019372e-02 -2.33652160e-01 -9.06470776e-01
-4.45639342e-01 -8.57078552e-01 -1.34751511e+00 -6.10130787e-01
1.64261594e-01 2.58640438e-01 8.37777853e-01 8.57239306e-01
1.08315572e-01 2.20535740e-01 4.74198222e-01 -1.06335425e+00
-8.58573318e-01 -8.36980581e-01 -1.28633511e+00 3.51222962e-01
3.84151310e-01 -7.85438895e-01 -8.77024531e-01 5.78261197e-01] | [7.370428562164307, 4.551607608795166] |
5453dc12-43c0-4e87-a9eb-00e78fdaffac | causal-inductive-synthesis-corpus | null | null | https://openreview.net/forum?id=rO24tIDmtSr | https://openreview.net/pdf?id=rO24tIDmtSr | Causal Inductive Synthesis Corpus | We introduce the Causal Inductive Synthesis Corpus (CISC) -- a manually constructed collection of interactive domains.
CISC domains abstract core causal concepts present in real world mechanisms and environments. We formulate two synthesis challenges of causal model discovery: the passive discovery of a model of a CISC domain from observed data, and active discovery while interacting with the domain. CISC problems are expressed in Autumn, a Turing-complete programming language for specifying causal probabilistic models. Autumn allows succinct expression for models that vary dynamically through time, respond to external input, have internal state and memory, exhibit probabilistic non-determinism, and have complex causal dependencies between variables.
| ['Armando Solar-Lezama', 'Joshua B. Tenenbaum', 'Kate Lin', 'Elizabeth Weeks', 'Ria Das', 'Zenna Tavares'] | 2020-10-13 | null | null | null | neurips-workshop-cap-2020-12 | ['model-discovery'] | ['miscellaneous'] | [ 1.78573668e-01 5.98130107e-01 -5.65890551e-01 -4.29321021e-01
-1.93332195e-01 -8.90264273e-01 1.47963858e+00 2.53618419e-01
2.16507971e-01 8.46490741e-01 7.19298899e-01 -6.74108267e-01
-6.41981542e-01 -9.02369380e-01 -8.36245537e-01 -3.22899491e-01
-9.51772988e-01 1.18840289e+00 4.03794616e-01 1.31613314e-01
1.43130884e-01 2.01938942e-01 -1.22449136e+00 3.54381353e-01
3.78652453e-01 2.46487245e-01 1.49525270e-01 9.83651459e-01
-7.33483583e-02 1.51490760e+00 -4.65198129e-01 9.85167325e-02
-4.79820907e-01 -5.91449499e-01 -1.17835772e+00 -5.31209707e-01
-6.38531864e-01 6.24949932e-02 -3.44625741e-01 7.20824122e-01
-2.43927374e-01 5.15462784e-03 1.15273428e+00 -1.86687303e+00
-5.96995592e-01 1.49019313e+00 3.33854593e-02 8.19576904e-02
7.34140515e-01 3.12381804e-01 1.20706618e+00 -2.82872498e-01
1.02805138e+00 1.93216133e+00 9.49613005e-02 8.44342887e-01
-1.77480006e+00 -5.76811492e-01 3.96906555e-01 2.21498802e-01
-1.09181345e+00 -2.64951348e-01 5.55880606e-01 -8.30430448e-01
1.01812410e+00 3.51907760e-01 4.84252572e-01 1.70409322e+00
2.24150449e-01 5.53459346e-01 1.05682588e+00 -4.59044069e-01
7.36219406e-01 -2.86376595e-01 1.59762487e-01 4.52376276e-01
-4.09142748e-02 8.38060915e-01 -1.05819404e+00 -8.42929661e-01
7.91707754e-01 -8.00127685e-01 1.61044568e-01 4.60675731e-02
-1.13011539e+00 5.19088566e-01 -2.98550546e-01 1.51959300e-01
-3.50749671e-01 1.12575150e+00 1.39248595e-01 1.04741834e-01
-1.31444782e-01 6.98381901e-01 -9.08876598e-01 -4.66117233e-01
-8.14385563e-02 9.44958568e-01 1.08236885e+00 1.32372236e+00
-1.27848210e-02 -3.19145828e-01 3.05084646e-01 8.67695883e-02
8.03728402e-01 6.51954472e-01 6.76573664e-02 -1.28665018e+00
1.60712585e-01 4.75192755e-01 3.93759727e-01 -9.75486517e-01
-4.33532119e-01 2.54040658e-01 -5.06803930e-01 -1.09123915e-01
3.30482334e-01 -2.20403403e-01 -6.39293194e-01 2.15763855e+00
1.71559855e-01 2.33914629e-01 1.72247097e-01 5.48476994e-01
5.77028453e-01 8.37788224e-01 8.57304573e-01 -4.81910765e-01
1.08129513e+00 3.01083297e-01 -9.49217916e-01 -3.11360300e-01
3.03345472e-01 -2.35102028e-01 6.42075539e-01 4.31571960e-01
-8.39867949e-01 1.75921053e-01 -1.00399566e+00 1.85888425e-01
-7.83547238e-02 -8.50684404e-01 8.80494416e-01 1.61061570e-01
-5.95497429e-01 4.04561281e-01 -1.08011794e+00 -4.38179187e-02
4.44228910e-02 2.05183685e-01 -1.03858532e-02 2.91733265e-01
-1.56968629e+00 7.55336523e-01 7.52836227e-01 -5.93987405e-01
-1.62942398e+00 -8.88821304e-01 -4.86877441e-01 -1.24324329e-01
6.13615096e-01 -4.83125478e-01 1.75873768e+00 -6.42454565e-01
-1.34006429e+00 3.47877979e-01 -2.79229254e-01 -4.81715888e-01
2.13845968e-01 -2.16422394e-01 -8.48009109e-01 -3.21258008e-01
9.68543217e-02 2.96678483e-01 2.89121538e-01 -1.49956822e+00
-7.51018703e-01 -1.38146788e-01 7.96059966e-02 -2.31301561e-01
6.10114872e-01 4.28423703e-01 -1.59479886e-01 -5.35115480e-01
-1.77136790e-02 -7.67294586e-01 -5.58168530e-01 -5.27967393e-01
-6.47154987e-01 -6.64031923e-01 7.20076025e-01 8.00636113e-02
1.31969464e+00 -1.67024148e+00 4.36170638e-01 3.63912165e-01
2.47256383e-01 -6.97110415e-01 4.21218947e-02 7.53556371e-01
-4.47812378e-01 8.56709540e-01 -2.27619912e-02 1.97436124e-01
3.15074265e-01 6.29688263e-01 -7.31721640e-01 2.52474010e-01
6.07437015e-01 7.25312471e-01 -1.23751402e+00 -9.28711891e-01
7.80442804e-02 -3.12137991e-01 -5.77907562e-01 5.43151855e-01
-1.36087739e+00 4.02292252e-01 -7.31524944e-01 4.32793200e-01
3.37973796e-02 -2.97530830e-01 8.56580377e-01 6.84183717e-01
-3.27651292e-01 7.59789288e-01 -1.47284591e+00 1.33496070e+00
-1.33883134e-01 5.12148619e-01 -4.48397458e-01 -5.20508766e-01
6.61201417e-01 8.17941725e-01 2.32282996e-01 -2.73203790e-01
-2.02522967e-02 -4.58146818e-02 3.31521854e-02 -6.53429627e-01
1.55202553e-01 -3.81819636e-01 -6.82435930e-01 8.06494176e-01
2.86523670e-01 -4.70931351e-01 1.26576096e-01 7.10065544e-01
1.45282960e+00 1.39686853e-01 5.53069174e-01 -6.49467945e-01
-1.56395480e-01 3.55516315e-01 1.00467682e+00 9.21419561e-01
1.03664868e-01 1.05853073e-01 1.38104880e+00 -4.53751951e-01
-8.78835261e-01 -1.42263651e+00 6.03079097e-03 1.08801913e+00
2.79180616e-01 -9.18828011e-01 -2.54334748e-01 -4.07688528e-01
-2.32828766e-01 1.50779915e+00 -8.63663316e-01 7.06743672e-02
-4.85165536e-01 -6.26344740e-01 8.24115694e-01 4.82794136e-01
-1.33685008e-01 -9.12431598e-01 -7.22299695e-01 6.00413442e-01
-3.35389674e-01 -9.08781886e-01 1.47467628e-01 5.51040888e-01
-7.18154609e-01 -1.39004505e+00 5.23146331e-01 -1.71102747e-01
1.64923131e-01 -6.93406820e-01 1.48200762e+00 -1.15887597e-01
-1.78629398e-01 4.06668752e-01 1.54330896e-03 -7.21628726e-01
-9.32896554e-01 -4.29115653e-01 2.32226610e-01 -6.74903035e-01
1.22017443e-01 -9.19058502e-01 1.09872445e-02 4.59364444e-01
-5.86856306e-01 4.07228768e-01 -1.62344694e-01 6.08574510e-01
2.78828472e-01 3.78893018e-01 2.86344558e-01 -1.12572694e+00
7.01770127e-01 -9.86971080e-01 -8.75545502e-01 4.00699347e-01
-5.99925995e-01 4.09395516e-01 3.24504912e-01 -6.33520603e-01
-1.53975022e+00 2.54097760e-01 6.56110406e-01 1.88921481e-01
-3.69655758e-01 8.27644289e-01 -3.78252268e-01 1.26842773e+00
1.19948387e+00 -1.78448066e-01 -5.43379128e-01 -1.62665948e-01
8.00040066e-01 2.90450291e-03 9.50978041e-01 -1.72240031e+00
6.78178489e-01 1.96233034e-01 1.16116613e-01 -2.18513742e-01
-3.91384631e-01 4.36439971e-03 -5.09631336e-01 -4.81514513e-01
6.63326740e-01 -7.20777631e-01 -1.06105804e+00 2.23577186e-01
-1.63147807e+00 -9.65354860e-01 -4.33716066e-02 2.40556210e-01
-7.49751568e-01 -5.31707525e-01 -3.23279947e-01 -1.34308982e+00
4.61045623e-01 -6.15775704e-01 5.92243016e-01 2.27555275e-01
-1.13257670e+00 -1.16428459e+00 4.50298041e-01 -6.32514060e-01
-1.49981290e-01 5.45740485e-01 1.57118928e+00 -4.90327209e-01
-5.63864827e-01 7.69749805e-02 1.11617476e-01 -6.43472075e-01
-1.07591361e-01 8.55970919e-01 -6.73439920e-01 6.31225348e-01
-6.00909650e-01 -2.30173811e-01 -1.51717023e-03 2.10458294e-01
1.04808891e+00 -6.30180776e-01 -7.84640431e-01 -3.03535104e-01
1.17662823e+00 7.00481713e-01 4.55777884e-01 -4.59883027e-02
-3.39400396e-02 9.94493663e-01 3.45910847e-01 7.05694795e-01
4.59231675e-01 4.26726371e-01 4.56668943e-01 4.96508896e-01
2.14046568e-01 -8.57548475e-01 1.75890833e-01 2.13340551e-01
-4.70744997e-01 -5.45745134e-01 -1.60561693e+00 8.24685276e-01
-2.19742727e+00 -1.31084037e+00 -8.52811456e-01 1.52217841e+00
1.45828545e+00 4.00229782e-01 -1.19465113e-01 -2.21559808e-01
5.47762036e-01 -2.26262465e-01 -4.76295829e-01 -2.91862428e-01
-2.64528036e-01 2.36633480e-01 4.63639528e-01 6.94328785e-01
-8.93372178e-01 8.42583954e-01 8.38201714e+00 3.91194999e-01
-3.52022409e-01 5.07122725e-02 4.21168447e-01 -2.42269575e-03
-8.62356424e-01 6.82029903e-01 -3.09600085e-01 3.50989640e-01
1.40269578e+00 -7.72242427e-01 3.89695615e-01 6.49156272e-01
4.62857515e-01 -3.90511334e-01 -1.78802502e+00 4.07286704e-01
-7.02904046e-01 -1.53146267e+00 -2.06357270e-01 6.42151982e-02
7.52098262e-01 -2.80171335e-01 -2.46664494e-01 1.25034556e-01
1.96234012e+00 -1.26480675e+00 1.29102707e+00 7.27097094e-01
5.30796230e-01 -4.52341735e-01 1.46724805e-01 4.23687726e-01
-8.27244282e-01 -2.62659788e-01 1.82060912e-01 -4.25917536e-01
4.08121496e-01 6.80001736e-01 -7.45895624e-01 2.41098970e-01
6.97681010e-01 4.07828718e-01 -1.25329167e-01 3.60574961e-01
-8.04570138e-01 1.18081951e+00 -3.80604595e-01 -3.76132667e-01
-1.28761813e-01 2.52806276e-01 6.87042594e-01 1.24696636e+00
-2.55452871e-01 8.22560251e-01 -1.43487155e-01 1.34050167e+00
3.50015551e-01 -7.51632214e-01 -6.60975456e-01 -2.95823365e-01
9.94916022e-01 3.07286918e-01 -2.40369529e-01 -5.41005731e-01
2.35290736e-01 1.38778925e-01 -1.71697050e-01 4.30815667e-01
-9.44600224e-01 2.21568912e-01 8.12013090e-01 -2.09358484e-02
-5.85466266e-01 -5.19874752e-01 -6.41831517e-01 -6.99207366e-01
-7.02388644e-01 -7.28444099e-01 5.75453997e-01 -7.91356444e-01
-1.22381628e+00 1.74990539e-02 8.44763696e-01 -4.54500884e-01
-8.49805653e-01 -2.83112377e-01 -8.41550112e-01 7.52778828e-01
-1.92965522e-01 -7.33982444e-01 3.24284345e-01 5.88084877e-01
3.83787274e-01 3.68039198e-02 1.15887284e+00 -4.32285756e-01
-4.07635301e-01 -3.44217032e-01 -5.32623053e-01 -1.99591562e-01
3.32255304e-01 -1.69061041e+00 5.65152526e-01 8.88321757e-01
-2.35245392e-01 1.04316199e+00 1.34548557e+00 -1.00229788e+00
-1.38981247e+00 -1.05316913e+00 1.21366107e+00 -9.14405882e-01
1.10614228e+00 -6.91643119e-01 -6.63907409e-01 9.18026388e-01
1.52129665e-01 -3.35654557e-01 5.93304396e-01 8.66395533e-01
-7.63061166e-01 4.58417207e-01 -6.96900547e-01 9.33254480e-01
1.32071841e+00 -6.71582580e-01 -1.16658258e+00 4.98397425e-02
1.04674363e+00 -1.78680211e-01 -7.14114726e-01 1.79909095e-01
6.02762580e-01 -3.34744334e-01 6.83168232e-01 -1.21123874e+00
8.83511782e-01 -4.07827199e-01 -2.81583428e-01 -1.10850441e+00
-3.93075168e-01 -9.93722796e-01 -2.72750974e-01 1.41695952e+00
8.33677411e-01 -2.82283515e-01 2.60723889e-01 1.38897216e+00
1.96901947e-01 2.26730466e-01 -8.68839025e-01 -5.92646599e-01
1.71151161e-01 -1.32060635e+00 7.87025273e-01 1.11858892e+00
7.21093774e-01 4.11446363e-01 5.29745109e-02 4.66758162e-01
8.39665353e-01 -1.03482164e-01 5.57238281e-01 -1.51432014e+00
-4.38253492e-01 -3.55350882e-01 9.59787965e-02 -5.80951512e-01
1.66504279e-01 -4.01889265e-01 6.46493495e-01 -1.27095401e+00
2.13497281e-01 -9.54916358e-01 2.14570299e-01 6.43332303e-01
2.00918555e-01 -9.29161847e-01 -1.93720087e-01 3.17671567e-01
-4.16985363e-01 4.10431981e-01 7.43154645e-01 -2.17835099e-01
-2.80073971e-01 -2.61238396e-01 -5.71562231e-01 9.08054948e-01
6.66501164e-01 -9.70196187e-01 -7.82499969e-01 -2.34330758e-01
8.52784336e-01 6.34184241e-01 7.33281016e-01 -2.74296880e-01
5.06870210e-01 -1.40554583e+00 -7.29226097e-02 -3.93809974e-01
-2.07971349e-01 -3.41164470e-01 8.08589935e-01 4.61297482e-01
-9.90270734e-01 -4.36539799e-02 2.10119292e-01 8.52147877e-01
1.00420982e-01 -5.28670549e-02 3.57288748e-01 -9.22353491e-02
-7.35130310e-01 -3.37831408e-01 -1.23014593e+00 2.28296563e-01
7.88633347e-01 6.68780625e-01 -5.27374208e-01 -3.54180306e-01
-8.84500682e-01 4.37532336e-01 -5.88169023e-02 7.46819377e-01
4.80441719e-01 -1.16951704e+00 -6.34112060e-01 -5.36284685e-01
2.14849025e-01 1.03596613e-01 -2.56816119e-01 2.25265160e-01
-1.03539415e-01 3.34673584e-01 1.78455651e-01 -4.23307180e-01
-7.84938812e-01 5.14105082e-01 2.31094465e-01 -6.85584396e-02
-2.98518002e-01 7.74682224e-01 3.85669827e-01 -4.09826517e-01
1.16268113e-01 -4.50139582e-01 1.49763063e-01 -3.24648440e-01
6.63554311e-01 1.47706658e-01 -6.31720066e-01 1.95026904e-01
-6.12215221e-01 -2.88202167e-01 5.84822357e-01 -9.56273913e-01
1.29725134e+00 4.73884819e-03 -3.96457165e-01 1.14460254e+00
3.34902048e-01 -5.11588514e-01 -1.07471037e+00 -2.64797479e-01
7.71101534e-01 9.81076341e-03 -1.60362244e-01 -1.40805542e+00
-9.76459533e-02 2.33739793e-01 1.40605912e-01 2.28224978e-01
5.83737493e-01 6.57678127e-01 -3.85179460e-01 1.23892091e-01
5.39655447e-01 -1.10035241e+00 3.41150239e-02 5.99523306e-01
1.38866246e+00 -6.01977289e-01 -1.69813827e-01 -3.81016910e-01
-4.41888899e-01 9.93988752e-01 4.53836083e-01 6.00324757e-02
9.27148879e-01 9.47164357e-01 -4.11589116e-01 -5.10561705e-01
-1.98835111e+00 7.67315179e-02 -1.73309639e-01 9.44791257e-01
3.31489146e-01 8.19306016e-01 -2.52198368e-01 1.00937140e+00
-2.19231099e-01 4.40659784e-02 6.29830420e-01 8.10297847e-01
4.23622988e-02 -1.17294979e+00 -4.46599782e-01 8.09006020e-02
-1.23438708e-01 -6.78608716e-02 -7.64431298e-01 1.02068007e+00
2.29589447e-01 1.41365552e+00 2.87181079e-01 -4.14381087e-01
1.61052585e-01 -6.83809351e-03 3.65241587e-01 -6.45555615e-01
-7.12436736e-02 -4.32159491e-02 6.09602392e-01 -6.63701952e-01
-4.75338638e-01 -1.14040625e+00 -1.62880254e+00 -4.01226819e-01
3.37766558e-02 2.46393129e-01 5.67146003e-01 1.03028882e+00
-4.85242158e-02 5.13317108e-01 3.11853081e-01 -2.64503639e-02
5.56892008e-02 -9.45475399e-01 -3.57962668e-01 2.39742193e-02
3.66152707e-03 -8.58511567e-01 -2.45957673e-01 8.54913712e-01] | [8.211679458618164, 5.854409217834473] |
7f479f63-efd1-4930-b996-adc441b4a405 | matching-based-term-semantics-pre-training | 2303.01341 | null | https://arxiv.org/abs/2303.01341v1 | https://arxiv.org/pdf/2303.01341v1.pdf | Matching-based Term Semantics Pre-training for Spoken Patient Query Understanding | Medical Slot Filling (MSF) task aims to convert medical queries into structured information, playing an essential role in diagnosis dialogue systems. However, the lack of sufficient term semantics learning makes existing approaches hard to capture semantically identical but colloquial expressions of terms in medical conversations. In this work, we formalize MSF into a matching problem and propose a Term Semantics Pre-trained Matching Network (TSPMN) that takes both terms and queries as input to model their semantic interaction. To learn term semantics better, we further design two self-supervised objectives, including Contrastive Term Discrimination (CTD) and Matching-based Mask Term Modeling (MMTM). CTD determines whether it is the masked term in the dialogue for each given term, while MMTM directly predicts the masked ones. Experimental results on two Chinese benchmarks show that TSPMN outperforms strong baselines, especially in few-shot settings. | ['Bo Xu', 'Shuang Xu', 'Jing Shi', 'Ziyi Ni', 'Minglun Han', 'Haoran Wu', 'Xiuyi Chen', 'Zefa Hu'] | 2023-03-02 | null | null | null | null | ['slot-filling'] | ['natural-language-processing'] | [ 7.54024923e-01 6.48444295e-01 -4.98520494e-01 -6.47317708e-01
-1.03184164e+00 3.73969018e-03 6.04780912e-01 3.68133932e-01
-6.09084845e-01 5.78735828e-01 5.65478265e-01 -5.85200131e-01
2.25490257e-02 -6.80945992e-01 -2.57139713e-01 -3.53867918e-01
1.26171872e-01 9.62078035e-01 1.67029560e-01 -5.85672975e-01
-3.12907279e-01 -2.79530525e-01 -1.07491827e+00 9.01190639e-01
1.18037879e+00 8.27571452e-01 4.66971666e-01 5.03815114e-01
-7.89232314e-01 1.19358289e+00 -7.42231429e-01 -5.45508623e-01
-2.57012039e-01 -7.83039570e-01 -1.68582058e+00 7.87012428e-02
-7.25148246e-02 -6.47039944e-03 -2.83672154e-01 1.10958743e+00
3.51617396e-01 2.76324272e-01 3.98991883e-01 -9.75563109e-01
-4.35119152e-01 9.81708646e-01 3.88711132e-02 9.76934880e-02
5.04859447e-01 -3.69801313e-01 1.21246469e+00 -4.24197048e-01
7.80510068e-01 1.42236340e+00 5.34235120e-01 1.21210599e+00
-1.01431310e+00 -2.30969086e-01 -8.35098848e-02 7.88163543e-02
-1.08156443e+00 -3.23919564e-01 3.76942962e-01 -1.56998754e-01
1.34253776e+00 7.50006378e-01 3.38198483e-01 1.06879270e+00
-1.00214861e-01 9.24708903e-01 7.51087904e-01 -6.39150858e-01
5.53787202e-02 2.52354413e-01 4.42305773e-01 7.38132894e-01
-3.92849416e-01 -1.21826038e-01 -3.20958227e-01 -5.73063493e-01
3.06861997e-01 7.23650530e-02 -3.28097254e-01 3.76557648e-01
-1.10189700e+00 9.48193133e-01 3.65085423e-01 6.55569494e-01
-2.10221246e-01 -1.14470415e-01 7.03264713e-01 4.96806026e-01
6.17799938e-01 6.91951752e-01 -5.83172202e-01 2.83017848e-02
-7.29179263e-01 2.61267930e-01 7.56876409e-01 8.97424281e-01
5.07908762e-01 -7.76172638e-01 -9.06723619e-01 1.32350504e+00
1.29825741e-01 2.64108092e-01 8.64593148e-01 -8.52530479e-01
4.34222221e-01 8.38663280e-01 -2.22112071e-02 -6.01369083e-01
-6.58637345e-01 -3.21983457e-01 -5.86044788e-01 -7.37250566e-01
1.67338848e-02 3.00586820e-02 -1.11580431e+00 1.89726317e+00
4.66088653e-01 1.34772941e-01 2.71629453e-01 8.14976037e-01
1.39545059e+00 4.57277298e-01 4.75570649e-01 -2.86544174e-01
2.08267498e+00 -1.27983737e+00 -1.19102144e+00 -5.35419405e-01
1.11296189e+00 -7.39314198e-01 9.20303941e-01 -2.98933625e-01
-9.49385524e-01 -2.24031106e-01 -5.72534204e-01 -4.95170832e-01
-2.18501180e-01 -1.71205387e-01 5.34531593e-01 3.21067333e-01
-8.69334161e-01 4.37763274e-01 -7.37648249e-01 -3.17201883e-01
1.85445845e-01 7.38808289e-02 -1.77711502e-01 -7.36103654e-02
-2.06947851e+00 9.75700855e-01 4.15795296e-01 -2.04401299e-01
-7.22659290e-01 -6.98352516e-01 -1.24420762e+00 1.29819676e-01
6.00217521e-01 -9.99526799e-01 1.79629195e+00 -6.41398668e-01
-1.06980753e+00 1.63702583e+00 -5.16676009e-01 -7.53821433e-01
3.54429841e-01 2.43476108e-01 -5.12662172e-01 1.70763865e-01
3.70315135e-01 7.42909193e-01 4.50217307e-01 -8.62343311e-01
-7.94203460e-01 -2.46671230e-01 3.57265741e-01 3.19306225e-01
-1.65717989e-01 1.93106428e-01 -4.86051232e-01 -6.68459892e-01
6.92110211e-02 -6.63381875e-01 -3.94752711e-01 -1.25215352e-01
-8.55807066e-01 -7.33676791e-01 2.33867727e-02 -7.35693395e-01
1.43610275e+00 -2.10879374e+00 -1.56108513e-01 -2.62371451e-01
4.12899971e-01 3.61878186e-01 -1.57403480e-02 6.11060441e-01
1.06866047e-01 -1.41750127e-01 -4.36793208e-01 -8.09768498e-01
1.34294018e-01 6.44222379e-01 -3.99304807e-01 -1.87545359e-01
1.51344970e-01 1.13657379e+00 -1.14639819e+00 -8.48718643e-01
-1.20490469e-01 2.69866407e-01 -2.19690144e-01 5.41772425e-01
-6.98216200e-01 1.69022352e-01 -4.71718967e-01 5.82554579e-01
4.23680335e-01 -7.27602601e-01 3.06304395e-01 -4.55496237e-02
4.39682543e-01 1.02484679e+00 -4.14940089e-01 2.00965977e+00
-5.04911184e-01 9.09015685e-02 8.24237429e-03 -1.04622281e+00
6.22803211e-01 8.44637871e-01 5.46377957e-01 -1.02875280e+00
1.56690255e-01 3.49419117e-01 -3.60973179e-01 -9.23113585e-01
5.05282998e-01 -5.85830927e-01 -2.86931336e-01 3.86181206e-01
1.30133376e-01 2.46058911e-01 9.43721682e-02 2.62286067e-01
1.14080679e+00 -5.39256752e-01 3.04567128e-01 -1.20737381e-01
5.72833002e-01 3.10966223e-01 6.34124577e-01 9.26517546e-01
-2.52820969e-01 4.32043761e-01 4.20995146e-01 -2.60168046e-01
-3.00949574e-01 -6.81482017e-01 -3.21298122e-01 1.45985353e+00
3.79882812e-01 -3.39168310e-01 -8.86035383e-01 -1.05379486e+00
-1.22538447e-01 8.65965605e-01 -8.54122341e-01 -3.80564034e-01
-4.96950597e-01 -6.60976112e-01 7.64594078e-01 3.46402258e-01
2.38972336e-01 -1.20013213e+00 -5.44943869e-01 3.97778511e-01
-9.70805645e-01 -1.22791576e+00 -9.54679430e-01 3.39164704e-01
-6.11570001e-01 -1.02612996e+00 -7.80890167e-01 -1.06385720e+00
4.58008230e-01 3.43435377e-01 1.50289619e+00 4.02036458e-01
-4.93832201e-01 -1.26682267e-01 -4.07844514e-01 -3.35205138e-01
-7.33479202e-01 1.41908988e-01 -5.52129686e-01 -1.92176104e-01
7.94528663e-01 -7.08746389e-02 -6.45134091e-01 2.35490158e-01
-1.01734340e+00 4.69399542e-01 3.08744431e-01 1.29594827e+00
3.34094793e-01 -4.61319059e-01 3.66946816e-01 -1.46754587e+00
6.45366848e-01 -5.45959890e-01 2.71585733e-02 7.60377824e-01
-6.82624519e-01 1.79949507e-01 1.86937869e-01 -3.07730585e-01
-1.12928212e+00 -4.36514169e-01 -5.84971786e-01 -1.06718846e-01
-8.05512443e-02 7.56321549e-01 1.48431361e-01 4.82370377e-01
7.14926898e-01 3.51746410e-01 -1.02126719e-02 -5.61085045e-01
1.53885409e-01 8.44300807e-01 5.36232233e-01 -2.25186095e-01
9.90809426e-02 4.56173837e-01 -5.24627626e-01 -4.64410782e-01
-1.36117089e+00 -8.37270319e-01 1.24006718e-02 2.54566431e-01
1.21076405e+00 -7.43624806e-01 -5.12217999e-01 1.19675018e-01
-1.25580299e+00 -2.65132546e-01 -2.99686372e-01 2.37969577e-01
-3.03171694e-01 1.14471719e-01 -1.04967332e+00 -7.48287320e-01
-5.94306171e-01 -9.55254674e-01 1.39322424e+00 5.41762002e-02
-4.55103755e-01 -1.31921601e+00 3.14498305e-01 8.95532668e-01
4.50727940e-01 -1.24668397e-01 1.24109733e+00 -1.18335509e+00
1.29577428e-01 -4.88820896e-02 -1.55523777e-01 -1.18426479e-01
3.65570396e-01 -8.35032403e-01 -1.16888440e+00 6.43303106e-03
3.20382863e-01 -5.36469340e-01 1.15289354e+00 2.24348754e-01
8.17211032e-01 -3.82640958e-01 -6.70923710e-01 4.57943141e-01
9.30936038e-01 3.63752335e-01 4.28115129e-01 1.20022483e-01
4.84349877e-01 9.10430253e-01 7.29808211e-01 1.74784765e-01
8.29668522e-01 7.75783718e-01 1.80212393e-01 -4.72104192e-01
-6.27930686e-02 -2.50680953e-01 -1.38378739e-01 6.46346331e-01
6.41530216e-01 -2.46516407e-01 -9.21707571e-01 5.57279527e-01
-1.91016066e+00 -7.01308250e-01 2.58585930e-01 1.80286956e+00
1.62479043e+00 -5.55049628e-02 -1.13560788e-01 -2.05665186e-01
7.57013917e-01 1.06050514e-01 -4.12677228e-01 -1.60411656e-01
8.26671794e-02 2.68423259e-01 1.96001437e-02 7.86089122e-01
-1.21615529e+00 8.31734419e-01 5.48512316e+00 1.04626203e+00
-1.00757539e+00 5.13469398e-01 7.74551213e-01 1.77718148e-01
-6.90262318e-01 -2.50231445e-01 -4.97849941e-01 3.47965181e-01
7.71808445e-01 -1.69434592e-01 9.78701487e-02 6.35740578e-01
-1.09471872e-01 2.58981049e-01 -1.15573990e+00 9.02802169e-01
1.14998808e-02 -1.43376410e+00 8.72759148e-02 -2.67638296e-01
4.13555086e-01 -1.03243351e-01 -2.88212508e-01 6.71204269e-01
1.61568031e-01 -1.16024208e+00 1.66621894e-01 4.51241463e-01
6.89678252e-01 -1.93676606e-01 9.45812643e-01 3.88819277e-01
-9.58898246e-01 1.89368069e-01 -2.05632031e-01 3.36013287e-01
3.59037369e-01 4.01111484e-01 -1.14290941e+00 6.57997072e-01
3.07899296e-01 3.35979342e-01 -1.39524683e-01 6.71532929e-01
-1.05709769e-01 4.45526510e-01 6.99534118e-02 3.26479152e-02
3.95349413e-01 2.46445626e-01 3.87894779e-01 1.40489066e+00
-2.35209748e-01 2.59325981e-01 4.10992026e-01 9.57955778e-01
-3.22535574e-01 2.73762852e-01 3.03241089e-02 -2.94082969e-01
5.16360462e-01 1.16213655e+00 -6.08709157e-01 -5.92297554e-01
-4.36865270e-01 1.22014248e+00 2.88969457e-01 8.71502161e-02
-5.72715759e-01 -2.95570493e-01 8.50470841e-01 -1.77749187e-01
-9.21239257e-02 7.40520477e-01 -1.78780735e-01 -1.22869790e+00
7.58351162e-02 -9.30248439e-01 9.81912494e-01 -3.04081410e-01
-1.41600478e+00 9.51142430e-01 -2.39724472e-01 -1.00093031e+00
-6.26142204e-01 -2.44902298e-01 -4.93789911e-01 9.38457847e-01
-1.70996225e+00 -1.22130740e+00 -1.43541768e-01 5.42132914e-01
8.10671806e-01 1.44249082e-01 1.33421791e+00 4.95574921e-01
-2.66457945e-01 8.75085711e-01 -3.05776536e-01 3.60485703e-01
5.99561095e-01 -1.36453605e+00 7.13145375e-01 1.81737885e-01
1.11112848e-01 6.97433352e-01 7.01060832e-01 -6.20958209e-01
-8.61228466e-01 -1.12510657e+00 1.77963376e+00 -1.66690886e-01
3.06754678e-01 -3.47139478e-01 -1.36029899e+00 3.14554304e-01
1.36300698e-01 -2.00542599e-01 1.05981433e+00 1.07477158e-01
-3.07688713e-01 3.74869168e-01 -1.16661060e+00 3.24630141e-01
1.03711212e+00 -1.04258835e+00 -1.06527615e+00 8.03747177e-01
1.37111413e+00 -7.04709768e-01 -6.68107450e-01 4.70199257e-01
1.56486273e-01 -5.57165921e-01 9.38533127e-01 -1.01565158e+00
3.20426971e-01 1.06583104e-01 -5.67908101e-02 -1.05365753e+00
1.24697775e-01 -5.61200142e-01 -1.11039810e-01 9.04238045e-01
6.78883374e-01 -5.04001498e-01 5.36509573e-01 8.28977287e-01
-2.49047965e-01 -1.01617634e+00 -1.16689003e+00 -5.60399532e-01
-3.29827778e-02 -2.82444388e-01 6.72979414e-01 1.39559400e+00
5.71937263e-01 5.75411379e-01 -3.33664536e-01 -1.21016279e-01
1.47178099e-01 3.08704734e-01 -1.31610096e-01 -1.05322397e+00
-5.39471865e-01 -3.98238212e-01 -6.37634844e-02 -1.12068939e+00
3.47507089e-01 -1.13488150e+00 3.42524529e-01 -1.76121414e+00
5.42337477e-01 -5.03719091e-01 -3.37167591e-01 6.64073050e-01
-7.61660933e-01 -1.00773498e-02 -2.05364943e-01 1.39954567e-01
-7.89027095e-01 5.64716041e-01 9.96098161e-01 -4.56789941e-01
-1.35224134e-01 1.63672835e-01 -7.30979204e-01 2.94939756e-01
3.23316514e-01 -5.96842289e-01 -5.12264788e-01 -6.79072320e-01
3.92122306e-02 6.85407460e-01 -6.44484395e-03 -3.00478280e-01
2.45880291e-01 -1.96970016e-01 -4.09372121e-01 -2.52364129e-01
3.39184552e-01 -4.87578452e-01 -2.17380777e-01 7.76437163e-01
-1.11503172e+00 -3.26746523e-01 -1.06310926e-01 2.57623225e-01
-4.21697289e-01 -5.24997592e-01 5.43357491e-01 -3.76341254e-01
-4.07185584e-01 2.37038538e-01 -3.02352935e-01 7.68574417e-01
4.49653238e-01 3.67568463e-01 -3.39993566e-01 -4.01871145e-01
-1.05078685e+00 7.70797312e-01 -1.81338936e-02 5.59421420e-01
5.87758422e-01 -1.15581346e+00 -7.21208870e-01 8.99565294e-02
5.58606505e-01 1.52782789e-02 5.84801733e-01 7.26080120e-01
-2.26290882e-01 6.94549978e-01 5.30812323e-01 -4.98225600e-01
-1.44044077e+00 5.65827787e-01 6.78068399e-01 -7.80614853e-01
-5.29611766e-01 1.37209201e+00 3.79090875e-01 -8.56280446e-01
5.89654565e-01 -3.95816535e-01 -3.57840091e-01 -6.02090731e-02
9.54043031e-01 -1.05241574e-01 2.31132060e-01 -4.81927484e-01
-4.00491863e-01 -1.05820142e-01 -4.03634012e-01 -9.38458741e-02
8.64646256e-01 -1.79255947e-01 -3.81934226e-01 3.92729521e-01
1.39255297e+00 -6.71496809e-01 -4.80020523e-01 -1.05500054e+00
4.87109989e-01 -1.94663748e-01 -5.70847951e-02 -9.31815326e-01
-6.03656530e-01 7.33352065e-01 4.48231310e-01 4.28845555e-01
9.89604115e-01 4.38051581e-01 1.52335691e+00 5.49350739e-01
1.55134797e-01 -1.04693687e+00 -8.51545632e-02 7.06137657e-01
5.06773233e-01 -1.31357372e+00 -7.97024071e-01 -6.74805462e-01
-7.15570927e-01 7.78865159e-01 5.39293170e-01 6.46563947e-01
5.36428988e-01 6.65176585e-02 6.03804529e-01 -6.07534409e-01
-1.01905084e+00 -4.48129565e-01 5.14252484e-01 1.73320234e-01
5.67784190e-01 1.16575681e-01 -3.89040530e-01 7.45058358e-01
-1.36576802e-01 -9.62434858e-02 -1.58667386e-01 1.08734345e+00
-3.88637871e-01 -1.16187918e+00 5.48609681e-02 6.16998196e-01
-7.87992775e-01 -5.26148796e-01 -3.89256865e-01 6.40813485e-02
8.54949877e-02 1.07610404e+00 4.42167893e-02 -4.15276557e-01
3.38821381e-01 3.83318573e-01 1.03496455e-01 -1.01902676e+00
-9.50055540e-01 5.06236032e-02 4.48471040e-01 -5.34164011e-01
-4.65331703e-01 -2.05118760e-01 -1.58212936e+00 2.14595750e-01
-4.68744010e-01 7.72765100e-01 2.71550179e-01 1.53157783e+00
2.44451195e-01 7.69332230e-01 5.17267704e-01 1.12866268e-01
-7.07817614e-01 -1.05185401e+00 -3.39859128e-01 9.36149240e-01
5.29363275e-01 -4.92591023e-01 -9.39360857e-02 -2.23967120e-01] | [8.805784225463867, 8.684979438781738] |
dcd2b604-1bc4-4225-b194-d48b052abed8 | flat-chinese-ner-using-flat-lattice | 2004.11795 | null | https://arxiv.org/abs/2004.11795v2 | https://arxiv.org/pdf/2004.11795v2.pdf | FLAT: Chinese NER Using Flat-Lattice Transformer | Recently, the character-word lattice structure has been proved to be effective for Chinese named entity recognition (NER) by incorporating the word information. However, since the lattice structure is complex and dynamic, most existing lattice-based models are hard to fully utilize the parallel computation of GPUs and usually have a low inference-speed. In this paper, we propose FLAT: Flat-LAttice Transformer for Chinese NER, which converts the lattice structure into a flat structure consisting of spans. Each span corresponds to a character or latent word and its position in the original lattice. With the power of Transformer and well-designed position encoding, FLAT can fully leverage the lattice information and has an excellent parallelization ability. Experiments on four datasets show FLAT outperforms other lexicon-based models in performance and efficiency. | ['Xuanjing Huang', 'Xipeng Qiu', 'Xiaonan Li', 'Hang Yan'] | 2020-04-24 | flat-chinese-ner-using-flat-lattice-1 | https://aclanthology.org/2020.acl-main.611 | https://aclanthology.org/2020.acl-main.611.pdf | acl-2020-6 | ['chinese-named-entity-recognition'] | ['natural-language-processing'] | [-0.39726076 -0.4772567 -0.22666173 -0.14534497 -0.8616117 -0.78905714
0.14397582 0.17945808 -0.69959575 0.678722 0.441427 -0.52778167
0.47739026 -1.0837616 -0.38216767 -0.4644934 0.10598721 0.47632322
0.47277325 -0.11466485 0.04483442 0.16543591 -0.8198652 0.1639546
0.9516016 0.49393138 0.44644564 0.44450155 -0.7826779 0.43926132
-0.28685912 -0.50645757 0.02485735 -0.24647877 -0.5891627 -0.51301175
0.02354327 -0.2752358 -0.3426643 1.1206056 0.61734027 -0.14042123
0.15951748 -0.47544593 -0.47897455 0.89986956 -0.4623898 0.24827838
0.27454108 -0.3621669 1.2219751 -0.8952845 0.5725737 1.1596755
1.0202479 0.41799092 -0.8123142 -0.7175571 0.2515077 0.08151329
-1.5144942 -0.03162261 0.20918424 0.04584518 1.3421485 0.09248265
0.66116583 0.672325 0.1205845 0.9192742 1.033034 -0.37226582
0.15394104 -0.39301178 0.36647668 0.9223104 0.457593 -0.27982602
-0.35640985 -0.43260357 0.79518896 0.05208779 0.07605705 0.24948402
-1.0038357 0.61539316 0.2571648 0.39361304 -0.23427019 -0.01005205
0.5960583 -0.13413794 0.33534172 0.08456051 -0.6449145 -0.38916165
-0.76379085 0.02332692 1.0642885 0.9568719 0.7074781 0.04160387
-0.35413554 0.64949936 0.25843126 0.55110925 0.6147678 -0.30430967
0.62229264 0.8343216 -0.16679454 -0.8069513 -0.4430945 -0.40633515
-1.1066632 -0.7561031 0.0834824 -0.27730146 -0.7131984 1.4966106
0.354708 0.6516305 0.06600585 0.5018132 0.71766806 0.86621064
0.2361056 -0.21694458 1.8874934 -1.0819905 -0.7464612 -0.24723917
0.8974184 -0.7313339 1.108074 0.05595429 -0.83922154 -0.15543847
-0.9862875 -0.3673633 -0.2647684 0.3596 0.87314856 0.92712307
-0.791092 0.3961991 -1.1494322 -0.08149674 0.04590618 0.2142287
-0.24471356 -0.27021274 -1.4353373 0.4286377 0.71652424 0.27276623
-0.22036646 -0.35897395 -0.96190315 0.4037759 0.20112723 -0.45725876
1.0611451 -0.10546107 -1.5500265 0.37495863 -0.6203145 -0.372369
0.1445148 -0.3083152 -0.5033069 -0.2256578 0.17208359 0.1652273
0.09005706 -0.4249196 -0.5365624 -0.25110108 -0.12937222 0.16266131
-0.8268579 0.17937757 -1.0249953 -0.68479145 0.27787057 -0.74554336
-0.6437793 -0.6476168 -0.4570815 -0.37042332 0.37224743 -0.6764317
2.0609803 -1.9766179 -0.54269934 0.16718939 0.04947264 0.6584318
-0.09179561 0.77611315 0.4077039 0.3280423 -0.14318936 -0.2572568
0.11356346 0.38732406 -0.43226856 0.17917077 0.05255478 1.0993503
-1.1318762 -0.70223826 -0.13175926 0.46070144 -0.7709147 0.12290069
-0.14437172 -0.03390434 -0.84985757 0.4720698 0.9627077 -0.19640768
0.75678045 -0.03251647 -0.36394328 0.763589 -1.185332 1.5097557
-0.31844455 0.0099327 -0.2186808 -0.32684055 1.1065013 0.44736302
0.10561783 -0.7682375 -0.26398712 0.4438096 -0.25305098 -0.00793671
0.64919657 0.04344017 -0.43346334 0.6198699 -0.23524626 0.36246565
0.42762494 0.21674794 1.3817575 0.05412105 0.55037344 -0.2306243
0.48860925 -0.03670518 1.3246012 0.6908231 0.13983929 0.3189261
0.57338387 -0.57781976 -1.0705142 -0.8783128 -0.04941204 1.0924001
-0.06703699 -1.0810825 -0.96981716 -0.58194417 -0.46915916 0.33471122
-0.01692538 0.10847721 -1.2333056 -0.9846561 1.3103082 0.7933085
0.84495574 -0.9467334 -0.20451233 0.6982504 -0.38685083 -1.2456813
-1.0791262 0.00867671 -1.0890685 -0.7699881 -0.37998298 -1.1860768
0.6958732 0.19860308 1.180809 0.20069042 -0.02125561 -0.3798247
-0.55224806 -0.12523326 -0.08485141 0.52840984 -0.06020284 -0.37918684
0.6273736 -0.48179212 -0.3973112 0.18466833 -1.0463148 0.19067018
0.76097935 1.1703783 0.7349835 0.07271421 0.2104351 -1.1009117
0.69802296 -0.31528887 -0.6976363 0.45537534 -0.68283486 0.45680243
0.8605333 -0.27166325 -1.1847308 0.1595721 -0.5645227 0.10738076
0.18959153 0.92982733 -0.54038817 0.41035938 0.13459192 0.45269862
-0.6214727 -0.94889766 0.19235697 0.652173 0.4903511 -0.85551053
0.52478516 0.26261336 -0.23015308 -0.4978391 -1.0194039 -0.6215176
-0.6279434 0.44859907 0.6536559 -1.0401813 -0.5068475 0.46799752
-1.3774198 -0.04601256 0.22729677 0.43262225 0.15135743 0.7448985
-1.3138939 -0.71515393 -0.72099316 -0.9087886 0.9376945 0.32204962
0.34570637 -1.0389049 0.3357004 -0.05079838 0.20463033 -0.3084912
1.0917331 -0.6605132 -0.47351435 -0.38766417 -0.39411163 -0.01144988
-0.14965165 -0.45251828 -0.54710174 -0.3836794 -0.18011412 -0.0438892
0.87454915 -0.21672413 0.9475253 -0.3807941 -0.36458543 0.7134798
1.5640715 -0.12132704 0.7205186 0.3224946 0.7707949 -0.06652018
0.43578014 0.53986925 0.67908955 0.5357096 -0.11796079 -0.01316948
0.2076503 -0.8077128 0.63831836 1.7969813 -0.04702703 -0.24442129
-1.0544877 0.4383545 -1.8773453 -0.8196928 -0.20417124 1.9848096
1.1427841 0.20391086 -0.31658614 -0.32536137 0.8284296 0.14438286
-0.41534948 -0.43766338 -0.3628436 0.30679145 0.8627321 0.2624177
-0.99985677 1.3980368 6.6220565 1.2548652 -0.7506552 0.2691747
0.31660584 0.5135891 -0.52788466 0.27856714 -1.4491833 0.37344208
0.97195727 -0.14871278 0.25969797 0.70360947 -0.05555548 0.30685556
-0.5737178 0.67850375 -0.20831369 -1.377309 0.16335261 0.1308705
0.6119123 0.23726591 -0.42483205 0.5182084 0.5898892 -0.8779203
0.3539348 0.29525003 1.0142659 -0.73712605 0.9800411 0.62874085
-1.9374254 0.17709477 -0.8090873 -0.18969555 0.4223625 0.6173938
-0.8044696 0.60933864 0.43960223 0.6035702 -0.25795293 0.9499537
-0.41080207 1.0177147 -0.45345798 -0.35082278 0.6376199 -0.36726907
0.34626392 1.820926 0.28721705 0.4034783 0.6226595 0.41268474
-0.10145137 0.3880992 -0.29497722 -0.19035926 0.9480503 1.3020328
-0.8435134 -0.23524536 -0.56854326 0.9904347 0.6670583 -0.02601437
-0.68029743 -0.54654783 0.5357557 -0.1294652 0.5707397 -0.61698407
-0.25932795 -1.4093807 0.09716423 -0.7301691 0.5631746 -0.327347
-1.2653832 0.8688615 -0.40211457 -1.1336813 0.00732685 -0.68692833
-0.5969214 0.89010787 -1.5156932 -1.1006509 0.16782603 0.42597806
0.35600564 -0.03722057 1.2971553 0.388575 -0.8630038 0.9131066
0.49840006 0.8032436 0.48585495 -1.3401269 1.0550004 0.94694734
0.30026293 1.2012533 0.04198628 -0.9603777 -1.6069596 -1.1576103
1.301174 -0.13856742 0.7024152 -0.5679622 -0.85291547 0.5865033
0.01991594 0.05160312 0.85430133 0.21577191 -0.64036894 0.0985784
-0.71039224 0.73733187 1.0892432 -0.6520329 -0.65657365 0.10245524
0.8755097 -0.5911479 -0.8420267 0.16510329 0.5451378 -0.6411892
0.83662355 -0.5657813 -0.06156934 -0.32870123 0.10996742 -0.92705154
-0.5703313 -0.5860372 -0.07558472 1.436248 0.4036291 -0.649399
0.79326653 0.27727357 -0.11428677 -0.79470515 -0.8824817 -0.73148304
0.04695459 -0.541958 1.0014341 0.79935205 0.15367644 0.7726622
-0.39459366 0.1131397 0.3947491 0.1172483 0.32736427 -1.1068913
-0.14348048 -0.17153572 -0.14356801 -1.5528753 0.14258942 -1.0531651
0.11464593 -1.5278494 0.46685523 -0.90427715 -0.52346677 0.74328583
-0.43401906 0.06799423 0.07227442 0.40038696 -0.8514258 0.41941056
1.0510563 0.17750254 -0.05015025 -0.22033937 -0.49477795 0.85652685
0.6234615 -0.66666 0.07301089 -0.8865318 0.7203212 0.20132431
-0.31705713 -0.6632664 0.7232225 0.08115945 0.14824742 -0.86755157
-0.01121352 -0.3988441 -0.0104331 0.5278069 -0.02425079 0.586826
0.13220268 0.6762009 -0.2813453 -0.33147517 0.23955332 -0.2111725
-0.7930333 0.6703631 -0.2849786 0.293938 0.4136125 -0.06083933
0.02622797 0.16094282 -0.1873946 0.26892078 0.24437714 0.20841509
0.5106056 -1.4273695 -0.7157987 0.24216583 -0.09418262 0.19605494
0.19631352 0.4768255 -0.82119346 0.6140155 0.09768125 -0.16335557
-1.0252624 0.23697636 0.01355945 -1.151555 -0.873381 0.61963063
0.05252471 -0.7182209 -0.07431734 -0.2639284 -0.31870896 -0.23774454
0.84948575 0.2429606 0.0254788 -0.39878377 -0.39733723 0.4622368
-0.09960271 -0.09576123 1.24105 0.01629872 -0.7080422 0.20966186
0.9315652 0.50107473 -0.7682635 -0.626259 0.51330763 -0.16123796
-0.08178051 -0.3918408 -0.8057689 0.7660162 0.0967529 -0.25325057
0.97047776 -0.4952688 1.6078974 0.74610215 0.7294647 -1.0352923
-0.31052205 1.3221047 -0.08108918 -0.586704 -0.2479599 -0.48375326
-0.5158269 1.2062596 0.5464436 -0.08838136 0.7473111 0.80193347
-0.03480504 0.18076894 -0.78468215 -0.2600479 0.0641873 0.21196057
0.57041264 0.2530224 -0.84139955 1.2958792 -0.10704892 -0.01373631
0.25286195 0.81922585 -0.5092094 -1.7630254 -0.19193281 0.2938118
-0.76092833 -0.7929649 0.08494632 0.33745003 0.11541048 0.6151846
0.08412484 -0.31057602 0.16441947 0.04709558 0.07511093 -0.9179877
-0.7463282 0.31477493 0.16257338 -0.46749187 -0.05321538 -0.66397876
-1.7542654 -0.38662726 -0.41262555 0.60065454 0.34413627 1.065114
0.5107691 0.2146049 0.44185168 -0.04745317 -0.62198013 -0.8703373
-0.6993507 -0.10306904 -0.44850114 -0.14009006 0.21848151 -0.3078351 ] | [9.794939994812012, 9.772170066833496] |
df55c60f-0772-4121-ac53-6554ed90c02d | learning-governing-physics-from-output-only | 2208.05609 | null | https://arxiv.org/abs/2208.05609v1 | https://arxiv.org/pdf/2208.05609v1.pdf | Learning governing physics from output only measurements | Extracting governing physics from data is a key challenge in many areas of science and technology. The existing techniques for equations discovery are dependent on both input and state measurements; however, in practice, we only have access to the output measurements only. We here propose a novel framework for learning governing physics of dynamical system from output only measurements; this essentially transfers the physics discovery problem from the deterministic to the stochastic domain. The proposed approach models the input as a stochastic process and blends concepts of stochastic calculus, sparse learning algorithms, and Bayesian statistics. In particular, we combine sparsity promoting spike and slab prior, Bayes law, and Euler Maruyama scheme to identify the governing physics from data. The resulting model is highly efficient and works with sparse, noisy, and incomplete output measurements. The efficacy and robustness of the proposed approach is illustrated on several numerical examples involving both complete and partial state measurements. The results obtained indicate the potential of the proposed approach in identifying governing physics from output only measurement. | ['Souvik Chakraborty', 'Tapas Tripura'] | 2022-08-11 | null | null | null | null | ['sparse-learning'] | ['methodology'] | [ 2.74746567e-01 -4.21017826e-01 2.68558592e-01 -1.10039942e-01
-7.52464473e-01 -6.81998253e-01 6.58485115e-01 4.90852147e-02
-1.08289540e-01 1.09413540e+00 8.97851288e-02 -9.00088064e-03
-6.04554653e-01 -7.22304046e-01 -5.22132277e-01 -1.15744233e+00
1.22822180e-01 4.28052783e-01 1.17103174e-01 2.12780192e-01
3.73523861e-01 7.10135460e-01 -1.28312111e+00 -1.92930803e-01
6.51870430e-01 1.12863195e+00 2.48027056e-01 7.35086262e-01
-3.61860067e-01 7.21815050e-01 -2.83424020e-01 3.65960568e-01
3.72551054e-01 -4.87118512e-01 -1.29786596e-01 9.70134437e-02
-1.46527752e-01 4.38670348e-03 -5.52833319e-01 1.05371487e+00
3.73330474e-01 1.66156337e-01 9.87211585e-01 -9.25304174e-01
-3.62678230e-01 2.65878171e-01 -5.09159863e-01 5.21818876e-01
2.02325821e-01 2.84009427e-01 5.20184219e-01 -5.96148968e-01
3.23463798e-01 9.47934270e-01 4.73199695e-01 7.17836544e-02
-1.27718878e+00 -6.13783598e-01 -3.56097706e-02 -2.95170527e-02
-1.37468004e+00 -5.93553126e-01 1.11814034e+00 -7.92867362e-01
4.46931988e-01 1.67327572e-03 5.65880418e-01 8.55175197e-01
2.51957029e-01 5.04924595e-01 1.80110681e+00 -2.89637208e-01
6.55346334e-01 2.50378311e-01 5.27940154e-01 6.18684471e-01
5.08417785e-01 5.72862744e-01 -5.12251377e-01 -4.43372458e-01
8.82193089e-01 -1.76259455e-05 -2.94101000e-01 -5.48457861e-01
-9.36608791e-01 4.89524990e-01 -2.63671726e-01 1.67413026e-01
-5.45833111e-01 1.67948350e-01 1.21353352e-02 1.64258569e-01
9.42773297e-02 1.92141488e-01 -2.48101830e-01 -1.70648769e-01
-1.19881380e+00 3.28192413e-01 1.14767015e+00 8.22220027e-01
5.54782987e-01 3.89392227e-01 -9.45627838e-02 2.44950116e-01
3.98821384e-01 1.21906698e+00 -1.50663324e-03 -1.01513410e+00
-6.36240793e-03 1.17770158e-01 5.69026768e-01 -7.39868402e-01
-1.37680933e-01 -6.44076943e-01 -9.81421649e-01 2.20871672e-01
3.97189200e-01 -4.44840342e-01 -8.94599736e-01 1.64397979e+00
4.27770108e-01 6.63345039e-01 2.70010650e-01 8.61461580e-01
6.02150619e-01 9.20082390e-01 -2.46857628e-01 -5.21579981e-01
8.73595655e-01 -9.71849859e-02 -1.05107784e+00 3.31118554e-01
-1.27389029e-01 -7.08122969e-01 5.11206031e-01 3.79091471e-01
-1.19315600e+00 -4.62601572e-01 -8.04308295e-01 3.46333206e-01
-8.42905641e-02 2.96048075e-01 4.99235809e-01 3.37157100e-01
-6.38202488e-01 6.68641388e-01 -1.17883158e+00 -3.43087614e-01
-3.10692494e-03 2.89786726e-01 5.29570766e-02 1.48773089e-01
-8.82527471e-01 6.15814149e-01 9.64057967e-02 2.88002849e-01
-1.11237049e+00 -7.61640429e-01 -4.45307583e-01 1.44907355e-01
2.95427918e-01 -6.96916521e-01 8.09821367e-01 -4.58466262e-03
-1.87569058e+00 1.10567130e-01 -4.21269238e-01 -3.15547824e-01
2.79923499e-01 -1.37722850e-01 -2.35781878e-01 1.15298249e-01
-1.47597477e-01 -3.87116998e-01 9.41526830e-01 -1.42066050e+00
-2.84130812e-01 -1.81719884e-01 -2.27404028e-01 -7.38143176e-02
3.66094001e-02 -5.02252281e-01 1.58250049e-01 -3.08347791e-01
5.26082277e-01 -6.79790914e-01 -2.03010753e-01 -7.96023980e-02
-3.42495263e-01 2.47873932e-01 9.62429762e-01 -4.40744430e-01
7.39844799e-01 -1.96595764e+00 2.82683730e-01 5.67981184e-01
1.58487841e-01 -7.12581202e-02 3.74404371e-01 6.47567034e-01
3.00566018e-01 -3.64404917e-01 -5.33242285e-01 -1.22418247e-01
7.80631928e-03 2.05450192e-01 -7.44746447e-01 6.24201298e-01
7.50286505e-02 3.95850360e-01 -7.49136806e-01 -3.59128743e-01
4.60648388e-01 4.39670414e-01 -2.37432793e-01 4.07204211e-01
3.94410826e-03 1.16416001e+00 -8.83576214e-01 5.21745265e-01
7.03222215e-01 -4.05994803e-01 -1.73822254e-01 -3.46456021e-01
-4.48938638e-01 9.05496441e-03 -1.86689699e+00 1.49528289e+00
-2.96883643e-01 2.21667483e-01 5.54370165e-01 -1.34448588e+00
9.31867003e-01 4.27777737e-01 6.65635884e-01 -2.17780247e-01
3.64916205e-01 1.60756007e-01 -1.40513768e-02 -7.81581640e-01
4.54648286e-02 -4.88968045e-01 3.02564967e-02 6.83034003e-01
2.38235742e-01 -4.69074041e-01 1.02304883e-01 2.07493216e-01
1.31056428e+00 2.13603020e-01 3.16690475e-01 -4.31512803e-01
6.44065619e-01 -3.73811051e-02 7.18937457e-01 8.68696868e-01
-2.16862038e-01 4.51085865e-02 2.47858316e-01 -8.46817940e-02
-1.12514818e+00 -1.47025001e+00 -2.66850591e-01 1.88754991e-01
2.96181887e-01 4.07026969e-02 -3.00360620e-01 6.76395558e-03
1.91605940e-01 4.61945444e-01 -3.55127692e-01 -1.15196206e-01
-4.48873937e-01 -7.99152315e-01 3.42428833e-01 2.38372535e-01
4.03438568e-01 -8.62706184e-01 -3.48654062e-01 3.81498545e-01
6.00490570e-02 -1.43235874e+00 1.24953136e-01 2.01054633e-01
-1.16431201e+00 -9.47251320e-01 -4.90579128e-01 -2.31267884e-01
4.69379395e-01 -1.57605812e-01 5.96468091e-01 -4.41609800e-01
-4.96322095e-01 6.68621659e-01 -2.05529276e-02 -2.17659459e-01
-3.09183389e-01 -3.44924837e-01 3.99252087e-01 2.17578456e-01
4.49123085e-02 -1.17792714e+00 -3.76770020e-01 -2.21813977e-01
-6.08712316e-01 -2.49238580e-01 7.68974006e-01 6.52877212e-01
5.45838237e-01 3.87695849e-01 3.95770788e-01 -6.80835187e-01
5.19425392e-01 -6.17590547e-01 -9.30335581e-01 -1.05506316e-01
-4.19915646e-01 3.57095748e-01 6.31921887e-01 -2.60868877e-01
-1.42924547e+00 2.21073747e-01 1.59715071e-01 -4.95563596e-01
-4.34013605e-01 3.74184549e-01 -2.35400535e-02 -2.27511659e-01
3.43121350e-01 8.33850861e-01 -1.31836459e-01 -7.21658647e-01
1.57020286e-01 4.63983804e-01 6.13101423e-01 -1.10675859e+00
1.06705415e+00 7.91811824e-01 4.93032426e-01 -1.08236742e+00
-6.99195147e-01 -5.54778755e-01 -5.80478013e-01 -1.04439914e-01
4.89263445e-01 -7.54050970e-01 -1.04987860e+00 5.16916752e-01
-9.33690310e-01 1.14978068e-01 -4.97251719e-01 9.35534239e-01
-4.61492419e-01 4.87198740e-01 -6.81876063e-01 -1.53861094e+00
-2.17226684e-01 -8.49806309e-01 1.07591367e+00 2.70738631e-01
5.06445579e-03 -9.93191600e-01 3.19525242e-01 4.18524146e-02
5.03736198e-01 3.47142696e-01 1.00488460e+00 -2.29731888e-01
-9.59428906e-01 -2.99890310e-01 -1.48465604e-01 1.77345470e-01
1.49519280e-01 -7.79728740e-02 -9.43617225e-01 -2.56283700e-01
7.50907421e-01 -1.92673802e-01 7.93760002e-01 6.26976013e-01
6.59095585e-01 1.87934756e-01 -3.70615780e-01 4.66005743e-01
1.77242482e+00 1.91563234e-01 2.13519275e-01 -4.21656311e-01
7.16599524e-01 5.22575438e-01 6.06215261e-02 8.27909052e-01
3.71125154e-02 1.77603334e-01 3.45315300e-02 2.88631797e-01
4.54225205e-02 2.73224302e-02 3.34750354e-01 1.38392317e+00
-2.74088476e-02 -4.98869270e-02 -7.51999378e-01 5.93290806e-01
-1.71465135e+00 -9.63169396e-01 -1.68941602e-01 2.18189931e+00
6.51670992e-01 2.67746057e-02 -3.96811634e-01 2.18019992e-01
6.96755469e-01 -1.97363406e-01 -7.22850084e-01 -7.51997977e-02
-3.07257306e-02 4.71546263e-01 4.27975565e-01 5.24147511e-01
-8.03082049e-01 5.01679182e-01 6.59274626e+00 6.82287157e-01
-9.59938645e-01 2.30561078e-01 1.54188350e-02 -1.38737679e-01
-1.69245154e-01 3.14010739e-01 -9.43105280e-01 5.82643032e-01
9.23706472e-01 -1.05717756e-01 3.82921815e-01 3.18046629e-01
6.29484355e-01 -4.97029364e-01 -9.10189986e-01 1.05221200e+00
-3.78809392e-01 -1.22573149e+00 -7.30689466e-02 2.46093958e-03
8.08665693e-01 -2.22646609e-01 -1.36516020e-01 5.15860058e-02
5.01037121e-01 -7.37975001e-01 4.69440967e-01 1.23828495e+00
3.26573014e-01 -1.87252939e-01 6.95136428e-01 7.19644904e-01
-1.24972796e+00 -1.08100899e-01 -1.82693288e-01 -3.98394763e-01
6.11516774e-01 1.20237517e+00 -2.08641127e-01 4.85104084e-01
4.67965722e-01 7.29497552e-01 2.04182670e-01 1.14642262e+00
-2.34868564e-03 1.05014503e+00 -7.82383680e-01 -7.44219571e-02
3.43451984e-02 -6.32218361e-01 9.84069347e-01 9.67293024e-01
6.68961525e-01 6.14570916e-01 4.73965168e-01 1.33637786e+00
4.13599014e-01 -2.73243636e-01 -7.84326911e-01 -2.40921929e-01
5.10987639e-01 1.14311433e+00 -7.96834052e-01 -3.99173021e-01
-3.16472083e-01 3.62742633e-01 -1.13282084e-01 6.18002295e-01
-7.06104577e-01 -9.97364894e-02 3.97028059e-01 5.26659675e-02
4.38652903e-01 -4.89247531e-01 -4.28274512e-01 -1.26782548e+00
-7.64443278e-02 -3.19457978e-01 -5.14331758e-02 -6.40874565e-01
-1.53670037e+00 1.10668212e-01 3.43101889e-01 -1.09524012e+00
3.20984535e-02 -5.04525661e-01 -6.81878090e-01 1.03839910e+00
-1.43714225e+00 -8.53230715e-01 -3.55754554e-01 7.17620373e-01
1.39748156e-01 -9.99355987e-02 7.05569386e-01 1.41984120e-01
-3.77951562e-01 -1.34962007e-01 8.24647725e-01 -1.55857116e-01
2.85043985e-01 -1.20606875e+00 -1.51768237e-01 6.88851118e-01
2.39251107e-02 6.78793848e-01 1.14619684e+00 -8.67573738e-01
-1.68996370e+00 -6.95677936e-01 3.03933501e-01 -1.01743706e-01
8.67468059e-01 -4.21940774e-01 -7.71645367e-01 2.23640680e-01
-1.02227936e-02 1.57352477e-01 4.80104715e-01 -2.05989257e-01
1.90903768e-01 -1.11271329e-01 -1.21663964e+00 1.32629916e-01
7.78823912e-01 -6.52446032e-01 -7.39820421e-01 2.57364750e-01
7.73334727e-02 -2.75458932e-01 -9.60985780e-01 3.21633697e-01
4.90406454e-01 -5.42278826e-01 7.36760616e-01 -3.43260348e-01
6.45486712e-02 -6.36875212e-01 -3.19475055e-01 -1.12908435e+00
-3.18991572e-01 -9.20714557e-01 -6.26511693e-01 1.23016703e+00
6.33861497e-02 -5.11011958e-01 7.20961094e-01 4.08979535e-01
1.10063143e-01 -5.05987406e-01 -7.28375793e-01 -8.05387378e-01
1.14568640e-02 -4.41506237e-01 2.53841013e-01 7.04320610e-01
-2.29331926e-01 2.23748401e-01 -4.10537869e-01 6.34643734e-01
1.24478102e+00 5.21507561e-01 5.64013362e-01 -1.39997339e+00
-7.83975542e-01 1.55756921e-01 -3.23012769e-01 -1.02536881e+00
2.51456443e-02 -5.54922163e-01 3.05226564e-01 -1.29937243e+00
3.54061157e-01 -5.05672514e-01 -5.37432611e-01 -2.12222040e-01
1.14255421e-01 -9.56269354e-02 7.98592195e-02 4.93028760e-01
-1.49912179e-01 7.97298014e-01 1.11103463e+00 1.58015549e-01
-9.89698842e-02 1.83927432e-01 -1.76167786e-01 7.83122957e-01
6.53643370e-01 -6.32521093e-01 -5.23575902e-01 -4.53938320e-02
-7.81515688e-02 4.52428460e-01 7.09985554e-01 -1.24330807e+00
5.28104246e-01 -1.96505725e-01 3.51765394e-01 -6.03364229e-01
5.50837994e-01 -7.92866528e-01 3.38535786e-01 2.02938318e-01
-1.98623344e-01 -2.51763433e-01 2.03667581e-01 1.04729116e+00
-9.11779776e-02 -1.76563069e-01 7.35401630e-01 -2.87375659e-01
-4.94509846e-01 1.69210166e-01 -4.93280560e-01 1.00331880e-01
7.74446607e-01 -1.84538290e-02 4.07482274e-02 -4.64412332e-01
-1.13182485e+00 -3.87119874e-02 7.68914521e-02 -3.98051590e-01
4.65218991e-01 -1.08532929e+00 -4.18923438e-01 2.64270246e-01
-4.35458839e-01 -3.03582430e-01 2.97161639e-01 1.13409984e+00
-2.19897851e-02 4.70789313e-01 -2.50387162e-01 -8.19777668e-01
-7.46050596e-01 3.23317081e-01 5.59846964e-03 -1.08356729e-01
-7.50692070e-01 4.31647748e-01 4.04682430e-03 -3.53559464e-01
3.24918889e-02 -5.24931312e-01 -1.34913862e-01 -3.61195505e-01
1.09667584e-01 5.79443574e-01 -2.69935995e-01 -4.60486203e-01
2.29258556e-02 9.41368639e-01 2.79131651e-01 -5.00019908e-01
1.19232798e+00 -3.90916586e-01 -2.17685431e-01 1.06706691e+00
9.10485327e-01 9.49437395e-02 -1.27767932e+00 -5.17506838e-01
-9.09903497e-02 -1.17624864e-01 1.93660602e-01 -6.02387667e-01
-8.90259027e-01 9.14941430e-01 7.28396654e-01 2.86802024e-01
7.96280861e-01 -8.09709430e-02 9.29487169e-01 6.60292387e-01
5.66818595e-01 -7.75547504e-01 -2.84100860e-01 5.42318821e-01
3.06523442e-01 -1.03074539e+00 1.03502057e-01 -4.02492523e-01
3.66189629e-02 9.23923492e-01 2.26184964e-01 -6.92479789e-01
1.24525237e+00 7.35954583e-01 -1.96115494e-01 -3.77166659e-01
-7.27867961e-01 -2.01341897e-01 1.48527846e-02 3.28146905e-01
2.24625662e-01 -1.75906673e-01 -1.79794133e-01 5.65421641e-01
5.96074276e-02 3.21759313e-01 5.05586922e-01 1.25446844e+00
-4.72446591e-01 -8.43498290e-01 -5.68008363e-01 6.65865123e-01
-3.85173589e-01 -7.94512860e-04 -3.21390145e-02 5.89157641e-01
6.22792495e-03 7.94222176e-01 -2.08911836e-01 4.90562730e-02
1.63969159e-01 2.31350377e-01 5.15443861e-01 -5.76406062e-01
-5.02422787e-02 3.65763426e-01 -2.75419086e-01 -4.74471837e-01
-5.67549646e-01 -8.60445440e-01 -1.25515497e+00 -1.80656821e-01
-3.65217984e-01 5.37970066e-01 6.56390309e-01 1.42326653e+00
3.54293972e-01 7.13590682e-01 4.05204058e-01 -7.97309041e-01
-8.14811289e-01 -9.63378072e-01 -9.44681346e-01 1.23965196e-01
4.59638983e-01 -1.11048245e+00 -6.76478505e-01 3.11723113e-01] | [6.624414443969727, 3.583439588546753] |
153181eb-9897-431e-aec8-7ecb88e2365b | tpt-an-empirical-term-selection-for-arabic | null | null | https://aclanthology.org/2021.icnlsp-1.26 | https://aclanthology.org/2021.icnlsp-1.26.pdf | TPT: An Empirical Term Selection for Arabic Text Categorization | null | ['Mohamed Lichouri', 'Mourad Abbas'] | null | null | null | null | icnlsp-2021-11 | ['text-categorization'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.256826877593994, 3.518962860107422] |
6408b0a2-2145-492d-a9e9-28c24002d2c4 | connecting-the-dots-loss-aversion-sybil | 2210.15181 | null | https://arxiv.org/abs/2210.15181v2 | https://arxiv.org/pdf/2210.15181v2.pdf | Connecting the Dots: Loss Aversion, Sybil Attacks, and Welfare Maximization | A celebrated known cognitive bias of individuals is that the pain of losing is psychologically higher than the pleasure of gaining. In robust decision making under uncertainty, this approach is typically associated with the selection of safety (aka security) level strategies. We consider a refined notion, which we term loss aversion, capturing the fact that when comparing two actions an agent should not care about payoffs in situations where they lead to identical payoffs, removing trivial equivalencies. We study the properties of loss aversion, its relations to other robust notions, and illustrate its use in auctions and other settings. Moreover, while loss aversion is a classical cognitive bias on the side of decision makers, the problem of economic design is to maximize social welfare when facing self-motivated participants. In online environments, such as the Web, participants' incentives take a novel form originating from the lack of clear agent identity -- the ability to create Sybil attacks, i.e., the ability of each participant to act using multiple identities. It is well-known that Sybil attacks are a major obstacle for welfare-maximization. Our major result proves that the celebrated VCG mechanism is welfare maximizing when agents are loss-averse, even under Sybil attacks. Altogether, our work shows a successful fundamental synergy between cognitive bias/robustness under uncertainty, economic design, and agents' strategic manipulations in online multi-agent systems. | ['Moshe Tennenholtz', 'Yotam Gafni'] | 2022-10-27 | null | null | null | null | ['decision-making-under-uncertainty', 'decision-making-under-uncertainty'] | ['medical', 'reasoning'] | [-4.18877341e-02 4.86380845e-01 -1.98556140e-01 1.45019636e-01
-1.61485538e-01 -8.90169382e-01 2.08186060e-01 2.10411370e-01
-1.07014000e+00 8.10851634e-01 -7.43999854e-02 -1.03938438e-01
-5.57667911e-01 -9.44799960e-01 -3.99865776e-01 -7.29650617e-01
-3.53035718e-01 1.33357331e-01 -4.03462261e-01 -6.59182668e-01
3.90092462e-01 1.52972072e-01 -1.22393072e+00 -5.50332546e-01
6.71987832e-01 1.12793887e+00 -2.03508794e-01 4.05057579e-01
2.68599093e-01 5.41054428e-01 -7.14420140e-01 -1.01349950e+00
6.80987954e-01 -1.87401906e-01 -3.76297057e-01 5.55357300e-02
-3.77574742e-01 -5.10693192e-01 1.11878663e-01 1.38187778e+00
6.56589985e-01 -2.68471539e-02 4.79916483e-01 -1.89968181e+00
-7.39053428e-01 9.01696086e-01 -5.89743972e-01 -2.39491668e-02
4.62204248e-01 6.68902248e-02 1.17899871e+00 -2.61999723e-02
5.06025076e-01 1.30874169e+00 3.67169440e-01 8.28500450e-01
-1.46302998e+00 -4.85846311e-01 2.34302238e-01 -1.70730338e-01
-1.18074322e+00 -2.08913356e-01 6.73669398e-01 -3.97035509e-01
2.74294227e-01 4.76166099e-01 5.28966129e-01 1.00703609e+00
2.05169797e-01 4.14581358e-01 1.40806568e+00 -3.98833990e-01
7.75219321e-01 3.91558230e-01 1.89253110e-02 4.85683493e-02
1.01381266e+00 5.47077000e-01 -5.44542611e-01 -5.25767028e-01
3.25966269e-01 -1.82615370e-01 -3.13648939e-01 -7.16059625e-01
-8.25878501e-01 1.02761149e+00 1.59863502e-01 -3.25583182e-02
-7.60126472e-01 3.88689749e-02 1.06815360e-01 9.54949141e-01
-5.99631807e-03 7.21443951e-01 -1.47998139e-01 5.33712171e-02
-5.03140949e-02 5.64839065e-01 9.17987645e-01 5.83294153e-01
4.05153841e-01 -9.70323458e-02 -2.88952217e-02 1.26273841e-01
3.93722147e-01 3.62480134e-01 2.53921568e-01 -1.34066021e+00
3.10843527e-01 3.06686372e-01 7.34814465e-01 -9.97092843e-01
-3.75481218e-01 -6.16826177e-01 -5.86479902e-01 9.31747079e-01
7.00821877e-01 -6.88844979e-01 4.22065556e-01 2.52460051e+00
1.88487917e-01 -5.25652528e-01 3.82471621e-01 8.46364200e-01
-1.32952690e-01 2.05696542e-02 9.63733122e-02 -7.12262034e-01
1.41436696e+00 -6.20919699e-03 -7.79935122e-01 -1.37709156e-01
2.88453221e-01 -3.46460879e-01 7.96931326e-01 3.90542537e-01
-1.34107757e+00 4.63542968e-01 -1.06797349e+00 5.63515723e-01
-2.69315332e-01 -8.29605043e-01 6.25395179e-01 1.30827963e+00
-7.74262905e-01 4.45693314e-01 6.89314008e-02 -3.10805112e-01
3.89111221e-01 1.93277389e-01 -3.64811309e-02 4.99304503e-01
-1.26193690e+00 8.98530304e-01 -6.44609183e-02 -2.39782900e-01
-7.19158351e-01 -6.52491629e-01 -3.59206975e-01 3.58507365e-01
1.07948983e+00 -8.65718126e-01 1.14039612e+00 -1.21568429e+00
-1.55539119e+00 9.90916908e-01 5.87221086e-01 -7.43519306e-01
1.22162938e+00 8.89511555e-02 -1.97586883e-02 -1.13910578e-01
2.26113573e-01 1.97538480e-01 8.33846867e-01 -1.28105617e+00
-4.36134875e-01 -5.04334271e-01 7.51809359e-01 5.46559513e-01
-2.49312147e-01 1.63677126e-01 7.91273952e-01 -5.79269469e-01
-3.39362293e-01 -9.70997334e-01 -4.11418676e-01 2.56769229e-02
-3.12314600e-01 -1.47899464e-01 3.22248861e-02 -4.35185991e-02
7.48463273e-01 -2.08130026e+00 -1.46633044e-01 3.22471112e-01
3.59442174e-01 -1.70564651e-01 -8.01133290e-02 2.99320877e-01
8.04374292e-02 5.55659056e-01 7.02252015e-02 1.40415691e-02
9.45108593e-01 -2.67434627e-01 -2.17352614e-01 6.25043511e-01
-2.52610117e-01 7.03661203e-01 -5.34945369e-01 -3.90331708e-02
-3.45233113e-01 -9.74332616e-02 -4.82886642e-01 -2.63242777e-02
-1.31712966e-02 -1.49406701e-01 -4.82181519e-01 3.15603644e-01
8.12323153e-01 2.16963321e-01 2.78705060e-01 5.26898324e-01
-3.17042261e-01 5.04669882e-02 -1.51748860e+00 1.03145385e+00
-7.00017065e-02 2.19170779e-01 7.21713543e-01 -6.62163615e-01
3.47394168e-01 2.46207073e-01 1.78681985e-01 -5.41112661e-01
5.73429108e-01 9.78381559e-02 2.91064143e-01 -1.52003765e-01
3.97065341e-01 -4.25046086e-01 -4.79157686e-01 1.24356985e+00
-4.60987091e-01 9.59734842e-02 9.42484215e-02 4.28776115e-01
6.96972907e-01 -4.26439971e-01 6.22070551e-01 -6.85492933e-01
2.72518426e-01 -3.71562630e-01 9.16332960e-01 1.22309375e+00
-6.95833206e-01 -1.05663374e-01 1.07969058e+00 -9.46662650e-02
-6.31321490e-01 -1.20264506e+00 1.71911314e-01 1.13182199e+00
4.87081707e-01 5.52711040e-02 -8.65287006e-01 -5.53986490e-01
3.63646537e-01 8.56618583e-01 -7.36580193e-01 -3.80048096e-01
1.79181591e-01 -1.06239128e+00 4.02831584e-01 -1.67883083e-01
6.84536457e-01 -7.75093317e-01 -1.21069157e+00 -4.71076481e-02
-2.82864034e-01 -6.21636808e-01 -6.72553718e-01 -1.90927699e-01
-2.68677026e-01 -1.05998993e+00 -5.85029304e-01 -4.07166108e-02
3.03400278e-01 3.37375700e-01 9.67113078e-01 2.09352389e-01
-2.15807572e-01 7.96828091e-01 -4.96338531e-02 -9.66321588e-01
-2.94438630e-01 -6.31863773e-01 6.86747909e-01 1.33283719e-01
4.74790961e-01 -5.09363949e-01 -6.54147923e-01 3.68526250e-01
-8.63511920e-01 -5.17823458e-01 2.96778381e-01 3.18096012e-01
-3.64313945e-02 4.92161691e-01 9.37675238e-01 -3.08656603e-01
1.26513612e+00 -3.80391836e-01 -8.64843547e-01 4.49597150e-01
-6.03732526e-01 -3.48663144e-02 2.03626469e-01 -5.51766813e-01
-1.10190701e+00 -4.41731393e-01 7.06999302e-01 4.29738313e-01
1.42473593e-01 1.90072671e-01 -5.43197274e-01 -3.28569978e-01
5.44845760e-01 -4.59658615e-02 4.68570709e-01 -1.14412904e-01
2.51207381e-01 5.74999094e-01 -5.89327849e-02 -9.32751298e-01
9.06686425e-01 5.11177838e-01 3.26400876e-01 -7.71508932e-01
-3.06511998e-01 3.75481606e-01 2.27693677e-01 -2.55118042e-01
6.77214921e-01 -7.04889357e-01 -1.88482428e+00 6.36325002e-01
-8.92140090e-01 7.15349987e-02 -5.11723220e-01 5.25727093e-01
-8.69540095e-01 5.36308110e-01 -3.23876083e-01 -1.54626071e+00
-3.60508193e-03 -7.13768184e-01 -1.20338470e-01 3.91538501e-01
-3.23168933e-02 -5.88702440e-01 -2.32942745e-01 3.67873222e-01
7.00996041e-01 1.68293148e-01 6.50827229e-01 -8.18564117e-01
-7.79132664e-01 -8.48284885e-02 2.14273095e-01 -2.73129791e-02
-1.40065238e-01 -3.22169065e-01 -6.63792193e-01 -3.41270685e-01
4.90843236e-01 -5.15531063e-01 3.15031171e-01 5.19471407e-01
2.16211468e-01 -7.07414329e-01 1.03848614e-01 -3.82472910e-02
1.40384638e+00 3.45847130e-01 3.51379722e-01 5.49163580e-01
-4.07028705e-01 1.27284825e+00 4.95642006e-01 9.05773222e-01
5.16803920e-01 7.31472790e-01 6.97080612e-01 4.47235733e-01
8.21568131e-01 -9.50902179e-02 5.10843396e-01 -3.11067313e-01
-2.53383964e-01 -1.59083873e-01 -2.05982000e-01 5.20288162e-02
-1.79226351e+00 -1.24475670e+00 5.30520022e-01 2.82955098e+00
5.93118966e-01 2.89905936e-01 7.24596798e-01 8.17414150e-02
1.12006950e+00 -3.33854765e-01 -5.50807953e-01 -6.24431193e-01
-6.01861894e-01 -5.89351952e-01 7.02811658e-01 7.26545632e-01
-6.91414118e-01 4.28989738e-01 5.98391628e+00 6.65681362e-01
-4.24945742e-01 2.72056818e-01 7.50705779e-01 -6.21702909e-01
-3.46663058e-01 -9.53215063e-02 -4.03605938e-01 4.59824532e-01
4.21168655e-01 -1.18961835e+00 7.18672335e-01 5.84856093e-01
3.71466756e-01 -3.23700130e-01 -1.00800371e+00 7.38046408e-01
-1.44282937e-01 -9.16536927e-01 -5.01236677e-01 4.96710986e-01
3.38856429e-01 -7.86923945e-01 3.62319559e-01 1.51395127e-01
8.29556882e-01 -8.62685144e-01 1.12260723e+00 2.83677876e-01
8.69240165e-02 -1.02476120e+00 5.33558369e-01 4.20718312e-01
-6.05827868e-01 -4.58263308e-01 -3.10873747e-01 -5.71200132e-01
1.13525182e-01 2.94042379e-01 -8.49864259e-02 3.65867376e-01
4.49766785e-01 -3.28535408e-01 -1.31965265e-01 9.23859179e-01
-4.24010456e-02 -3.41709852e-01 -4.46922868e-01 -2.82034010e-01
4.04362679e-02 -5.20407736e-01 8.50571752e-01 4.13427860e-01
3.21217149e-01 2.20159382e-01 -1.74741656e-01 1.22852826e+00
-1.24764636e-01 1.01378836e-01 -7.04661727e-01 -1.50497686e-02
5.91742277e-01 1.11285245e+00 -6.93648934e-01 1.43576488e-01
-2.17852846e-01 5.89374006e-01 -1.55636340e-01 2.10120142e-01
-6.94836915e-01 -2.23650753e-01 1.20053911e+00 -7.13793263e-02
-1.93752080e-01 2.26934150e-01 -4.68359739e-01 -1.04038262e+00
1.51989892e-01 -8.89696717e-01 6.34007812e-01 -3.98133963e-01
-1.49806178e+00 2.48034354e-02 -2.06400707e-01 -9.18425322e-01
-5.85723668e-02 -4.23724711e-01 -6.59956872e-01 7.54467726e-01
-1.40603697e+00 -4.83581007e-01 3.44917923e-01 4.97803122e-01
-9.80780125e-02 -1.45810768e-01 5.73147178e-01 -2.32449964e-01
-5.13053000e-01 6.18423462e-01 1.69677287e-03 -2.56047964e-01
3.96956146e-01 -1.19136715e+00 -3.54409993e-01 7.10677028e-01
-2.67002791e-01 5.53080976e-01 1.02810264e+00 -4.55848992e-01
-1.38590097e+00 -2.89716810e-01 7.11275041e-01 -2.73161978e-01
7.91117787e-01 -4.00917530e-01 -1.98364794e-01 2.88242131e-01
2.04990909e-01 -6.03108585e-01 7.92891204e-01 7.68275335e-02
-3.42304468e-01 -1.36123627e-01 -1.66120982e+00 1.20167398e+00
1.11152565e+00 -3.54895324e-01 -6.43717527e-01 6.67942036e-03
7.16014922e-01 5.08394420e-01 -4.29105371e-01 -1.31166682e-01
6.63773835e-01 -1.16234207e+00 1.05698204e+00 -6.91864431e-01
-9.34406668e-02 -8.84389132e-02 -3.44263673e-01 -1.36851037e+00
-2.79333979e-01 -1.28281844e+00 5.11752546e-01 1.13737130e+00
2.38010064e-01 -9.96039629e-01 7.68277526e-01 1.22719717e+00
8.34399343e-01 6.83047920e-02 -1.18743622e+00 -1.15489340e+00
5.32204747e-01 -2.03264445e-01 6.24459982e-01 8.67250919e-01
7.96989977e-01 1.04851745e-01 -4.14705336e-01 1.47778660e-01
1.36646545e+00 -8.15507919e-02 3.65222275e-01 -1.50519383e+00
-3.04763556e-01 -8.05810332e-01 -2.78145939e-01 -3.24117422e-01
2.47001812e-01 -4.59896237e-01 -3.23944390e-01 -5.89489996e-01
1.06274620e-01 -3.55297297e-01 -3.65433723e-01 -6.24183938e-02
2.27220684e-01 -1.01779640e-01 7.66912341e-01 -2.02720165e-02
-5.31478047e-01 3.90004903e-01 8.72832239e-01 -9.31308270e-02
-3.90655130e-01 4.09113228e-01 -1.46427333e+00 7.89243221e-01
8.75897706e-01 -3.47930849e-01 -3.22498322e-01 2.30223358e-01
8.39370787e-01 3.94482672e-01 5.15263617e-01 -4.20084119e-01
3.03063929e-01 -6.85502350e-01 -3.76642853e-01 2.25201935e-01
2.55902201e-01 -1.11021805e+00 1.58978060e-01 9.02872264e-01
-6.66582644e-01 -3.66457142e-02 -4.16776150e-01 6.90530777e-01
5.05751252e-01 -5.70219994e-01 7.26205230e-01 -3.71813059e-01
-1.11133471e-01 -9.76251513e-02 -8.40428174e-01 1.72691017e-01
1.39733863e+00 -1.73441321e-02 -6.93702400e-01 -9.40812409e-01
-6.50719643e-01 3.18495870e-01 6.78937256e-01 2.57750601e-01
3.10138434e-01 -1.07292521e+00 -8.18997741e-01 -1.39587373e-01
5.78086339e-02 -7.52605021e-01 3.48530412e-01 8.01699996e-01
6.43952936e-02 1.66051194e-01 -5.79111636e-01 2.74103522e-01
-1.18411112e+00 8.88131142e-01 5.00245452e-01 6.18261881e-02
-8.27910006e-03 5.17677128e-01 4.81021464e-01 4.86604236e-02
4.02584881e-01 3.05479497e-01 -5.13220578e-02 4.14678544e-01
6.96241498e-01 6.89895213e-01 -3.28404456e-01 -3.41471821e-01
-4.07402873e-01 1.86443418e-01 3.34084816e-02 -6.71946347e-01
9.82258677e-01 -7.32937455e-01 -8.84344280e-02 1.17167167e-01
3.29953432e-01 2.69059241e-01 -1.10923779e+00 -2.07169354e-01
1.33672178e-01 -5.92300773e-01 -2.44735360e-01 -9.23312306e-01
-7.75133967e-01 3.04409802e-01 4.19950873e-01 1.04072559e+00
9.11162376e-01 -3.51516694e-01 -1.75045818e-01 6.33612275e-01
7.90346563e-01 -1.58484459e+00 -2.61903331e-02 1.25033176e-02
8.28384399e-01 -1.11547780e+00 -2.03299850e-01 -4.70788658e-01
-8.79693091e-01 5.91162562e-01 3.69949013e-01 -3.54278535e-02
6.03967249e-01 2.09024981e-01 2.95041967e-03 1.96284860e-01
-8.22345793e-01 -4.59406435e-01 -5.61426878e-01 9.45111036e-01
-2.34981194e-01 4.81532812e-01 -8.11546981e-01 1.19878578e+00
-2.33019546e-01 -1.40975565e-01 1.39149272e+00 9.03258920e-01
-5.54218352e-01 -1.09763288e+00 -7.17084765e-01 6.99931532e-02
-7.94140041e-01 1.67602468e-02 -6.91973388e-01 6.70467556e-01
1.12983445e-03 1.34122956e+00 9.49463546e-02 2.14737579e-01
3.02177191e-01 1.39524806e-02 2.89126188e-01 -1.03724577e-01
-4.99608308e-01 -6.96950480e-02 -1.92827005e-02 -3.73755991e-01
-4.58999246e-01 -7.42716014e-01 -5.34896970e-01 -6.49947822e-01
-8.48660246e-02 2.82219112e-01 4.83171582e-01 6.60925567e-01
7.49046952e-02 -7.97068849e-02 8.48857820e-01 -2.18190297e-01
-1.27353704e+00 -3.36076260e-01 -1.28481233e+00 4.30446446e-01
3.65002602e-01 -7.52316296e-01 -9.31487024e-01 -6.07909083e-01] | [4.322923183441162, 3.0443131923675537] |
13ad3d4f-57f4-4904-b9ca-50f01a4d88b1 | jiuzhang-2-0-a-unified-chinese-pre-trained | 2306.11027 | null | https://arxiv.org/abs/2306.11027v1 | https://arxiv.org/pdf/2306.11027v1.pdf | JiuZhang 2.0: A Unified Chinese Pre-trained Language Model for Multi-task Mathematical Problem Solving | Although pre-trained language models~(PLMs) have recently advanced the research progress in mathematical reasoning, they are not specially designed as a capable multi-task solver, suffering from high cost for multi-task deployment (\eg a model copy for a task) and inferior performance on complex mathematical problems in practical applications. To address these issues, in this paper, we propose \textbf{JiuZhang~2.0}, a unified Chinese PLM specially for multi-task mathematical problem solving. Our idea is to maintain a moderate-sized model and employ the \emph{cross-task knowledge sharing} to improve the model capacity in a multi-task setting. Specially, we construct a Mixture-of-Experts~(MoE) architecture for modeling mathematical text, so as to capture the common mathematical knowledge across tasks. For optimizing the MoE architecture, we design \emph{multi-task continual pre-training} and \emph{multi-task fine-tuning} strategies for multi-task adaptation. These training strategies can effectively decompose the knowledge from the task data and establish the cross-task sharing via expert networks. In order to further improve the general capacity of solving different complex tasks, we leverage large language models~(LLMs) as complementary models to iteratively refine the generated solution by our PLM, via in-context learning. Extensive experiments have demonstrated the effectiveness of our model. | ['Guoping Hu', 'Cong Liu', 'Shijin Wang', 'Jing Sha', 'Ji-Rong Wen', 'Yuanhang Zhou', 'Zhipeng Chen', 'Zheng Gong', 'Beichen Zhang', 'Kun Zhou', 'Wayne Xin Zhao'] | 2023-06-19 | null | null | null | null | ['mathematical-reasoning'] | ['natural-language-processing'] | [ 1.88803189e-02 -3.18168312e-01 1.06144079e-03 -3.95035177e-01
-9.16220307e-01 -5.57158887e-01 1.37290001e-01 -4.15736020e-01
-4.66964662e-01 6.18166864e-01 -1.43216029e-01 -4.40117896e-01
-2.23636493e-01 -5.30680954e-01 -7.72322416e-01 -3.54981214e-01
3.30700845e-01 6.37388945e-01 -6.95544258e-02 -4.00271595e-01
1.57889947e-01 4.71133925e-02 -8.57628763e-01 5.96623421e-01
1.79154038e+00 7.19421387e-01 1.17121017e+00 5.25465608e-01
-7.31154144e-01 9.86169755e-01 -7.71575928e-01 -7.61332214e-01
-6.21491149e-02 -7.87007883e-02 -1.02037358e+00 -1.01322524e-01
1.60286024e-01 -1.42074257e-01 -1.09910972e-01 1.03732812e+00
4.72731590e-01 2.69939750e-01 5.91156363e-01 -1.20411968e+00
-9.58748102e-01 1.04545081e+00 -5.84235013e-01 2.69960523e-01
1.38551459e-01 -7.73214102e-02 8.17574501e-01 -9.60936010e-01
2.19757631e-01 1.48241472e+00 6.29634559e-01 7.55790532e-01
-9.39246535e-01 -1.08954525e+00 5.78233957e-01 2.93459535e-01
-1.45546699e+00 -3.72273177e-01 9.01790619e-01 -2.96183348e-01
9.78102565e-01 1.79039501e-02 2.33938158e-01 1.07652700e+00
4.34538305e-01 1.20776439e+00 1.01627135e+00 -2.81641543e-01
-3.53051186e-01 3.32336992e-01 -4.51734141e-02 9.02325809e-01
-9.65383276e-02 -7.11369872e-01 -6.30574286e-01 4.82399464e-02
1.02864361e+00 -6.98983818e-02 -1.47561312e-01 1.61975160e-01
-1.38197112e+00 5.82865417e-01 9.68662053e-02 6.44206583e-01
-4.84039495e-03 2.50114858e-01 4.34159189e-01 6.33954525e-01
6.81786358e-01 6.21754348e-01 -9.72801983e-01 1.51179522e-01
-6.15461528e-01 4.49859023e-01 6.02543294e-01 1.32905257e+00
6.39755368e-01 2.36207768e-01 -3.33256751e-01 1.19400787e+00
1.94042012e-01 4.59509045e-01 5.09995461e-01 -8.59654903e-01
1.24434125e+00 6.22853160e-01 -1.31624028e-01 -6.53682232e-01
-5.11563957e-01 -6.98897779e-01 -1.05426741e+00 -4.86537755e-01
2.79598266e-01 -4.71544057e-01 -6.12564087e-01 1.99241602e+00
6.35714605e-02 2.68755972e-01 1.84781268e-01 5.67208648e-01
5.54016471e-01 8.41762722e-01 4.93718505e-01 -1.17463917e-01
1.22613001e+00 -1.24306893e+00 -9.34022069e-01 -5.66256464e-01
9.09835696e-01 -8.66766870e-01 1.23598278e+00 6.20296478e-01
-1.27667964e+00 -8.39648306e-01 -6.94861770e-01 -3.78069162e-01
-3.92231882e-01 4.75262731e-01 7.43081391e-01 3.78894806e-01
-8.96624923e-01 3.78500819e-01 -5.19912899e-01 2.11241484e-01
3.28591853e-01 3.92373949e-01 4.43267561e-02 -3.15567970e-01
-1.45595551e+00 1.22709918e+00 7.31547296e-01 2.60019839e-01
-9.97621536e-01 -1.03571129e+00 -8.37674737e-01 1.17236227e-01
6.55520380e-01 -9.73650217e-01 1.22392166e+00 -7.78800130e-01
-1.73079288e+00 5.20461798e-01 -3.52197379e-01 5.14548682e-02
3.67859125e-01 -2.61703640e-01 -4.78126794e-01 -2.68967837e-01
1.50022924e-01 3.71084571e-01 9.91529346e-01 -1.25091183e+00
-6.89905524e-01 -1.25374377e-01 2.27366418e-01 4.17270422e-01
-5.79234004e-01 4.39151853e-01 -5.71105838e-01 -1.16445267e+00
-5.50137684e-02 -6.60778880e-01 -3.28978598e-01 -4.76431459e-01
-1.46207929e-01 -7.58271575e-01 5.27061820e-01 -8.26900661e-01
1.62656367e+00 -1.81247699e+00 7.54323781e-01 -1.72397539e-01
4.36307251e-01 4.47885692e-01 -4.95040029e-01 4.77000547e-04
8.78807977e-02 3.69062304e-01 -7.20969662e-02 -7.96612620e-01
9.31132399e-03 2.49171361e-01 -1.67595401e-01 -1.99168131e-01
3.71257961e-01 1.24141669e+00 -7.48939574e-01 -7.58094609e-01
-7.53426552e-03 4.90151197e-02 -5.47257781e-01 3.23554158e-01
-7.36782312e-01 4.77388710e-01 -9.97746944e-01 4.65461880e-01
6.57982647e-01 -5.55256248e-01 2.73479968e-01 2.44714925e-03
2.02759400e-01 2.15031192e-01 -1.28837967e+00 2.33154106e+00
-1.01603508e+00 1.66837975e-01 3.21356118e-01 -1.24065244e+00
8.21280181e-01 3.69094640e-01 2.09060550e-01 -4.99528259e-01
8.27193037e-02 2.14275897e-01 1.29506528e-01 -7.32968867e-01
3.64175349e-01 -2.48825341e-01 -2.17859492e-01 4.55647081e-01
1.99733749e-01 -4.24154460e-01 1.02004871e-01 1.70711070e-01
7.04694092e-01 1.42361030e-01 -8.82384926e-02 -2.85782069e-01
8.95069242e-01 -1.39207644e-02 5.54242730e-01 5.76174319e-01
1.37499496e-01 1.37627544e-03 1.33989632e-01 -2.06813172e-01
-8.05532157e-01 -6.15390658e-01 1.54723534e-02 1.87556601e+00
-6.51547499e-03 -6.23246312e-01 -6.63980126e-01 -3.89681369e-01
-1.30379885e-01 8.08483422e-01 -1.96578562e-01 -1.44442588e-01
-1.07687163e+00 -1.02151072e+00 6.36250734e-01 4.04267102e-01
6.82498276e-01 -9.66155887e-01 -1.31762981e-01 5.77933133e-01
-4.65322495e-01 -1.27569532e+00 -5.18946528e-01 2.03289255e-01
-7.29521513e-01 -6.14833355e-01 -7.59283304e-01 -1.05598676e+00
4.94728357e-01 2.68230021e-01 1.18701088e+00 4.88528535e-02
1.02817439e-01 1.29353791e-01 -2.55841821e-01 -4.81907666e-01
-3.02477658e-01 3.29658985e-01 9.55860838e-02 -2.95292467e-01
8.12211931e-02 -5.05570889e-01 6.34472966e-02 2.17261344e-01
-8.18265736e-01 5.28169155e-01 7.72277534e-01 7.43115187e-01
1.58408165e-01 3.68016422e-01 8.11584651e-01 -8.84834051e-01
1.08456349e+00 -5.05122900e-01 -3.39305401e-01 9.00648654e-01
-3.69344532e-01 1.52613714e-01 1.12992895e+00 -8.55582714e-01
-1.53589487e+00 -4.60245997e-01 1.99386433e-01 -6.88052118e-01
2.70832688e-01 9.96641159e-01 -2.61864513e-01 -1.84440568e-01
2.43594185e-01 4.08391654e-01 -4.25900787e-01 -7.34475732e-01
3.69358689e-01 4.30774301e-01 2.60416269e-01 -1.40294755e+00
8.49264085e-01 -2.73751050e-01 -2.90708303e-01 -3.07159841e-01
-1.16348910e+00 -1.93909228e-01 -5.48158944e-01 7.01118559e-02
8.57529521e-01 -1.22955739e+00 -8.31356049e-01 7.38336444e-01
-1.69469190e+00 -6.64093792e-01 5.70617020e-01 3.14024895e-01
-3.31520706e-01 3.34410131e-01 -7.39590406e-01 -7.53064692e-01
-3.17160934e-01 -1.26793563e+00 8.11617553e-01 1.66475609e-01
1.37018129e-01 -1.43151534e+00 -2.92581826e-01 8.21194649e-01
6.28342569e-01 -4.22724783e-01 1.54251885e+00 -6.06661618e-01
-6.49966061e-01 3.66996437e-01 -4.75834876e-01 2.45908186e-01
-1.83967710e-01 -3.07269931e-01 -7.24839985e-01 -1.67894274e-01
1.79788142e-01 -7.83226728e-01 7.17551947e-01 3.14054638e-01
1.71645844e+00 -4.96881939e-02 -2.94823766e-01 6.62783027e-01
1.22701156e+00 1.56215400e-01 2.11637080e-01 1.87957287e-01
1.04907656e+00 4.41991031e-01 4.43191409e-01 3.32285941e-01
9.06229377e-01 5.22170246e-01 -1.49613112e-01 8.11502934e-02
4.93617542e-02 -6.31124377e-02 5.37869334e-01 1.35613370e+00
-1.21876992e-01 -1.73239946e-01 -1.22949159e+00 4.06398267e-01
-1.87837577e+00 -3.91085505e-01 -4.26205993e-02 1.60551894e+00
1.27573764e+00 5.14272526e-02 -3.43839616e-01 -5.61432660e-01
4.81704384e-01 2.76476182e-02 -5.36065698e-01 -1.84860811e-01
-3.83554995e-02 3.89901489e-01 5.38439900e-02 4.70445544e-01
-9.58277106e-01 1.54695225e+00 5.91396570e+00 1.50395131e+00
-9.07729566e-01 4.86290663e-01 3.94975483e-01 6.31213188e-02
-4.99219894e-01 -3.58904660e-01 -1.09362459e+00 2.91285902e-01
7.58971810e-01 -3.78085047e-01 7.33430743e-01 7.21815228e-01
8.07093829e-02 1.61978260e-01 -9.66675997e-01 9.93045509e-01
2.11541131e-01 -1.16699195e+00 3.64192843e-01 -3.84836316e-01
8.74969780e-01 -1.50499463e-01 5.81627376e-02 9.14231837e-01
4.72776502e-01 -1.13132465e+00 6.57938302e-01 4.53686416e-01
7.50730932e-01 -4.74795043e-01 4.20978814e-01 9.33522284e-01
-1.47590685e+00 -9.72708911e-02 -5.33517480e-01 -1.31684810e-01
1.05202578e-01 3.58756274e-01 -3.56399804e-01 9.35352683e-01
5.39634526e-01 6.59189105e-01 -5.67667484e-01 4.80596691e-01
1.59661155e-02 1.27044201e-01 -1.08391017e-01 -7.17579350e-02
2.24478766e-01 -2.54009455e-01 2.38428637e-01 1.18991721e+00
2.29174584e-01 3.49432290e-01 6.87775075e-01 1.25967395e+00
-9.22802240e-02 9.70552713e-02 -2.23142177e-01 -1.19660944e-01
4.77298349e-01 1.18447983e+00 -1.65834919e-01 -5.23912072e-01
-5.54975033e-01 1.01140094e+00 9.21450734e-01 5.76205552e-01
-1.02146602e+00 -2.60054141e-01 4.90094811e-01 -4.52006996e-01
-3.69437933e-02 -6.06710553e-01 -5.63663721e-01 -1.64971411e+00
1.10589847e-01 -1.09469736e+00 2.50520945e-01 -6.93501890e-01
-1.53704762e+00 4.61467922e-01 2.61218756e-01 -6.58465147e-01
3.82810012e-02 -8.47958028e-01 -6.69561505e-01 1.41056287e+00
-1.85563803e+00 -1.48393238e+00 -4.63676043e-02 9.52311695e-01
1.05298316e+00 -3.64389926e-01 6.92515314e-01 4.62387770e-01
-1.07491362e+00 6.71073675e-01 -3.31391506e-02 1.18954238e-02
7.74363697e-01 -1.13787711e+00 2.00230196e-01 6.43122733e-01
-2.29876384e-01 9.28474069e-01 2.84949392e-01 -7.05556929e-01
-1.69355249e+00 -1.08567393e+00 1.03053272e+00 -6.55702174e-01
9.63312745e-01 -5.77033222e-01 -1.23501825e+00 9.57841158e-01
4.63611148e-02 -5.56436300e-01 6.42126441e-01 3.84287745e-01
-2.76706100e-01 -3.94452363e-02 -7.31848776e-01 6.15214944e-01
1.05412996e+00 -5.92446387e-01 -8.95709693e-01 6.01555228e-01
1.12243092e+00 -6.20560110e-01 -1.18447161e+00 3.79293859e-01
1.08447626e-01 -2.78697819e-01 1.06308591e+00 -7.70336092e-01
7.27914155e-01 -5.80851957e-02 7.86176473e-02 -1.40605223e+00
-4.30042922e-01 -6.78665936e-01 -7.11908415e-02 1.18347514e+00
6.03930056e-01 -3.98888260e-01 4.52498794e-01 8.01698744e-01
-5.11954010e-01 -6.85126364e-01 -8.15418780e-01 -6.26336515e-01
6.32091284e-01 -7.64674187e-01 4.39623147e-01 1.28300142e+00
5.78340665e-02 6.32610619e-01 -5.49436331e-01 1.69942677e-01
4.20465082e-01 5.82720852e-03 7.42349982e-01 -1.06076050e+00
-4.17729706e-01 -5.69909453e-01 7.90588439e-01 -1.43323147e+00
7.84858763e-01 -1.08977199e+00 -1.43233046e-01 -1.31816983e+00
3.39066148e-01 -9.98539448e-01 -4.09515381e-01 7.11592674e-01
-7.35409081e-01 -4.73699719e-01 3.64559293e-01 1.57416821e-01
-9.22115922e-01 6.07281208e-01 1.80728507e+00 -8.98986906e-02
-1.73893958e-01 -7.59836435e-02 -7.67354369e-01 6.90678716e-01
5.70265234e-01 -4.14117306e-01 -5.15466869e-01 -1.28146434e+00
4.17265743e-01 3.79696101e-01 1.19849905e-01 -6.80814922e-01
4.82903600e-01 -5.93163431e-01 4.25547242e-01 -3.80319506e-01
3.16996455e-01 -8.19768786e-01 -5.29151931e-02 1.88322291e-01
-3.85940880e-01 1.99765906e-01 5.01767457e-01 3.09358805e-01
-5.60005987e-03 -4.97143179e-01 4.39304739e-01 -6.50944829e-01
-7.91709423e-01 3.15132171e-01 -2.60861754e-01 1.35570422e-01
7.49852777e-01 3.41229707e-01 -4.77394909e-01 4.56016436e-02
-5.22345603e-01 9.16078389e-01 -1.75678372e-01 4.89554971e-01
5.12809873e-01 -1.14429092e+00 -8.55853498e-01 -4.97114658e-02
-1.25032216e-01 5.13665020e-01 5.21017671e-01 7.21405625e-01
-1.04496352e-01 7.11906135e-01 -1.39071107e-01 -2.09390357e-01
-8.71602118e-01 5.52198052e-01 3.10849607e-01 -9.48626518e-01
-2.39719793e-01 1.13432252e+00 3.43792886e-01 -8.23461771e-01
2.78128147e-01 -5.33509374e-01 -2.03261524e-01 -1.39400661e-01
3.95449609e-01 2.16903999e-01 -1.95281237e-01 5.55172265e-02
-1.27934724e-01 6.61578596e-01 -4.00835693e-01 2.99475412e-03
1.28224921e+00 -2.01879919e-01 -5.13437450e-01 5.19790709e-01
8.55442584e-01 -2.77767628e-01 -9.19680834e-01 -6.28412187e-01
5.39375506e-02 -1.00673050e-01 6.07532673e-02 -1.08527505e+00
-8.36828411e-01 8.79346490e-01 -2.20898449e-01 -2.15125397e-01
1.12129247e+00 -2.00176150e-01 8.26710284e-01 7.94651031e-01
4.03116286e-01 -1.02210116e+00 4.92233872e-01 9.83115435e-01
8.62563848e-01 -1.16923249e+00 -7.06078336e-02 -6.61130011e-01
-9.45186138e-01 1.15635216e+00 1.04678392e+00 1.48520723e-01
4.72170144e-01 2.36630738e-01 -1.59482330e-01 -1.42532423e-01
-1.00134027e+00 1.02436608e-02 4.89858896e-01 3.37872118e-01
5.35740793e-01 9.65275615e-03 -1.97625924e-02 1.18345082e+00
-1.31537035e-01 3.22589099e-01 2.81511899e-03 7.57690310e-01
-4.27867800e-01 -1.25589609e+00 -2.64279932e-01 3.47645104e-01
-2.45416626e-01 -5.01087010e-01 -3.21367718e-02 5.14882445e-01
4.47040856e-01 8.79433811e-01 -2.89811254e-01 -2.85290122e-01
1.87359631e-01 3.75627935e-01 7.09270656e-01 -7.94187307e-01
-9.78912354e-01 1.03273250e-01 7.39321038e-02 -4.67173941e-02
-3.25944513e-01 -1.93497598e-01 -1.11901772e+00 -4.56242710e-01
-3.98567647e-01 7.68302009e-02 4.38650012e-01 1.40236783e+00
1.99234605e-01 9.69761789e-01 3.42235655e-01 -5.77918649e-01
-7.22689450e-01 -1.13871968e+00 -3.19202989e-01 1.28698274e-01
-1.09058328e-01 -5.94635785e-01 5.22129573e-02 1.23706445e-01] | [10.636126518249512, 8.242810249328613] |
897891e3-34c0-4319-8d1c-de1a4d439dc9 | distributional-variational-autoencoder-to | 2302.11294 | null | https://arxiv.org/abs/2302.11294v2 | https://arxiv.org/pdf/2302.11294v2.pdf | Distributional Learning of Variational AutoEncoder: Application to Synthetic Data Generation | The Gaussianity assumption has been consistently criticized as a main limitation of the Variational Autoencoder (VAE), despite its efficiency in computational modeling. In this paper, we propose a new approach that expands the model capacity (i.e., expressive power of distributional family) without sacrificing the computational advantages of the VAE framework. Our VAE model's decoder is composed of an infinite mixture of asymmetric Laplacian distribution, which possesses general distribution fitting capabilities for continuous variables. Our model is represented by a special form of a nonparametric M-estimator for estimating general quantile functions, and we theoretically establish the relevance between the proposed model and quantile estimation. We apply the proposed model to synthetic data generation, and particularly, our model demonstrates superiority in easily adjusting the level of data privacy. | ['Jong-June Jeon', 'SeungHwan An'] | 2023-02-22 | null | null | null | null | ['synthetic-data-generation', 'synthetic-data-generation'] | ['medical', 'miscellaneous'] | [-3.71101618e-01 2.93609947e-01 -1.78340271e-01 -3.05683702e-01
-6.86581910e-01 -5.52474976e-01 2.97856539e-01 -3.31040770e-01
-3.47030163e-01 9.39417064e-01 3.85800116e-02 -4.39017534e-01
-1.36495516e-01 -9.04198408e-01 -8.32998335e-01 -8.45620573e-01
1.72598884e-01 1.33118302e-01 -3.83555144e-01 5.44730127e-02
-2.04120085e-01 3.69224638e-01 -1.29244280e+00 -3.28973800e-01
1.25587642e+00 1.07281172e+00 -3.80709954e-02 2.97075659e-01
7.62099726e-03 5.15806913e-01 -5.60499489e-01 -7.41031647e-01
8.01823810e-02 -2.64241338e-01 -1.58431858e-01 -1.00596398e-01
4.90608104e-02 -4.91378546e-01 -3.94947290e-01 1.22859550e+00
5.88912189e-01 1.21803045e-01 1.11544776e+00 -1.72590494e+00
-1.07219577e+00 6.45124137e-01 -3.57296288e-01 -7.69456029e-02
-3.50789189e-01 -3.49219084e-01 1.01359141e+00 -8.48276973e-01
2.00171292e-01 1.14550245e+00 7.74705231e-01 6.47264004e-01
-1.33939004e+00 -8.32100749e-01 -1.59959644e-01 -1.93732575e-01
-1.86512613e+00 -4.71274048e-01 7.99511909e-01 -4.95726764e-01
4.17463869e-01 8.94569457e-02 3.84252787e-01 1.43597639e+00
1.77912027e-01 9.41685140e-01 6.85371935e-01 -8.84699523e-02
6.62259281e-01 6.35236442e-01 -5.85561730e-02 4.53846246e-01
4.50719208e-01 3.43173184e-02 -2.35338911e-01 -6.09063089e-01
1.10464978e+00 -1.97222471e-01 -2.97338367e-01 -5.28039455e-01
-6.18288457e-01 1.19631398e+00 1.52583392e-02 -2.24033371e-02
-3.92616898e-01 4.41466898e-01 2.22071931e-01 2.05997303e-01
6.94722354e-01 -4.20657033e-03 -2.91362822e-01 -1.63556948e-01
-1.00156868e+00 2.19889268e-01 7.74617255e-01 1.29639482e+00
5.00691533e-01 5.02348006e-01 -3.39489192e-01 6.46493196e-01
4.97746259e-01 6.32431388e-01 4.40335244e-01 -9.87552643e-01
2.47798607e-01 7.16687217e-02 3.08190674e-01 -9.37953174e-01
-8.85644257e-02 -6.32871151e-01 -1.14009905e+00 -3.91327664e-02
3.42103690e-01 -6.23338223e-01 -2.72061557e-01 2.21409798e+00
2.14028463e-01 8.81561339e-02 2.27387100e-01 5.68787992e-01
4.94092792e-01 6.02211416e-01 1.37033641e-01 -2.31760800e-01
9.08032298e-01 -4.71691579e-01 -1.03025901e+00 4.68720287e-01
1.78460360e-01 -1.73849627e-01 1.05992723e+00 2.97711939e-01
-1.03499615e+00 -4.71201003e-01 -8.86907637e-01 -1.06734835e-01
-2.32564151e-01 2.48738289e-01 6.63076460e-01 1.11158454e+00
-1.07605493e+00 5.09063125e-01 -7.80851185e-01 -1.47670597e-01
5.66292048e-01 2.90088087e-01 -1.73382819e-01 6.66442156e-01
-1.27936769e+00 3.85490209e-01 6.01226330e-01 -2.65135262e-02
-8.28342199e-01 -7.43642151e-01 -9.91083384e-01 6.49789751e-01
1.96995854e-01 -8.65315199e-01 1.05102944e+00 -8.43578398e-01
-1.84905040e+00 2.77433455e-01 -1.21423997e-01 -5.78601599e-01
9.67894435e-01 -1.02674954e-01 -3.43645841e-01 -4.97085191e-02
-1.88757241e-01 4.31841046e-01 1.10490143e+00 -1.14622140e+00
-2.97415167e-01 -2.27749377e-01 -1.47247151e-01 -2.42113173e-01
-7.76267588e-01 -3.53169233e-01 -2.62521803e-01 -7.59301126e-01
-3.10933948e-01 -5.69382310e-01 -1.04429819e-01 1.77860439e-01
-4.75416660e-01 -2.36612573e-01 5.78495324e-01 -5.02258897e-01
1.59162903e+00 -2.37696671e+00 -9.60665196e-02 4.62885022e-01
1.35386318e-01 -1.32835269e-01 8.52394104e-02 5.16698360e-01
9.96420011e-02 3.26071054e-01 -3.48506540e-01 -6.83365464e-01
5.75177491e-01 2.38039747e-01 -5.86923182e-01 6.75651908e-01
7.40244538e-02 8.72178376e-01 -5.82121193e-01 -6.38004661e-01
-3.27582285e-02 6.58686042e-01 -8.59708011e-01 3.01683038e-01
-7.42570385e-02 3.24757367e-01 -5.79568148e-01 4.88506496e-01
1.00316036e+00 -3.87042642e-01 9.68925133e-02 -2.53688619e-02
-4.11145836e-02 -4.68462765e-01 -1.06799030e+00 1.35391581e+00
-2.68369466e-01 4.28565025e-01 1.68752521e-01 -9.48056877e-01
9.49246943e-01 4.26936656e-01 3.63199979e-01 -5.91443069e-02
3.69632572e-01 -3.16357170e-03 -2.43873075e-01 -4.57935542e-01
4.44528729e-01 -1.21025749e-01 -4.05762941e-02 1.69440046e-01
9.33353901e-02 2.43411034e-01 -1.51707307e-01 -6.18138835e-02
2.19016626e-01 1.16368182e-01 3.78677189e-01 -5.84804177e-01
5.03550768e-01 -6.10705793e-01 5.61572969e-01 8.68903577e-01
-3.08334321e-01 5.05110562e-01 8.38751614e-01 1.53142989e-01
-1.27139771e+00 -1.25958574e+00 -5.40264666e-01 8.98460746e-01
-1.98198818e-02 -1.76731721e-01 -9.35536146e-01 -4.42944705e-01
3.59194189e-01 1.28215444e+00 -8.84777606e-01 -1.86044276e-01
6.06738999e-02 -8.00766587e-01 1.08184505e+00 7.22651720e-01
5.66214979e-01 -4.78116006e-01 -4.23913836e-01 -7.82511458e-02
-1.11418799e-01 -7.76888072e-01 -5.32962143e-01 -6.19713515e-02
-7.07044005e-01 -7.36414135e-01 -8.59050632e-01 -4.07891423e-01
3.86388958e-01 -3.46230716e-01 8.13505828e-01 -6.07502222e-01
4.05335337e-01 5.46347439e-01 -2.80349236e-02 -5.83135068e-01
-5.24814010e-01 7.74009004e-02 2.13620007e-01 5.11675060e-01
1.28674686e-01 -8.17120731e-01 -5.74750900e-01 1.02587000e-01
-1.13200521e+00 -3.05515736e-01 3.66224855e-01 1.03373718e+00
6.30814195e-01 4.58213165e-02 1.03333151e+00 -9.21169817e-01
1.11095393e+00 -8.97543550e-01 -7.23664880e-01 2.25124076e-01
-8.23577583e-01 1.23987116e-01 8.56599152e-01 -4.99428570e-01
-1.24033749e+00 -2.80120581e-01 -2.94299781e-01 -7.59878635e-01
1.17892176e-01 6.07303262e-01 -4.70774144e-01 1.29384980e-01
2.40143061e-01 5.44462025e-01 2.74153113e-01 -5.60764790e-01
5.10629117e-01 8.52424502e-01 5.63551247e-01 -8.25820446e-01
5.66620529e-01 5.10958552e-01 3.91196534e-02 -8.83985579e-01
-3.77626717e-01 3.28241810e-02 -2.60718971e-01 8.19728747e-02
7.43957162e-01 -9.88950551e-01 -1.27360320e+00 3.50066215e-01
-1.03680491e+00 -4.57662903e-02 -3.74030024e-01 5.57912648e-01
-8.39795589e-01 5.28516352e-01 -4.49909508e-01 -1.30434215e+00
-4.02203292e-01 -9.27240312e-01 8.37267578e-01 9.98372808e-02
2.18425050e-01 -1.24082899e+00 4.66730632e-03 -1.01191752e-01
6.22153878e-01 3.81137282e-01 1.00146902e+00 -8.27653587e-01
-1.67474449e-01 -1.02271341e-01 -2.14967445e-01 7.28435278e-01
-9.37141776e-02 9.34669077e-02 -1.15525138e+00 -5.53230166e-01
1.27159864e-01 -1.67680636e-01 7.87361264e-01 5.87301612e-01
1.64394736e+00 -4.33637857e-01 -8.20247158e-02 8.58625233e-01
1.44452167e+00 7.33988583e-02 5.06075084e-01 -2.61819735e-03
3.37523520e-01 1.88752323e-01 7.00017959e-02 9.67621982e-01
3.21023673e-01 2.49084264e-01 5.07631600e-01 1.37043342e-01
5.10239422e-01 -6.63662195e-01 4.42765296e-01 8.89238179e-01
2.00518686e-02 -5.64131320e-01 -3.85586649e-01 4.65529710e-01
-1.87942505e+00 -9.57390547e-01 1.75400764e-01 2.26158428e+00
8.13194215e-01 -3.57504994e-01 1.46727324e-01 -3.42791319e-01
7.40071416e-01 3.17458324e-02 -7.83424735e-01 -3.50192487e-01
-2.23492131e-01 4.49608453e-02 5.61242163e-01 2.79533088e-01
-1.11606848e+00 6.20159388e-01 7.29021883e+00 1.25642192e+00
-7.68643677e-01 1.84748173e-01 5.18029273e-01 1.05384439e-01
-6.73040807e-01 -2.94319063e-01 -6.02182627e-01 8.48163545e-01
7.79910028e-01 -6.93644464e-01 3.85481685e-01 1.17226708e+00
1.37706831e-01 4.28246468e-01 -1.15359867e+00 1.01318896e+00
-9.04103890e-02 -9.03182328e-01 1.35723829e-01 2.99470782e-01
5.45415640e-01 -5.01430333e-01 6.63138747e-01 5.10990441e-01
3.88070755e-02 -1.19291341e+00 8.29455614e-01 6.44241691e-01
1.01722443e+00 -1.18403006e+00 7.02366352e-01 5.37378967e-01
-7.28043854e-01 -1.97188646e-01 -8.00257742e-01 1.03564017e-01
-4.09277044e-02 4.72562492e-01 -4.27866340e-01 4.99922156e-01
4.06300277e-01 3.65066946e-01 -2.19583333e-01 7.34855890e-01
6.11372180e-02 8.15317988e-01 -4.09667194e-01 -8.03733524e-03
9.53901783e-02 -6.73716486e-01 5.77778697e-01 1.13917482e+00
8.10548246e-01 -2.23550066e-01 -1.63824856e-01 1.36084044e+00
-2.33416170e-01 2.36856088e-01 -6.10308886e-01 -1.46081150e-01
7.27657974e-01 8.57313216e-01 6.20026281e-03 -1.96445182e-01
-4.85726804e-01 7.30555594e-01 3.94705147e-01 7.38570511e-01
-1.13314438e+00 -2.94755280e-01 7.29727387e-01 -2.79092163e-01
6.77581072e-01 -8.83514211e-02 -3.07741046e-01 -1.39972305e+00
-1.39028341e-01 -5.47614157e-01 4.13309634e-01 -4.20730442e-01
-1.65951347e+00 3.80788535e-01 6.00359440e-02 -1.17402577e+00
-3.67385149e-01 -5.76671124e-01 -5.54104626e-01 8.13875377e-01
-1.29148304e+00 -1.10744572e+00 2.97412444e-02 8.84510756e-01
-1.28454819e-01 -3.38112295e-01 8.51869762e-01 4.49036270e-01
-6.98579550e-01 1.37781870e+00 9.03943419e-01 1.59638062e-01
4.32977557e-01 -1.39158607e+00 5.57401776e-02 5.61908960e-01
-1.74310684e-01 8.43201637e-01 8.41056764e-01 -4.12363112e-01
-1.28105915e+00 -1.20360100e+00 4.40016538e-01 -2.39414677e-01
7.16637373e-01 -5.49580991e-01 -8.05113971e-01 9.90758717e-01
1.67353094e-01 -3.33836734e-01 9.59803462e-01 4.83550355e-02
-3.17668766e-01 3.69564816e-02 -1.44187117e+00 6.05299890e-01
6.68079615e-01 -5.32796681e-01 -3.24589074e-01 -3.42631410e-03
8.77781987e-01 -1.33296743e-01 -1.11604726e+00 2.48874784e-01
7.55273283e-01 -9.35067534e-01 8.25204730e-01 -7.03489661e-01
3.74592036e-01 -4.29307669e-02 -4.34915692e-01 -1.14657652e+00
-2.86090374e-01 -7.49248862e-01 -4.38364565e-01 1.48483765e+00
2.29573801e-01 -9.05016065e-01 6.43394709e-01 7.34251678e-01
1.79377571e-01 -7.49563515e-01 -1.21925211e+00 -8.62863600e-01
5.93977511e-01 -5.76045394e-01 8.25907409e-01 8.23344171e-01
-1.73116103e-01 -8.34697410e-02 -8.55340838e-01 3.17200005e-01
8.23098898e-01 1.56013332e-02 7.49222457e-01 -1.06660104e+00
-5.58086157e-01 -4.78332669e-01 -1.28787309e-01 -1.23358297e+00
4.47383463e-01 -8.79136562e-01 -3.42045516e-01 -7.80350864e-01
2.84045011e-01 -7.53226280e-02 -5.09010971e-01 -3.86741315e-03
5.09098470e-02 -2.04264700e-01 7.22372681e-02 5.58759049e-02
-1.75372615e-01 1.28772724e+00 1.07880497e+00 -6.80084201e-03
-1.67845845e-01 2.84811616e-01 -7.50916839e-01 6.75709784e-01
6.80570662e-01 -3.12924296e-01 -7.97218382e-01 -1.77032948e-01
1.02221869e-01 2.78637260e-01 4.79471356e-01 -6.80351138e-01
2.16351733e-01 -7.74208680e-02 4.37414587e-01 -5.31600714e-01
2.58795947e-01 -9.66545641e-01 2.52272457e-01 1.82986595e-02
-5.33374727e-01 -1.24861978e-01 2.01350138e-01 9.60437894e-01
-2.14240357e-01 -2.32094198e-01 6.19834661e-01 3.29789400e-01
-2.93783665e-01 5.69859803e-01 -3.32495451e-01 1.21034108e-01
7.70806968e-01 -9.03230067e-03 -1.12386718e-01 -7.19842315e-01
-7.09242940e-01 2.36000180e-01 4.02883470e-01 -2.23277248e-02
5.00165403e-01 -1.47008860e+00 -7.51490831e-01 3.20404708e-01
-2.91553419e-02 -2.45323852e-01 4.20995951e-01 7.35309243e-01
-1.62245199e-01 2.26812348e-01 -4.95799892e-02 -2.53719896e-01
-5.78089178e-01 9.33639228e-01 4.08921927e-01 -1.64789229e-03
-5.20186543e-01 5.88520944e-01 6.27072096e-01 -2.49522582e-01
3.88722748e-01 -2.15976119e-01 -1.40381366e-01 -2.38425992e-02
2.05115497e-01 6.10871494e-01 -4.95762080e-01 -4.02146459e-01
-1.53609946e-01 1.28807351e-01 2.92766362e-01 -2.32275218e-01
1.12691450e+00 -3.44261557e-01 -7.41560757e-02 5.02901554e-01
1.32252896e+00 1.71079502e-01 -1.47225976e+00 -3.00081354e-03
-3.18204939e-01 -2.55915314e-01 -3.09871826e-02 -4.47306931e-01
-1.06625438e+00 9.31083918e-01 4.76709902e-01 2.13195845e-01
9.57970798e-01 -9.91984010e-02 7.34384835e-01 3.12330306e-01
3.06124032e-01 -1.03734386e+00 -4.94001716e-01 1.89584002e-01
7.64292181e-01 -1.01841462e+00 -2.73800015e-01 -1.00599647e-01
-7.93291271e-01 1.05223632e+00 3.36687684e-01 -5.54125197e-02
9.62603509e-01 2.42702454e-01 -1.85518131e-01 7.22138584e-02
-5.73387027e-01 5.90430982e-02 2.22470701e-01 7.30538666e-01
2.40607470e-01 1.90838635e-01 -4.08349127e-01 1.43894053e+00
-3.81800443e-01 4.19391990e-02 5.93257904e-01 2.63709217e-01
-1.81239560e-01 -6.11270845e-01 -6.99610412e-02 1.65909931e-01
-5.88202119e-01 -1.28034011e-01 1.26013324e-01 1.00363088e+00
-2.69270297e-02 8.97505105e-01 2.00433940e-01 -1.74277291e-01
1.16749652e-01 2.11331367e-01 1.06548548e-01 1.16701514e-01
-1.26667231e-01 2.17036217e-01 -3.87078136e-01 -2.69618392e-01
-1.37897983e-01 -5.18188655e-01 -7.21968532e-01 -4.12092686e-01
-2.82368779e-01 4.36810702e-01 4.71894085e-01 6.19579971e-01
5.05530775e-01 2.89953500e-01 6.87096119e-01 -1.58213198e-01
-1.42029524e+00 -9.23490345e-01 -1.34735882e+00 3.59959573e-01
3.01133841e-01 -7.32491851e-01 -4.75572735e-01 -1.41147226e-01] | [7.243322849273682, 3.9516007900238037] |
95276fa4-4457-4250-b9a3-64e1dbc74b6e | online-video-instance-segmentation-via-robust | 2207.05580 | null | https://arxiv.org/abs/2207.05580v1 | https://arxiv.org/pdf/2207.05580v1.pdf | Online Video Instance Segmentation via Robust Context Fusion | Video instance segmentation (VIS) aims at classifying, segmenting and tracking object instances in video sequences. Recent transformer-based neural networks have demonstrated their powerful capability of modeling spatio-temporal correlations for the VIS task. Relying on video- or clip-level input, they suffer from high latency and computational cost. We propose a robust context fusion network to tackle VIS in an online fashion, which predicts instance segmentation frame-by-frame with a few preceding frames. To acquire the precise and temporal-consistent prediction for each frame efficiently, the key idea is to fuse effective and compact context from reference frames into the target frame. Considering the different effects of reference and target frames on the target prediction, we first summarize contextual features through importance-aware compression. A transformer encoder is adopted to fuse the compressed context. Then, we leverage an order-preserving instance embedding to convey the identity-aware information and correspond the identities to predicted instance masks. We demonstrate that our robust fusion network achieves the best performance among existing online VIS methods and is even better than previously published clip-level methods on the Youtube-VIS 2019 and 2021 benchmarks. In addition, visual objects often have acoustic signatures that are naturally synchronized with them in audio-bearing video recordings. By leveraging the flexibility of our context fusion network on multi-modal data, we further investigate the influence of audios on the video-dense prediction task, which has never been discussed in existing works. We build up an Audio-Visual Instance Segmentation dataset, and demonstrate that acoustic signals in the wild scenarios could benefit the VIS task. | ['Yan Lu', 'Bhiksha Raj', 'Xiaohao Xu', 'Jinglu Wang', 'Xiang Li'] | 2022-07-12 | null | null | null | null | ['video-instance-segmentation'] | ['computer-vision'] | [ 5.53048134e-01 -3.23401570e-01 -3.17962378e-01 -3.27001065e-01
-9.78963077e-01 -4.03653413e-01 5.55771828e-01 -1.10848583e-01
-2.47560844e-01 4.01386976e-01 3.51813912e-01 2.10552499e-01
-3.79573330e-02 -4.49343413e-01 -1.09387589e+00 -8.13443363e-01
-2.30655059e-01 -9.21584889e-02 4.33474630e-01 9.13856402e-02
2.02600192e-02 2.32578710e-01 -2.00281930e+00 7.26204276e-01
3.11975271e-01 1.59519017e+00 5.22188008e-01 6.89923048e-01
3.48178670e-02 9.49974537e-01 -4.12223190e-01 -2.93227673e-01
4.62258846e-01 -3.21083844e-01 -4.93151397e-01 2.73265362e-01
8.72292221e-01 -3.43222320e-01 -7.64393866e-01 8.59842718e-01
3.08561891e-01 2.16433749e-01 2.75161117e-01 -1.26901555e+00
-3.31771314e-01 7.41419077e-01 -5.59564948e-01 3.34989578e-01
4.18496519e-01 1.57721624e-01 1.20149267e+00 -1.04571176e+00
8.22305441e-01 9.71606851e-01 5.89968979e-01 3.69425803e-01
-1.01163793e+00 -6.86203301e-01 5.26314676e-01 8.45053971e-01
-1.41865849e+00 -5.04701972e-01 1.05680060e+00 -4.80373859e-01
6.65433943e-01 3.81880254e-01 9.19880569e-01 1.35890102e+00
-4.93155383e-02 1.08667672e+00 6.78733885e-01 -3.53979021e-02
8.27502385e-02 -3.24739695e-01 -8.54650959e-02 3.32904071e-01
-4.17365491e-01 2.33159855e-01 -1.01122820e+00 2.56718904e-01
4.05427098e-01 2.89332241e-01 -6.11636400e-01 -2.36086756e-01
-1.51130295e+00 3.87020737e-01 3.25921327e-01 2.17418134e-01
-5.31548321e-01 3.31181288e-01 5.74669003e-01 8.78678784e-02
3.71204853e-01 -7.08621666e-02 -5.32589614e-01 -1.83976844e-01
-1.50643480e+00 1.34085342e-01 4.72114950e-01 1.10970688e+00
5.68282306e-01 1.40201688e-01 -6.56667471e-01 6.84078157e-01
9.24366117e-02 4.03004229e-01 2.81649470e-01 -1.14817178e+00
5.97517133e-01 1.59476563e-01 -4.65294458e-02 -1.03690135e+00
1.21788436e-03 -4.96548325e-01 -7.28053808e-01 -2.82835603e-01
2.35386670e-01 2.05504626e-01 -7.45245814e-01 1.82465971e+00
4.34798032e-01 1.14907551e+00 -2.45452881e-01 1.11439168e+00
7.51841605e-01 8.11676204e-01 1.01052202e-01 -4.18858200e-01
1.43414342e+00 -9.19855654e-01 -7.11521447e-01 4.32807058e-02
2.50469297e-01 -6.16426885e-01 8.39715302e-01 4.85026002e-01
-1.11024857e+00 -7.50726223e-01 -1.05987108e+00 -2.84236111e-02
-5.53914942e-02 1.60401866e-01 1.26911551e-01 2.70026535e-01
-8.87954891e-01 6.32695913e-01 -7.73060918e-01 1.52043179e-01
6.48674130e-01 3.95737775e-02 -3.16197962e-01 -2.38264829e-01
-1.23541546e+00 1.43121108e-01 3.22995007e-01 1.69788584e-01
-1.26047206e+00 -9.62383032e-01 -9.45022881e-01 8.24088678e-02
8.82592499e-01 -4.50071126e-01 9.82931972e-01 -1.09923899e+00
-1.15974653e+00 3.93342286e-01 -5.31603515e-01 -8.41949582e-01
5.27512312e-01 -3.21491063e-01 -3.90138060e-01 5.37963390e-01
4.92642494e-03 6.99177504e-01 1.26699853e+00 -1.09833479e+00
-1.00880027e+00 -1.01474993e-01 2.26031244e-01 1.42806008e-01
-3.62876892e-01 -1.18301995e-01 -1.02964449e+00 -9.80390012e-01
-9.71082821e-02 -8.63132834e-01 1.79458082e-01 7.22994357e-02
-2.87958413e-01 -5.94077213e-03 1.18391979e+00 -7.96595454e-01
1.39788151e+00 -2.41225886e+00 3.22450280e-01 -1.61830097e-01
1.91388726e-01 3.46119881e-01 -2.66323864e-01 1.99617535e-01
3.36050205e-02 -1.81708366e-01 -2.31237099e-01 -5.27969182e-01
6.12365864e-02 2.16841981e-01 -5.59123099e-01 3.39227647e-01
2.33619884e-01 9.14153576e-01 -9.04650509e-01 -6.23554289e-01
4.16767657e-01 8.59064639e-01 -7.94042706e-01 1.82499856e-01
-3.60557497e-01 5.73598564e-01 -9.50490683e-02 7.81511247e-01
4.06467885e-01 -1.86941817e-01 2.05168054e-01 -7.65595615e-01
3.75487395e-02 1.11137316e-01 -1.09846163e+00 1.90713191e+00
-3.95510525e-01 1.00154459e+00 1.30069897e-01 -1.28570139e+00
4.97174025e-01 4.20470566e-01 8.61673295e-01 -8.18116963e-01
-1.03321001e-01 -1.18757728e-02 -4.61705267e-01 -5.03880084e-01
7.44404674e-01 1.83529273e-01 7.86701888e-02 -9.60197970e-02
1.34057552e-01 3.58175635e-01 2.66134828e-01 1.31133452e-01
1.02222800e+00 3.91831785e-01 -6.98084012e-02 2.02783108e-01
6.86541855e-01 -3.90033811e-01 8.06788743e-01 6.43555522e-01
-2.45735675e-01 8.94060850e-01 3.35706294e-01 -3.70044827e-01
-8.64822567e-01 -1.00830770e+00 -4.30075079e-02 1.23098004e+00
3.36976171e-01 -5.81385791e-01 -6.83250248e-01 -5.91115952e-01
-1.85103282e-01 4.26894963e-01 -5.55880487e-01 7.20346048e-02
-7.25638926e-01 -2.05326930e-01 2.79316902e-01 5.77602029e-01
3.03069979e-01 -7.87867367e-01 -5.25642514e-01 5.39321363e-01
-5.34802854e-01 -1.68180215e+00 -7.01492131e-01 -1.08266480e-01
-4.96392548e-01 -1.15108085e+00 -8.05983901e-01 -6.27613544e-01
1.34363562e-01 3.60336691e-01 1.06569374e+00 -3.42981424e-03
-2.67733514e-01 7.18577802e-01 -5.34187019e-01 -1.48517871e-02
2.19912156e-02 -2.89778829e-01 -1.12818673e-01 5.44365525e-01
1.27403080e-01 -6.77308619e-01 -9.05262351e-01 4.10244823e-01
-9.95552540e-01 2.54867703e-01 2.52606839e-01 8.71573925e-01
1.02532423e+00 -5.96096441e-02 4.27223980e-01 -5.28864980e-01
-1.32885009e-01 -5.10150790e-01 -5.07464111e-01 2.49215767e-01
-7.27104694e-02 -3.61758143e-01 5.21756113e-01 -5.40339530e-01
-6.88782930e-01 2.74988085e-01 3.61532415e-03 -1.15792596e+00
-7.73404390e-02 3.62072170e-01 -1.89001963e-01 9.63202864e-02
1.11075498e-01 5.19680262e-01 -3.77049237e-01 -3.46175343e-01
3.91400844e-01 4.56620991e-01 8.15359950e-01 -5.63971519e-01
6.07077956e-01 6.43287361e-01 -4.32745256e-02 -8.03047001e-01
-8.80599022e-01 -4.74965304e-01 -4.21764433e-01 -4.73647982e-01
1.05332565e+00 -1.32035208e+00 -6.30358934e-01 2.84485579e-01
-1.07138216e+00 -3.24747592e-01 -4.12650287e-01 3.98458749e-01
-6.96381927e-01 5.65675080e-01 -6.66476548e-01 -6.63579404e-01
-9.15198997e-02 -1.34703243e+00 1.34599125e+00 -7.71197379e-02
1.20322116e-01 -6.23426557e-01 -3.13272834e-01 4.34040785e-01
1.37769267e-01 3.31803322e-01 4.29683834e-01 -4.68354285e-01
-1.03136206e+00 3.42979166e-03 -1.28668562e-01 3.92439187e-01
-7.04266354e-02 1.19870804e-01 -1.11499953e+00 -1.79530546e-01
-3.33785601e-02 -1.37576964e-02 1.15475261e+00 5.89670002e-01
1.57059145e+00 -4.51887310e-01 -2.80864060e-01 7.62889504e-01
1.38704574e+00 1.95760638e-01 5.18451273e-01 -9.15063396e-02
1.01905668e+00 5.19825697e-01 9.91201460e-01 5.66622913e-01
4.27864224e-01 9.87048388e-01 6.13293231e-01 3.38022560e-01
-3.30925107e-01 -1.86422303e-01 4.90058422e-01 8.84417951e-01
-1.08044140e-01 -3.91502053e-01 -6.33753240e-01 8.05061460e-01
-1.93392515e+00 -1.32404149e+00 2.27468848e-01 2.03816581e+00
7.05516636e-01 -1.99368596e-02 1.59139022e-01 3.09491932e-01
9.13863957e-01 5.06615818e-01 -3.76656175e-01 2.52916217e-01
-1.45785853e-01 7.19548538e-02 2.38523170e-01 2.64184088e-01
-1.24854672e+00 6.24001324e-01 5.16955185e+00 1.15508962e+00
-1.31001818e+00 3.66290748e-01 6.35502219e-01 -7.14859128e-01
-1.92256302e-01 -1.25477389e-01 -6.77245080e-01 8.26819897e-01
8.67986381e-01 6.36723712e-02 4.51801836e-01 6.89554691e-01
3.17670882e-01 1.31902903e-01 -1.33611608e+00 1.20553124e+00
1.23947121e-01 -1.65399826e+00 3.31936553e-02 -6.08466081e-02
7.26451814e-01 3.48057970e-02 1.67138338e-01 2.86168247e-01
-4.28608477e-01 -7.33093381e-01 1.21834338e+00 6.81267798e-01
7.86778629e-01 -6.83285475e-01 4.36788738e-01 4.63761576e-02
-1.77559984e+00 -2.84915388e-01 -1.25332162e-01 2.16224760e-01
5.39440155e-01 6.56712234e-01 -4.04853493e-01 7.71124363e-01
8.38938296e-01 1.31737697e+00 -3.20843995e-01 9.86219585e-01
9.24416706e-02 7.97184885e-01 -3.37659895e-01 5.07250249e-01
2.53448159e-01 -1.54193444e-02 7.28573501e-01 1.33245516e+00
4.68525261e-01 3.44354734e-02 2.53416926e-01 6.77255154e-01
-6.58103377e-02 -6.83365688e-02 -2.60244459e-01 -2.58762408e-02
5.07542133e-01 1.05466449e+00 -5.27441263e-01 -5.12479067e-01
-7.05732822e-01 8.56518030e-01 -1.28820896e-01 4.03835684e-01
-1.27059555e+00 1.29217491e-01 9.08621669e-01 2.78942920e-02
9.45599139e-01 -1.81233093e-01 3.84093039e-02 -1.19952261e+00
3.30001980e-01 -9.29776609e-01 2.46862575e-01 -6.80505216e-01
-1.02695525e+00 4.79625881e-01 -2.12738421e-02 -1.62870908e+00
-2.23659232e-01 -3.74271423e-01 -3.82409543e-01 2.87300885e-01
-1.64216483e+00 -1.25652397e+00 -3.40388179e-01 8.36443067e-01
9.57979977e-01 -6.71286508e-02 2.44763285e-01 7.96571493e-01
-5.57163596e-01 7.02376425e-01 -5.14137298e-02 1.28128529e-01
6.48199081e-01 -8.62371147e-01 8.37820023e-02 1.00382102e+00
6.35615826e-01 2.48394549e-01 6.15690589e-01 -6.17372513e-01
-1.54477119e+00 -1.35073423e+00 4.96015102e-01 -3.34193617e-01
6.93727255e-01 -4.14082736e-01 -9.56893563e-01 5.85418582e-01
2.10839193e-02 4.85980541e-01 4.26431596e-01 -3.59007865e-01
-4.03180540e-01 -2.82466769e-01 -6.30958200e-01 5.29719472e-01
1.32867682e+00 -9.52185631e-01 -3.52601916e-01 1.45182401e-01
1.07837951e+00 -5.68082571e-01 -9.39242184e-01 5.67958593e-01
7.08817184e-01 -1.12091076e+00 1.27622116e+00 -2.65111268e-01
7.57764399e-01 -4.52089876e-01 -5.38742661e-01 -6.88054144e-01
-4.50716130e-02 -6.10069752e-01 -6.21390402e-01 1.42845941e+00
-2.34454107e-02 -5.40787913e-02 7.49423325e-01 3.21972728e-01
-2.49985605e-01 -9.31401312e-01 -1.17646980e+00 -6.71099424e-01
-5.24914086e-01 -1.05443740e+00 6.76659763e-01 8.23893487e-01
-3.62893581e-01 -2.10990906e-01 -8.16729307e-01 2.72607148e-01
6.52311683e-01 2.44848609e-01 5.36925912e-01 -8.33778203e-01
-5.33774137e-01 -2.58824259e-01 -7.02691317e-01 -1.27107477e+00
2.43531048e-01 -7.09450722e-01 5.10219075e-02 -1.01703417e+00
1.33890258e-02 -1.59278557e-01 -6.88923478e-01 2.91502446e-01
-1.44492656e-01 5.97870171e-01 6.25310898e-01 2.52607077e-01
-9.65562165e-01 6.02777719e-01 1.03052628e+00 -3.58300954e-01
3.52746695e-02 -1.74306959e-01 -2.18371809e-01 5.26174784e-01
3.38635564e-01 -3.55397195e-01 -4.92154896e-01 -3.28683525e-01
5.36516458e-02 4.02921617e-01 7.00718999e-01 -1.20745671e+00
3.64346236e-01 -8.41625556e-02 1.76713556e-01 -9.01068568e-01
7.06495523e-01 -1.15166938e+00 4.39640969e-01 1.15516469e-01
-4.24343050e-01 -2.64943898e-01 7.14051202e-02 1.07637894e+00
-5.39386928e-01 1.15655079e-01 4.79448587e-01 2.24962592e-01
-1.07451308e+00 7.72603452e-01 -1.85849726e-01 5.82267009e-02
1.03465831e+00 -2.93990999e-01 -8.52013752e-03 -3.64222854e-01
-7.52156973e-01 8.05690289e-02 3.16652477e-01 6.48392618e-01
6.14575565e-01 -1.46150064e+00 -5.38425982e-01 1.31956831e-01
1.57251388e-01 -5.21402843e-02 7.10522592e-01 1.10952210e+00
-1.07083648e-01 1.62619829e-01 -2.69118659e-02 -1.08217931e+00
-1.35501349e+00 7.10589290e-01 8.58326107e-02 5.58267124e-02
-8.23324442e-01 7.97321618e-01 4.78060663e-01 4.19939786e-01
4.81228441e-01 -4.61333305e-01 -2.43424311e-01 3.39798808e-01
8.01622987e-01 2.04948783e-01 -1.12125129e-01 -9.57361937e-01
-2.03828722e-01 6.00286186e-01 6.34003878e-02 3.49307172e-02
1.33733618e+00 -3.70846123e-01 3.72882664e-01 5.45520425e-01
1.36039102e+00 -2.07567401e-02 -1.89473677e+00 -4.76767242e-01
-2.61420816e-01 -6.01906598e-01 6.66026166e-03 -3.45832199e-01
-1.50338209e+00 9.51537192e-01 5.53028643e-01 1.12202920e-01
1.37638354e+00 -1.26121789e-01 1.13256001e+00 -3.22048888e-02
3.05425614e-01 -1.08249176e+00 1.90673813e-01 3.04745406e-01
8.06781411e-01 -1.16731012e+00 -1.30054250e-01 -4.46510136e-01
-5.56946516e-01 8.98215771e-01 2.20716238e-01 1.67360867e-03
5.96396863e-01 2.42735460e-01 -2.38624483e-01 6.32591844e-02
-8.47390115e-01 -2.07187653e-01 5.76570272e-01 6.05533242e-01
8.37854445e-02 -6.68970644e-02 8.09043124e-02 7.56393313e-01
2.26760302e-02 5.02918251e-02 2.30824441e-01 5.54847479e-01
-1.72873974e-01 -8.26778769e-01 -2.93300807e-01 3.59487593e-01
-6.31284177e-01 -1.69595152e-01 8.96875784e-02 4.72523749e-01
2.79329330e-01 8.48127365e-01 3.47668082e-02 -6.91414118e-01
9.91171598e-02 8.79902020e-02 3.02878171e-01 -2.77541041e-01
-5.81386745e-01 2.83904701e-01 -1.07762568e-01 -1.15574586e+00
-7.67337203e-01 -7.59373307e-01 -1.03509581e+00 -6.10422306e-02
-9.17559043e-02 -2.34622672e-01 5.23705900e-01 1.03110099e+00
4.91003245e-01 7.10202456e-01 5.45050025e-01 -1.42963076e+00
-1.31026775e-01 -4.24573451e-01 -4.52101558e-01 6.27435565e-01
6.49866045e-01 -6.64549172e-01 -4.27663594e-01 3.88241142e-01] | [9.410284996032715, 0.23920592665672302] |
967ea30e-13ff-48a9-b052-8ce3678dedc4 | moving-object-detection-for-event-based-2 | 2109.14979 | null | https://arxiv.org/abs/2109.14979v3 | https://arxiv.org/pdf/2109.14979v3.pdf | Moving Object Detection for Event-based vision using Graph Spectral Clustering | Moving object detection has been a central topic of discussion in computer vision for its wide range of applications like in self-driving cars, video surveillance, security, and enforcement. Neuromorphic Vision Sensors (NVS) are bio-inspired sensors that mimic the working of the human eye. Unlike conventional frame-based cameras, these sensors capture a stream of asynchronous 'events' that pose multiple advantages over the former, like high dynamic range, low latency, low power consumption, and reduced motion blur. However, these advantages come at a high cost, as the event camera data typically contains more noise and has low resolution. Moreover, as event-based cameras can only capture the relative changes in brightness of a scene, event data do not contain usual visual information (like texture and color) as available in video data from normal cameras. So, moving object detection in event-based cameras becomes an extremely challenging task. In this paper, we present an unsupervised Graph Spectral Clustering technique for Moving Object Detection in Event-based data (GSCEventMOD). We additionally show how the optimum number of moving objects can be automatically determined. Experimental comparisons on publicly available datasets show that the proposed GSCEventMOD algorithm outperforms a number of state-of-the-art techniques by a maximum margin of 30%. | ['Ananda S. Chowdhury', 'Thierry Bouwmans', 'Jhony H. Giraldo', 'Shashant R', 'Anindya Mondal'] | 2021-09-30 | moving-object-detection-for-event-based-1 | https://www.researchgate.net/publication/354462913_Moving_Object_Detection_for_Event-based_Vision_using_Graph_Spectral_Clustering | https://www.researchgate.net/publication/354462913_Moving_Object_Detection_for_Event-based_Vision_using_Graph_Spectral_Clustering | international-conference-on-computer-vision-5 | ['moving-object-detection', 'event-based-vision'] | ['computer-vision', 'computer-vision'] | [ 5.35901070e-01 -5.65835893e-01 6.08955957e-02 -2.09751025e-01
-1.54111773e-01 -4.26042885e-01 4.96715426e-01 1.87772691e-01
-7.03640640e-01 5.63990891e-01 -3.89751732e-01 1.36575133e-01
-2.58155726e-02 -5.00635207e-01 -5.42639852e-01 -9.65763509e-01
5.93956895e-02 -2.06453726e-01 1.11013150e+00 2.82534093e-01
2.69942343e-01 3.98189723e-01 -1.86066139e+00 8.42272416e-02
4.50787306e-01 1.19271004e+00 4.98763621e-01 5.87680459e-01
7.71689266e-02 6.63707972e-01 -6.39227033e-01 4.43062425e-04
8.33260119e-02 -5.30939341e-01 -1.76553684e-03 4.68557999e-02
3.35459471e-01 -2.13029549e-01 -5.65567493e-01 1.29951930e+00
4.47086811e-01 1.74947292e-01 3.01746219e-01 -1.44186866e+00
-1.40732005e-01 4.63155061e-02 -6.60494387e-01 8.74777675e-01
2.16259703e-01 1.66087613e-01 3.30045730e-01 -5.76599598e-01
6.94370866e-01 8.49749267e-01 2.69252777e-01 7.52014697e-01
-1.06187594e+00 -5.34454465e-01 9.48204026e-02 6.84527278e-01
-1.26189744e+00 -5.28251708e-01 8.66522014e-01 -3.33123177e-01
8.77940834e-01 2.33649626e-01 8.37870121e-01 1.12076795e+00
5.42810857e-01 6.51022315e-01 9.08623695e-01 -7.64633790e-02
7.90954590e-01 -1.35421664e-01 1.63532749e-01 4.20837790e-01
4.59586859e-01 -1.30844727e-01 -7.34468877e-01 3.91665613e-03
6.32177353e-01 4.00533915e-01 -5.42808175e-01 -2.22757727e-01
-1.34709644e+00 3.04554790e-01 9.86499190e-02 1.38977811e-01
-3.80614519e-01 5.11874318e-01 2.75752962e-01 2.87664589e-03
-1.77290663e-01 -3.17970932e-01 3.65445204e-02 -3.09036374e-01
-7.63458014e-01 -6.45989105e-02 4.59055573e-01 8.26872587e-01
4.24679786e-01 2.31156528e-01 1.49201779e-02 2.55722255e-01
3.47148210e-01 6.65953636e-01 8.52205455e-01 -8.40601504e-01
4.84545194e-02 6.94072604e-01 -2.34805327e-02 -9.65708017e-01
-4.39342022e-01 6.45586774e-02 -9.28876221e-01 3.01925838e-01
3.60214412e-01 6.48512086e-03 -7.55054593e-01 1.51532638e+00
3.47958475e-01 5.27302444e-01 -4.31245379e-02 1.14839792e+00
8.16262662e-01 6.42545164e-01 1.43548865e-02 -5.91124713e-01
1.47422409e+00 -2.67899662e-01 -9.04268146e-01 -4.67535168e-01
-1.12162463e-01 -4.32578444e-01 4.56354529e-01 3.20912540e-01
-6.65830553e-01 -4.56230700e-01 -1.24350011e+00 2.94728190e-01
-3.88551325e-01 -5.29441610e-02 6.28036499e-01 7.24812686e-01
-9.60573971e-01 2.20362440e-01 -1.16986549e+00 -6.34820819e-01
4.23500836e-01 4.03537065e-01 -1.98283494e-01 1.60862476e-01
-7.90672600e-01 6.28439248e-01 2.43112326e-01 9.81620164e-04
-8.53425205e-01 -2.86061227e-01 -6.05795264e-01 -1.69549257e-01
3.08345765e-01 -3.80374819e-01 1.05639446e+00 -8.77367914e-01
-1.39567482e+00 7.80796289e-01 -3.79519582e-01 -7.10750043e-01
3.47798675e-01 2.89376080e-01 -6.24216259e-01 5.45024812e-01
-2.12431684e-01 5.91673374e-01 9.58800554e-01 -5.56284487e-01
-6.69030190e-01 -4.92725670e-01 -1.34752780e-01 -5.25743421e-03
-3.14127833e-01 1.90753758e-01 -5.59238017e-01 -2.76451021e-01
5.31896092e-02 -7.97852159e-01 -9.23586488e-02 3.30169380e-01
-3.97468060e-02 -1.14151940e-01 1.38581669e+00 8.66802707e-02
9.47282672e-01 -2.37939024e+00 -1.38321415e-01 -2.19503030e-01
2.87701637e-01 4.29762602e-01 1.30814925e-01 1.52717322e-01
2.04103470e-01 -4.13359493e-01 -1.42511159e-01 1.72718465e-01
-5.12632728e-01 1.95947988e-03 -1.06112555e-01 7.21767247e-01
1.84856862e-01 7.06798911e-01 -9.12327886e-01 -4.65953708e-01
6.56503320e-01 5.29260218e-01 -2.71721873e-02 -2.85028189e-01
-1.41753629e-01 4.71302569e-01 -3.40249419e-01 6.16762340e-01
4.57920820e-01 -3.08167785e-01 -1.60021763e-02 -3.18587333e-01
-2.60036826e-01 -2.01895431e-01 -1.28675354e+00 1.43680048e+00
2.31534794e-01 1.19851124e+00 -1.19082578e-01 -9.85686064e-01
7.89681196e-01 2.39128873e-01 5.94457805e-01 -8.55150700e-01
4.26577836e-01 -1.72876060e-01 1.17415614e-01 -6.06865764e-01
2.87062556e-01 1.91672802e-01 2.41305575e-01 1.92899466e-01
-6.53426275e-02 3.34738910e-01 2.89293379e-01 1.02321126e-01
1.41256428e+00 -1.31115213e-01 1.41306162e-01 -1.01919509e-01
4.10526931e-01 -1.21362939e-01 8.39610338e-01 5.92985749e-01
-7.01608837e-01 4.42404926e-01 1.01013832e-01 -3.40794057e-01
-5.60636103e-01 -1.06650424e+00 -1.55413017e-01 4.71175194e-01
7.23027647e-01 -2.22630590e-01 -8.33260059e-01 -4.94822077e-02
-3.48779470e-01 2.77848512e-01 -2.87852347e-01 -2.53896922e-01
-3.13803792e-01 -1.04566061e+00 3.03562433e-01 4.74652350e-01
6.25779152e-01 -8.79121780e-01 -1.73111176e+00 3.33407968e-01
2.87337638e-02 -1.67049515e+00 -3.38568628e-01 9.57110897e-02
-8.74934435e-01 -1.15333354e+00 -4.66556162e-01 -7.26583481e-01
7.14450181e-01 7.08725870e-01 4.84865278e-01 -3.34853083e-01
-7.66265512e-01 5.44111848e-01 -2.16150224e-01 -7.18701184e-01
1.22361988e-01 -6.53858960e-01 1.99376732e-01 5.24686575e-01
6.57129645e-01 -5.72805107e-01 -9.84661400e-01 2.94148833e-01
-1.05319047e+00 2.53257155e-02 4.91673946e-01 3.31076413e-01
7.17844367e-01 3.05133909e-01 5.59092164e-01 -4.94528562e-01
2.81443238e-01 -2.73959368e-01 -9.73780811e-01 1.09399594e-01
-2.37437144e-01 -4.14024830e-01 6.11141384e-01 -8.25193822e-01
-8.64107549e-01 4.07037705e-01 5.94594061e-01 -6.14328980e-01
-2.50266254e-01 1.01348748e-02 -6.55418932e-02 -2.18738616e-01
6.16692543e-01 4.98747736e-01 -1.01091317e-03 7.47193843e-02
-9.14519355e-02 7.25052476e-01 8.54202390e-01 2.43091822e-01
3.14555794e-01 9.82179940e-01 1.35991737e-01 -1.24090171e+00
-2.04894021e-02 -7.07114458e-01 -1.89527228e-01 -7.49584734e-01
8.78714442e-01 -8.62765431e-01 -9.33404386e-01 8.49641204e-01
-1.18268001e+00 3.80120729e-03 -1.17702885e-02 6.50322199e-01
-3.43640327e-01 2.35852689e-01 -3.85235816e-01 -1.10285306e+00
-3.43359083e-01 -1.06199706e+00 8.68719816e-01 8.77423346e-01
-2.06461791e-02 -7.18000889e-01 -2.36443996e-01 1.47236645e-01
3.53403300e-01 5.57470143e-01 4.66934144e-01 -2.80706316e-01
-9.12996054e-01 -3.54178429e-01 -8.41702074e-02 -9.74084884e-02
2.00134516e-01 1.27627077e-02 -1.04839981e+00 -7.22037703e-02
1.94480017e-01 2.63152897e-01 7.09291101e-01 6.65221334e-01
9.39926744e-01 1.99571520e-01 -5.80991328e-01 4.08182114e-01
1.65971017e+00 7.22186387e-01 7.39829600e-01 1.83115620e-02
5.27820170e-01 3.14233214e-01 3.73601407e-01 4.44263637e-01
1.35224029e-01 4.52330410e-01 5.62166631e-01 2.67436504e-01
-1.16585016e-01 2.63785183e-01 5.73981643e-01 5.11728883e-01
3.08161048e-04 -3.54510307e-01 -7.46176541e-01 6.33543015e-01
-2.10558105e+00 -1.11176467e+00 -5.43741047e-01 2.49442029e+00
3.58475268e-01 2.20958158e-01 2.10410416e-01 2.64900863e-01
1.13575697e+00 -1.73577726e-01 -9.95581269e-01 -6.28587650e-03
-2.75561690e-01 -3.49230282e-02 6.18942916e-01 -2.83631831e-01
-9.54823554e-01 5.37355423e-01 5.51252890e+00 4.03895527e-01
-1.35425365e+00 -9.28424858e-03 2.93793768e-01 -4.65415239e-01
2.54031599e-01 -1.04351632e-01 -7.98473597e-01 9.94670331e-01
1.10816312e+00 -2.82241166e-01 3.47975910e-01 5.32762706e-01
5.07663667e-01 -6.68099999e-01 -1.16665375e+00 1.56739318e+00
1.55329332e-01 -1.24560928e+00 -2.71307796e-01 8.05165805e-03
4.60734725e-01 1.56689540e-01 -1.24189980e-01 -4.07875746e-01
-2.85861373e-01 -4.91814047e-01 6.99805796e-01 4.83108699e-01
6.26969218e-01 -6.00781262e-01 5.14136493e-01 3.01707566e-01
-1.08237064e+00 -1.12550341e-01 -4.35993075e-01 -1.11175306e-01
2.67196536e-01 7.96065688e-01 -3.10857564e-01 1.60007372e-01
9.75913882e-01 6.82249725e-01 -3.04214120e-01 1.40544617e+00
2.80109525e-01 4.77176547e-01 -4.28639680e-01 -4.45775419e-01
-9.31283236e-02 -9.55776423e-02 8.07760596e-01 8.99203360e-01
3.05724382e-01 3.65394711e-01 -1.41981915e-01 6.99998260e-01
9.32059810e-02 -4.10337150e-01 -7.48162806e-01 -1.20404109e-01
5.63958764e-01 1.30631590e+00 -1.30965197e+00 -1.67926848e-01
-6.72311544e-01 1.06295109e+00 -2.67414153e-01 2.81268805e-01
-8.20324302e-01 -6.06308222e-01 5.27923644e-01 3.25423591e-02
4.37305957e-01 -2.13669717e-01 1.62947163e-01 -9.71794963e-01
1.47639036e-01 -3.67949158e-01 2.06341505e-01 -7.26301312e-01
-8.29557955e-01 3.22103471e-01 -1.97802797e-01 -1.51182938e+00
4.31598304e-03 -6.51013136e-01 -6.73548102e-01 1.99750409e-01
-1.20675302e+00 -5.81246674e-01 -6.70451343e-01 7.39131987e-01
5.97300708e-01 -3.03470343e-03 3.93213004e-01 2.61565268e-01
-7.72185504e-01 1.13426983e-01 1.92098066e-01 1.92964911e-01
4.90418583e-01 -8.60789776e-01 1.39379740e-01 1.15990102e+00
1.04147501e-01 2.57360131e-01 6.24078929e-01 -5.49480736e-01
-2.07183695e+00 -1.09731042e+00 2.65270114e-01 -1.70008734e-01
5.44669151e-01 -3.37233156e-01 -7.94384122e-01 2.60655105e-01
4.27128151e-02 3.43240768e-01 5.16926825e-01 -8.09277952e-01
1.06346309e-01 -5.54518700e-01 -1.07301009e+00 5.61018944e-01
9.32272494e-01 -2.69694626e-01 -4.03696984e-01 1.24160722e-01
2.75466383e-01 -2.52269536e-01 -2.94528067e-01 2.48718783e-01
5.60826957e-01 -1.03553379e+00 7.57560611e-01 2.63131838e-02
-1.99137002e-01 -8.18440795e-01 -1.05613284e-02 -7.89565086e-01
-7.93088898e-02 -7.77246892e-01 -2.69004643e-01 1.26263535e+00
-1.82234049e-01 -7.58876979e-01 7.19916463e-01 4.93741184e-01
1.63461745e-01 -3.39975893e-01 -1.22146058e+00 -8.21874678e-01
-8.92476261e-01 -3.68972957e-01 1.79659333e-02 6.40857518e-01
1.00539677e-01 3.15098256e-01 5.88623527e-03 2.41829067e-01
8.92383039e-01 1.25203133e-01 2.80561239e-01 -1.36094403e+00
-3.10671441e-02 -2.81166345e-01 -1.33322287e+00 -7.12209284e-01
-2.99093515e-01 -5.14937758e-01 1.67054564e-01 -1.33817470e+00
4.02677238e-01 2.11586848e-01 -4.25112575e-01 1.27390727e-01
-7.61147588e-02 4.12663400e-01 -9.95026436e-03 1.33311808e-01
-9.47404385e-01 2.44727284e-01 7.26398706e-01 -1.11709356e-01
-1.50995195e-01 -2.69203663e-01 -2.22143382e-01 7.78818130e-01
6.59728169e-01 -3.69791865e-01 -5.66232622e-01 -4.24698621e-01
-2.98824217e-02 9.58234593e-02 6.48957610e-01 -1.40366614e+00
1.00208473e+00 -8.49491581e-02 5.01295805e-01 -5.14602363e-01
3.50982130e-01 -8.85381699e-01 4.14403707e-01 4.89770204e-01
1.91826791e-01 6.61180839e-02 2.57204086e-01 1.12570274e+00
-1.20544508e-01 -1.12044752e-01 8.71792495e-01 -3.32157314e-02
-1.21304989e+00 1.84596106e-01 -9.72700775e-01 -1.02531701e-01
1.45754468e+00 -7.12082744e-01 -4.97437119e-01 -1.87830806e-01
-1.31299645e-01 9.04540718e-02 4.54593509e-01 4.81097639e-01
8.27924669e-01 -1.03766048e+00 -2.22290322e-01 1.73295975e-01
1.25281960e-01 3.75771523e-02 3.17062467e-01 1.01811779e+00
-3.55278999e-01 2.43795320e-01 -3.22578728e-01 -1.05979848e+00
-1.31132710e+00 5.04409611e-01 1.31772697e-01 6.87860250e-01
-8.32653403e-01 5.23889899e-01 1.17758587e-01 6.04295850e-01
2.46795446e-01 -3.44516158e-01 -9.85735506e-02 -4.95507978e-02
9.01178777e-01 4.69267875e-01 1.20846473e-01 -5.06518006e-01
-7.06424713e-01 6.16351068e-01 5.55111356e-02 -5.00519089e-02
1.14418542e+00 -1.88922405e-01 4.65425737e-02 6.68261349e-01
7.85367072e-01 -3.83202553e-01 -1.27708733e+00 -7.80560970e-02
6.76284656e-02 -2.69938469e-01 2.77279049e-01 -2.68195570e-01
-1.09485805e+00 7.58836448e-01 1.00943208e+00 2.45597154e-01
1.49122107e+00 -7.14679658e-02 7.80640304e-01 3.13274086e-01
5.45624197e-01 -1.23080730e+00 7.25089163e-02 1.63093756e-03
2.80712783e-01 -1.11917055e+00 -1.92366853e-01 -3.52324098e-01
-3.99121672e-01 1.06965268e+00 6.36688769e-01 -1.40187547e-01
4.31382060e-01 4.83180881e-01 -1.89439170e-02 -1.36475682e-01
-9.28301394e-01 -3.24875861e-01 -1.02153644e-01 7.42962182e-01
-7.75478482e-02 -1.57380581e-01 -2.18505785e-01 4.06519443e-01
3.09955150e-01 1.87054142e-01 6.86803401e-01 1.03228319e+00
-5.66346288e-01 -5.24212420e-01 -4.44389552e-01 5.74002206e-01
-5.52941561e-01 2.98066586e-01 -3.38365942e-01 3.39169323e-01
1.00979604e-01 1.22257173e+00 4.59315240e-01 -3.25497270e-01
1.69246465e-01 -2.12677140e-02 5.43948829e-01 -1.84205741e-01
2.10274179e-02 1.14666216e-01 -3.36576253e-01 -6.31074786e-01
-8.20464492e-01 -6.96601391e-01 -1.55328763e+00 7.78050497e-02
-2.54645586e-01 -3.69660795e-01 9.23917592e-01 8.44710290e-01
5.60126126e-01 5.38579166e-01 3.60685438e-01 -6.36997283e-01
1.92921087e-02 -4.67504770e-01 -4.95911598e-01 3.91998380e-01
3.94790471e-01 -5.65245509e-01 -3.03895682e-01 4.49726373e-01] | [8.608266830444336, -1.207703709602356] |
dbf1ce25-36a8-4a7e-aaf9-0cb5920d3792 | visual-speech-recognition-in-a-driver | null | null | https://eurasip.org/Proceedings/Eusipco/Eusipco2022/pdfs/0001131.pdf | https://eurasip.org/Proceedings/Eusipco/Eusipco2022/pdfs/0001131.pdf | Visual Speech Recognition in a Driver Assistance System | Visual speech recognition or automated lipreading is a field of growing attention. Video data proved its usefulness in multimodal speech recognition, especially when acoustic data is heavily noised or even inaccessible. In this paper, we present a novel method for visual speech recognition. We benchmark it on the famous LRW lip-reading dataset by outperforming the existing approaches. After a comprehensive evaluation, we adapt the developed method and test it on the collected RUSAVIC corpus we recorded in-the-wild for vehicle driver. The results obtained demonstrate not only the high performance of the proposed method, but also the fundamental possibility of recognizing speech only by using video modality, even in such difficult natural conditions as driving. | ['Alexey Karpov', 'Alexandr Axyonov', 'Alexey Kashevnik', 'Dmitry Ryumin', 'Denis Ivanko'] | 2022-08-29 | null | null | null | 30th-european-signal-processing-conference | ['lipreading'] | ['computer-vision'] | [ 4.01644170e-01 5.31684011e-02 -2.11146489e-01 -6.66758418e-02
-8.25962305e-01 -4.43374842e-01 1.08098698e+00 -3.18871707e-01
-7.85273135e-01 8.70303750e-01 1.45756125e-01 -5.33435464e-01
1.29080012e-01 3.40852095e-03 -4.40792233e-01 -7.86192358e-01
2.92547852e-01 2.68635005e-01 2.45445460e-01 -1.18853338e-01
4.23718959e-01 7.22660899e-01 -2.64162135e+00 1.25888899e-01
4.94190544e-01 1.01484895e+00 5.18673301e-01 8.48085880e-01
-2.67869771e-01 7.46099114e-01 -6.06338620e-01 -3.87762010e-01
-3.71301770e-02 -1.17277093e-01 -5.89192569e-01 2.98942029e-01
6.08596087e-01 -2.36849502e-01 -3.45443070e-01 9.18797910e-01
8.32733512e-01 1.60733476e-01 5.98555744e-01 -1.33909929e+00
-2.21679553e-01 4.95556593e-02 -5.42134419e-02 2.27161139e-01
5.68811595e-01 4.99153197e-01 4.87163156e-01 -1.04806399e+00
7.56931186e-01 9.83777761e-01 2.41093546e-01 8.76548290e-01
-9.15351331e-01 -4.50439513e-01 -3.06232154e-01 9.09981966e-01
-1.37314773e+00 -1.35060132e+00 8.85030746e-01 -4.07056510e-01
9.63895798e-01 3.62372071e-01 3.87576401e-01 1.47317803e+00
-2.57972300e-01 9.76242721e-01 1.29805720e+00 -5.28875113e-01
3.66486907e-02 4.56406713e-01 1.29545793e-01 3.02747160e-01
-1.10791456e-02 2.29863927e-01 -6.73374116e-01 1.23558059e-01
-1.17280312e-01 -3.62054616e-01 -4.78548199e-01 -3.09316665e-01
-1.27935958e+00 3.87984425e-01 -3.72924566e-01 3.97906810e-01
-2.18194857e-01 -1.61330312e-01 6.17462695e-01 1.66432962e-01
2.07267419e-01 -2.41214484e-01 -4.71860617e-02 -6.09963119e-01
-1.09732223e+00 -1.38557732e-01 5.38974941e-01 8.72248292e-01
3.63826603e-01 1.76542863e-01 -1.02629587e-01 1.06147552e+00
3.11823010e-01 8.84611547e-01 8.43186140e-01 -4.38239038e-01
5.10834634e-01 8.36916715e-02 6.94361925e-02 -6.28473163e-01
-3.04172307e-01 1.56824335e-01 -5.10326326e-01 5.21032393e-01
4.04232830e-01 1.47832364e-01 -1.02901173e+00 1.40750074e+00
1.35939121e-01 2.31231764e-01 4.38513637e-01 8.03158820e-01
1.42607951e+00 5.89139342e-01 1.47778749e-01 -6.70630753e-01
1.31823516e+00 -8.73466909e-01 -1.17161608e+00 6.98555261e-02
2.86013424e-01 -1.05127549e+00 1.02994323e+00 6.22806489e-01
-9.70786929e-01 -7.69440651e-01 -9.91222203e-01 6.27463162e-02
-5.59347749e-01 6.01724505e-01 1.68845803e-01 1.08731985e+00
-1.31814718e+00 -7.22253770e-02 -3.31794500e-01 -5.46013772e-01
2.20799595e-01 2.36641496e-01 -8.09331238e-01 7.55078122e-02
-9.56621170e-01 9.04506505e-01 2.19798103e-01 3.84072959e-01
-9.09394801e-01 2.13804208e-02 -9.83209312e-01 -1.31040320e-01
4.88274217e-01 -8.28059688e-02 1.23236561e+00 -7.59547830e-01
-1.83612573e+00 1.14886701e+00 -6.40157998e-01 -5.04847348e-01
8.75576317e-01 1.89682543e-01 -9.19164181e-01 4.63597208e-01
-4.74948615e-01 5.91136038e-01 1.26221955e+00 -1.34263825e+00
-4.06931221e-01 -3.11090916e-01 -3.67807806e-01 9.51608643e-02
-1.35610774e-01 1.91936754e-02 -4.43704337e-01 -2.40249261e-01
-2.30729088e-01 -6.99684978e-01 5.89431226e-01 -2.29407355e-01
-3.34578931e-01 -5.69491744e-01 1.10170007e+00 -7.81154275e-01
7.69206703e-01 -2.29057431e+00 -1.42052010e-01 -6.32025814e-03
4.53699864e-02 9.35851216e-01 -1.50759622e-01 4.43645149e-01
-3.35203782e-02 -6.75915852e-02 -9.46230292e-02 -6.56379819e-01
4.67995070e-02 1.49284631e-01 -1.33792177e-01 6.49613619e-01
-5.07207215e-02 7.82294869e-01 -4.25080895e-01 -7.61540174e-01
6.82427824e-01 6.50395453e-01 1.20539866e-01 3.12005669e-01
1.91218391e-01 2.59405732e-01 -8.65471829e-03 7.04241753e-01
8.10295939e-01 5.09842217e-01 -2.13202685e-01 -2.21317410e-01
-3.57165545e-01 -9.79469717e-02 -1.01722836e+00 1.41475189e+00
-4.92929697e-01 1.41045260e+00 3.22699755e-01 -1.13167226e+00
1.04984629e+00 6.08106613e-01 1.00102790e-01 -9.44910586e-01
1.88608006e-01 2.92667776e-01 -2.09529847e-01 -1.32581294e+00
6.28286898e-01 -1.90242127e-01 5.87927938e-01 -3.02332751e-02
2.17003733e-01 -9.65553671e-02 2.02515781e-01 -4.96007688e-02
5.33758223e-01 -9.13038477e-02 3.54091257e-01 -1.25322985e-02
1.30518222e+00 -3.57640326e-01 -2.80621499e-01 6.06006801e-01
-6.39641225e-01 5.62459409e-01 3.86588335e-01 1.57450229e-01
-8.27068269e-01 -1.00787461e+00 -3.71896237e-01 6.95717573e-01
1.08398557e-01 -1.11401863e-01 -7.22925901e-01 -6.99551463e-01
-2.20326304e-01 6.32706344e-01 -4.11501884e-01 2.24760562e-01
-2.49950349e-01 -1.31638050e-01 7.13516295e-01 5.07858098e-02
4.70232576e-01 -1.37660992e+00 -3.43877703e-01 -3.55576366e-01
-2.14788273e-01 -1.50452816e+00 -8.45230147e-02 -2.87546903e-01
-2.74854749e-01 -9.55141962e-01 -1.15677047e+00 -9.81870413e-01
2.57805526e-01 4.24500406e-01 8.41230750e-01 1.24692865e-01
-2.68977970e-01 7.23479927e-01 -3.64353031e-01 -3.54955047e-01
-8.63251030e-01 -2.14324608e-01 9.25632864e-02 4.44147676e-01
3.78563672e-01 -9.18843597e-02 -3.26357454e-01 5.73708653e-01
-7.19491065e-01 -2.95575351e-01 6.42628968e-01 7.60894477e-01
1.84708744e-01 -3.17841381e-01 1.01437414e+00 -3.70544493e-01
7.66892552e-01 -2.53076218e-02 -7.33558178e-01 2.70114750e-01
-3.49070489e-01 -1.53801739e-01 3.30919385e-01 -3.17809194e-01
-1.13651431e+00 1.17247730e-01 -8.29456508e-01 -4.64689404e-01
-8.80809188e-01 2.48270947e-02 -4.63066220e-01 -2.67226040e-01
3.25257003e-01 6.03252947e-01 4.86717343e-01 -5.65741718e-01
1.58121586e-01 1.39690280e+00 8.07001650e-01 3.65483761e-02
5.49703300e-01 3.69093329e-01 -1.22577347e-01 -1.73335135e+00
3.63405526e-01 -7.90091515e-01 -3.10853660e-01 -6.91850305e-01
9.15661693e-01 -5.91336668e-01 -1.32574201e+00 7.59417653e-01
-1.09827733e+00 1.04626333e-02 -4.74010110e-02 6.93698525e-01
-6.80219710e-01 8.98522198e-01 6.94600726e-03 -1.47585011e+00
5.73804900e-02 -1.58008432e+00 1.09380937e+00 4.91040908e-02
3.71582478e-01 -6.78837001e-01 5.69193922e-02 7.81125605e-01
3.63648593e-01 -4.30449516e-01 2.84925103e-01 -8.44701231e-01
-3.00576806e-01 -3.06350708e-01 -3.54098737e-01 7.05788970e-01
-5.67551516e-02 5.80562726e-02 -1.68156910e+00 1.60690024e-02
-1.34372413e-01 -5.26519477e-01 1.03668058e+00 2.00168997e-01
1.01779401e+00 1.62213475e-01 -1.61241457e-01 2.36725628e-01
1.07860470e+00 3.22142631e-01 9.58395064e-01 1.93103049e-02
3.12369376e-01 8.09765577e-01 5.96210301e-01 3.30592208e-02
4.11286158e-03 1.00964093e+00 4.22593951e-01 -1.70160368e-01
-5.46479344e-01 -2.17751265e-01 4.02349770e-01 5.90310514e-01
-1.36478007e-01 -4.88716394e-01 -8.38679373e-01 5.81200302e-01
-1.54419649e+00 -1.14245760e+00 -1.73500702e-01 2.28032112e+00
2.81249464e-01 6.62297234e-02 3.17684799e-01 5.09576201e-01
7.57367015e-01 2.21992850e-01 2.19429024e-02 -5.53876698e-01
-4.06491995e-01 1.21003486e-01 3.71632159e-01 8.03956032e-01
-9.91366625e-01 8.37389588e-01 6.24116611e+00 1.11152494e+00
-1.37213302e+00 1.61261499e-01 3.11710894e-01 8.41604397e-02
-3.97410756e-03 -6.29708409e-01 -7.12398052e-01 4.03543055e-01
1.03518701e+00 1.40605390e-01 2.32690573e-01 6.30538523e-01
4.44405556e-01 -5.17294765e-01 -6.55901492e-01 1.60068023e+00
7.21908867e-01 -1.09078252e+00 -1.17818750e-01 1.52414348e-02
2.92582586e-02 1.90253090e-02 2.47513950e-01 1.33497551e-01
-7.82148898e-01 -1.26143479e+00 6.78497195e-01 4.26030576e-01
9.38355803e-01 -6.36740863e-01 9.13621187e-01 3.43249887e-01
-1.03379846e+00 1.43893585e-01 -1.68422446e-01 2.17344627e-01
2.51001388e-01 4.86523919e-02 -1.24958611e+00 3.89160812e-01
3.07114154e-01 5.80654860e-01 -7.19302654e-01 1.17642713e+00
-4.59778197e-02 6.28207862e-01 -1.84059694e-01 -4.01078820e-01
3.06659844e-03 1.14380255e-01 8.52367818e-01 1.46695006e+00
4.13519681e-01 -3.91904503e-01 -4.34585959e-01 3.08143228e-01
1.16614424e-01 4.80049998e-01 -1.21954036e+00 -1.08318634e-01
5.31049445e-02 1.20170784e+00 -5.63262999e-01 -4.18579906e-01
-5.32479525e-01 7.25910366e-01 -2.88274258e-01 5.87927163e-01
-6.43006206e-01 -3.33659232e-01 5.29004693e-01 -8.06850269e-02
3.44248563e-01 -2.87927866e-01 1.16658777e-01 -1.07751274e+00
2.05357298e-01 -7.61004627e-01 -4.77776788e-02 -9.78838861e-01
-7.63429284e-01 9.45701897e-01 6.82016760e-02 -1.42053711e+00
-4.66575503e-01 -1.08253264e+00 -1.99018478e-01 8.27659726e-01
-1.82089365e+00 -1.05655241e+00 -2.83287495e-01 8.41190040e-01
8.85266781e-01 -7.30850816e-01 5.51251173e-01 4.21676487e-01
-3.50228339e-01 7.24779189e-01 4.44134735e-02 -2.97609478e-01
7.18079805e-01 -7.20261931e-01 -3.18313912e-02 8.26344252e-01
2.80469447e-01 1.66716129e-01 9.15084839e-01 -2.54318923e-01
-1.40568018e+00 -3.13851297e-01 1.13545442e+00 -2.34854952e-01
4.06692296e-01 -5.41651547e-01 -8.37467790e-01 1.22281544e-01
6.87750697e-01 -1.24149173e-01 3.57150823e-01 -3.70517522e-01
-6.10115081e-02 -2.61319160e-01 -1.22970319e+00 3.79474372e-01
7.83684254e-01 -6.97285712e-01 -8.87426078e-01 1.70903653e-01
2.78537482e-01 -3.87823470e-02 -2.23705292e-01 3.26865494e-01
6.65932953e-01 -1.30980325e+00 1.01302993e+00 -3.09407145e-01
-1.46461442e-01 -3.12476814e-01 -1.46483347e-01 -9.74724472e-01
6.12385154e-01 -8.07189465e-01 2.56809384e-01 1.47563744e+00
4.72412527e-01 -8.02463114e-01 6.35546088e-01 2.77006626e-02
-9.79732499e-02 -2.41950184e-01 -1.41897178e+00 -8.20079744e-01
-4.51799542e-01 -9.23189640e-01 2.26797640e-01 4.98712897e-01
1.04079902e-01 1.37517601e-01 -5.71080148e-01 -2.83851549e-02
5.51008165e-01 -4.39683706e-01 9.60155547e-01 -1.02528691e+00
-2.99150180e-02 -5.43144107e-01 -1.08270633e+00 -8.82339776e-01
6.32841229e-01 -6.51646912e-01 2.37921104e-01 -1.20670152e+00
-1.17109753e-01 2.87165165e-01 -6.31013885e-02 3.14963758e-02
1.77672207e-01 5.64123273e-01 2.43358135e-01 -7.30749443e-02
-3.97453755e-01 4.53369081e-01 1.05407310e+00 -3.67216885e-01
5.88856675e-02 2.58527070e-01 3.08710393e-02 4.96901602e-01
6.82157934e-01 8.78133625e-02 -5.70210099e-01 1.39528915e-01
-2.72736520e-01 2.05809847e-01 5.39656937e-01 -9.74304199e-01
3.47349226e-01 2.28873029e-01 -7.64120668e-02 -8.22700083e-01
7.89211512e-01 -8.96626651e-01 -1.66498527e-01 1.60617456e-01
-2.95515358e-01 -1.66336950e-02 3.82345349e-01 5.00058532e-01
-5.63654482e-01 -4.34230179e-01 7.73541689e-01 2.65448123e-01
-1.14372230e+00 -2.62164414e-01 -7.33935893e-01 -1.92636684e-01
1.03791082e+00 -4.70772117e-01 -5.18790901e-01 -5.25412142e-01
-9.15179491e-01 -2.25722566e-01 3.50586683e-01 6.45738959e-01
9.22896147e-01 -9.50542092e-01 -6.24098361e-01 6.84987903e-01
5.04019201e-01 -8.13844621e-01 5.36750138e-01 1.11925018e+00
-4.03850168e-01 6.09974205e-01 -3.01436156e-01 -8.58941734e-01
-1.87994874e+00 8.62833500e-01 1.74782068e-01 6.62766993e-01
-5.98131001e-01 3.74172419e-01 6.73660189e-02 5.48887067e-02
6.53400064e-01 -2.57422686e-01 -6.93195403e-01 2.06617013e-01
6.18285298e-01 4.18762803e-01 4.59611923e-01 -1.23250890e+00
-4.88911062e-01 6.35917008e-01 3.91621351e-01 -4.95325804e-01
6.46838963e-01 -4.40974832e-01 2.34247342e-01 6.25384271e-01
1.28847909e+00 3.90494168e-01 -7.99873888e-01 2.59182245e-01
-1.95470408e-01 -4.76374984e-01 6.33556545e-02 -6.99421525e-01
-7.27179945e-01 1.30294776e+00 1.00987840e+00 5.34023702e-01
1.01851368e+00 9.80236754e-02 3.29940140e-01 5.01910985e-01
7.70221353e-02 -1.05669975e+00 -3.32375258e-01 1.30418807e-01
9.17786062e-01 -1.47151351e+00 -4.52107817e-01 -3.15774143e-01
-9.02808070e-01 1.22981071e+00 5.69175817e-02 5.82240164e-01
5.82329273e-01 2.50796616e-01 5.06704330e-01 2.71421254e-01
-6.14490211e-01 -8.47151101e-01 3.44383389e-01 1.06509960e+00
2.90563017e-01 -6.59602657e-02 -5.23946285e-01 3.58314328e-02
1.38315735e-02 2.18024001e-01 5.53486407e-01 5.11001587e-01
-3.54079664e-01 -9.40556467e-01 -5.22050142e-01 -4.22914959e-02
-2.94369221e-01 7.17164502e-02 -4.80232775e-01 9.23943520e-01
2.51772888e-02 1.44842696e+00 -1.26477450e-01 -3.44035506e-01
4.34839845e-01 5.05566359e-01 3.53193760e-01 1.59266382e-01
-1.80904493e-01 2.22217664e-01 4.37858820e-01 -5.00007927e-01
-7.33933032e-01 -8.71528745e-01 -7.48323798e-01 5.16489800e-03
-2.06352368e-01 6.72893748e-02 1.24242389e+00 9.58227217e-01
1.01683542e-01 1.26263961e-01 5.47979534e-01 -9.87769008e-01
-3.37486774e-01 -1.08253539e+00 -5.60627401e-01 4.16821271e-01
9.50185418e-01 -9.92128491e-01 -6.77643895e-01 1.74936622e-01] | [14.326522827148438, 5.042632102966309] |
53da3c43-b130-472e-a40b-69cf53e68d31 | exploiting-multimodal-synthetic-data-for | 2306.12152 | null | https://arxiv.org/abs/2306.12152v1 | https://arxiv.org/pdf/2306.12152v1.pdf | Exploiting Multimodal Synthetic Data for Egocentric Human-Object Interaction Detection in an Industrial Scenario | In this paper, we tackle the problem of Egocentric Human-Object Interaction (EHOI) detection in an industrial setting. To overcome the lack of public datasets in this context, we propose a pipeline and a tool for generating synthetic images of EHOIs paired with several annotations and data signals (e.g., depth maps or instance segmentation masks). Using the proposed pipeline, we present EgoISM-HOI a new multimodal dataset composed of synthetic EHOI images in an industrial environment with rich annotations of hands and objects. To demonstrate the utility and effectiveness of synthetic EHOI data produced by the proposed tool, we designed a new method that predicts and combines different multimodal signals to detect EHOIs in RGB images. Our study shows that exploiting synthetic data to pre-train the proposed method significantly improves performance when tested on real-world data. Moreover, the proposed approach outperforms state-of-the-art class-agnostic methods. To support research in this field, we publicly release the datasets, source code, and pre-trained models at https://iplab.dmi.unict.it/egoism-hoi. | ['Giovanni Maria Farinella', 'Antonino Furnari', 'Francesco Ragusa', 'Rosario Leonardi'] | 2023-06-21 | null | null | null | null | ['human-object-interaction-detection'] | ['computer-vision'] | [ 4.32779908e-01 2.64110059e-01 5.06434083e-01 -3.07274044e-01
-5.63707769e-01 -5.92078805e-01 5.77948749e-01 -3.99779618e-01
-2.28871971e-01 4.64190662e-01 -7.88843036e-02 3.19174558e-01
4.01423015e-02 -6.16264701e-01 -9.03670192e-01 -5.01504719e-01
2.87297487e-01 7.68714428e-01 2.26314545e-01 -1.82973966e-01
1.57984212e-01 4.46690232e-01 -1.88119888e+00 5.97473264e-01
5.99737704e-01 1.05976176e+00 4.00980890e-01 8.23780477e-01
5.00386059e-01 6.66073084e-01 -5.73670089e-01 -3.35561782e-01
3.92673165e-01 -3.39841515e-01 -7.11597919e-01 2.85736024e-01
5.22584021e-01 -2.81650394e-01 -2.50275642e-01 8.09681535e-01
7.82631218e-01 6.95549622e-02 5.97689927e-01 -1.75566888e+00
-5.85218668e-02 4.40143347e-01 -2.15636015e-01 -4.59360093e-01
6.88374460e-01 7.12401032e-01 6.10221565e-01 -9.60289240e-01
1.14846218e+00 1.20307744e+00 4.71262276e-01 5.66503048e-01
-1.27846301e+00 -4.60228980e-01 -2.06018999e-01 5.89540228e-02
-1.31652927e+00 -4.43908572e-01 9.32171822e-01 -6.35654569e-01
7.15809405e-01 2.91839480e-01 6.33490443e-01 1.86957419e+00
-1.67874902e-01 1.17778850e+00 1.24352980e+00 -4.57879722e-01
3.30450177e-01 5.02802968e-01 -1.64775297e-01 6.19627953e-01
1.89648867e-02 -5.39031178e-02 -6.41763091e-01 1.89872578e-01
8.61875951e-01 -2.21091032e-01 -1.47760764e-01 -8.89784992e-01
-1.75498736e+00 2.90121228e-01 2.80651331e-01 2.25495234e-01
-5.48005939e-01 1.31644085e-01 4.36360091e-01 -6.72873780e-02
6.34159818e-02 3.48344982e-01 -3.55158687e-01 -1.15096256e-01
-5.37738144e-01 6.14149034e-01 7.36451626e-01 1.28076267e+00
7.03389645e-01 -3.80843282e-01 -2.87079275e-01 6.11138165e-01
1.61253303e-01 5.18674493e-01 1.58147678e-01 -1.04383552e+00
6.88966811e-01 6.75535262e-01 4.28032398e-01 -8.75633419e-01
-7.35403597e-01 -1.75211892e-01 -2.89363146e-01 3.84913862e-01
8.89528453e-01 -1.42623410e-01 -7.40420640e-01 1.52605271e+00
5.16042709e-01 -1.60778522e-01 2.19134808e-01 1.27223766e+00
8.00813258e-01 2.25798488e-01 -1.21270992e-01 4.12511379e-01
1.29575121e+00 -1.04438150e+00 -5.71151555e-01 -2.49929175e-01
5.74580073e-01 -7.45695353e-01 1.39248586e+00 6.75797641e-01
-1.06628680e+00 -6.92662776e-01 -8.18736553e-01 1.94161683e-01
-4.93656099e-01 7.42811084e-01 4.95093793e-01 5.72047114e-01
-6.52143896e-01 4.65728104e-01 -6.70580566e-01 -7.16480553e-01
2.82943070e-01 3.12784702e-01 -6.03842854e-01 -4.60806899e-02
-8.82668614e-01 4.67765212e-01 6.11299872e-01 2.51757145e-01
-1.20098364e+00 -3.21309119e-01 -7.87821472e-01 -3.53303313e-01
6.62528276e-01 -7.74150789e-01 1.28250635e+00 -1.01055133e+00
-1.55768669e+00 1.03567624e+00 3.08222324e-01 -2.43846878e-01
1.15646970e+00 -4.20103937e-01 -1.58354983e-01 2.25667283e-01
1.36283785e-02 8.84748578e-01 7.40840912e-01 -1.68138063e+00
-6.00476027e-01 -5.21245480e-01 2.83499330e-01 2.42969915e-02
-2.05601811e-01 -1.79171965e-01 -5.20474911e-01 -3.63023371e-01
-2.47047156e-01 -1.34622204e+00 4.29889448e-02 -2.09593102e-01
-9.47821677e-01 6.40765131e-02 7.41721928e-01 -6.05486989e-01
5.12827754e-01 -2.12364292e+00 3.61505121e-01 2.63337523e-01
8.59268308e-02 1.11859448e-01 -2.94562299e-02 4.03770268e-01
5.14660031e-02 -5.23281217e-01 -3.19573611e-01 -4.92808819e-01
4.43959147e-01 -3.35891247e-02 2.96066284e-01 4.15957361e-01
6.95176199e-02 8.41328740e-01 -1.03473055e+00 -5.85849464e-01
7.75399387e-01 6.08751833e-01 -4.11941886e-01 4.20628786e-01
-4.28961694e-01 9.95663464e-01 -2.56440282e-01 8.09381664e-01
6.64117038e-01 1.21119857e-01 2.10881174e-01 -5.48332155e-01
2.02355571e-02 -3.22949171e-01 -1.44306672e+00 2.10381055e+00
-6.44681454e-01 4.89971280e-01 3.64070758e-02 -5.59426785e-01
8.27203631e-01 4.21501428e-01 5.77073932e-01 -4.74631965e-01
5.67305982e-01 2.87961245e-01 -2.11881280e-01 -8.34869921e-01
5.53944051e-01 3.45918268e-01 -2.16280654e-01 1.00023873e-01
3.79012734e-01 -3.00089210e-01 3.98738980e-01 -6.06106147e-02
1.01960242e+00 7.50397265e-01 -9.56159160e-02 1.59209639e-01
5.91226578e-01 1.37771994e-01 2.34428003e-01 6.61869407e-01
-2.38357082e-01 9.30933297e-01 5.28838277e-01 -2.83977151e-01
-1.05014181e+00 -1.01480138e+00 1.39049411e-01 9.75645125e-01
2.69230038e-01 -1.61525607e-01 -1.29517508e+00 -9.52027380e-01
-3.01055349e-02 7.35647738e-01 -7.57156789e-01 1.83275923e-01
-3.76742333e-01 -5.86803317e-01 4.78396118e-01 4.70213354e-01
6.29993796e-01 -1.35106862e+00 -1.09269977e+00 -3.72858196e-02
-3.78645331e-01 -1.54535496e+00 5.71138300e-02 -7.94876274e-03
-5.09325027e-01 -1.25931525e+00 -8.95923913e-01 -5.08945107e-01
6.13533258e-01 -3.48180085e-02 1.22314715e+00 -6.69780374e-01
-6.77031398e-01 8.04423809e-01 -4.85000342e-01 -4.86029297e-01
-3.48149747e-01 2.00622797e-01 -1.60539940e-01 3.24097306e-01
4.89654057e-02 -3.85276616e-01 -8.30497026e-01 7.11785674e-01
-7.79199421e-01 2.33086154e-01 6.11722767e-01 4.87027854e-01
4.35765326e-01 -3.61584544e-01 1.36997774e-01 -9.89142895e-01
3.61010611e-01 -1.79126486e-01 -6.08661771e-01 8.77475664e-02
-2.58205421e-02 -1.68463200e-01 2.35642865e-01 -3.37678283e-01
-1.23973787e+00 7.77322471e-01 6.55397549e-02 -4.97140229e-01
-6.81280434e-01 -7.70927891e-02 -6.47109449e-01 -4.70786653e-02
1.02688968e+00 -1.71040237e-01 -2.59953886e-01 -4.33981419e-01
6.02868140e-01 7.91703999e-01 7.51664996e-01 -5.66613257e-01
5.40269077e-01 6.36332214e-01 -4.45835106e-02 -6.67388260e-01
-5.57887077e-01 -3.01795393e-01 -1.02892923e+00 -6.70486629e-01
1.03390431e+00 -8.03436041e-01 -7.87577271e-01 5.60443699e-01
-1.22316015e+00 -6.36025548e-01 -3.45774919e-01 5.51655829e-01
-1.14480364e+00 8.10323656e-02 -2.07449883e-01 -9.67478812e-01
-2.53871661e-02 -1.27955449e+00 1.51138210e+00 2.43439320e-02
-3.22097450e-01 -6.49540901e-01 -5.75219020e-02 7.47958601e-01
4.14076783e-02 8.74819100e-01 4.06163335e-02 -5.22059321e-01
-7.40683794e-01 -4.23091739e-01 -9.32242423e-02 3.64194542e-01
-1.07691936e-01 -1.26572460e-01 -1.49497581e+00 9.58442166e-02
-4.03188556e-01 -5.25649488e-01 4.38373029e-01 6.19554780e-02
1.05606675e+00 6.25953749e-02 -3.00501049e-01 3.46291900e-01
1.14978802e+00 -1.14910491e-01 6.65406644e-01 5.97281396e-01
1.03562629e+00 1.09360945e+00 1.17830229e+00 7.14561582e-01
3.28617841e-01 1.01413572e+00 8.83226931e-01 -2.05657631e-01
-2.06685454e-01 2.70312708e-02 3.75340641e-01 1.07451566e-01
-3.42846632e-01 -3.92535657e-01 -9.47519124e-01 6.60378873e-01
-2.07428789e+00 -7.61468530e-01 -4.23525393e-01 2.05487633e+00
3.03334415e-01 3.13029774e-02 5.32295585e-01 2.76284963e-01
7.07602203e-01 -1.66044578e-01 -3.91552389e-01 -7.72629753e-02
-8.23906884e-02 -2.11571306e-01 5.78202307e-01 1.75895259e-01
-1.30791974e+00 9.56498981e-01 4.77034807e+00 2.97985405e-01
-8.18601847e-01 2.38371581e-01 1.61593720e-01 -5.62274419e-02
2.43755355e-01 -4.09026742e-01 -4.96823519e-01 3.35831344e-01
7.63919294e-01 4.36507523e-01 4.57358748e-01 1.26130998e+00
3.04902434e-01 -8.12801793e-02 -1.26820958e+00 1.22385371e+00
1.87619328e-01 -8.39395404e-01 -4.27678555e-01 5.68105206e-02
7.08328307e-01 -6.93818480e-02 1.18549466e-01 6.05294257e-02
1.23388372e-01 -6.55932367e-01 1.10264075e+00 6.81248128e-01
5.39578736e-01 -5.69726646e-01 9.66853440e-01 1.36279061e-01
-9.87810910e-01 -1.55011266e-01 -1.72623433e-02 2.93835402e-01
1.06140189e-01 3.66226137e-01 -1.20844376e+00 7.59569168e-01
6.93572760e-01 5.26804924e-01 -8.26867163e-01 9.19933975e-01
-4.06515747e-01 1.89918756e-01 -2.90406227e-01 2.37505093e-01
6.47440553e-02 -2.17055455e-02 5.83347738e-01 1.41488075e+00
3.39165300e-01 -6.00093484e-01 -5.51504595e-03 1.08167636e+00
1.77176639e-01 6.89027607e-02 -9.21037972e-01 -4.81269434e-02
-1.76145472e-02 1.58668649e+00 -8.74674737e-01 -3.24250191e-01
-8.87742788e-02 1.51121378e+00 -4.64633144e-02 3.27162772e-01
-1.11010218e+00 -6.51637137e-01 3.34105521e-01 3.03878307e-01
1.92467764e-01 -8.95008519e-02 -1.13081977e-01 -1.04619324e+00
1.32900104e-01 -8.77811968e-01 2.55432487e-01 -1.27618158e+00
-9.05146837e-01 7.45801270e-01 2.56346434e-01 -1.36894321e+00
-5.76201975e-01 -1.05273461e+00 -2.89529324e-01 5.37741303e-01
-8.76708627e-01 -1.73879325e+00 -1.16498816e+00 7.16790736e-01
5.24271786e-01 -1.44826442e-01 5.99990726e-01 4.20636535e-01
-5.30500531e-01 3.24198037e-01 -1.67872414e-01 8.46138597e-03
8.68150473e-01 -1.35810280e+00 3.76268744e-01 4.90929067e-01
4.27085347e-02 2.96641409e-01 9.29427028e-01 -4.78341460e-01
-1.49542749e+00 -1.02621388e+00 1.92178294e-01 -8.97723556e-01
4.41555083e-01 -7.86198616e-01 -2.83479422e-01 8.50614071e-01
1.09359518e-01 9.68122929e-02 2.53399700e-01 -3.85145754e-01
1.04519278e-01 -6.46707937e-02 -1.28924227e+00 5.11214793e-01
1.32527208e+00 -3.42373699e-01 -2.68848240e-01 6.15207791e-01
2.48665884e-01 -6.30212843e-01 -9.26062167e-01 5.53596914e-01
7.79076636e-01 -1.34948444e+00 9.79061007e-01 -7.60079101e-02
3.82897735e-01 -5.56085765e-01 -3.47524583e-01 -1.20305026e+00
2.90246010e-01 -5.21603167e-01 2.22986192e-03 1.35222042e+00
2.69663751e-01 -3.28405529e-01 9.77090418e-01 4.41981584e-01
-2.29942605e-01 -2.29175106e-01 -6.03878975e-01 -7.47050643e-01
-7.00023711e-01 -7.21471131e-01 3.88924062e-01 5.81223249e-01
-1.17584154e-01 1.25259653e-01 -4.19909775e-01 1.39470726e-01
7.00910032e-01 -1.04333982e-01 1.67507184e+00 -1.07838142e+00
-3.21560621e-01 7.98200518e-02 -7.29946792e-01 -3.38759542e-01
7.44464770e-02 -7.49321699e-01 1.98111102e-01 -1.61291695e+00
7.04669505e-02 -1.91807985e-01 -3.80971096e-02 4.54420179e-01
1.41868502e-01 7.01704919e-01 3.44200313e-01 3.43005992e-02
-8.03718626e-01 4.39241022e-01 1.11693966e+00 -6.92219660e-02
-2.87340492e-01 -1.08299986e-01 -7.77095109e-02 8.56186271e-01
8.95156264e-01 -3.41662616e-01 -3.53220880e-01 -6.92456216e-02
-5.37557229e-02 -2.85850585e-01 1.01539755e+00 -1.48657918e+00
-1.70007318e-01 3.00170302e-01 4.93448138e-01 -6.69689536e-01
7.21208334e-01 -9.99221325e-01 3.83665800e-01 4.34898704e-01
-1.08079635e-01 -1.98158219e-01 9.33288783e-02 3.53444874e-01
3.79357254e-03 -1.45689413e-01 4.70962137e-01 -3.23438495e-01
-7.67072022e-01 -2.54400253e-01 -2.11594909e-01 -1.73350245e-01
1.32769275e+00 -2.21986189e-01 -1.71009779e-01 -2.90002108e-01
-8.70309174e-01 3.91446464e-02 7.18826115e-01 5.67804754e-01
5.09805799e-01 -1.08984292e+00 -4.73085046e-01 1.73376516e-01
7.17645884e-01 1.85510528e-03 3.62902761e-01 1.04885697e+00
-7.59474456e-01 2.82591373e-01 -6.27582252e-01 -8.48690569e-01
-1.25778341e+00 5.28648198e-01 3.77048254e-01 -7.81434327e-02
-6.00733817e-01 5.78601897e-01 3.85972708e-01 -9.08523381e-01
2.64521778e-01 -4.04082000e-01 1.65590574e-03 -7.16823041e-02
1.63512826e-01 6.33490860e-01 1.07798606e-01 -8.25151443e-01
-3.60297918e-01 4.74264771e-01 5.68295062e-01 -4.38712120e-01
1.09057176e+00 -2.02405062e-02 2.42496446e-01 5.80024779e-01
8.59124541e-01 -1.02444202e-01 -1.50443947e+00 2.32371867e-01
-9.41535980e-02 -6.58505380e-01 -2.92168409e-01 -9.55836415e-01
-9.20301437e-01 8.61034572e-01 1.08253455e+00 -5.64912148e-02
8.92846584e-01 1.65654197e-01 4.61782008e-01 3.71680647e-01
6.48208439e-01 -1.26930976e+00 3.77043694e-01 6.72647133e-02
1.04732716e+00 -1.42772353e+00 -3.84730339e-01 -6.82228148e-01
-1.06561863e+00 9.54389811e-01 7.34881341e-01 4.82817218e-02
1.17893144e-01 2.80819312e-02 1.87641948e-01 -4.05664086e-01
-1.15748286e-01 -4.59859669e-01 8.34968016e-02 9.06826317e-01
2.95666099e-01 4.63590249e-02 7.61882737e-02 5.28787613e-01
-2.29407325e-01 3.10086340e-01 3.63647789e-01 9.70229506e-01
8.60606506e-02 -9.40404236e-01 -6.99868381e-01 -7.42139593e-02
-1.89980671e-01 4.94863659e-01 -6.84047818e-01 1.05386508e+00
4.88018483e-01 9.57018793e-01 -2.31368691e-01 -6.18225217e-01
8.57931674e-01 6.63642958e-02 7.37807810e-01 -3.57474864e-01
-6.76723838e-01 5.56062423e-02 3.64347845e-01 -9.57860887e-01
-5.46639979e-01 -7.20902681e-01 -1.02003515e+00 1.47168517e-01
-1.09275460e-01 -2.55468130e-01 1.14239371e+00 6.24955237e-01
3.85558814e-01 9.68627393e-01 3.95981640e-01 -1.76727211e+00
-1.19078547e-01 -1.07323682e+00 -5.34568310e-01 8.20367575e-01
-7.85281286e-02 -9.87542212e-01 -2.06604064e-01 2.25052074e-01] | [7.618464946746826, -0.823140561580658] |
ac163c87-e6cc-4911-87e5-e552cc78ae6b | wavemix-lite-a-resource-efficient-neural-1 | null | null | https://openreview.net/forum?id=y_icnxeeUcl | https://openreview.net/pdf?id=y_icnxeeUcl | WaveMix-Lite: A Resource-efficient Neural Network for Image Analysis | Gains in the ability to generalize on image analysis tasks for neural networks have come at the cost of increased number of parameters and layers, dataset sizes, training and test computations, and GPU RAM. We introduce a new architecture -- WaveMix-Lite -- that can generalize on par with contemporary transformers and convolutional neural networks (CNNs) while needing fewer resources. WaveMix-Lite uses 2D-discrete wavelet transform to efficiently mix spatial information from pixels. WaveMix-Lite seems to be a versatile and scalable architectural framework that can be used for multiple vision tasks, such as image classification and semantic segmentation, without requiring significant architectural changes, unlike transformers and CNNs. It is able to meet or exceed several accuracy benchmarks while training on a single GPU. For instance, it achieves state-of-the-art accuracy on five EMNIST datasets, outperforms CNNs and transformers in ImageNet-1K and Places-365, and achieves an mIoU of 77\% on Cityscapes validation set, while using less than one-fifth the number parameters and half the GPU RAM of comparable CNNs or transformers. Our experiments show that while the convolutional elements of neural architectures exploit the shift-invariance property of images, new types of layers (e.g., wavelet transform) can exploit additional properties of images, such as scale-invariance and finite spatial extents of objects. | ['Amit', 'Pranav; Sethi', 'Jeevan'] | 2022-10-13 | null | null | null | iclr-2022-10 | ['scene-classification'] | ['computer-vision'] | [ 9.34417173e-02 -1.97700962e-01 -4.49622311e-02 -4.71402228e-01
-3.72168362e-01 -5.67054451e-01 4.53439981e-01 -5.72783537e-02
-9.19603765e-01 1.68709740e-01 -5.86942077e-01 -6.08688295e-01
2.05718596e-02 -9.33085322e-01 -7.59284258e-01 -7.15895236e-01
-4.69357334e-02 3.21313113e-01 8.02665591e-01 -3.91102284e-01
-1.54693916e-01 8.13047528e-01 -1.75380015e+00 5.16146362e-01
4.04289544e-01 1.57630563e+00 1.81679368e-01 6.61689460e-01
-6.63132146e-02 5.38194895e-01 -3.02527964e-01 -4.15690541e-01
4.32379901e-01 1.40252009e-01 -6.57027066e-01 -3.04712564e-01
1.11110187e+00 -2.02681780e-01 -3.48867118e-01 8.83433521e-01
6.28112972e-01 -2.57903822e-02 3.32171589e-01 -1.21858656e+00
-4.09383178e-01 2.46327162e-01 -3.84040982e-01 4.01979148e-01
-4.16987509e-01 2.96305478e-01 6.68448567e-01 -9.10680234e-01
3.91052574e-01 1.12094605e+00 1.42767906e+00 5.22421360e-01
-1.11766756e+00 -8.33363116e-01 -1.50933936e-01 1.96991801e-01
-1.20244360e+00 -2.47035533e-01 2.58212268e-01 -2.36803249e-01
1.46830165e+00 3.99056435e-01 7.16413975e-01 9.37772274e-01
2.49169290e-01 6.34402752e-01 1.02568460e+00 -4.11388464e-02
1.75018921e-01 -1.49800301e-01 1.60949767e-01 9.66324985e-01
3.34519148e-01 -2.41758481e-01 -4.32036489e-01 1.53901905e-01
9.63990033e-01 -9.65567976e-02 8.29618052e-02 -4.73697595e-02
-1.20133364e+00 7.50030756e-01 7.27700293e-01 2.98190862e-01
-2.47929066e-01 6.22153223e-01 7.77968466e-01 3.33978832e-01
5.73560238e-01 3.50789577e-01 -8.16785395e-01 -1.99115232e-01
-1.12947512e+00 1.43510789e-01 5.24489284e-01 9.43078458e-01
8.94708872e-01 3.09789866e-01 6.23688176e-02 7.29791582e-01
-2.44080186e-01 5.33569694e-01 6.07689857e-01 -9.77287889e-01
4.80184823e-01 6.88468039e-01 -4.71789479e-01 -7.10432827e-01
-8.00816834e-01 -5.25739253e-01 -1.19975281e+00 5.21606147e-01
4.81073558e-01 -1.16725102e-01 -1.23110366e+00 1.48871481e+00
1.29166707e-01 3.05335641e-01 -4.58639843e-04 6.57833576e-01
1.03607202e+00 5.96955180e-01 1.28195614e-01 3.80926609e-01
1.68227005e+00 -9.81256604e-01 -1.29747698e-02 -5.63230395e-01
8.90893638e-01 -6.31587565e-01 1.15704608e+00 3.67466092e-01
-1.22019958e+00 -8.23733270e-01 -1.17844713e+00 -3.34057689e-01
-6.94107115e-01 2.30195686e-01 8.48495781e-01 9.34063077e-01
-1.38570619e+00 7.45487690e-01 -1.17958701e+00 -4.43127841e-01
8.30911696e-01 5.83978653e-01 -3.99259865e-01 -4.33767363e-02
-6.42158866e-01 6.34441793e-01 5.39353311e-01 -1.42278150e-01
-3.76179188e-01 -1.10515225e+00 -5.73157728e-01 1.46769270e-01
-6.85992539e-02 -7.16424525e-01 1.21119726e+00 -1.05877113e+00
-1.40614009e+00 8.53185952e-01 2.37259850e-01 -7.97737658e-01
5.03437161e-01 -8.55969563e-02 -1.84026256e-01 3.12634408e-01
-9.36607197e-02 1.21267772e+00 7.40279853e-01 -4.27010566e-01
-7.45087862e-01 -3.79960030e-01 -1.75984442e-01 -7.22909346e-02
-6.30553365e-01 -1.32256001e-01 -7.23047614e-01 -6.31115615e-01
3.24956119e-01 -1.17240012e+00 -2.33556584e-01 4.19414520e-01
5.69134764e-02 -1.61708370e-01 1.35728621e+00 -3.13366413e-01
5.38233697e-01 -2.23056579e+00 -2.71484703e-01 1.05287798e-01
1.63245425e-01 7.37641215e-01 -2.00791836e-01 -6.79258630e-02
-1.86519206e-01 3.34080346e-02 -1.88672155e-01 -4.18752313e-01
-3.39922577e-01 4.26445961e-01 -5.71334213e-02 4.48650032e-01
2.25840598e-01 1.02412140e+00 -5.01460850e-01 -2.00206712e-01
3.09473544e-01 6.01238191e-01 -7.60881186e-01 -4.26561803e-01
-7.55556747e-02 1.07474230e-01 -2.09810287e-01 4.22866970e-01
8.61677229e-01 -3.49937409e-01 -1.01837054e-01 -3.89076412e-01
-1.51518866e-01 2.55896635e-02 -9.96979535e-01 1.70225859e+00
-6.23110175e-01 1.04363024e+00 7.79310241e-02 -1.18211174e+00
8.30525279e-01 2.63621837e-01 3.12275857e-01 -9.85549092e-01
1.90232649e-01 6.07017100e-01 -1.14048034e-01 -4.19816315e-01
3.95228505e-01 2.37003446e-01 1.31558388e-01 9.87775400e-02
3.52122784e-01 -1.68330371e-01 3.04150432e-01 -2.92630941e-01
1.05666506e+00 -1.49997383e-01 -2.24209607e-01 -6.86870575e-01
1.99866891e-01 1.92455307e-01 2.84826130e-01 7.84197450e-01
1.34303406e-01 5.41943610e-01 3.44548315e-01 -1.10579979e+00
-1.28375602e+00 -9.64176655e-01 -3.80520701e-01 1.25142491e+00
-1.00147292e-01 -2.42406026e-01 -9.33141291e-01 -3.36321026e-01
-1.52470231e-01 2.97665387e-01 -5.76091707e-01 6.98080063e-02
-7.59951293e-01 -9.68600988e-01 1.12228823e+00 9.64258611e-01
1.30291069e+00 -7.52689481e-01 -1.21450007e+00 1.80839106e-01
1.48854852e-01 -1.48269022e+00 -6.82085529e-02 4.38736260e-01
-1.27660108e+00 -8.66811931e-01 -7.94496775e-01 -1.05102038e+00
4.07951891e-01 2.64893919e-01 1.15284157e+00 -3.76399420e-02
-5.82084179e-01 8.10715780e-02 -1.10282429e-01 -3.55077416e-01
9.55040082e-02 5.91705620e-01 -1.94081083e-01 -2.38089278e-01
1.76623151e-01 -7.70267367e-01 -8.48451078e-01 3.96615237e-01
-9.16056275e-01 2.78813869e-01 5.87482870e-01 7.43715107e-01
4.04400766e-01 2.94336844e-02 2.32253894e-01 -8.06382596e-01
2.34940246e-01 -9.41663012e-02 -6.33462727e-01 -3.72165032e-02
-4.24622715e-01 -3.50186788e-02 7.39858627e-01 -5.89544356e-01
-7.94039965e-01 1.08617991e-01 -5.02285600e-01 -3.94906998e-01
-1.35871217e-01 2.34199017e-01 2.85128832e-01 -4.62009698e-01
8.77516806e-01 6.78002313e-02 7.57142678e-02 -4.37242061e-01
2.21723348e-01 3.39810580e-01 8.10257435e-01 -5.62176824e-01
3.82835865e-01 7.20968843e-01 2.87504762e-01 -1.01395035e+00
-5.51288068e-01 -5.06289721e-01 -5.79614758e-01 9.28277224e-02
1.09734690e+00 -9.68475223e-01 -7.36910880e-01 8.57954681e-01
-1.02333224e+00 -6.61596417e-01 -3.12125862e-01 4.57568914e-01
-4.16445345e-01 -1.01666860e-01 -8.59490991e-01 -2.07689822e-01
-5.01047373e-01 -1.34980178e+00 9.11435187e-01 3.32038790e-01
1.74448863e-01 -9.88648117e-01 -5.29254854e-01 1.45564750e-01
7.86720335e-01 3.44998747e-01 8.92646790e-01 -4.57392186e-01
-4.81112212e-01 -2.12002605e-01 -6.47511184e-01 5.66842735e-01
-3.30954373e-01 -1.32722661e-01 -1.12922215e+00 -4.14743930e-01
-1.25904173e-01 -3.28667730e-01 1.16469479e+00 4.91656959e-01
1.44892585e+00 -1.15516394e-01 -1.69928640e-01 1.19010532e+00
1.55871022e+00 9.92114171e-02 6.33867025e-01 4.75607246e-01
5.44980884e-01 2.81597674e-01 -3.65377516e-02 1.48732111e-01
1.75353557e-01 5.63505054e-01 5.70581436e-01 -5.45561075e-01
-3.71947825e-01 3.49170327e-01 7.02839252e-03 6.63902819e-01
-4.06725645e-01 -3.82985696e-02 -1.14278018e+00 4.17919844e-01
-1.72293448e+00 -6.25464141e-01 -3.88420522e-01 1.89278257e+00
4.82100725e-01 3.31715971e-01 1.16423942e-01 2.16342181e-01
3.47227544e-01 4.83508818e-02 -4.58969057e-01 -6.79674029e-01
-2.04839244e-01 7.82940686e-01 1.09940362e+00 -2.16530692e-02
-1.03390086e+00 1.09975195e+00 6.68788052e+00 1.07214689e+00
-1.55670357e+00 2.93139845e-01 8.41136813e-01 -1.53228760e-01
3.59030455e-01 -4.78083909e-01 -7.55325794e-01 4.67252359e-02
1.03677702e+00 5.02568185e-01 2.51973182e-01 1.04212177e+00
-2.70752370e-01 1.06635898e-01 -8.89453769e-01 1.06833673e+00
-1.87429205e-01 -1.75137830e+00 -5.01186065e-02 7.21003041e-02
7.30864704e-01 8.66522789e-01 4.19619322e-01 2.47926697e-01
8.18025544e-02 -9.99296069e-01 8.99160981e-01 -4.48141955e-02
1.09203374e+00 -7.77459323e-01 1.01778138e+00 3.08548778e-01
-1.41965115e+00 -8.76856148e-02 -5.27826011e-01 -9.71942991e-02
-4.29067641e-01 4.27215427e-01 -5.41047871e-01 2.04496205e-01
1.33515322e+00 5.31775296e-01 -8.69584918e-01 9.51170564e-01
2.23544598e-01 6.03409588e-01 -6.88943267e-01 2.11331137e-02
8.69443893e-01 2.99929529e-02 4.22621071e-02 1.64622021e+00
6.16182029e-01 -3.17577980e-02 -1.52536020e-01 4.98696357e-01
-1.35775253e-01 -1.44859165e-01 -3.53439718e-01 3.80051523e-01
2.16232613e-01 1.38691783e+00 -1.21895611e+00 -5.88006318e-01
-4.90118027e-01 8.83476019e-01 1.03359990e-01 2.08083361e-01
-9.28552389e-01 -4.68528390e-01 8.61614108e-01 9.73364785e-02
6.01252556e-01 -4.82443333e-01 -5.49623609e-01 -7.56197989e-01
1.01362178e-02 -6.92984343e-01 1.58427343e-01 -7.03183651e-01
-8.27309787e-01 9.91321385e-01 -6.58633113e-02 -9.39493895e-01
1.06192574e-01 -1.30603993e+00 -6.48496926e-01 5.99700093e-01
-1.49989235e+00 -1.27667880e+00 -6.43368483e-01 8.26159060e-01
5.57021499e-01 -2.55428165e-01 7.27113664e-01 3.91383469e-01
-4.95377332e-01 7.18391776e-01 1.98375806e-01 3.60983372e-01
3.00651610e-01 -1.09146571e+00 9.82289910e-01 5.66377699e-01
1.42192483e-01 2.82040328e-01 3.15535069e-01 9.32047889e-03
-1.27348149e+00 -1.26458204e+00 3.99991781e-01 -9.29022506e-02
5.42299628e-01 -5.10057449e-01 -7.85455823e-01 6.75415158e-01
1.41043633e-01 6.29015565e-01 3.94522041e-01 1.51830446e-02
-7.29306459e-01 -3.70198995e-01 -1.14769602e+00 5.75187147e-01
1.14354372e+00 -4.10965681e-01 -1.67544290e-01 3.67064714e-01
6.79021716e-01 -6.90545738e-01 -7.90714681e-01 3.46883506e-01
5.88098168e-01 -1.21467054e+00 1.29242277e+00 -2.34705672e-01
2.19599709e-01 5.70066869e-02 -1.76605687e-01 -1.05111575e+00
-3.25393826e-01 -4.65264797e-01 3.58433425e-01 7.58853436e-01
4.30543184e-01 -1.05565596e+00 1.09601021e+00 3.07004899e-01
-4.10553724e-01 -8.37633848e-01 -1.34766030e+00 -1.03537965e+00
1.98465154e-01 -8.29760790e-01 5.93638837e-01 7.94254780e-01
-6.98461533e-01 1.10717624e-01 2.76751071e-01 8.53503570e-02
4.63078052e-01 -2.39034265e-01 5.11542320e-01 -1.24212110e+00
-1.15252651e-01 -7.63369083e-01 -6.52528882e-01 -1.04938138e+00
-2.25501209e-01 -8.92607331e-01 -3.13476980e-01 -1.36681962e+00
-3.16670835e-01 -7.27996945e-01 -1.20796986e-01 9.77713227e-01
4.06276286e-01 1.03841615e+00 2.98710316e-01 1.69775769e-01
-4.03713524e-01 1.07293241e-01 1.16480458e+00 -1.29779071e-01
-4.62214388e-02 -9.45694745e-02 -2.63481766e-01 1.03005731e+00
8.36747944e-01 -3.18241179e-01 -4.04953420e-01 -9.26633596e-01
2.20629871e-01 -2.79793113e-01 6.61036670e-01 -1.69321859e+00
3.70607108e-01 2.53353655e-01 4.27030474e-01 -5.03540099e-01
5.77421248e-01 -7.55883992e-01 1.10968873e-01 7.50265181e-01
7.39215240e-02 5.09115219e-01 9.08113420e-01 5.49323969e-02
-1.12646990e-01 -2.88331900e-02 9.95434344e-01 -2.06897721e-01
-9.13045049e-01 3.58309805e-01 -3.94388378e-01 3.46002467e-02
8.54847312e-01 -6.44243956e-01 -5.72767496e-01 3.01839620e-01
-4.26478416e-01 -1.99320897e-01 1.94764391e-01 3.97751510e-01
3.48014653e-01 -1.14510489e+00 -6.88496709e-01 5.11803806e-01
-6.07600883e-02 3.53023201e-01 3.17478418e-01 7.69830883e-01
-1.24715292e+00 4.81772482e-01 -5.70572078e-01 -1.07175946e+00
-1.25788236e+00 1.41153544e-01 4.47033137e-01 -1.01209313e-01
-1.00080085e+00 1.12319934e+00 2.44477451e-01 -4.08717513e-01
5.79124608e-04 -8.74519587e-01 2.55167723e-01 -3.06493845e-02
5.05308568e-01 3.89242381e-01 6.65344596e-01 -3.00537825e-01
-4.04761463e-01 7.72152722e-01 3.42487916e-02 9.58416983e-02
1.57247758e+00 4.26896065e-01 -6.96496814e-02 -3.73000139e-03
1.35151851e+00 -6.71756327e-01 -1.47350216e+00 -1.88533232e-01
-1.45585462e-01 -1.23903014e-01 5.30911684e-01 -6.50711954e-01
-1.52864039e+00 1.09259140e+00 9.43250656e-01 2.15455696e-01
1.31571054e+00 -1.44963652e-01 1.04265213e+00 5.65453112e-01
3.38589221e-01 -1.06364655e+00 9.74882171e-02 6.52490377e-01
5.74137568e-01 -9.31599557e-01 -2.28225172e-01 -3.31799686e-01
-1.87662154e-01 1.32357049e+00 6.46048725e-01 -3.34672660e-01
6.31798744e-01 6.92859888e-01 -6.77576885e-02 -1.74271971e-01
-6.23632848e-01 -7.66063407e-02 2.95742720e-01 6.21470392e-01
1.71826631e-01 2.50547733e-02 3.12800109e-01 1.29427776e-01
-4.29032654e-01 -1.29991561e-01 6.71272576e-02 8.12454283e-01
-3.12997282e-01 -6.50427997e-01 -2.09454477e-01 4.37364280e-01
-4.51747060e-01 -2.54255056e-01 7.44892731e-02 1.09006119e+00
5.64643145e-01 5.12620807e-01 6.38561249e-01 -4.07047182e-01
2.81587124e-01 -9.82182473e-02 5.12660027e-01 -3.06164831e-01
-9.83710527e-01 -8.87684748e-02 -3.63307931e-02 -6.51572466e-01
-3.31758559e-01 -4.37709510e-01 -1.05779636e+00 -7.92001724e-01
-1.30261675e-01 -3.46518517e-01 1.04767454e+00 7.54962265e-01
3.90366435e-01 7.22027898e-01 -8.95347632e-03 -1.12974715e+00
-2.75857925e-01 -8.24152827e-01 -4.20092463e-01 1.17017575e-01
2.06109330e-01 -3.49410802e-01 -1.10443234e-01 -1.10465549e-02] | [9.075857162475586, 1.6443334817886353] |
d87f64d0-cd6f-4d6d-9b1d-5f3ecc8f2c48 | rapid-extraction-of-respiratory-waveforms | 2212.12578 | null | https://arxiv.org/abs/2212.12578v1 | https://arxiv.org/pdf/2212.12578v1.pdf | Rapid Extraction of Respiratory Waveforms from Photoplethysmography: A Deep Encoder Approach | Much of the information of breathing is contained within the photoplethysmography (PPG) signal, through changes in venous blood flow, heart rate and stroke volume. We aim to leverage this fact, by employing a novel deep learning framework which is a based on a repurposed convolutional autoencoder. Our model aims to encode all of the relevant respiratory information contained within photoplethysmography waveform, and decode it into a waveform that is similar to a gold standard respiratory reference. The model is employed on two photoplethysmography data sets, namely Capnobase and BIDMC. We show that the model is capable of producing respiratory waveforms that approach the gold standard, while in turn producing state of the art respiratory rate estimates. We also show that when it comes to capturing more advanced respiratory waveform characteristics such as duty cycle, our model is for the most part unsuccessful. A suggested reason for this, in light of a previous study on in-ear PPG, is that the respiratory variations in finger-PPG are far weaker compared with other recording locations. Importantly, our model can perform these waveform estimates in a fraction of a millisecond, giving it the capacity to produce over 6 hours of respiratory waveforms in a single second. Moreover, we attempt to interpret the behaviour of the kernel weights within the model, showing that in part our model intuitively selects different breathing frequencies. The model proposed in this work could help to improve the usefulness of consumer PPG-based wearables for medical applications, where detailed respiratory information is required. | ['Danilo P. Mandic', 'Harry J. Davies'] | 2022-12-22 | null | null | null | null | ['photoplethysmography-ppg'] | ['medical'] | [ 1.63287446e-01 1.80983782e-01 3.44045222e-01 -2.17056423e-01
-5.22442520e-01 -5.91117263e-01 8.47491995e-02 -1.72745466e-01
-2.78218627e-01 6.28417134e-01 4.05721366e-01 -9.74646062e-02
-7.67439604e-02 -5.85378587e-01 -3.66247833e-01 -9.17783201e-01
-1.65538937e-01 -7.77594447e-02 -9.43749547e-02 -7.90722966e-02
-2.47973911e-02 5.67758799e-01 -1.20855224e+00 7.19430000e-02
4.90958780e-01 1.19881141e+00 -1.86817437e-01 1.11792696e+00
4.03174490e-01 5.23526967e-01 -7.34582663e-01 -3.26302350e-01
2.55580544e-01 -5.83231747e-01 -2.44722724e-01 -4.07566905e-01
5.26351869e-01 -6.12510562e-01 -6.09906793e-01 3.66616607e-01
1.06794512e+00 -3.92440148e-02 5.20896673e-01 -7.21618414e-01
-2.34037384e-01 4.44276392e-01 -7.52406642e-02 3.72253299e-01
2.02279612e-01 4.27395433e-01 7.67546654e-01 -3.72968614e-01
2.32138753e-01 4.36504722e-01 1.16971898e+00 6.13613486e-01
-1.15039480e+00 -4.25801128e-01 -6.76107883e-01 -6.40323758e-02
-1.24974394e+00 -6.41448796e-01 9.05429959e-01 -3.30020905e-01
8.62047732e-01 3.58863711e-01 8.04035187e-01 1.11724997e+00
4.83035684e-01 8.02303925e-02 1.10708559e+00 -2.78799772e-01
1.98952302e-01 2.50133842e-01 -2.10301280e-01 3.25265229e-01
2.30087966e-01 3.13837707e-01 -5.39944410e-01 -1.29410058e-01
8.65438044e-01 -2.32592136e-01 -8.04678559e-01 1.23713121e-01
-8.58567357e-01 5.91710091e-01 1.95134029e-01 6.79778814e-01
-5.98347068e-01 5.53012073e-01 2.11456254e-01 9.91913229e-02
2.16259375e-01 5.92060506e-01 -3.91295552e-01 -7.39780545e-01
-1.10172427e+00 -4.23343778e-02 1.29330945e+00 3.57951671e-01
3.02649617e-01 1.19314946e-01 -2.12119415e-01 6.15615964e-01
3.83513629e-01 5.77235401e-01 6.23651385e-01 -1.17319965e+00
1.77067876e-01 2.09164061e-02 1.51172757e-01 -8.06228161e-01
-6.15515828e-01 -6.79260492e-02 -5.85049152e-01 -4.84235287e-02
6.42091870e-01 -3.65019619e-01 -5.24485707e-01 1.53078628e+00
9.42189246e-02 4.16909307e-01 -9.47860107e-02 1.12869906e+00
8.72932494e-01 2.38544360e-01 -1.60720013e-02 -2.50695884e-01
1.54543912e+00 -2.05286965e-01 -7.87681520e-01 -7.60086402e-02
2.22450450e-01 -5.78002334e-01 7.22161055e-01 3.92060399e-01
-1.11379600e+00 -6.38817549e-01 -1.04322219e+00 1.59681328e-02
-1.18288338e-01 -1.22439126e-02 4.90879416e-01 1.15430391e+00
-1.03883529e+00 1.18541336e+00 -8.76224756e-01 -1.50411248e-01
1.48811236e-01 2.26004541e-01 -6.81675449e-02 3.66790801e-01
-1.47024596e+00 9.02154028e-01 -1.02095135e-01 2.81170249e-01
-2.00075239e-01 -1.10162985e+00 -7.41108894e-01 2.60911375e-01
-2.29714543e-01 -5.94783902e-01 1.16738069e+00 -2.06913516e-01
-1.89577317e+00 5.82976282e-01 -5.95952421e-02 -5.24893641e-01
4.99102592e-01 -2.32360847e-02 -6.23068929e-01 6.76924407e-01
-7.10530639e-01 2.28505701e-01 1.01140261e+00 -6.91967070e-01
1.47936344e-01 -2.81339496e-01 -3.78357500e-01 -2.07468837e-01
-2.73252904e-01 -3.17945719e-01 -1.84953377e-01 -3.53693724e-01
-1.91294953e-01 -8.66704166e-01 1.58772349e-01 -7.17648417e-02
-2.02693380e-02 9.31706801e-02 4.31980550e-01 -7.43786156e-01
1.38800991e+00 -2.01736665e+00 -2.56995946e-01 2.86890239e-01
4.28849250e-01 4.22746569e-01 6.14570566e-02 5.16898572e-01
-1.21827774e-01 2.01923683e-01 -3.03824455e-01 -1.92789793e-01
-4.84707132e-02 8.48015547e-02 -3.27783555e-01 6.77747786e-01
8.98015723e-02 1.04668510e+00 -6.49920106e-01 -1.84236810e-01
6.09216332e-01 1.04837978e+00 -1.35991782e-01 4.81337905e-01
2.85198003e-01 4.91251141e-01 -1.74886957e-02 3.99819940e-01
5.41860342e-01 -1.69618949e-02 6.78669512e-02 -7.59522855e-01
-1.96722195e-01 5.19494653e-01 -7.78044164e-01 1.58935475e+00
-5.95834553e-01 1.03194976e+00 -2.33607337e-01 -6.95986569e-01
9.50081527e-01 7.75903761e-01 8.40454876e-01 -8.99104655e-01
4.24281269e-01 1.70126960e-01 3.25986505e-01 -9.06271040e-01
1.09213822e-01 -5.50261915e-01 4.37202692e-01 8.28428090e-01
4.94255647e-02 -4.72894877e-01 -9.67017338e-02 -2.41115987e-01
1.05855453e+00 5.02359532e-02 1.25238433e-01 -1.34197623e-01
2.80728102e-01 -6.43468082e-01 4.80642319e-02 8.65775645e-01
-5.30886769e-01 1.10739887e+00 4.32444036e-01 -4.21851367e-01
-1.10948884e+00 -9.13446546e-01 -4.71044242e-01 2.57027596e-01
-2.18228981e-01 -2.24159047e-01 -7.33561039e-01 -1.07923694e-01
1.25675276e-01 5.11296570e-01 -7.53252804e-01 -1.96422920e-01
-6.58111751e-01 -7.10424900e-01 9.58589494e-01 5.69790184e-01
1.45302907e-01 -1.26749623e+00 -1.15710068e+00 5.42305112e-01
-4.68273349e-02 -1.28089690e+00 -3.47824544e-01 1.42476529e-01
-9.48492944e-01 -1.09864068e+00 -9.48583782e-01 6.72014616e-03
-9.56648365e-02 -2.93214560e-01 1.09973335e+00 -2.04847842e-01
-7.95427799e-01 7.11995006e-01 -2.30531514e-01 -5.97697973e-01
-4.07410294e-01 -9.65660885e-02 3.78802717e-02 1.54799670e-01
5.11596143e-01 -1.00201011e+00 -1.16500378e+00 -1.55579187e-02
-6.70296550e-01 -4.48589504e-01 4.72767204e-01 3.19591641e-01
2.45851487e-01 -3.13378751e-01 5.65164328e-01 -6.00144565e-01
8.27625275e-01 -4.52145576e-01 -2.39388183e-01 -2.06973881e-01
-6.38008833e-01 -4.31527831e-02 6.74986243e-01 -5.30405581e-01
-6.64587915e-01 -3.36010098e-01 -3.07531983e-01 -4.37450379e-01
-1.63359165e-01 1.91073075e-01 5.18488348e-01 -2.17072904e-01
9.31843519e-01 4.03240025e-01 3.69615227e-01 -3.16702485e-01
1.18123353e-01 8.10405135e-01 6.54526711e-01 -2.41858616e-01
6.26034737e-01 4.39778298e-01 1.09729990e-01 -1.04827058e+00
-2.78770417e-01 -4.57053155e-01 -4.56335753e-01 -2.01948047e-01
6.92287147e-01 -6.49013519e-01 -1.27279484e+00 4.50507581e-01
-8.40825438e-01 -4.93645549e-01 -4.78810161e-01 7.42393970e-01
-7.10858881e-01 4.46783602e-01 -8.33351016e-01 -1.27126741e+00
-6.96476460e-01 -5.68256676e-01 8.65152597e-01 4.90358114e-01
-4.44652051e-01 -1.21964312e+00 3.76848906e-01 1.71373233e-01
1.09549224e+00 4.80753094e-01 5.28775394e-01 -5.22573829e-01
-1.92781135e-01 -4.42592114e-01 -3.82235758e-02 4.92132276e-01
3.52693439e-01 -4.15275656e-02 -1.55551398e+00 -2.04674333e-01
4.64215487e-01 -2.30794206e-01 6.49922550e-01 6.19315863e-01
1.15322936e+00 -1.02740861e-01 1.80233091e-01 5.65462589e-01
1.25920105e+00 2.69867420e-01 9.53738868e-01 -2.89159924e-01
5.21724164e-01 5.48997581e-01 1.72000304e-02 5.76214373e-01
1.30483791e-01 5.39060056e-01 1.35151327e-01 -2.66550213e-01
-2.16617048e-01 5.58964815e-03 1.63132891e-01 7.61528134e-01
-2.68389046e-01 -3.50429825e-02 -5.29309809e-01 4.61048782e-01
-1.25906432e+00 -9.79893327e-01 -2.94562802e-02 2.44980621e+00
1.00010443e+00 -2.35230729e-01 2.64423102e-01 2.20675111e-01
3.88337880e-01 2.60902762e-01 -6.48832977e-01 -7.39139855e-01
3.23213637e-01 8.53281260e-01 6.15981042e-01 3.03414881e-01
-6.32878184e-01 2.85678562e-02 6.97813940e+00 -1.09007865e-01
-1.66571128e+00 -2.41960064e-01 2.96191335e-01 -6.81032464e-02
-3.33553880e-01 -3.53825599e-01 -3.49883109e-01 8.87038112e-01
1.59245551e+00 -1.95359290e-01 6.16924584e-01 4.11424011e-01
5.75778365e-01 -4.25646007e-02 -1.20463490e+00 1.23467267e+00
2.29437649e-01 -1.07044387e+00 -5.86087525e-01 2.96287723e-02
-3.33026052e-02 4.09002192e-02 -4.63920832e-02 3.42901275e-02
-5.49501002e-01 -1.31321180e+00 9.56205353e-02 7.22099066e-01
1.07811832e+00 -4.83968586e-01 6.65025353e-01 -1.70991644e-02
-9.20212269e-01 7.58106709e-02 -2.65876889e-01 1.03605874e-01
9.13007334e-02 8.21774662e-01 -1.06890500e+00 3.22997898e-01
4.63268012e-01 4.10771102e-01 -2.72936374e-01 1.22899675e+00
-8.90057683e-02 9.07946169e-01 -6.76026523e-01 -2.75715981e-02
-1.66934446e-01 -2.66384613e-02 3.94736618e-01 1.18478787e+00
4.64075357e-01 4.22964573e-01 -6.79630399e-01 1.23030031e+00
1.83428936e-02 -2.67240256e-01 -6.88461661e-01 -3.05608481e-01
5.08805215e-01 1.34078109e+00 -1.27170444e-01 -1.17066808e-01
-3.77869308e-01 6.49884641e-01 -3.73595595e-01 4.94482964e-01
-7.84002364e-01 -7.88829982e-01 4.71495211e-01 2.57166624e-01
2.12322056e-01 -2.26583090e-02 -3.80039155e-01 -9.50797379e-01
1.26140490e-01 -5.74447453e-01 4.82101515e-02 -8.63344073e-01
-1.10952032e+00 5.65176189e-01 -2.37186357e-01 -1.08912694e+00
-6.03784502e-01 -5.71560740e-01 -7.83710599e-01 1.23565125e+00
-1.69053912e+00 -4.98868346e-01 -6.15635216e-01 4.36238974e-01
-2.74339803e-02 3.40070218e-01 9.83969212e-01 4.09055561e-01
-2.34837443e-01 6.00398302e-01 -4.58798081e-01 7.15256333e-02
8.47141385e-01 -1.44689524e+00 4.35612589e-01 4.82236564e-01
3.49934548e-02 8.90161812e-01 8.51076782e-01 -1.03199951e-01
-1.56084716e+00 -6.63084865e-01 7.89077342e-01 -6.29446208e-01
5.71978986e-01 -7.60052130e-02 -1.00899839e+00 1.44436702e-01
1.06824443e-01 3.20766717e-01 9.57479000e-01 -2.92309165e-01
-1.74553871e-01 -3.45753670e-01 -1.34258068e+00 1.60408556e-01
3.55138659e-01 -8.50898683e-01 -7.66995788e-01 -1.51406899e-02
4.59941834e-01 -6.12493992e-01 -1.43304014e+00 2.52955437e-01
1.11028433e+00 -1.03179681e+00 7.89859474e-01 -7.92418942e-02
4.09533173e-01 -9.61597934e-02 2.86675870e-01 -1.36113286e+00
-3.92922983e-02 -1.03273356e+00 -5.17410934e-01 9.44976628e-01
3.10212839e-02 -9.16988254e-01 7.12698042e-01 1.08910429e+00
2.79820804e-02 -7.04680026e-01 -8.61178100e-01 -5.20962596e-01
-7.80112110e-03 -4.99939650e-01 3.01292330e-01 7.22194731e-01
2.98634470e-02 -4.93944883e-02 -5.49257338e-01 -2.66425405e-02
6.35513663e-01 -1.23807276e-02 3.45447123e-01 -1.15470648e+00
-5.05716145e-01 -1.26539424e-01 -2.03337535e-01 -5.83949208e-01
-3.58530462e-01 -3.58830035e-01 1.26108587e-01 -1.02229381e+00
-2.49448121e-01 -2.44551927e-01 -5.92709541e-01 4.37462896e-01
-9.31250080e-02 6.96236789e-01 1.43905118e-01 -1.13316342e-01
2.51416117e-01 6.63022101e-02 1.16209996e+00 2.88209498e-01
-6.33559883e-01 7.15538859e-03 -6.95126534e-01 2.50061542e-01
7.51585364e-01 -3.22764188e-01 -2.99404204e-01 7.95329288e-02
5.95296659e-02 3.74991030e-01 4.10218179e-01 -1.06658471e+00
2.45128021e-01 2.25140944e-01 6.31333947e-01 -1.24812759e-01
3.83833796e-01 -8.14334989e-01 2.63129264e-01 4.08060491e-01
-2.52551883e-01 -2.55985439e-01 3.78356308e-01 2.26833463e-01
-7.29772896e-02 -1.73065290e-01 8.22516680e-01 -1.81330770e-01
-1.14077032e-01 2.82375008e-01 -2.84664094e-01 1.14824802e-01
3.56955826e-01 -5.04029870e-01 -3.51918131e-01 -6.00302756e-01
-5.79522014e-01 -3.04483324e-01 1.43977255e-01 2.45124772e-01
4.43774819e-01 -1.07018387e+00 -5.95198810e-01 3.96157563e-01
6.12202808e-02 -5.19400299e-01 4.72940266e-01 1.20192611e+00
-5.47753453e-01 5.21847606e-01 -8.79993215e-02 -5.52423477e-01
-8.21799517e-01 3.81405652e-02 8.52175593e-01 1.35210916e-01
-8.61337960e-01 4.49286580e-01 -2.83575416e-01 2.84740657e-01
6.43610060e-02 -4.15373176e-01 -2.21102268e-01 -1.03572821e-02
7.84530878e-01 2.15672418e-01 2.63750046e-01 -4.36519124e-02
-3.28392953e-01 7.88775623e-01 3.85210335e-01 -1.90674998e-02
1.24091840e+00 -4.19694111e-02 2.02484533e-01 4.21547771e-01
1.27075577e+00 2.69606858e-01 -1.29981792e+00 2.87623286e-01
-3.99350137e-01 -4.04102743e-01 9.63268504e-02 -1.05963171e+00
-1.07421064e+00 1.07601094e+00 9.17071640e-01 5.14860511e-01
1.22868586e+00 -4.43030238e-01 1.03424644e+00 2.20269203e-01
5.88625558e-02 -9.28806841e-01 1.32639427e-02 -1.18733816e-01
6.31540895e-01 -8.25344086e-01 -5.60858697e-02 -3.09523921e-02
-4.17999625e-01 1.46806288e+00 7.70642534e-02 -1.91886842e-01
3.74622136e-01 4.51985717e-01 3.97898585e-01 -2.05743447e-01
-5.46080709e-01 -1.45168491e-02 2.69397944e-01 7.50775516e-01
7.46788979e-01 -9.84330028e-02 -4.07429665e-01 4.36012894e-01
-2.79798895e-01 4.16633844e-01 7.52152622e-01 3.47409666e-01
-2.11190775e-01 -9.21280324e-01 -2.06819363e-02 4.34010744e-01
-8.57024252e-01 -1.54413760e-01 -1.18412212e-01 7.19222426e-01
-2.28991702e-01 9.34728205e-01 1.53201580e-01 -1.71738803e-01
3.98339152e-01 3.78817916e-01 6.95663273e-01 -3.20484370e-01
-7.29654253e-01 1.99567396e-02 -3.27274650e-02 -7.00339496e-01
-3.45836431e-01 -1.48344547e-01 -1.03263628e+00 -1.60674781e-01
-3.72883528e-02 6.87045902e-02 9.55164075e-01 7.41111219e-01
3.80975038e-01 5.73787689e-01 6.24299109e-01 -7.98943520e-01
-5.20164490e-01 -1.02094758e+00 -7.75162160e-01 5.79833269e-01
6.63916826e-01 -1.39216498e-01 -8.24348629e-01 -5.67892753e-02] | [13.914237976074219, 2.9271230697631836] |
db3543a5-cc1a-4825-85ad-8372b357d681 | real-time-semantic-segmentation-using | 2303.15623 | null | https://arxiv.org/abs/2303.15623v1 | https://arxiv.org/pdf/2303.15623v1.pdf | Real-Time Semantic Segmentation using Hyperspectral Images for Mapping Unstructured and Unknown Environments | Autonomous navigation in unstructured off-road environments is greatly improved by semantic scene understanding. Conventional image processing algorithms are difficult to implement and lack robustness due to a lack of structure and high variability across off-road environments. The use of neural networks and machine learning can overcome the previous challenges but they require large labeled data sets for training. In our work we propose the use of hyperspectral images for real-time pixel-wise semantic classification and segmentation, without the need of any prior training data. The resulting segmented image is processed to extract, filter, and approximate objects as polygons, using a polygon approximation algorithm. The resulting polygons are then used to generate a semantic map of the environment. Using our framework. we show the capability to add new semantic classes in run-time for classification. The proposed methodology is also shown to operate in real-time and produce outputs at a frequency of 1Hz, using high resolution hyperspectral images. | ['Swaminathan Gopalswamy', 'Reza Langari', 'Anant Bhamri', 'Anthony Medellin'] | 2023-03-27 | null | null | null | null | ['real-time-semantic-segmentation'] | ['computer-vision'] | [ 1.01550901e+00 9.64078456e-02 4.46012110e-01 -5.13047755e-01
-2.87948042e-01 -6.49906576e-01 3.94442648e-01 3.18645269e-01
-5.00127077e-01 7.32788682e-01 -5.48946857e-01 -3.73397380e-01
-4.00130451e-01 -1.36697173e+00 -5.95400572e-01 -5.11921465e-01
-5.56325093e-02 6.27799928e-01 3.43673795e-01 -2.00314924e-01
2.41512209e-01 8.66957307e-01 -2.15151644e+00 1.72886103e-01
1.31984544e+00 1.03289843e+00 6.92230582e-01 7.70039380e-01
-2.81926990e-01 3.39836150e-01 -4.33457762e-01 4.75181162e-01
5.34152985e-01 5.07700220e-02 -5.72067440e-01 4.55034971e-01
3.95271868e-01 -1.89523578e-01 2.35518720e-02 1.29614079e+00
2.70089179e-01 3.76123846e-01 7.24888325e-01 -8.17329645e-01
1.04026012e-01 1.20311849e-01 -1.86818853e-01 -4.47221160e-01
4.40719351e-02 -1.43217063e-02 3.78756911e-01 -7.08126426e-01
4.68509883e-01 8.54295313e-01 8.57846141e-01 -5.20813502e-02
-1.30380285e+00 -3.73203635e-01 -1.10671185e-02 1.77072585e-01
-1.36174417e+00 -2.31343582e-01 5.93694568e-01 -4.32895154e-01
1.03199017e+00 2.50835329e-01 7.93638468e-01 1.77714720e-01
-1.55565903e-01 2.46009454e-01 1.37822425e+00 -5.45443535e-01
5.85510552e-01 1.85620174e-01 8.38228166e-02 7.01904774e-01
3.78978431e-01 2.11356506e-02 -5.60566895e-02 2.68479675e-01
6.92643702e-01 6.39118953e-03 -3.56761575e-01 -1.76415354e-01
-7.57989645e-01 6.96351230e-01 6.65395379e-01 1.47489995e-01
-7.04541266e-01 -1.39535740e-02 6.59255162e-02 -1.29846940e-02
4.89495605e-01 4.54372853e-01 -4.37052995e-01 1.98290363e-01
-1.09259856e+00 1.52275980e-01 6.42807186e-01 7.66557813e-01
1.27708089e+00 1.92086205e-01 5.79940081e-01 9.08885479e-01
1.09659910e-01 8.45379233e-01 3.86900753e-01 -1.06977212e+00
9.21607614e-02 5.27460039e-01 1.90618664e-01 -1.01516771e+00
-8.03385675e-01 -2.89420784e-01 -4.92720991e-01 8.33428562e-01
3.79144073e-01 -1.82474509e-01 -1.39746845e+00 9.56954062e-01
2.70354122e-01 -6.18429072e-02 4.16357785e-01 6.21683538e-01
5.16344309e-01 1.01499057e+00 -4.30970490e-02 -1.46858662e-03
1.09166229e+00 -7.97700286e-01 -4.97089714e-01 -3.97865713e-01
3.95168096e-01 -5.66749394e-01 9.20757830e-01 6.55916035e-01
-5.36874235e-01 -5.35615504e-01 -1.15663135e+00 1.69621766e-01
-1.02758729e+00 4.00377780e-01 6.25853658e-01 8.13566089e-01
-9.97603714e-01 7.02826142e-01 -7.77826786e-01 -5.26483059e-01
4.00455922e-01 5.80738664e-01 -2.80654132e-01 8.16328377e-02
-8.05200636e-01 9.35117066e-01 8.32558811e-01 3.26114893e-01
-3.49675983e-01 -4.64755028e-01 -8.85170400e-01 2.43761744e-02
2.28645548e-01 -2.99194485e-01 1.03812027e+00 -1.16686916e+00
-1.65811419e+00 5.48324525e-01 -1.09702811e-01 -4.08970833e-01
4.37585205e-01 -1.12961516e-01 -2.53646761e-01 4.75702018e-01
1.17892504e-01 7.76244938e-01 6.69619322e-01 -1.42218876e+00
-1.08204520e+00 -4.31632906e-01 8.09667632e-02 3.76161635e-01
-1.09424613e-01 -7.15948105e-01 -1.24560691e-01 -2.22355217e-01
6.34925127e-01 -9.26061392e-01 -4.93378848e-01 2.24627350e-02
-2.07000747e-01 3.44493508e-01 1.13046145e+00 -6.80240989e-01
3.96784395e-01 -1.87012935e+00 -4.23068970e-01 7.22251892e-01
-2.40551203e-01 3.81484151e-01 1.83715997e-03 2.38451719e-01
4.68674228e-02 -1.51160464e-01 -6.87935174e-01 1.57414123e-01
-2.92729795e-01 2.83471078e-01 -4.66175117e-02 3.11352879e-01
1.06136218e-01 3.04726601e-01 -7.58795142e-01 -1.94386005e-01
8.00077379e-01 5.07277966e-01 -2.14465186e-01 -1.47450030e-01
-3.63820314e-01 2.32280195e-01 -3.46975356e-01 4.47231859e-01
9.18221414e-01 1.55400231e-01 1.28579006e-01 -2.89384127e-01
-5.82780719e-01 4.40733619e-02 -1.46041083e+00 1.39119959e+00
-6.86959445e-01 8.97784591e-01 2.98478782e-01 -1.25951326e+00
1.25853479e+00 -2.31849384e-02 3.94253671e-01 -7.24806488e-01
1.44418433e-01 5.18740356e-01 -3.61522704e-01 -6.36595905e-01
8.16270888e-01 -8.15396756e-02 4.28116351e-01 2.59389937e-01
-3.24595004e-01 -7.62178004e-01 1.78359538e-01 -4.08749282e-01
5.24108827e-01 3.28007936e-01 -4.05864082e-02 -3.72093618e-01
6.34839892e-01 7.72338629e-01 5.34045734e-02 7.79105961e-01
2.13125065e-01 4.90075648e-01 -1.32049978e-01 -5.21776855e-01
-1.13651919e+00 -8.29580963e-01 -3.79080415e-01 7.44837344e-01
4.15503561e-01 4.00460064e-01 -8.84070635e-01 1.52580524e-02
2.43249116e-04 7.01315701e-01 -1.89721346e-01 3.10327619e-01
-2.35613972e-01 -9.30018544e-01 2.02077106e-01 1.90641955e-01
7.68879533e-01 -1.10794795e+00 -1.08401477e+00 5.88138044e-01
1.26794642e-02 -1.19937098e+00 6.72289789e-01 4.23898995e-01
-1.07203794e+00 -9.20884669e-01 -3.34542394e-01 -8.75348508e-01
8.69906425e-01 4.37883556e-01 5.20581663e-01 -2.39815623e-01
-6.28930628e-01 2.17352882e-01 -2.88299471e-01 -6.21947587e-01
-2.12720007e-01 7.24514062e-03 -2.13446662e-01 -5.06643280e-02
3.61955278e-02 -6.57179534e-01 -4.63599741e-01 9.47333947e-02
-9.80137467e-01 3.79459202e-01 5.59361875e-01 6.00899875e-01
8.29029560e-01 7.69999504e-01 5.44999659e-01 -1.05782616e+00
2.55230814e-01 -9.85720158e-02 -1.25648069e+00 6.52766898e-02
-5.06891668e-01 -6.46337271e-02 7.04016268e-01 -1.09727450e-01
-1.30365562e+00 6.94045961e-01 -3.62263173e-02 2.66861200e-01
-6.73271656e-01 7.69983947e-01 4.16472089e-03 -5.26249111e-01
9.01348829e-01 1.86993718e-01 2.10777566e-01 -1.64184824e-01
3.83437932e-01 9.86078918e-01 7.69254446e-01 -2.41732627e-01
7.30325043e-01 8.52009296e-01 1.97273597e-01 -1.49353552e+00
-4.36040312e-01 -5.03511786e-01 -7.02224374e-01 -4.76098180e-01
7.33439684e-01 -8.72812569e-01 -2.45215133e-01 5.18575788e-01
-8.10843706e-01 -5.94770908e-01 -1.77285925e-01 5.73644221e-01
-6.16071403e-01 2.63502419e-01 -4.02464345e-02 -1.02812231e+00
-3.48547280e-01 -9.46221471e-01 7.58936048e-01 4.97675657e-01
2.14960754e-01 -8.42752278e-01 -4.93534714e-01 3.48167926e-01
3.66120696e-01 3.76715600e-01 7.04437435e-01 -1.65221244e-01
-7.05371916e-01 -4.01370555e-01 -5.09020388e-01 3.41774225e-01
1.97145611e-01 1.19720004e-01 -1.30492401e+00 2.30240077e-01
-8.40536878e-02 -1.59723654e-01 7.24226415e-01 6.86746657e-01
1.19652498e+00 9.25094336e-02 -3.13650817e-01 6.09284878e-01
1.97904193e+00 4.51732188e-01 6.99564457e-01 5.75027704e-01
5.66785157e-01 1.08462489e+00 8.48679066e-01 3.22837293e-01
1.21259749e-01 2.34354883e-01 6.38601780e-01 -3.34794194e-01
7.72389099e-02 3.37242126e-01 -2.46035621e-01 1.18748941e-01
-2.38254398e-01 -9.37776789e-02 -1.14117849e+00 6.24044299e-01
-1.76657462e+00 -9.55208063e-01 -6.61910176e-01 2.26299095e+00
3.00732940e-01 1.39227659e-01 -2.69594908e-01 3.44388485e-01
7.10148990e-01 -2.61652291e-01 -3.42319608e-01 -2.96054959e-01
2.12667007e-02 4.54013050e-01 1.18397903e+00 8.44977975e-01
-1.29315484e+00 1.05960739e+00 5.89058113e+00 5.09553492e-01
-1.45760095e+00 -3.98250371e-01 3.24591547e-01 3.88307542e-01
-2.97743790e-02 -1.30405813e-01 -3.81410003e-01 5.40841222e-02
9.30992305e-01 1.52438685e-01 7.09517956e-01 7.28089035e-01
6.14702344e-01 -7.62001455e-01 -3.54775995e-01 8.46870661e-01
-1.57420576e-01 -1.15555024e+00 -1.18373401e-01 -1.63925096e-01
7.16812491e-01 1.80280149e-01 -2.48334378e-01 -2.06789002e-01
1.22566015e-01 -1.00672376e+00 7.11376786e-01 5.61443031e-01
6.12716079e-01 -8.79293382e-01 6.84907079e-01 3.74726444e-01
-1.27689242e+00 -2.18793556e-01 -5.88955402e-01 -2.75384396e-01
-1.12984022e-02 6.10962749e-01 -1.03583384e+00 6.67823553e-01
5.06659985e-01 3.43091637e-01 -4.01860625e-01 1.22205746e+00
-1.24766894e-01 3.08236182e-01 -8.08523834e-01 -1.43910408e-01
4.26766723e-01 -5.94866097e-01 2.23690689e-01 1.08755398e+00
5.97595394e-01 1.42490789e-01 3.33509356e-01 5.69305956e-01
5.46213567e-01 1.79705009e-01 -7.36539423e-01 6.36852831e-02
3.52512807e-01 1.23403919e+00 -1.28520024e+00 -5.54052889e-01
-2.51973271e-01 8.80481541e-01 -1.97789937e-01 4.63047892e-01
-4.94882673e-01 -8.73143673e-01 3.17132771e-01 6.61131218e-02
3.22018296e-01 -3.94430101e-01 -7.06798613e-01 -6.03597224e-01
-7.80685619e-02 -4.92238939e-01 -7.74534419e-02 -1.13843787e+00
-7.58562803e-01 4.81715381e-01 -1.43659636e-01 -1.01118767e+00
-2.78638918e-02 -1.02158642e+00 -2.18476489e-01 1.04759216e+00
-1.57766163e+00 -1.09726906e+00 -1.04220307e+00 4.06880617e-01
5.41973293e-01 2.05301568e-02 1.08241320e+00 8.68651047e-02
-7.91602731e-02 -4.31953430e-01 4.62015867e-01 -2.58708507e-01
1.35059252e-01 -1.32523608e+00 6.14509284e-02 8.63605678e-01
-2.10926428e-01 -3.99900600e-02 9.10845041e-01 -6.86959147e-01
-1.02773881e+00 -1.43417454e+00 4.81781304e-01 2.90558964e-01
4.51706439e-01 -8.28951672e-02 -8.82020831e-01 1.96217552e-01
-2.01491490e-01 -1.20637402e-01 4.63548839e-01 -2.96267509e-01
9.07793120e-02 -3.04788023e-01 -1.34195507e+00 4.24477696e-01
7.15586245e-01 -5.81279874e-01 -3.86443853e-01 5.35198808e-01
2.37822622e-01 -2.93665498e-01 -4.89307374e-01 5.18952250e-01
6.84157073e-01 -1.06677473e+00 8.17528129e-01 1.89394832e-01
2.05802545e-01 -6.40344322e-01 2.06540022e-02 -1.17321789e+00
-9.64164808e-02 -7.50271305e-02 9.61186171e-01 6.44243240e-01
7.43792474e-01 -9.12972212e-01 1.11819792e+00 4.84167695e-01
-4.68626052e-01 -6.06393181e-02 -5.79588830e-01 -6.90269232e-01
-4.27508563e-01 -5.77048242e-01 4.20671105e-01 6.99113965e-01
-2.82151699e-01 -2.45539546e-02 1.87667593e-01 9.19131935e-01
6.72383606e-01 3.44735354e-01 6.86290443e-01 -1.51432633e+00
7.07839355e-02 -3.00139248e-01 -5.84837198e-01 -5.22392094e-01
-1.09522603e-02 -6.98387146e-01 5.69546998e-01 -2.11901021e+00
-4.53447402e-01 -7.26151645e-01 2.39933565e-01 5.18600047e-01
1.96245790e-01 4.99424100e-01 -2.30196446e-01 7.09615201e-02
9.04418975e-02 3.02024096e-01 7.65155733e-01 -2.09061757e-01
-5.64177990e-01 -7.31766447e-02 -4.63772938e-02 7.71132290e-01
1.08682299e+00 -3.61331254e-01 -6.68269277e-01 -3.33666354e-01
-7.47215897e-02 -3.36685181e-01 2.99740374e-01 -1.63401055e+00
1.29597262e-01 -1.25085384e-01 4.79669422e-01 -7.79677927e-01
5.04769742e-01 -1.29154229e+00 4.65327531e-01 6.31323516e-01
1.64796159e-01 -4.72827137e-01 3.84218782e-01 4.57082927e-01
-1.77721858e-01 -6.46597266e-01 8.47369969e-01 -2.61833608e-01
-1.13741207e+00 -4.46263850e-02 -8.19302320e-01 -7.41299450e-01
1.08187485e+00 -7.51545548e-01 -1.39209569e-01 -2.10661933e-01
-8.82788301e-01 4.75808121e-02 5.90246618e-01 -2.68745661e-01
3.69292468e-01 -6.14012301e-01 -3.14249814e-01 2.76598215e-01
3.53767425e-02 3.70604157e-01 2.39765942e-01 2.52842933e-01
-1.54052806e+00 2.51026839e-01 -5.43051839e-01 -7.15265512e-01
-1.10270905e+00 9.70411599e-02 6.01913452e-01 2.52143264e-01
-7.24901438e-01 4.92306441e-01 -1.35278583e-01 -6.48066103e-01
-8.09275582e-02 -4.95398253e-01 -3.77800852e-01 6.43967539e-02
3.04252446e-01 2.63901383e-01 3.11726123e-01 -4.77813035e-01
-5.34083508e-03 6.90381765e-01 6.45033777e-01 -3.75425041e-01
1.49834454e+00 -9.31860879e-02 -1.29813701e-01 2.22792834e-01
7.16219425e-01 -9.42887366e-02 -1.26837039e+00 1.06128938e-01
9.33560580e-02 -4.30647850e-01 2.46706322e-01 -7.94461429e-01
-7.92504907e-01 7.71029234e-01 8.47193837e-01 3.64505082e-01
1.36244333e+00 -6.17483616e-01 3.63614976e-01 7.91185200e-01
4.74288195e-01 -1.49369550e+00 -5.45293570e-01 6.11048818e-01
4.53376204e-01 -1.14189553e+00 1.26379147e-01 -9.39296544e-01
-1.85167909e-01 1.65195954e+00 2.87436485e-01 -9.41290185e-02
6.11558914e-01 1.57213867e-01 4.08780098e-01 -1.72467798e-01
8.29169154e-02 -6.03689849e-01 -7.63827264e-02 1.02530992e+00
4.75502461e-02 2.89046079e-01 -6.49250895e-02 -1.13677569e-01
-4.41433311e-01 7.73047097e-03 8.56313229e-01 1.04395056e+00
-1.11595547e+00 -8.54649723e-01 -6.73197448e-01 5.04203796e-01
3.49831171e-02 7.04312176e-02 -1.28688067e-01 7.01268673e-01
4.48610127e-01 1.04411232e+00 2.38308445e-01 -1.34847522e-01
3.12130302e-01 1.31332219e-01 2.73390234e-01 -6.65844798e-01
-1.83097467e-01 1.89377125e-02 4.14464563e-01 -3.06889027e-01
-5.15609503e-01 -5.24491608e-01 -1.48835969e+00 1.26322076e-01
-2.16576159e-01 1.26143143e-01 1.41946101e+00 8.34932566e-01
1.93691790e-01 4.35332179e-01 3.67991954e-01 -1.28415978e+00
-5.16483895e-02 -8.56041133e-01 -7.40483761e-01 1.09412350e-01
1.83557332e-01 -5.47407866e-01 -2.60021538e-01 4.04180914e-01] | [9.274158477783203, -1.5775575637817383] |
d3c43917-ef30-462c-bff7-b66307426df0 | hybrid-rl-using-both-offline-and-online-data | 2210.06718 | null | https://arxiv.org/abs/2210.06718v3 | https://arxiv.org/pdf/2210.06718v3.pdf | Hybrid RL: Using Both Offline and Online Data Can Make RL Efficient | We consider a hybrid reinforcement learning setting (Hybrid RL), in which an agent has access to an offline dataset and the ability to collect experience via real-world online interaction. The framework mitigates the challenges that arise in both pure offline and online RL settings, allowing for the design of simple and highly effective algorithms, in both theory and practice. We demonstrate these advantages by adapting the classical Q learning/iteration algorithm to the hybrid setting, which we call Hybrid Q-Learning or Hy-Q. In our theoretical results, we prove that the algorithm is both computationally and statistically efficient whenever the offline dataset supports a high-quality policy and the environment has bounded bilinear rank. Notably, we require no assumptions on the coverage provided by the initial distribution, in contrast with guarantees for policy gradient/iteration methods. In our experimental results, we show that Hy-Q with neural network function approximation outperforms state-of-the-art online, offline, and hybrid RL baselines on challenging benchmarks, including Montezuma's Revenge. | ['Wen Sun', 'Akshay Krishnamurthy', 'J. Andrew Bagnell', 'Ayush Sekhari', 'Yifei Zhou', 'Yuda Song'] | 2022-10-13 | null | null | null | null | ['montezumas-revenge'] | ['playing-games'] | [-4.02232766e-01 5.84072992e-02 -5.44183254e-01 1.78368479e-01
-1.15507317e+00 -8.51182818e-01 5.96139967e-01 1.48219094e-01
-8.21263671e-01 1.17998636e+00 2.62198776e-01 -6.34023547e-01
-4.24922913e-01 -7.73601234e-01 -1.13374054e+00 -6.93881989e-01
-6.81453049e-01 6.14948392e-01 -1.29116476e-01 -1.85817957e-01
8.81798118e-02 7.96323717e-02 -1.38996899e+00 -3.72493893e-01
8.32608163e-01 1.20017612e+00 5.54569997e-02 8.63236964e-01
3.42144817e-01 1.07787430e+00 -3.25446546e-01 -2.78538138e-01
9.82391953e-01 -4.07034248e-01 -4.76517171e-01 1.02181800e-01
2.04408795e-01 -1.10721791e+00 -3.78175884e-01 1.10526061e+00
6.51082873e-01 4.59505737e-01 2.91906714e-01 -1.50107372e+00
-3.64932716e-01 6.95929468e-01 -5.21507382e-01 -3.15927863e-02
5.78386903e-01 5.77371955e-01 1.30224586e+00 -5.24757206e-01
6.52298987e-01 1.21881628e+00 3.86525750e-01 3.62348408e-01
-1.23449206e+00 -5.98491549e-01 4.45364982e-01 2.17240021e-01
-5.21240056e-01 -3.09475601e-01 3.03263843e-01 -2.75014579e-01
6.07847452e-01 -1.39850244e-01 8.58094931e-01 1.20151353e+00
-3.32488120e-01 1.34741855e+00 1.57734978e+00 -2.87875682e-01
7.12603688e-01 3.59134711e-02 -2.75912404e-01 7.65003145e-01
1.07977852e-01 8.61423492e-01 -6.12433672e-01 -4.79266346e-01
8.36969793e-01 -1.83355790e-02 -1.19641609e-01 -8.26746285e-01
-1.08735263e+00 9.38919663e-01 1.94496557e-01 -4.15496200e-01
-7.26070166e-01 3.57809335e-01 5.47338068e-01 8.66225600e-01
3.25741261e-01 4.14142966e-01 -4.83628124e-01 -5.38249910e-01
-5.50141096e-01 7.26422012e-01 1.09222198e+00 8.61254334e-01
6.87902093e-01 -8.77848938e-02 -2.32564375e-01 2.70379931e-01
5.11005744e-02 7.07621515e-01 1.76488802e-01 -1.60642922e+00
7.78479695e-01 -7.85742551e-02 1.00645494e+00 -4.30772185e-01
-2.08978832e-01 -4.71479714e-01 -3.65714878e-01 3.16174299e-01
7.02865481e-01 -6.13796830e-01 -3.36069584e-01 1.99446106e+00
6.09431446e-01 2.70824373e-01 2.28842750e-01 8.51893425e-01
-3.92093137e-02 5.25348127e-01 -2.93840706e-01 -5.93744099e-01
6.98605180e-01 -1.02953851e+00 -6.48534656e-01 1.89930424e-02
5.17102480e-01 -1.10526323e-01 1.33603358e+00 6.54212356e-01
-1.27540767e+00 -3.20972614e-02 -8.42020750e-01 3.04537773e-01
-8.19926783e-02 -2.71040678e-01 8.04937720e-01 5.08476377e-01
-1.08604121e+00 7.60694206e-01 -8.70283127e-01 -1.97244391e-01
4.53569263e-01 2.59202540e-01 -1.46411315e-01 -4.65266481e-02
-1.12905359e+00 7.09567666e-01 2.86771506e-01 -2.25212649e-01
-1.45635045e+00 -6.45624995e-01 -5.44808626e-01 2.30678897e-02
1.27865553e+00 -5.59213400e-01 1.91088283e+00 -9.93468344e-01
-2.07035947e+00 1.66852087e-01 3.02410662e-01 -8.55158865e-01
1.25969779e+00 -4.67741162e-01 2.15359211e-01 2.49999747e-01
-8.06002878e-03 1.11718051e-01 6.72403634e-01 -1.02342010e+00
-1.00860107e+00 -2.76363432e-01 6.30686820e-01 5.58723927e-01
-1.53130338e-01 -1.73918754e-01 -5.32464050e-02 -1.14457779e-01
-6.62755847e-01 -8.35801661e-01 -5.65768838e-01 1.39347285e-01
-2.10965835e-02 -2.26105183e-01 4.30011451e-01 -4.51039374e-01
7.15845346e-01 -1.88983357e+00 -5.98277785e-02 3.86128545e-01
5.55607677e-02 -2.88743805e-02 -2.53971815e-01 8.48446548e-01
5.42013884e-01 -9.33948904e-02 7.29591176e-02 -5.12558043e-01
6.82885289e-01 3.76396984e-01 -5.76888382e-01 8.48068655e-01
-4.65815395e-01 9.07227755e-01 -1.47656572e+00 -3.29953939e-01
9.02267247e-02 -4.09951180e-01 -9.10736024e-01 5.60975254e-01
-5.72348833e-01 5.24321616e-01 -5.57372689e-01 3.37561816e-01
4.63902980e-01 -3.04504365e-01 5.51300228e-01 6.51995540e-01
-1.56220764e-01 3.64297807e-01 -1.30021465e+00 1.55565202e+00
-5.78498125e-01 1.61177143e-01 5.20411432e-01 -9.35310781e-01
1.54726028e-01 2.96384871e-01 6.64915681e-01 -9.67243731e-01
6.96092844e-04 1.84961408e-01 -3.45022708e-01 -3.80265385e-01
4.65667933e-01 9.03078243e-02 2.63187056e-03 9.62050557e-01
9.35645252e-02 9.51747969e-02 3.89605880e-01 3.11917812e-01
1.13301945e+00 4.03726280e-01 4.22063231e-01 -5.46962284e-02
-9.92231071e-02 -2.58481532e-01 6.11837924e-01 1.49147677e+00
-5.53719580e-01 -2.15838015e-01 9.26735997e-01 -1.00672856e-01
-1.03743553e+00 -1.19506097e+00 4.15308595e-01 1.57532942e+00
1.25859514e-01 -2.46699706e-01 -5.81539392e-01 -8.56636941e-01
4.57288593e-01 3.46664131e-01 -7.88818240e-01 2.50272840e-01
-2.82293856e-01 -4.27017957e-01 1.00153849e-01 3.73780996e-01
5.43610513e-01 -8.63587737e-01 -7.51552463e-01 3.86857390e-01
-1.52657572e-02 -1.00646770e+00 -5.69567323e-01 1.44679043e-02
-6.95093036e-01 -1.18455875e+00 -3.75055999e-01 -4.57146987e-02
2.99551338e-01 1.93510294e-01 1.16483951e+00 -3.51360589e-01
2.39958391e-01 9.02266800e-01 -3.78297269e-01 -3.73973876e-01
-1.98155805e-01 -5.70062958e-02 4.23343152e-01 -1.57566547e-01
-2.09500462e-01 -5.81949472e-01 -9.38233972e-01 -2.57421527e-02
-8.31490576e-01 -3.62723708e-01 5.76967359e-01 1.12498212e+00
3.92711282e-01 -1.52384862e-01 6.89760745e-01 -9.24442828e-01
8.12446892e-01 -7.33625531e-01 -1.35410607e+00 9.61320326e-02
-6.74208760e-01 1.60755485e-01 1.07002151e+00 -4.20751721e-01
-7.88656831e-01 -1.29814968e-01 1.67861566e-01 -3.90112549e-01
2.87351906e-01 4.12372887e-01 1.43010631e-01 -8.06745961e-02
5.68046153e-01 2.09644720e-01 3.78635257e-01 -1.77345380e-01
7.03480780e-01 4.17429417e-01 3.29285294e-01 -1.09110963e+00
8.93756568e-01 5.46638727e-01 -7.92476237e-02 -4.22172695e-01
-9.24016833e-01 -3.82549345e-01 3.20081748e-02 -2.83436239e-01
2.49880940e-01 -9.94193912e-01 -1.46378469e+00 3.17327052e-01
-4.90211219e-01 -1.04851663e+00 -6.60474181e-01 5.08890390e-01
-1.30358803e+00 3.68727356e-01 -6.61431849e-01 -1.30047369e+00
-5.16141020e-02 -9.94893074e-01 6.37521029e-01 2.02576488e-01
6.10585451e-01 -1.01390207e+00 3.57430935e-01 9.88819972e-02
3.14678341e-01 1.26787648e-01 2.39590853e-01 -5.14906228e-01
-8.55184257e-01 2.61620730e-01 -4.83076386e-02 2.30471432e-01
-2.67195672e-01 -3.19254339e-01 -8.41634154e-01 -9.25568342e-01
-3.28088760e-01 -1.19197345e+00 6.71062469e-01 3.70309383e-01
1.25483704e+00 -9.37295914e-01 2.89720267e-01 5.82762480e-01
1.44059312e+00 -4.72022370e-02 1.65675983e-01 4.87043679e-01
3.78989764e-02 3.24677110e-01 1.04243302e+00 1.17887688e+00
5.20799398e-01 4.14605260e-01 6.62737310e-01 1.35028183e-01
7.99845159e-01 -8.66383433e-01 8.47173274e-01 3.12234282e-01
3.24575952e-03 -1.16997343e-02 -3.00599068e-01 4.61038381e-01
-2.36257315e+00 -1.12375855e+00 7.87254035e-01 2.60195518e+00
1.14929795e+00 -3.25634815e-02 8.23083878e-01 -3.57325047e-01
1.79043829e-01 1.13770001e-01 -1.13107443e+00 -2.50313073e-01
4.88685369e-02 2.19146833e-01 1.05461991e+00 4.91970509e-01
-1.05340898e+00 7.30157852e-01 6.70706987e+00 7.32081950e-01
-6.60442770e-01 3.48402351e-01 5.40791035e-01 -4.09734428e-01
-6.79182336e-02 -7.46643096e-02 -3.99129987e-01 3.15009862e-01
1.08374548e+00 -3.64321560e-01 1.28021121e+00 1.00399506e+00
3.90647858e-01 -3.08631957e-01 -1.23053122e+00 7.46368527e-01
-4.42353845e-01 -1.20186150e+00 -6.82556450e-01 4.01706547e-01
1.00057876e+00 3.60806167e-01 3.94225001e-01 7.16400743e-01
1.32851851e+00 -8.08734953e-01 7.76826143e-01 3.13699663e-01
7.84189522e-01 -8.96178484e-01 5.07873237e-01 6.40066683e-01
-8.09633315e-01 -7.67704308e-01 -1.81504250e-01 -2.21364751e-01
-2.39391103e-01 1.11232236e-01 -6.59943879e-01 5.87563872e-01
4.82586324e-01 5.22492349e-01 7.88705200e-02 1.08089423e+00
-2.62129366e-01 7.35536158e-01 -5.15795350e-01 -1.92105800e-01
6.05982840e-01 -5.14054477e-01 4.42693561e-01 7.10677266e-01
8.62876475e-02 -7.98034668e-03 9.24659848e-01 7.13039815e-01
-2.57756025e-01 1.97345659e-01 -6.31388426e-01 -1.76904052e-01
6.17150664e-01 1.05433893e+00 -1.16029605e-01 -2.50005424e-01
-4.20071870e-01 7.05522418e-01 6.07683301e-01 6.10016286e-01
-6.55230105e-01 -5.09754010e-02 7.86976814e-01 -3.28284234e-01
3.52595836e-01 -3.39188457e-01 4.08222795e-01 -1.22804379e+00
1.23324804e-01 -1.13731956e+00 5.44553161e-01 -1.61026001e-01
-1.48956275e+00 -3.03472579e-01 -3.49695422e-02 -1.04531944e+00
-8.08316290e-01 -6.81630671e-01 -3.11104774e-01 3.72024179e-01
-1.73343909e+00 -7.04531431e-01 2.21443191e-01 6.42837226e-01
2.00715393e-01 -4.58076745e-02 4.88770992e-01 -2.88754925e-02
-3.60693485e-01 5.75656831e-01 8.91045868e-01 -9.60943326e-02
6.65321469e-01 -1.85211563e+00 7.45090516e-03 6.73029363e-01
-3.66062857e-02 2.27488652e-01 7.18658328e-01 -2.04876974e-01
-2.04396653e+00 -9.25296187e-01 -2.54143238e-01 -1.28844842e-01
1.08913565e+00 -4.74987388e-01 -2.43003979e-01 8.60924184e-01
4.04724255e-02 2.79449582e-01 3.86908084e-01 2.56582320e-01
-2.77986795e-01 -2.20647082e-01 -1.15613353e+00 6.71192169e-01
1.12094104e+00 -6.15485251e-01 -2.00744241e-01 6.98089004e-01
8.27118218e-01 -6.08245015e-01 -7.90186226e-01 -1.13303676e-01
5.85238457e-01 -1.04142213e+00 7.94211805e-01 -9.69858229e-01
1.04672924e-01 9.14043486e-02 -1.25156417e-01 -1.56503820e+00
1.18285380e-01 -1.45741248e+00 -5.72039008e-01 5.26549876e-01
1.37308940e-01 -9.75906789e-01 7.19027102e-01 3.71514440e-01
2.40431681e-01 -7.64137208e-01 -9.35452580e-01 -1.07418883e+00
3.25637072e-01 -3.56889129e-01 6.76117182e-01 5.39136648e-01
4.86328825e-02 -3.16579551e-01 -6.12276435e-01 1.38549060e-01
1.06336641e+00 3.45879287e-01 1.12163889e+00 -7.05795288e-01
-9.36131477e-01 -4.31997806e-01 2.53005147e-01 -1.39762414e+00
4.74859416e-01 -3.64130050e-01 6.86709955e-02 -1.21921813e+00
3.63962114e-01 -5.17060518e-01 -4.75420356e-01 2.78491944e-01
-1.96176302e-03 -2.66648442e-01 4.07152772e-01 -3.72154526e-02
-1.35451710e+00 1.04187834e+00 1.38387299e+00 2.20651776e-01
-3.06605339e-01 2.45009452e-01 -6.10522807e-01 3.31792712e-01
6.85072660e-01 -5.69610186e-02 -5.57276011e-01 1.02263968e-02
5.22374034e-01 6.01731300e-01 3.27053070e-01 -5.68981946e-01
1.77207217e-01 -7.16059625e-01 -1.40902460e-01 -2.15831533e-01
6.62425905e-02 -6.77828431e-01 -4.45139378e-01 4.68028665e-01
-7.54168808e-01 2.08817217e-02 -1.96177483e-01 9.90520895e-01
2.69727468e-01 9.40161645e-02 6.54804468e-01 -2.32122093e-01
-3.23520362e-01 7.57042527e-01 -1.77861124e-01 5.57472587e-01
9.72423911e-01 2.78624982e-01 -4.17315990e-01 -1.09102285e+00
-4.25153553e-01 9.19758677e-01 4.70064461e-01 -1.04232274e-01
4.22497153e-01 -1.23506343e+00 -6.73248589e-01 -1.04512721e-01
4.47621904e-02 -3.17458093e-01 6.41794130e-02 8.25381041e-01
-5.75334057e-02 1.34860292e-01 -1.10752434e-01 -8.39365870e-02
-5.81039488e-01 7.17198610e-01 4.67463493e-01 -5.81618488e-01
-5.42933583e-01 3.73137265e-01 -1.14584155e-02 -7.21204698e-01
7.22386301e-01 -3.15656960e-01 2.07229614e-01 -1.14425734e-01
5.38296223e-01 5.60027063e-01 -2.34563172e-01 9.23512876e-02
3.39513481e-01 -2.06844583e-01 -8.81319121e-03 -7.68670022e-01
1.28126323e+00 -1.87159479e-01 2.91923016e-01 5.06679893e-01
8.46786618e-01 4.21091244e-02 -1.97192264e+00 -5.93427896e-01
-7.17532933e-02 -6.07260525e-01 4.53659110e-02 -8.67800355e-01
-8.55555475e-01 4.00963455e-01 5.59042811e-01 2.94229001e-01
9.68810439e-01 -2.74432302e-01 7.86128342e-01 8.60836864e-01
9.57324862e-01 -1.63568294e+00 1.49435028e-01 6.08013451e-01
5.40505290e-01 -1.34221578e+00 -2.14425456e-02 4.41727281e-01
-6.83119178e-01 6.69134736e-01 3.81265700e-01 -4.27295327e-01
4.76734519e-01 4.19648513e-02 -2.34352291e-01 2.90932983e-01
-1.30521762e+00 -5.34325898e-01 -3.56678635e-01 4.22676444e-01
-3.07397783e-01 3.41271847e-01 -6.85440749e-02 2.95373708e-01
-2.16507107e-01 1.02073759e-01 6.20415330e-01 1.10936975e+00
-4.26222473e-01 -1.10723734e+00 -1.97007358e-01 5.15039444e-01
-5.28624535e-01 2.70494968e-01 -9.73889828e-02 8.24173331e-01
-3.87377590e-01 1.02172995e+00 3.64584439e-02 -7.01340334e-03
3.69006544e-02 -3.46130431e-01 6.84533954e-01 -2.16606244e-01
-5.67528784e-01 1.21652437e-02 8.31267387e-02 -1.24431586e+00
-3.53672147e-01 -7.60424674e-01 -9.52243209e-01 -4.91185486e-01
-1.75311357e-01 2.54103571e-01 4.85551476e-01 9.86100793e-01
2.65922725e-01 -4.63192612e-02 1.18395591e+00 -7.04940677e-01
-1.84408355e+00 -8.00895035e-01 -8.83634508e-01 2.99151719e-01
8.38017225e-01 -8.28817427e-01 -4.45274621e-01 -6.04241908e-01] | [4.131948947906494, 2.3615708351135254] |
8685a436-aa58-4d85-bb8b-3372d578c14b | machine-translation-between-spoken-languages | 2210.05404 | null | https://arxiv.org/abs/2210.05404v2 | https://arxiv.org/pdf/2210.05404v2.pdf | Machine Translation between Spoken Languages and Signed Languages Represented in SignWriting | This paper presents work on novel machine translation (MT) systems between spoken and signed languages, where signed languages are represented in SignWriting, a sign language writing system. Our work seeks to address the lack of out-of-the-box support for signed languages in current MT systems and is based on the SignBank dataset, which contains pairs of spoken language text and SignWriting content. We introduce novel methods to parse, factorize, decode, and evaluate SignWriting, leveraging ideas from neural factored MT. In a bilingual setup--translating from American Sign Language to (American) English--our method achieves over 30 BLEU, while in two multilingual setups--translating in both directions between spoken languages and signed languages--we achieve over 20 BLEU. We find that common MT techniques used to improve spoken language translation similarly affect the performance of sign language translation. These findings validate our use of an intermediate text representation for signed languages to include them in natural language processing research. | ['Sarah Ebling', 'Mathias Müller', 'Amit Moryossef', 'Zifan Jiang'] | 2022-10-11 | null | null | null | null | ['sign-language-translation'] | ['computer-vision'] | [ 3.33524406e-01 2.94962466e-01 -5.07568419e-01 -9.34623480e-01
-1.27143073e+00 -1.06029379e+00 9.05949295e-01 -7.79273510e-01
-4.94816482e-01 7.86992133e-01 9.13190424e-01 -5.76017916e-01
3.93390536e-01 -1.81033969e-01 -4.81102675e-01 -1.91292822e-01
5.87183535e-01 7.79311836e-01 -1.53308108e-01 -4.88742709e-01
1.84841871e-01 2.59827793e-01 -9.44817662e-01 8.45691919e-01
5.96109331e-01 2.25370005e-01 -3.37633729e-01 9.86009777e-01
-4.27061439e-01 8.19280624e-01 -5.62525392e-01 -5.93335569e-01
5.55680692e-01 -1.02118492e+00 -9.99377847e-01 -7.41096437e-02
1.00870430e+00 -5.48867404e-01 -1.18307255e-01 6.95549071e-01
6.20302320e-01 -6.77570939e-01 7.60684192e-01 -9.87152874e-01
-8.33906054e-01 9.74615276e-01 -1.08037606e-01 -3.64551961e-01
5.44335008e-01 3.27971190e-01 9.63071823e-01 -8.72547865e-01
1.27570820e+00 1.41309595e+00 6.38853490e-01 1.01577938e+00
-6.09665215e-01 -4.44757283e-01 -5.99492081e-02 -5.30465961e-01
-7.97224760e-01 -9.15331602e-01 1.00822985e-01 -3.55761915e-01
1.30038905e+00 5.66060133e-02 4.79353130e-01 1.00031161e+00
1.60729066e-01 1.09707093e+00 1.43151820e+00 -9.65791583e-01
-2.27938175e-01 -4.82251085e-02 1.64159134e-01 1.02504921e+00
1.66804001e-01 1.76109090e-01 -1.12031400e+00 1.03669383e-01
6.83189154e-01 -9.44608569e-01 -3.88251655e-02 1.35483444e-01
-1.70439088e+00 4.28301960e-01 -3.42895478e-01 4.74488437e-01
-3.13482173e-02 2.68069059e-01 6.51247025e-01 8.63959432e-01
1.17001332e-01 2.89515346e-01 -6.05531573e-01 -6.74984336e-01
-1.02821243e+00 8.70383233e-02 1.34062374e+00 1.12815869e+00
2.00719699e-01 6.77747503e-02 -2.83618063e-01 7.22867787e-01
6.78246856e-01 1.27250934e+00 8.57047677e-01 -7.74863899e-01
1.07261050e+00 3.00113380e-01 -1.69871643e-03 -1.27652213e-01
3.07277385e-02 4.54735339e-01 -9.06509161e-02 3.00073326e-01
7.72623122e-01 -3.21728766e-01 -1.41141498e+00 1.37314844e+00
-1.16371997e-01 -7.21147001e-01 4.07759577e-01 1.06193626e+00
6.16150975e-01 2.34416470e-01 2.20744237e-02 3.65406536e-02
1.31253862e+00 -1.07352769e+00 -6.95134401e-01 -4.51048821e-01
7.85534382e-01 -1.27000868e+00 1.25307024e+00 2.24964589e-01
-1.37171352e+00 -1.66763723e-01 -1.17087984e+00 -3.81999642e-01
-3.44325840e-01 4.47538972e-01 4.54030216e-01 8.73701215e-01
-1.42950368e+00 3.29400480e-01 -8.98819029e-01 -9.13569152e-01
-5.82726933e-02 2.94936091e-01 -5.31867504e-01 -1.22013241e-01
-7.31158972e-01 1.44593465e+00 -5.24317026e-02 1.04662262e-01
-3.45851332e-01 1.35892751e-02 -7.84240901e-01 -6.58140302e-01
-3.23995769e-01 -5.82768142e-01 2.02849197e+00 -1.16638148e+00
-2.02898479e+00 1.43852127e+00 -6.48714244e-01 -3.89349237e-02
9.33132291e-01 -3.62922788e-01 -4.00455296e-01 -7.55774751e-02
8.07533339e-02 5.79991698e-01 7.62065530e-01 -9.02741551e-01
-6.07271314e-01 -2.50927955e-01 -5.16631842e-01 2.78995782e-01
7.24561363e-02 8.21102858e-01 -3.38234842e-01 -5.50142109e-01
2.88045108e-01 -9.78197992e-01 2.59591401e-01 2.36486807e-01
2.96506565e-04 7.80374706e-02 5.29510021e-01 -1.17947555e+00
9.87119973e-01 -1.61025333e+00 7.67549723e-02 3.20523322e-01
-1.48023173e-01 4.04414594e-01 -5.25387526e-01 6.47293270e-01
4.19034839e-01 1.64981619e-01 -4.18629736e-01 -5.07633746e-01
4.62456495e-01 6.76116645e-01 -4.17474926e-01 3.15762728e-01
4.40296084e-01 1.53893375e+00 -8.58207226e-01 -6.14718199e-01
-1.38420999e-01 2.69256413e-01 -7.45131895e-02 -2.22519338e-01
-1.80399880e-01 2.57027566e-01 -2.62504816e-01 1.06484890e+00
3.14052165e-01 3.64960849e-01 5.47623754e-01 2.66269565e-01
-3.96144301e-01 7.63889730e-01 -6.67068660e-01 1.77445698e+00
-4.47183222e-01 9.70046222e-01 3.19969863e-01 -1.96534678e-01
9.51911390e-01 5.69078743e-01 -1.42673291e-02 -6.87691689e-01
2.65067726e-01 1.11242115e+00 4.29730453e-02 -7.16090798e-01
5.01939297e-01 -3.86246264e-01 -1.94102436e-01 1.18955314e+00
7.09940568e-02 -6.52619481e-01 5.31267226e-01 -1.51366070e-01
9.63221908e-01 6.33510172e-01 1.52364671e-01 -1.95392266e-01
1.99723303e-01 4.07478094e-01 1.28820166e-01 5.37861645e-01
-2.70134300e-01 6.11041903e-01 8.14418346e-02 -2.54491150e-01
-1.17576218e+00 -1.12083161e+00 2.78938562e-01 1.17762744e+00
-5.01343131e-01 -6.22490086e-02 -8.35714221e-01 -8.12563896e-01
-9.51150954e-02 6.05020106e-01 -5.22712283e-02 2.45449752e-01
-1.20794237e+00 -2.80505598e-01 1.58056796e+00 7.61561334e-01
3.47182631e-01 -1.26834035e+00 -6.52041912e-01 2.81776518e-01
-2.69744784e-01 -1.03470576e+00 -8.93025815e-01 -2.38404833e-02
-8.65449071e-01 -9.86303449e-01 -1.11928391e+00 -1.41512227e+00
6.47160351e-01 -2.63898075e-01 1.06422663e+00 -1.36554703e-01
8.36150497e-02 5.63352287e-01 -4.48769510e-01 -5.88978767e-01
-1.24960649e+00 -5.85341491e-02 1.49908498e-01 -3.92450422e-01
8.09426010e-01 4.42801900e-02 1.33950859e-01 3.33827168e-01
-9.09625888e-01 8.10197964e-02 7.85328269e-01 8.08317304e-01
-3.98856997e-02 -1.29613936e+00 1.30958423e-01 -3.62591505e-01
9.73488748e-01 4.35941190e-01 -2.24076644e-01 7.02399313e-01
-5.68868935e-01 6.04837418e-01 1.16612822e-01 -4.72525120e-01
-1.04748929e+00 1.60721049e-01 9.25021917e-02 3.27464491e-01
3.58873680e-02 5.84024847e-01 2.66369025e-04 -2.02060178e-01
7.18561471e-01 3.08242053e-01 4.10943627e-01 -3.09285372e-01
6.81417167e-01 1.20402980e+00 7.10322618e-01 -7.08765745e-01
6.92642808e-01 1.92880422e-01 -3.42854649e-01 -7.45004356e-01
-6.51774928e-02 -3.18520457e-01 -1.12528753e+00 -2.43787423e-01
6.19852781e-01 -7.66000450e-01 -1.84775412e-01 8.85371447e-01
-1.37101638e+00 -6.89149976e-01 -4.12674099e-01 7.09542155e-01
-8.20624352e-01 4.05190885e-01 -1.19519031e+00 -5.97409785e-01
-4.71910000e-01 -9.44789469e-01 1.59774745e+00 -9.77762118e-02
-8.76043022e-01 -6.96665943e-01 4.63463962e-01 4.07293767e-01
6.21197343e-01 -2.09856391e-01 8.06910455e-01 -3.55515361e-01
-2.42864728e-01 -2.38233060e-01 -3.83156270e-01 4.90389228e-01
2.70013094e-01 1.18588284e-01 -8.03432822e-01 -9.82626081e-02
-5.31381488e-01 -6.41658962e-01 6.78658962e-01 -1.51346140e-02
-5.80702782e-01 -4.23169732e-01 1.17339619e-01 2.63099253e-01
8.34546030e-01 -4.93488312e-02 5.36492944e-01 1.06824458e-01
4.08558846e-01 7.20125616e-01 5.03969073e-01 4.16490026e-02
5.33742785e-01 3.75072211e-01 -6.13664210e-01 -1.30529448e-01
-6.71334207e-01 -5.17786562e-01 1.10724449e+00 1.54144442e+00
-2.65754551e-01 -1.91178471e-01 -1.16201842e+00 6.91976428e-01
-1.77986526e+00 -6.07880950e-01 -8.73139426e-02 1.95859277e+00
1.24845815e+00 -2.88044065e-01 9.01555568e-02 -1.66759297e-01
3.28535616e-01 -2.79893968e-02 -2.70229757e-01 -9.07825291e-01
-4.36228961e-01 5.29094517e-01 7.77097285e-01 9.63608027e-01
-7.68398166e-01 1.52292216e+00 7.27408600e+00 -5.76957017e-02
-1.59371316e+00 -1.08047143e-01 -1.07961409e-01 -2.53044944e-02
-2.60464519e-01 -2.90707988e-03 -8.89956176e-01 5.31747825e-02
1.14930046e+00 -9.55802202e-02 4.91330594e-01 2.33830139e-01
4.62326348e-01 1.55085698e-01 -1.42117894e+00 7.77898192e-01
4.97839153e-01 -9.42012310e-01 2.97365248e-01 1.85448173e-02
7.58239388e-01 4.35956597e-01 -2.30976522e-01 1.92119598e-01
8.41609001e-01 -9.47806418e-01 1.20887446e+00 5.41502178e-01
1.25849736e+00 1.09046079e-01 6.15442693e-01 1.03581078e-01
-1.01397777e+00 4.18176413e-01 1.46312892e-01 -2.85847094e-02
7.39865303e-01 -3.66965607e-02 -1.06309533e+00 1.50008112e-01
9.77097154e-02 4.49475229e-01 -3.60748142e-01 6.17666960e-01
-6.91359043e-01 8.85680377e-01 -4.79889095e-01 -4.17250544e-01
3.52744222e-01 -8.38179439e-02 3.76523107e-01 1.68395245e+00
4.40476507e-01 -1.05302140e-01 1.65243864e-01 3.02030146e-01
1.02937415e-01 2.51401305e-01 -7.78945148e-01 -6.67715371e-01
-8.83026570e-02 3.42875719e-01 -5.27978301e-01 -6.92908466e-01
-3.72139156e-01 1.53049254e+00 -2.55044967e-01 4.59869623e-01
-3.23939025e-01 -4.06554371e-01 4.82896626e-01 -7.59289488e-02
8.95706122e-04 -7.31252134e-01 -6.33840144e-01 -1.38590360e+00
5.86213529e-01 -1.42818761e+00 -1.04805984e-01 -7.21769154e-01
-1.28286684e+00 4.90369976e-01 -1.42517820e-01 -1.25711083e+00
-1.04123640e+00 -1.02870417e+00 -1.96280628e-01 1.08866656e+00
-1.37687397e+00 -1.69828153e+00 9.25312340e-02 3.18630636e-01
4.87333268e-01 -3.14527690e-01 8.88928831e-01 2.47998834e-01
3.90063375e-01 6.90131903e-01 9.56061929e-02 6.18993938e-01
1.24107194e+00 -8.90837252e-01 1.06514776e+00 8.92271161e-01
4.67581719e-01 7.82284141e-01 3.03255051e-01 -9.65586901e-01
-1.37658787e+00 -6.12660706e-01 2.08152699e+00 -7.19898045e-01
6.47228122e-01 -1.23122945e-01 -1.77492872e-01 7.97030210e-01
5.04440546e-01 -6.14003956e-01 4.62790549e-01 -3.61235172e-01
-6.84327006e-01 5.00332654e-01 -1.05770266e+00 8.51967931e-01
1.40076256e+00 -1.17205572e+00 -1.02018547e+00 2.41845801e-01
3.17066193e-01 -3.43443662e-01 -5.48000395e-01 2.11922258e-01
1.44904625e+00 -2.27628484e-01 2.38631278e-01 -9.76158917e-01
5.49895704e-01 -3.70896786e-01 -3.39625597e-01 -1.36977625e+00
2.11475745e-01 -8.49118829e-01 2.36408323e-01 9.97862458e-01
7.70153284e-01 -7.81287074e-01 7.78885126e-01 8.46243382e-01
-2.22713128e-01 -7.99205080e-02 -8.85778010e-01 -9.23106670e-01
3.08679491e-01 -1.98143646e-01 3.36255670e-01 8.20703268e-01
4.72761095e-01 3.14537644e-01 -1.94099084e-01 -4.05415982e-01
2.83156306e-01 -6.19001575e-02 8.88032615e-01 -8.02496433e-01
-3.64146471e-01 -6.08366013e-01 -3.65943164e-01 -1.22331846e+00
6.70379847e-02 -1.15382791e+00 5.86558700e-01 -1.79696381e+00
-2.70115018e-01 1.61347941e-01 3.23435456e-01 1.15827596e+00
3.13567221e-01 3.71649951e-01 6.21918738e-01 2.61183560e-01
-1.18706241e-01 1.78469256e-01 1.16282070e+00 -3.63630801e-01
-2.49602363e-01 -3.22447598e-01 -3.67708892e-01 6.67485654e-01
6.86401069e-01 -2.22498551e-01 1.71168908e-01 -1.03849685e+00
-8.61284360e-02 -9.29111838e-02 -5.75976595e-02 -6.96845114e-01
1.81971669e-01 -1.44747540e-01 -4.47123460e-02 -3.34343404e-01
8.18227232e-03 -3.98649156e-01 -4.18975532e-01 5.59230983e-01
-5.04227757e-01 4.26235467e-01 1.66857004e-01 -2.05137476e-01
-4.48044270e-01 -1.55104205e-01 4.79436904e-01 -3.19482177e-01
-7.25417614e-01 -4.43427056e-01 -8.13187838e-01 1.21518038e-01
4.56118017e-01 -4.55676585e-01 -6.24603987e-01 -5.28414428e-01
-3.57817560e-01 1.25884011e-01 5.65641820e-01 6.06823206e-01
4.94124234e-01 -1.18196940e+00 -1.01194036e+00 6.08596623e-01
2.34969720e-01 -5.83803773e-01 -8.43188286e-01 6.10918343e-01
-1.13837230e+00 7.48215079e-01 -3.85044068e-01 -2.58190840e-01
-1.58421516e+00 -6.12779975e-01 2.98921198e-01 -2.43873760e-01
-1.84603646e-01 7.60267973e-01 -7.13360727e-01 -1.05855405e+00
1.92584425e-01 -8.19220364e-01 5.73685050e-01 -2.30333090e-01
4.63966370e-01 1.18940577e-01 1.77045137e-01 -8.83999228e-01
-3.55014384e-01 8.54929924e-01 1.52788416e-01 -1.33491325e+00
8.85371983e-01 2.45370530e-02 -3.41735303e-01 5.14173985e-01
9.74768639e-01 4.55986887e-01 -7.56098449e-01 -5.32683954e-02
4.74083692e-01 -9.42048728e-02 -5.47231913e-01 -1.30678117e+00
-3.38365108e-01 8.11763108e-01 4.75286424e-01 -9.27235067e-01
6.52053237e-01 -1.29591227e-01 1.03953540e+00 9.99745309e-01
7.16645956e-01 -1.41861677e+00 -4.32857990e-01 1.23154283e+00
9.38938379e-01 -1.13411617e+00 -2.56280541e-01 5.79187274e-02
-8.99612486e-01 1.26257110e+00 1.17074728e-01 1.63513020e-01
3.00400525e-01 6.52135551e-01 1.17117941e+00 -1.03854753e-01
-3.40249509e-01 -8.43999311e-02 3.33061486e-01 5.38662314e-01
8.66340399e-01 2.20847636e-01 -8.53400409e-01 8.66205022e-02
-6.00852072e-01 5.07187545e-01 3.92763674e-01 1.55070150e+00
-2.32825264e-01 -1.70440710e+00 -5.26583433e-01 2.15907916e-01
-2.32339725e-01 -3.46401334e-01 -1.49898934e+00 7.89106905e-01
-1.64196745e-01 9.83409166e-01 -2.94455469e-01 -2.77504355e-01
3.67653012e-01 7.99212396e-01 9.50314045e-01 -5.50300300e-01
-8.88000965e-01 -1.67818502e-01 6.64208770e-01 -5.19347668e-01
-6.79202914e-01 -8.40943873e-01 -1.42777276e+00 -3.91406901e-02
5.43953255e-02 -1.37690991e-01 1.06977010e+00 1.17083192e+00
8.15086588e-02 -1.38958558e-01 -1.85150310e-01 -7.53526390e-01
-9.39564407e-01 -1.14156199e+00 -4.54019606e-01 1.91545054e-01
2.64188081e-01 2.76242137e-01 -1.48373827e-01 6.12810254e-01] | [9.195640563964844, -6.522155284881592] |
85a238b8-04aa-40fc-b1e3-92cda034d53a | dota-2-with-large-scale-deep-reinforcement | 1912.06680 | null | https://arxiv.org/abs/1912.06680v1 | https://arxiv.org/pdf/1912.06680v1.pdf | Dota 2 with Large Scale Deep Reinforcement Learning | On April 13th, 2019, OpenAI Five became the first AI system to defeat the world champions at an esports game. The game of Dota 2 presents novel challenges for AI systems such as long time horizons, imperfect information, and complex, continuous state-action spaces, all challenges which will become increasingly central to more capable AI systems. OpenAI Five leveraged existing reinforcement learning techniques, scaled to learn from batches of approximately 2 million frames every 2 seconds. We developed a distributed training system and tools for continual training which allowed us to train OpenAI Five for 10 months. By defeating the Dota 2 world champion (Team OG), OpenAI Five demonstrates that self-play reinforcement learning can achieve superhuman performance on a difficult task. | ['Susan Zhang', 'Filip Wolski', 'Jie Tang', 'Henrique Pondé de Oliveira Pinto', 'Rafal Józefowicz', 'Chris Hesse', 'Przemysław Dębiak', 'Vicki Cheung', 'Tim Salimans', 'Szymon Sidor', 'Quirin Fischer', 'Brooke Chan', 'Scott Gray', 'Ilya Sutskever', 'Shariq Hashme', 'Michael Petrov', 'Jonathan Raiman', 'Jeremy Schlatter', 'Jakub Pachocki', 'David Farhi', 'Christy Dennison', 'Jonas Schneider', 'Greg Brockman', 'Christopher Berner', 'Catherine Olsson'] | 2019-12-13 | null | null | null | null | ['dota-2'] | ['playing-games'] | [-3.83695334e-01 -1.30386446e-02 -1.19173983e-02 1.08661212e-01
-5.63283563e-01 -7.07883716e-01 2.71600991e-01 -3.42915565e-01
-7.94112742e-01 1.07103229e+00 -4.47276801e-01 -3.05794865e-01
-1.77054048e-01 -5.51388502e-01 -6.95421040e-01 -1.73821911e-01
-7.09360719e-01 6.61620438e-01 4.72420663e-01 -8.32557440e-01
5.82291074e-02 9.90788639e-02 -1.40583849e+00 8.37446675e-02
3.87295216e-01 9.37666595e-01 1.05797760e-01 1.23857725e+00
6.37791038e-01 1.23193228e+00 -8.68822455e-01 -2.97889024e-01
8.02492201e-01 -3.35129112e-01 -8.45588028e-01 -3.34409267e-01
2.50859201e-01 -6.43663645e-01 -6.99503422e-01 6.27212107e-01
4.22227114e-01 2.32857317e-01 -9.17828083e-02 -1.62264502e+00
-1.98372722e-01 7.91840196e-01 -1.48942590e-01 6.73480988e-01
3.53483170e-01 1.00655174e+00 1.05009890e+00 -2.44534776e-01
6.33472204e-01 1.06217825e+00 9.36801195e-01 8.41157556e-01
-9.36959028e-01 -7.91236222e-01 -1.00292608e-01 4.37059522e-01
-1.04633188e+00 -3.07136983e-01 1.14358015e-01 -3.33639026e-01
1.26322758e+00 -1.16245681e-03 1.16215622e+00 1.20003140e+00
6.52327895e-01 1.00911534e+00 1.17323685e+00 1.04083478e-01
4.70477164e-01 -7.10307896e-01 -4.36689347e-01 6.51017129e-01
-3.36932950e-02 7.01083183e-01 -5.05405426e-01 5.30288666e-02
1.01131582e+00 -2.18890831e-01 2.54768372e-01 -1.40757889e-01
-1.45592082e+00 8.26147199e-01 3.20606798e-01 5.45719601e-02
-4.85880941e-01 8.96820724e-01 6.99407995e-01 7.77453661e-01
-2.25555390e-01 1.17855394e+00 -5.33357680e-01 -1.00197721e+00
-4.24364716e-01 9.07560885e-01 9.18154180e-01 7.26968706e-01
5.33724010e-01 6.50376737e-01 2.96629220e-01 2.43378907e-01
-9.63203311e-02 6.53033495e-01 5.70757449e-01 -1.68356788e+00
1.39803797e-01 2.19388515e-01 3.81611735e-01 -8.14330637e-01
-6.30070269e-01 -2.87504941e-01 -4.32532728e-01 6.35184526e-01
5.75187922e-01 -9.90283191e-01 -7.34014869e-01 1.49176836e+00
1.79838464e-01 2.01868787e-01 3.68365884e-01 1.05194187e+00
1.07758142e-01 7.81908154e-01 -6.10755868e-02 9.12388265e-02
8.81460190e-01 -9.69817519e-01 -3.77210796e-01 -6.32630467e-01
4.38086212e-01 -3.69208872e-01 7.82774746e-01 8.04481745e-01
-1.38546991e+00 -6.05590701e-01 -1.16464388e+00 6.20069146e-01
-5.82354777e-02 -6.20205224e-01 1.02529252e+00 3.29276174e-01
-8.27810466e-01 9.56356347e-01 -1.36551714e+00 -6.34813309e-02
3.12844962e-01 6.29366279e-01 -2.92894304e-01 2.06846371e-01
-1.42528558e+00 1.08938575e+00 5.54730058e-01 -2.79557586e-01
-1.61167490e+00 -5.94569385e-01 -6.37826920e-01 -1.85506850e-01
8.81314039e-01 -5.34937739e-01 1.81878543e+00 -9.32782710e-01
-1.96166158e+00 3.91139865e-01 1.08089399e+00 -1.27203882e+00
5.67029297e-01 -7.56543756e-01 -5.37539661e-01 1.72598824e-01
1.90082118e-01 6.64020956e-01 6.96232975e-01 -5.82994044e-01
-8.41468394e-01 -3.88788842e-02 3.47148746e-01 4.00483549e-01
1.55760825e-01 -7.14722322e-04 1.11558340e-01 -3.37470591e-01
-4.04439807e-01 -1.08373070e+00 -5.38022399e-01 -3.53298396e-01
7.29013607e-02 -1.85454160e-01 5.25282800e-01 -1.31037518e-01
8.61818373e-01 -2.05408072e+00 3.05578560e-01 -1.00108400e-01
2.73605138e-01 5.46502948e-01 -1.32401571e-01 5.20981312e-01
1.62984103e-01 -9.48583558e-02 2.33805075e-01 4.30760533e-01
2.02816442e-01 5.30543149e-01 -6.35142699e-02 2.61301309e-01
3.22669856e-02 9.46749091e-01 -1.30122459e+00 -3.32830399e-01
-8.22909735e-03 -3.43865335e-01 -8.53661597e-01 3.34442914e-01
-4.94055867e-01 3.35515767e-01 -4.41982687e-01 3.15959841e-01
2.80498024e-02 -2.15241253e-01 1.09631971e-01 3.78985941e-01
-3.03752422e-01 -1.28206626e-01 -1.11997902e+00 1.70374489e+00
-6.61775842e-02 5.25717854e-01 1.11821920e-01 -9.05235708e-01
6.58924758e-01 3.48308444e-01 1.03012466e+00 -1.00552118e+00
4.33918446e-01 9.74683240e-02 4.34782088e-01 -3.33567709e-01
5.52888393e-01 -2.81413823e-01 -6.61750913e-01 4.56031859e-01
3.64827812e-01 -8.29643130e-01 5.35810649e-01 1.89770445e-01
1.66310012e+00 3.41125637e-01 2.09380999e-01 -2.54609942e-01
-8.32865834e-02 5.27590215e-01 9.05819237e-01 1.01692855e+00
-8.76762033e-01 2.31312945e-01 3.61671776e-01 -1.22568619e+00
-1.43461537e+00 -1.10142589e+00 4.94450063e-01 1.19652462e+00
9.15937498e-02 -6.89488709e-01 -6.64710343e-01 -3.83917183e-01
1.59721568e-01 3.62529635e-01 -5.21714687e-01 -3.75198007e-01
-7.48743653e-01 -1.16098024e-01 8.54657531e-01 4.43969935e-01
8.15198004e-01 -1.30329657e+00 -1.48124886e+00 6.00419819e-01
7.78786391e-02 -1.01081908e+00 -3.84394735e-01 3.87699544e-01
-2.62573481e-01 -1.11460924e+00 -4.54641014e-01 -4.27236855e-01
-3.10825557e-01 -1.43399835e-01 1.27073133e+00 -5.14296405e-02
-5.64019084e-01 5.26482046e-01 -3.04636389e-01 -6.15860641e-01
-5.31644702e-01 6.44741580e-02 6.74018443e-01 -6.55010760e-01
4.02015354e-03 -6.04516804e-01 -6.70178056e-01 5.52204370e-01
-4.63326335e-01 3.48451614e-01 4.17826861e-01 1.08596766e+00
1.68330699e-01 -1.31857857e-01 8.49410534e-01 -5.87198734e-01
6.66531622e-01 -3.94728810e-01 -8.69722724e-01 -1.92917511e-02
-3.79241884e-01 -4.09688838e-02 7.82379270e-01 -6.27298772e-01
-7.69648135e-01 1.34526998e-01 -7.28233382e-02 -5.90287983e-01
1.49811879e-01 2.47851327e-01 7.11037874e-01 -2.82023363e-02
1.32156587e+00 6.49046749e-02 2.39431247e-01 3.70830059e-01
3.25480938e-01 2.69012868e-01 9.24312532e-01 -7.53660440e-01
7.83156216e-01 -1.13935530e-01 -1.18930705e-01 -5.44297934e-01
-6.72202826e-01 -1.67168543e-01 -2.65398294e-01 -4.83294457e-01
8.28280866e-01 -1.13122511e+00 -1.43948865e+00 5.55036604e-01
-6.14183486e-01 -1.12149024e+00 -8.10251415e-01 4.37889248e-01
-1.16048014e+00 -1.36441857e-01 -8.29413831e-01 -4.19115484e-01
-2.04694062e-01 -8.52143109e-01 3.82498354e-01 5.16961873e-01
-4.26494658e-01 -5.45032740e-01 6.75124347e-01 3.62261981e-01
5.68839908e-01 2.85330832e-01 2.22938001e-01 -4.14937794e-01
-4.91516203e-01 -1.18366510e-01 2.54567683e-01 4.41111475e-01
-1.60558298e-01 -4.94315624e-02 -2.89020985e-01 -4.93248224e-01
-4.69546676e-01 -1.15146279e+00 1.01955354e-01 1.24360733e-01
5.98511696e-01 -1.24478482e-01 2.14508638e-01 3.23999494e-01
1.12666678e+00 4.61975873e-01 4.10494506e-01 5.73007107e-01
3.50615740e-01 -2.44635567e-01 8.07490647e-01 9.89330709e-01
2.01037467e-01 2.94009060e-01 6.21220946e-01 4.56460565e-01
1.98743254e-01 -2.28531659e-01 5.09142339e-01 8.33740056e-01
-4.12185341e-01 3.36722508e-02 -1.09967494e+00 5.30757010e-01
-1.94066548e+00 -1.29356265e+00 4.96535867e-01 1.79708052e+00
9.00500655e-01 6.52693450e-01 5.12140810e-01 -1.57805815e-01
2.95950383e-01 -1.05719708e-01 -1.12125862e+00 -5.67266405e-01
-1.78041756e-02 1.66711241e-01 6.34587705e-01 1.30298197e-01
-1.11557722e+00 1.26898718e+00 8.38963127e+00 7.25211740e-01
-9.39486921e-01 -1.35659888e-01 3.97479326e-01 -3.61338258e-01
1.97770312e-01 3.71437147e-02 -5.17907679e-01 3.36725891e-01
1.34631002e+00 -7.78892279e-01 9.88701761e-01 1.11149120e+00
-2.24777713e-01 -1.70462802e-01 -9.16685045e-01 8.70837748e-01
-1.98574826e-01 -1.47472000e+00 -6.43975019e-01 -9.39370915e-02
1.12431538e+00 5.19885600e-01 -6.86169416e-03 1.03215384e+00
1.43220294e+00 -1.16759014e+00 6.65656030e-01 2.74865746e-01
5.17199278e-01 -9.62798297e-01 4.17649448e-01 6.09901369e-01
-8.16733599e-01 -6.70062304e-01 -2.64431357e-01 -6.26575649e-01
7.83065036e-02 -4.36918467e-01 -8.31579328e-01 9.27653238e-02
8.64720821e-01 5.14830649e-01 -2.12698624e-01 9.23281729e-01
-2.20491514e-02 6.18122101e-01 -4.86029714e-01 -2.42287606e-01
5.99084198e-01 3.34449410e-01 5.84979415e-01 5.41607440e-01
-1.92980357e-02 4.54391181e-01 5.16192496e-01 3.80589485e-01
8.40699673e-02 -3.59984577e-01 -7.12302446e-01 -4.40979987e-01
1.62173182e-01 1.04311359e+00 -4.77790117e-01 -3.50982219e-01
-2.39422798e-01 8.02868307e-01 3.73724580e-01 -1.72597513e-01
-1.06079602e+00 -4.17190939e-01 7.74861395e-01 -1.01144850e-01
2.70012498e-01 -6.57871723e-01 2.78068542e-01 -1.13795269e+00
-6.32529557e-01 -1.46357369e+00 4.40413356e-01 -8.60394299e-01
-1.16624749e+00 7.27567852e-01 4.07848172e-02 -1.19401217e+00
-7.36899018e-01 -5.42922556e-01 -4.96513993e-01 1.07500404e-01
-6.11000180e-01 -5.85689008e-01 4.59268736e-03 8.82901728e-01
7.93880463e-01 -6.23222232e-01 7.93086648e-01 6.47970587e-02
-4.83627528e-01 2.82826841e-01 9.47419554e-02 3.26587906e-05
4.81426746e-01 -1.33987153e+00 8.44095230e-01 6.11419201e-01
3.87703069e-02 1.05030179e-01 1.05844915e+00 -6.56231999e-01
-1.85523164e+00 -3.91769946e-01 -2.29981780e-01 -3.62414569e-01
1.16095030e+00 -6.17363304e-02 -3.88674319e-01 9.47535634e-01
4.21753705e-01 1.42942116e-01 3.19707870e-01 8.98620859e-02
3.13201994e-02 -1.09173745e-01 -7.29276061e-01 6.18741989e-01
9.63266373e-01 -2.64707029e-01 -7.86118686e-01 2.90821135e-01
7.19669580e-01 -7.90840685e-01 -1.06064367e+00 2.05193639e-01
5.82011938e-01 -7.41381288e-01 7.61027336e-01 -9.01825368e-01
4.86614287e-01 -2.38309074e-02 3.06471922e-02 -1.67361510e+00
-3.84626687e-01 -1.28698754e+00 -1.01594225e-01 3.13477784e-01
1.62330970e-01 -2.51872629e-01 1.04204249e+00 5.55717051e-01
-1.76417336e-01 -4.80904073e-01 -9.88775849e-01 -1.06100714e+00
2.04081491e-01 -3.18967432e-01 3.09122711e-01 6.04872823e-01
5.28572261e-01 3.12293649e-01 -9.64900434e-01 -2.31694341e-01
3.36259484e-01 1.21270923e-03 1.10347342e+00 -9.26114023e-01
-7.20074892e-01 -1.14733510e-01 -6.33609831e-01 -9.06730294e-01
-1.81157157e-01 -2.55401880e-01 4.17227924e-01 -9.21176612e-01
-3.13099056e-01 -3.74847829e-01 -2.71705180e-01 7.69685030e-01
1.59488782e-01 2.81409979e-01 5.11426568e-01 1.22115605e-01
-1.41803694e+00 5.23203015e-01 1.49516153e+00 -8.14584270e-02
-8.29501152e-02 6.75199740e-03 -5.06181777e-01 8.25048923e-01
9.14827287e-01 -3.15235436e-01 -4.77378786e-01 -4.86157447e-01
5.96815944e-01 5.42899311e-01 1.42070532e-01 -1.97223580e+00
5.19936621e-01 -7.32758939e-01 1.91375121e-01 -9.00622383e-02
5.01805425e-01 -7.33332515e-01 2.04296187e-01 8.98344994e-01
-1.65022671e-01 4.50186133e-01 6.36318147e-01 4.11547244e-01
-1.11282334e-01 1.80503547e-01 7.16286719e-01 -5.03054976e-01
-1.21423590e+00 3.34138423e-01 -9.70286191e-01 5.36303282e-01
1.61915052e+00 -8.38449448e-02 -2.33305335e-01 -5.01082122e-01
-9.71706510e-01 8.14214706e-01 1.04244001e-01 3.86470050e-01
4.44385469e-01 -1.15949321e+00 -7.16938317e-01 9.09252688e-02
-1.78025380e-01 -1.52159363e-01 6.46178246e-01 7.77824819e-02
-1.06048965e+00 9.08161476e-02 -1.33149147e+00 -4.41911548e-01
-1.07520342e+00 5.38555682e-01 5.30863941e-01 -4.58279639e-01
-7.90855825e-01 7.09833026e-01 -1.37638137e-01 -3.53353173e-01
-7.63637992e-03 1.20754004e-03 1.48499414e-01 -5.55142462e-01
4.55304235e-01 2.22918808e-01 -4.54658061e-01 -2.02255026e-01
-2.48775512e-01 -5.03355600e-02 -1.28927991e-01 -3.89443427e-01
1.53994048e+00 2.50012189e-01 7.14316428e-01 4.84538704e-01
3.93768013e-01 -6.43639147e-01 -1.92329335e+00 1.64871793e-02
-2.86599547e-01 -3.10732812e-01 -2.22921282e-01 -9.89856422e-01
-7.19512522e-01 5.70637882e-01 6.09370410e-01 2.41375551e-01
8.18293095e-01 -4.70545769e-01 9.64677572e-01 1.02570069e+00
8.47109556e-01 -1.48637211e+00 7.08663285e-01 9.85218823e-01
7.85329521e-01 -1.14995503e+00 4.16804515e-02 5.38276434e-01
-1.05870771e+00 1.05009079e+00 1.21971154e+00 -7.34839082e-01
3.20334375e-01 4.91987586e-01 5.46940304e-02 -2.95848727e-01
-1.40668190e+00 1.12402484e-01 -4.36783463e-01 5.16294181e-01
-1.96511000e-01 2.22999081e-01 2.32924953e-01 5.05991399e-01
-4.78447109e-01 2.67378449e-01 7.24800289e-01 1.36090326e+00
-5.98822117e-01 -9.05360103e-01 -1.56220943e-01 2.04224899e-01
-3.76004547e-01 3.74906600e-01 1.33132428e-01 1.01369059e+00
1.79704964e-01 6.47342384e-01 1.64459959e-01 -6.73693120e-01
4.49902236e-01 -1.76003039e-01 5.98580956e-01 -3.53890151e-01
-8.71452987e-01 -2.99370199e-01 1.94494992e-01 -1.00774014e+00
1.49833020e-02 -5.92543364e-01 -1.31137800e+00 -1.00294232e+00
-9.00843889e-02 4.23209548e-01 5.68726838e-01 8.11374545e-01
3.47873211e-01 5.67732334e-01 5.71352303e-01 -8.84836376e-01
-7.71865845e-01 -5.65278828e-01 -5.22915721e-01 3.27304751e-01
1.37050033e-01 -5.03322542e-01 -1.98590700e-02 -5.53384237e-02] | [3.60996413230896, 1.5209455490112305] |
fc0e4945-4062-4373-82db-00e63f843649 | memory-enriched-computation-and-learning-in | 2205.11276 | null | https://arxiv.org/abs/2205.11276v1 | https://arxiv.org/pdf/2205.11276v1.pdf | Memory-enriched computation and learning in spiking neural networks through Hebbian plasticity | Memory is a key component of biological neural systems that enables the retention of information over a huge range of temporal scales, ranging from hundreds of milliseconds up to years. While Hebbian plasticity is believed to play a pivotal role in biological memory, it has so far been analyzed mostly in the context of pattern completion and unsupervised learning. Here, we propose that Hebbian plasticity is fundamental for computations in biological neural systems. We introduce a novel spiking neural network architecture that is enriched by Hebbian synaptic plasticity. We show that Hebbian enrichment renders spiking neural networks surprisingly versatile in terms of their computational as well as learning capabilities. It improves their abilities for out-of-distribution generalization, one-shot learning, cross-modal generative association, language processing, and reward-based learning. As spiking neural networks are the basis for energy-efficient neuromorphic hardware, this also suggests that powerful cognitive neuromorphic systems can be build based on this principle. | ['Robert Legenstein', 'Ozan Özdenizci', 'Thomas Limbacher'] | 2022-05-23 | null | null | null | null | ['one-shot-learning'] | ['methodology'] | [ 3.21945488e-01 -2.46167183e-01 1.38478026e-01 -4.85742986e-02
3.16338181e-01 -5.32911301e-01 7.22647309e-01 2.92586029e-01
-7.07560778e-01 1.04205132e+00 -2.12187961e-01 -1.53046563e-01
-3.51138055e-01 -1.02420831e+00 -8.44076753e-01 -1.00746942e+00
-2.16253653e-01 1.81086034e-01 9.86125588e-01 -3.56252134e-01
5.56296408e-01 7.96502650e-01 -1.94707847e+00 3.55212033e-01
5.68699956e-01 1.09146237e+00 4.48980361e-01 5.43878615e-01
-1.86748222e-01 4.01684433e-01 -3.87309968e-01 -1.04725212e-01
-8.25945884e-02 -5.23197770e-01 -3.66929531e-01 -6.23241127e-01
-4.87138718e-01 4.47178096e-01 -4.15572643e-01 7.35307515e-01
4.59740162e-01 2.05621168e-01 5.97418725e-01 -7.53784895e-01
-6.58165455e-01 3.54047179e-01 1.11034930e-01 4.96785402e-01
-1.78719029e-01 1.82494476e-01 5.23675978e-01 -1.01450872e+00
5.70079982e-01 7.19913423e-01 6.81546986e-01 9.24576998e-01
-1.53123415e+00 -6.77261293e-01 -3.63870114e-01 2.10998610e-01
-1.01812327e+00 -5.84871590e-01 3.08426470e-01 -3.29103470e-01
1.34847403e+00 1.67000666e-03 1.33590543e+00 8.14912617e-01
1.02090156e+00 3.48142296e-01 1.35659647e+00 -3.64084214e-01
6.72802925e-01 -2.67059207e-01 1.24421820e-01 5.04867375e-01
3.40815991e-01 3.98866892e-01 -1.16926050e+00 1.58440083e-01
8.88237834e-01 3.19028914e-01 -1.74941123e-01 -1.28410801e-01
-1.11836386e+00 3.13780189e-01 5.28978109e-01 6.49887741e-01
-2.70937979e-01 5.00328660e-01 1.48150250e-01 2.74895459e-01
-2.83510327e-01 6.01838529e-01 -1.92790732e-01 -1.65620431e-01
-8.57178390e-01 -6.66719601e-02 5.34030735e-01 3.49439621e-01
8.67203772e-01 3.58782858e-01 -1.72289982e-02 8.63030314e-01
2.69431211e-02 7.02245116e-01 9.64168310e-01 -8.09213698e-01
-7.35921144e-01 7.97537446e-01 -4.12748486e-01 -5.12035668e-01
-3.60729754e-01 -2.41761610e-01 -1.15585172e+00 5.92194617e-01
1.33770585e-01 4.85422552e-01 -9.00905371e-01 1.77289772e+00
-2.23825991e-01 8.43956620e-02 2.53313817e-02 5.81463635e-01
5.87509871e-01 6.12232983e-01 1.51258379e-01 -3.22093070e-01
9.39779222e-01 -3.73320729e-01 -4.44602281e-01 -1.69030547e-01
-1.77074775e-01 -1.18475989e-01 7.25164294e-01 3.63165736e-01
-1.06394315e+00 -3.27202052e-01 -1.07042372e+00 1.53100863e-01
-7.90799141e-01 -7.46744514e-01 6.90479755e-01 6.99248135e-01
-1.42070115e+00 9.96954799e-01 -8.21266532e-01 -7.00215280e-01
5.95419466e-01 6.56056166e-01 -4.31700468e-01 1.96053416e-01
-8.18811536e-01 9.22059476e-01 4.51711088e-01 -3.08906764e-01
-8.72977793e-01 -5.26934266e-01 -5.58387786e-02 2.11319506e-01
-3.63195956e-01 -8.24345469e-01 8.22673082e-01 -7.33025014e-01
-1.78354740e+00 1.13541424e+00 -3.66589576e-01 -7.94713497e-01
-4.77632731e-01 3.04376930e-01 -3.41698319e-01 -1.24305142e-02
-4.48830664e-01 9.92904425e-01 5.55499673e-01 -7.06578910e-01
-1.87671766e-01 -5.21469235e-01 -4.38359052e-01 -4.19776201e-01
-3.72444481e-01 -1.51886299e-01 1.41892791e-01 -7.13003218e-01
4.03692901e-01 -1.04371548e+00 -1.59875327e-03 1.98118776e-01
4.07864392e-01 -5.40640652e-02 5.40917993e-01 -7.34838843e-02
7.52482533e-01 -2.14678454e+00 1.18678145e-01 2.57162213e-01
7.95304552e-02 3.11731637e-01 7.76794693e-03 5.28531194e-01
5.29623181e-02 -7.12735429e-02 -6.07441664e-01 4.04257715e-01
-3.52435321e-01 2.60341078e-01 -5.98991334e-01 1.68557614e-01
4.06487584e-01 1.36329436e+00 -6.53589189e-01 8.52175802e-02
-1.83676273e-01 4.41762567e-01 -3.61659974e-01 6.23894930e-02
-2.29028478e-01 7.10400164e-01 9.66503695e-02 6.62125409e-01
1.88923642e-01 -2.75221765e-01 9.19504985e-02 3.65810782e-01
-4.41929281e-01 1.34971842e-01 -2.67776340e-01 1.98378825e+00
3.00152656e-02 8.06129813e-01 -2.74024993e-01 -1.04806376e+00
1.23594534e+00 1.40276581e-01 3.48110706e-01 -1.36826730e+00
6.66538626e-02 4.60711896e-01 4.92317557e-01 6.72205687e-02
2.25892738e-01 -4.25573260e-01 2.01840520e-01 4.89567935e-01
4.26377088e-01 -2.24644765e-01 2.64048249e-01 -1.13501407e-01
1.32564712e+00 -1.20058849e-01 1.02329776e-01 -8.22344303e-01
1.02569439e-01 -2.24024355e-01 4.87367272e-01 6.64313495e-01
-1.36509910e-01 3.44674557e-01 5.44187948e-02 -4.35553014e-01
-1.07506025e+00 -1.58458126e+00 -1.28795445e-01 1.25828886e+00
1.83684900e-01 -1.63221005e-02 -4.72577155e-01 6.22472465e-01
-2.46458009e-01 3.01134199e-01 -6.45849466e-01 -6.16351902e-01
-4.19644415e-01 -8.38664591e-01 8.32457721e-01 3.24848384e-01
5.61429381e-01 -1.71737289e+00 -1.24848950e+00 4.81470168e-01
4.71527129e-01 -5.74524343e-01 7.54166469e-02 9.64406908e-01
-1.38459659e+00 -7.20980167e-01 -6.20249629e-01 -7.82842398e-01
3.93648744e-01 6.94988295e-02 1.00035870e+00 1.04401372e-01
-8.07513595e-01 2.33205333e-01 -2.76430994e-02 -4.78915751e-01
-1.75860107e-01 -8.53722021e-02 3.70167673e-01 -3.81845683e-01
2.41807625e-01 -1.56754959e+00 -8.19330812e-01 -1.01468094e-01
-1.05637360e+00 -9.24997702e-02 8.29891980e-01 8.10921788e-01
1.05158138e+00 -3.03077936e-01 8.72714698e-01 -4.55762804e-01
6.92970335e-01 -3.16761166e-01 -3.71068835e-01 1.87933430e-01
-7.77415633e-01 3.55228245e-01 4.64907974e-01 -4.67514753e-01
-5.58103204e-01 -1.84609056e-01 -1.55320510e-01 1.59430757e-01
-1.36474237e-01 2.81294972e-01 1.73517823e-01 -4.43702906e-01
8.41666162e-01 1.04079306e+00 6.58087060e-02 -3.14693041e-02
7.98908323e-02 3.69534306e-02 6.54584885e-01 -5.63763380e-01
2.15729237e-01 5.04210055e-01 3.46035630e-01 -1.06142926e+00
-2.68010329e-03 -8.81466717e-02 -4.71085966e-01 -2.55258620e-01
7.95292139e-01 -4.33785498e-01 -9.79567587e-01 7.24357903e-01
-1.06051624e+00 -5.95603287e-01 -5.76789856e-01 2.18639914e-02
-6.63485289e-01 -3.10722828e-01 -8.12068939e-01 -7.57148862e-01
-5.30876935e-01 -5.40250480e-01 2.55161852e-01 7.12681293e-01
-2.01240897e-01 -6.15085185e-01 6.45302117e-01 -5.74124455e-01
8.58442545e-01 -4.08423245e-02 1.19620073e+00 -3.10280621e-01
-6.79607272e-01 3.47299099e-01 -3.55514814e-03 -5.60586452e-02
-1.41781792e-01 -1.12747975e-01 -1.00949574e+00 -1.73139185e-01
-1.64765850e-01 -4.09506261e-01 1.54196763e+00 2.35811606e-01
1.03243423e+00 7.21069127e-02 -4.48831260e-01 4.08639044e-01
1.29996181e+00 6.26605034e-01 1.08391488e+00 1.99500471e-01
-2.20725108e-02 4.92323309e-01 -4.01477277e-01 1.34098053e-01
-8.27220604e-02 2.22315490e-01 4.13062364e-01 4.32696790e-01
-1.78703025e-01 -3.19834403e-03 3.10849786e-01 1.11584139e+00
-2.52835661e-01 2.22320914e-01 -8.43012154e-01 5.90398312e-01
-1.66016269e+00 -1.28572357e+00 2.94425279e-01 2.36931682e+00
9.90194023e-01 2.31044397e-01 -5.81005253e-02 2.98083812e-01
6.12762332e-01 -4.50204849e-01 -9.48302329e-01 -5.84887743e-01
-8.23231339e-01 1.00755107e+00 2.81437099e-01 -1.25936702e-01
-3.88380527e-01 8.96271527e-01 7.06482887e+00 5.74154496e-01
-1.67189419e+00 1.98615372e-01 3.38691920e-01 -7.56074861e-02
-4.01132762e-01 -8.85175541e-02 -4.25513029e-01 5.89170694e-01
1.37605870e+00 -5.14585488e-02 6.92596912e-01 3.36688221e-01
-2.68323779e-01 -4.18758124e-01 -9.73845780e-01 1.15133131e+00
-3.88107866e-01 -1.88453233e+00 1.49205159e-02 8.90200026e-04
7.49371350e-01 3.05654913e-01 3.26647937e-01 8.72379690e-02
7.40991682e-02 -1.45537567e+00 6.38523757e-01 9.83475864e-01
6.48065507e-01 -7.09760427e-01 1.16912082e-01 4.46624398e-01
-8.03264380e-01 -2.72546977e-01 -5.30826628e-01 -4.10561353e-01
-1.28784180e-01 8.44949841e-01 -1.80271521e-01 -3.95313352e-01
7.99268126e-01 4.34240997e-01 -4.05624717e-01 1.48778069e+00
3.41922827e-02 4.77695405e-01 -2.49395639e-01 -3.73833209e-01
-2.48597264e-01 -1.29334778e-01 3.10027450e-01 1.08184659e+00
6.92819476e-01 3.16925257e-01 -5.03115416e-01 1.10961413e+00
-2.92622745e-01 -2.17080638e-01 -7.57163823e-01 -4.19454992e-01
7.08795846e-01 1.00069451e+00 -1.33564138e+00 -4.73956354e-02
2.29748577e-01 1.02977288e+00 4.32266802e-01 1.50021002e-01
-3.23446184e-01 -4.45605814e-01 4.94364202e-01 2.44877622e-01
3.40608448e-01 -6.09613359e-01 -5.57217240e-01 -6.69604957e-01
-3.67210180e-01 -2.09143907e-01 -2.03294203e-01 -6.83093667e-01
-9.64661896e-01 5.57780683e-01 -8.36954713e-01 -6.01112485e-01
-2.17387173e-02 -7.92062461e-01 -7.42249966e-01 4.95653808e-01
-1.24461973e+00 -6.73676968e-01 -1.09310888e-01 7.27161884e-01
1.42525613e-01 -2.96856225e-01 1.35254192e+00 5.36056235e-02
-2.34050170e-01 1.57583326e-01 5.11489987e-01 -2.31972784e-01
4.94901478e-01 -9.27731991e-01 1.84912324e-01 6.30798101e-01
3.58906209e-01 7.99093425e-01 4.42123264e-01 -3.97478074e-01
-1.50676775e+00 -9.56728101e-01 6.16998732e-01 2.89536994e-02
5.14055431e-01 -4.12582487e-01 -1.08388305e+00 3.90322804e-02
1.44101113e-01 -2.45120097e-02 1.00469828e+00 -1.30701020e-01
-7.26387382e-01 -4.01107639e-01 -1.14400089e+00 6.22223854e-01
1.05393624e+00 -6.93964541e-01 -5.42949677e-01 -2.26128444e-01
4.00456727e-01 3.17346931e-01 -6.56726599e-01 4.27377433e-01
1.12157929e+00 -1.29139113e+00 6.65291727e-01 -3.35548848e-01
3.18350255e-01 -2.86244333e-01 -1.65723965e-01 -1.25670147e+00
-6.08969629e-01 -4.59620625e-01 -2.00886413e-01 7.58606255e-01
4.05543476e-01 -9.93903577e-01 5.80039084e-01 2.30353087e-01
-2.91388541e-01 -7.08107531e-01 -1.30307794e+00 -9.37414765e-01
3.04461122e-01 -2.72612721e-02 2.77376205e-01 4.00525570e-01
2.65377253e-01 5.50449416e-02 2.08231166e-01 -5.13487756e-01
3.59365255e-01 2.06414431e-01 6.14689030e-02 -1.51529396e+00
-3.99035633e-01 -1.06413102e+00 -5.93920112e-01 -5.82013726e-01
9.74524468e-02 -9.94410098e-01 3.96397620e-01 -1.25651658e+00
4.19907242e-01 -3.34785670e-01 -7.84937620e-01 3.89387846e-01
4.52661246e-01 7.17261851e-01 -2.84366142e-02 4.03423935e-01
-4.97426867e-01 3.14839333e-01 8.41220021e-01 -3.43669616e-02
-4.41886894e-02 -2.87320524e-01 -4.67288166e-01 3.99559051e-01
9.50735688e-01 -5.42339444e-01 -2.12057739e-01 -3.07328135e-01
4.53788161e-01 -3.24726611e-01 3.91104639e-01 -1.44818664e+00
6.25516951e-01 -1.17553368e-01 5.55607080e-01 -4.97428738e-02
4.55263346e-01 -4.62373435e-01 3.81332546e-01 9.55645204e-01
-3.30781579e-01 9.84354094e-02 3.90554488e-01 6.09718144e-01
-1.04454838e-01 3.89957018e-02 1.09804881e+00 -2.71436363e-01
-9.19407547e-01 2.12165028e-01 -1.11944425e+00 -9.07895863e-02
9.04711723e-01 -7.88964391e-01 -6.28989279e-01 1.63469523e-01
-7.08742321e-01 -5.16825795e-01 6.16845787e-01 1.76663905e-01
9.88929570e-01 -1.23093033e+00 -1.62594140e-01 2.74681985e-01
7.36427829e-02 -4.93036866e-01 1.86248824e-01 8.18209827e-01
-3.81372511e-01 5.83205700e-01 -1.14424002e+00 -5.76288760e-01
-7.31531084e-01 4.82789606e-01 9.98382568e-02 -3.66418660e-02
-2.09868789e-01 9.69091952e-01 9.80115086e-02 -9.97074321e-03
-4.87442277e-02 1.08689992e-02 5.55004515e-02 -3.05886626e-01
5.65064013e-01 1.60415113e-01 8.96330699e-02 -2.20250830e-01
-5.22422194e-01 5.85019350e-01 3.46221536e-01 -2.25200787e-01
1.53320539e+00 1.50423661e-01 -8.65338564e-01 1.09423208e+00
5.28133750e-01 -3.17335814e-01 -9.36633527e-01 1.39973193e-01
3.35615933e-01 2.03169137e-01 -5.05064547e-01 -8.25268507e-01
-6.91807985e-01 1.39535761e+00 7.37133265e-01 -1.87032651e-02
1.22587562e+00 -1.17953278e-01 7.59670138e-01 8.11115980e-01
9.35111880e-01 -9.83457565e-01 6.57007635e-01 9.70915556e-01
7.08595455e-01 -6.65450752e-01 -5.40979087e-01 2.21604750e-01
1.23539463e-01 1.28474855e+00 4.41661984e-01 -5.98752022e-01
8.73581231e-01 5.85524023e-01 -6.37561798e-01 9.74715203e-02
-1.02056026e+00 -2.39420339e-01 2.38412991e-02 9.12389159e-01
6.81119561e-01 9.47649404e-02 -3.61487418e-01 6.31572187e-01
4.05101776e-02 2.00502530e-01 4.37806487e-01 9.65487182e-01
-1.31084430e+00 -1.22847068e+00 -6.87788725e-02 4.62326765e-01
-3.36532682e-01 -3.14133286e-01 -3.16691726e-01 9.16108415e-02
2.04704762e-01 4.28375393e-01 1.95116058e-01 -4.26092565e-01
-2.14239985e-01 4.48045731e-01 1.11160028e+00 -6.00289047e-01
-7.00386584e-01 -3.12469959e-01 -5.51649392e-01 -4.03540760e-01
-3.64924252e-01 -1.00855939e-01 -1.82368577e+00 -3.15039933e-01
1.84354812e-01 -2.44920887e-02 7.93935895e-01 8.67307067e-01
7.16819584e-01 5.63703954e-01 1.36907371e-02 -8.30433011e-01
9.82333794e-02 -6.05907381e-01 -8.35748792e-01 1.20738417e-01
-1.11898802e-01 -4.86756504e-01 3.21929634e-01 2.21006021e-01] | [8.158235549926758, 2.5761733055114746] |
9b038787-fd05-44f7-9d8d-c67194d7e4fb | semantic-relationships-guided-representation | 1904.09939 | null | http://arxiv.org/abs/1904.09939v1 | http://arxiv.org/pdf/1904.09939v1.pdf | Semantic Relationships Guided Representation Learning for Facial Action Unit Recognition | Facial action unit (AU) recognition is a crucial task for facial expressions
analysis and has attracted extensive attention in the field of artificial
intelligence and computer vision. Existing works have either focused on
designing or learning complex regional feature representations, or delved into
various types of AU relationship modeling. Albeit with varying degrees of
progress, it is still arduous for existing methods to handle complex
situations. In this paper, we investigate how to integrate the semantic
relationship propagation between AUs in a deep neural network framework to
enhance the feature representation of facial regions, and propose an AU
semantic relationship embedded representation learning (SRERL) framework.
Specifically, by analyzing the symbiosis and mutual exclusion of AUs in various
facial expressions, we organize the facial AUs in the form of structured
knowledge-graph and integrate a Gated Graph Neural Network (GGNN) in a
multi-scale CNN framework to propagate node information through the graph for
generating enhanced AU representation. As the learned feature involves both the
appearance characteristics and the AU relationship reasoning, the proposed
model is more robust and can cope with more challenging cases, e.g.,
illumination change and partial occlusion. Extensive experiments on the two
public benchmarks demonstrate that our method outperforms the previous work and
achieves state of the art performance. | ['Liang Lin', 'Yirui Zeng', 'Xin Zhu', 'Guanbin Li', 'Qing Wang'] | 2019-04-22 | null | null | null | null | ['facial-action-unit-detection'] | ['computer-vision'] | [ 2.43757680e-01 8.14308152e-02 -2.04946741e-01 -4.67296451e-01
-6.00188002e-02 -1.60467147e-03 3.73022288e-01 -3.46666247e-01
4.38139290e-02 2.76959330e-01 1.57919571e-01 1.87581390e-01
8.07741135e-02 -9.78005111e-01 -5.40744185e-01 -8.67682099e-01
1.24072485e-01 -1.62182882e-01 2.68158503e-02 -5.40605068e-01
-2.02415973e-01 6.77970409e-01 -1.51650596e+00 9.20762047e-02
6.11500084e-01 1.36252701e+00 -3.93990964e-01 4.50061075e-02
-2.62370139e-01 1.08482742e+00 -2.91069806e-01 -6.33658588e-01
4.60574590e-02 -6.26734078e-01 -6.03462517e-01 3.84421587e-01
2.22271129e-01 -1.95879117e-01 -5.10994136e-01 1.11995959e+00
4.87889946e-01 2.27741703e-01 5.83599389e-01 -1.57076764e+00
-7.89358556e-01 3.39208126e-01 -9.37685966e-01 -3.24204541e-03
2.93237478e-01 -1.32218167e-01 9.37255681e-01 -8.35419834e-01
5.20805299e-01 1.53063571e+00 4.51986074e-01 5.40643334e-01
-7.33534455e-01 -1.03340101e+00 4.39221978e-01 5.07632792e-01
-1.58929932e+00 -3.39319587e-01 1.21094227e+00 -1.50874600e-01
6.99155748e-01 -3.31858657e-02 7.01804698e-01 9.68400955e-01
5.66813946e-02 9.14093733e-01 7.99650252e-01 -2.23332986e-01
-2.79379427e-01 -9.29082260e-02 -2.30913647e-02 1.10943389e+00
-1.65954486e-01 -1.98719487e-01 -3.85228574e-01 -5.69298565e-02
9.66342568e-01 2.58095711e-01 -1.97189197e-01 -2.27507100e-01
-5.57411313e-01 7.72007108e-01 9.54351485e-01 4.39319342e-01
-4.52961892e-01 2.24686846e-01 3.97809327e-01 2.33361274e-01
7.16262996e-01 -1.82761312e-01 -1.21535711e-01 2.09831551e-01
-3.73763502e-01 -4.81885485e-02 4.18869048e-01 7.66062200e-01
1.10595226e+00 2.58199722e-01 -1.24004841e-01 1.07008886e+00
4.71231997e-01 2.65706122e-01 2.04638675e-01 -6.20322347e-01
1.86255634e-01 1.20957327e+00 -4.28657621e-01 -1.73304832e+00
-3.82237375e-01 -2.82306373e-01 -9.88533378e-01 1.63535830e-02
7.43253753e-02 -1.89430460e-01 -6.14781022e-01 1.89845705e+00
5.17908812e-01 5.79340875e-01 4.33011390e-02 8.60463738e-01
1.20682752e+00 5.91880143e-01 2.02750444e-01 -1.90171868e-01
1.27447975e+00 -1.12515128e+00 -9.68577147e-01 1.42225465e-02
6.29539192e-01 -5.26420176e-01 4.20684010e-01 5.84507361e-02
-7.76572883e-01 -5.73655605e-01 -7.20642447e-01 1.49284573e-02
-4.32835370e-01 1.85222700e-01 9.45731938e-01 2.81748801e-01
-1.09558630e+00 2.92304844e-01 -4.60647583e-01 -4.04146820e-01
8.32502127e-01 5.05922019e-01 -6.85986876e-01 -2.88187563e-01
-1.28471613e+00 5.93211651e-01 1.88530371e-01 6.73858166e-01
-5.68917513e-01 -1.90310255e-01 -1.16399574e+00 3.23389731e-02
5.16310394e-01 -3.88539225e-01 7.20457494e-01 -1.57833827e+00
-1.45346367e+00 6.62573218e-01 -6.69462383e-02 9.35414284e-02
7.28938654e-02 6.15985394e-02 -6.33050323e-01 1.51901081e-01
-3.48961025e-01 5.70732355e-01 9.24540758e-01 -1.14936066e+00
-3.63482356e-01 -7.83608913e-01 1.29713982e-01 2.31882200e-01
-6.55681849e-01 3.55696470e-01 -6.47227526e-01 -5.53211212e-01
7.54565969e-02 -8.72182727e-01 -2.20124468e-01 2.22752064e-01
7.02909529e-02 -6.70831144e-01 1.01287794e+00 -6.27487361e-01
1.18744016e+00 -2.21976161e+00 3.91867459e-01 2.74999022e-01
3.75657141e-01 3.83577257e-01 -4.02532607e-01 2.48747557e-01
-1.67384535e-01 -1.07884131e-01 -1.31147532e-02 -2.84648210e-01
-9.14027914e-02 3.68964493e-01 2.68680733e-02 4.86708373e-01
4.56881076e-01 1.10594189e+00 -8.60909104e-01 -6.69423223e-01
7.12909475e-02 9.05468166e-01 -4.16907549e-01 3.18915248e-01
4.03036596e-03 4.70702440e-01 -7.87850678e-01 1.02991557e+00
7.23733366e-01 -1.92937881e-01 5.30819874e-03 -4.61250693e-01
3.77882540e-01 -4.31710333e-01 -1.01277173e+00 1.67675757e+00
-3.76177311e-01 4.64752465e-01 3.40010494e-01 -1.13483477e+00
1.33740783e+00 2.96359420e-01 6.50462270e-01 -8.53226066e-01
6.95467830e-01 -9.49982703e-02 -2.13584080e-02 -5.28411567e-01
1.25755638e-01 -7.67663792e-02 2.01782033e-01 1.89117014e-01
2.43073031e-01 3.09434056e-01 -2.13847354e-01 7.43793249e-02
8.64093125e-01 3.21264923e-01 3.07975560e-01 2.02117741e-01
9.00714934e-01 -5.67248404e-01 6.69896007e-01 -7.36394227e-02
-4.30137664e-01 2.69705236e-01 6.59672022e-01 -5.36781132e-01
-3.10040444e-01 -6.25366211e-01 5.25455065e-02 1.28133619e+00
2.61685938e-01 -4.00232852e-01 -6.97825193e-01 -9.21237469e-01
-6.80956542e-02 4.01169844e-02 -1.02835274e+00 -5.12803555e-01
-4.42035764e-01 -6.70684695e-01 5.48965037e-01 6.60692692e-01
9.19554174e-01 -1.43614340e+00 -4.61036824e-02 4.41148542e-02
-6.99400008e-02 -1.20352149e+00 -2.34781444e-01 -3.38241756e-01
-3.99748087e-01 -1.12952256e+00 -5.10448694e-01 -1.05778098e+00
8.91555727e-01 3.19146067e-01 7.82155335e-01 3.18792492e-01
-4.91178393e-01 5.70360541e-01 -6.93736970e-01 -3.69363874e-01
3.61772105e-02 -3.23537469e-01 -3.16619754e-01 8.79450560e-01
5.76807439e-01 -5.76125622e-01 -6.29178107e-01 3.60861897e-01
-9.86015141e-01 -2.18024194e-01 6.07104897e-01 8.01208496e-01
4.56173718e-01 2.98432540e-02 6.41543627e-01 -8.15957546e-01
5.82273543e-01 -6.87629223e-01 -2.65696883e-01 3.24590564e-01
-2.17464745e-01 -2.64976591e-01 4.16740566e-01 -3.16054553e-01
-1.23525953e+00 4.50373702e-02 -3.29513639e-01 -7.85453022e-01
-1.44550845e-01 4.78613436e-01 -4.87055004e-01 -5.66928864e-01
2.54091710e-01 2.31677145e-01 2.98029184e-01 -4.64611687e-02
4.47613418e-01 4.86909062e-01 2.18997300e-01 -5.25687873e-01
5.78553677e-01 4.67254102e-01 2.15129271e-01 -6.64121449e-01
-7.07478523e-01 -2.54489303e-01 -6.33653879e-01 -5.81695914e-01
9.65710223e-01 -9.56055939e-01 -6.62653506e-01 5.37083507e-01
-1.10843778e+00 -3.83839794e-02 1.47097304e-01 1.06992222e-01
-2.72672415e-01 4.73920226e-01 -5.94097674e-01 -6.33017063e-01
-3.53596985e-01 -1.22183621e+00 1.20835030e+00 5.46671331e-01
1.30804449e-01 -1.09462476e+00 -1.51010752e-01 3.79876137e-01
3.45136762e-01 6.27341568e-01 7.44957805e-01 -1.56844631e-01
-3.16095591e-01 -4.06602979e-01 -6.55092001e-01 4.44756955e-01
5.73128819e-01 1.45672724e-01 -9.31654215e-01 9.46970377e-03
-3.03252876e-01 -6.27808869e-01 8.17736089e-01 1.16687706e-02
1.47596920e+00 -2.98382103e-01 -2.19664544e-01 7.41114318e-01
1.20609057e+00 1.33149505e-01 7.69969761e-01 -3.62349451e-02
1.04916632e+00 9.25578356e-01 6.48461223e-01 6.52279854e-01
4.59766418e-01 6.63854301e-01 8.62958014e-01 -4.56561357e-01
-1.08669169e-01 -8.84985849e-02 3.30496222e-01 6.79679036e-01
-5.66266418e-01 1.32496823e-02 -4.57944870e-01 2.60614514e-01
-2.13048410e+00 -8.93043399e-01 2.12521762e-01 1.39387882e+00
5.36113560e-01 -4.81381983e-01 -2.08283365e-02 -2.12459400e-01
6.24134839e-01 5.37264585e-01 -4.76181567e-01 -4.35361922e-01
-2.77398556e-01 2.42344335e-01 -8.02101418e-02 9.73797217e-02
-9.91677165e-01 1.40500033e+00 5.01082277e+00 7.29153812e-01
-1.17542303e+00 2.15104595e-02 6.93547666e-01 3.60063702e-01
-7.22194761e-02 -2.39443138e-01 -4.75662678e-01 1.52936250e-01
3.72367233e-01 5.70866354e-02 4.43539530e-01 8.80249739e-01
-6.19545542e-02 3.66105348e-01 -6.90060556e-01 1.24632061e+00
5.02945423e-01 -9.68230188e-01 3.84505689e-01 -2.36696154e-02
6.18046463e-01 -3.34958166e-01 9.90870874e-03 4.29292500e-01
2.47728020e-01 -1.30312872e+00 1.29243359e-01 6.53661907e-01
6.54490113e-01 -8.74596000e-01 9.59798455e-01 -1.87410265e-02
-1.63463807e+00 -2.73240283e-02 -5.13918698e-01 -9.94358286e-02
-2.52355099e-01 -5.88018745e-02 -4.54549015e-01 8.91666651e-01
7.78269410e-01 1.30860293e+00 -5.92671514e-01 4.39071298e-01
-3.90128702e-01 3.73005182e-01 1.35455383e-02 -6.15183786e-02
3.97649199e-01 -5.08179486e-01 -7.98790529e-02 9.61833656e-01
8.16172436e-02 3.93011183e-01 8.07811245e-02 7.01946080e-01
-3.16509902e-01 5.75539052e-01 -8.16686034e-01 -1.10675238e-01
-4.89460975e-02 1.76255369e+00 -6.17825627e-01 -1.63001269e-02
-8.07110012e-01 1.02757573e+00 7.85924554e-01 4.79805946e-01
-9.23604906e-01 -1.77917019e-01 7.78038383e-01 -2.31637225e-01
2.30262265e-01 -4.67553623e-02 4.15347189e-01 -1.15620947e+00
-9.10797343e-02 -8.03429604e-01 5.75983942e-01 -7.40982592e-01
-1.33108735e+00 8.88889134e-01 -2.76092172e-01 -1.05244648e+00
3.13380323e-02 -6.65324807e-01 -7.62674928e-01 6.10514402e-01
-1.81446922e+00 -1.60871482e+00 -7.28100955e-01 1.21490133e+00
3.31135720e-01 -1.27117679e-01 8.06609154e-01 4.19550359e-01
-8.85400355e-01 8.10857296e-01 -5.90891123e-01 5.68520367e-01
5.48818052e-01 -5.99830508e-01 -3.75370711e-01 5.92465401e-01
8.21787640e-02 3.68043154e-01 1.05208009e-01 -3.69496614e-01
-1.53795660e+00 -1.40242183e+00 4.76238370e-01 5.72675876e-02
7.28944957e-01 -3.33235741e-01 -9.30393457e-01 7.81571984e-01
2.15931177e-01 7.11344004e-01 8.06093633e-01 -1.41021768e-02
-5.00357091e-01 -3.90446872e-01 -8.76477182e-01 6.41958952e-01
1.31321084e+00 -4.51075137e-01 -1.67557538e-01 1.40641421e-01
5.08730114e-01 -1.50287732e-01 -9.84338284e-01 7.23148704e-01
5.09012818e-01 -7.94523835e-01 6.99118912e-01 -7.56909847e-01
4.14658159e-01 -2.68775403e-01 -2.03386620e-01 -1.14869463e+00
-3.57373744e-01 -4.16481733e-01 1.70908589e-02 1.48479736e+00
-1.72852054e-02 -7.07293093e-01 7.21679807e-01 4.28207248e-01
-9.96991619e-03 -1.12232530e+00 -7.95252800e-01 -2.72945523e-01
-2.88092494e-01 -2.71511674e-01 7.37003088e-01 1.09468210e+00
-1.96198896e-01 4.25144285e-01 -4.33257669e-01 2.14127570e-01
3.91790599e-01 2.37183541e-01 8.22945178e-01 -1.15519154e+00
1.37324736e-01 -6.91843748e-01 -8.96068394e-01 -6.88742399e-01
7.39919305e-01 -9.64963973e-01 -2.80694127e-01 -1.61782992e+00
1.23909876e-01 -3.28260183e-01 -6.63577557e-01 9.43196833e-01
-2.89145112e-01 3.69567573e-01 9.99864861e-02 -1.07517779e-01
-8.89227390e-01 1.17077923e+00 1.51112819e+00 -3.48373920e-01
1.11661395e-02 -2.43335843e-01 -7.68183470e-01 8.17133725e-01
6.06995761e-01 2.78387647e-02 -4.69270349e-01 -3.35946470e-01
2.06325591e-01 1.13321468e-01 4.29064840e-01 -6.70967042e-01
2.12065831e-01 -7.54263625e-02 4.21742141e-01 -2.55199164e-01
3.56732368e-01 -1.15564919e+00 -1.97188333e-01 6.22401461e-02
-2.09292863e-02 -8.58711079e-02 1.63050309e-01 6.18297696e-01
-6.90560997e-01 3.52785051e-01 6.50945246e-01 -1.55409008e-01
-1.16085887e+00 9.68963563e-01 1.37447745e-01 -3.37172896e-01
1.25243664e+00 -3.19488019e-01 -1.39479965e-01 -5.11749566e-01
-6.90456212e-01 3.50109786e-01 8.35190639e-02 7.61455715e-01
9.21736598e-01 -1.58510876e+00 -6.48746252e-01 3.49155992e-01
4.36823308e-01 -4.75056134e-02 5.08210957e-01 8.51263523e-01
-3.11107546e-01 -1.67478938e-02 -4.18418139e-01 -4.64331090e-01
-1.38139927e+00 3.07104528e-01 4.98935550e-01 -5.07085845e-02
-3.14188361e-01 9.77743447e-01 5.66258729e-01 -3.91912371e-01
2.45409682e-01 2.08931521e-01 -8.65476370e-01 2.31811553e-01
4.85537082e-01 4.32302430e-02 -2.08289072e-01 -1.30554307e+00
-3.75968099e-01 9.78716254e-01 2.92056035e-02 7.16227233e-01
1.37036788e+00 -2.03960165e-02 -6.12580597e-01 1.04601130e-01
1.43690300e+00 -2.98399240e-01 -1.18210447e+00 -5.56647539e-01
-3.72000992e-01 -3.27107370e-01 1.69722941e-02 -2.73974150e-01
-1.78843939e+00 8.48883986e-01 4.34147537e-01 -1.42696813e-01
1.52951121e+00 1.45377949e-01 6.25737488e-01 1.58221498e-01
4.37356234e-01 -8.05859029e-01 4.46070820e-01 3.20039153e-01
9.61206317e-01 -1.25924253e+00 -5.96098639e-02 -7.64714122e-01
-6.84129059e-01 1.29533041e+00 1.11653078e+00 -2.64362484e-01
9.68460381e-01 -4.51783463e-02 7.95927346e-02 -4.55937892e-01
-4.43115771e-01 -3.44845027e-01 3.25551629e-01 4.62561429e-01
3.80867094e-01 -5.49697205e-02 -2.10209519e-01 7.14472950e-01
1.40460610e-01 -4.77520414e-02 -1.07870325e-01 7.62860894e-01
-2.32425109e-01 -9.95996892e-01 2.10010465e-02 2.84503281e-01
-3.47885996e-01 1.08520016e-01 -6.62343144e-01 8.89193058e-01
4.31813419e-01 9.30450201e-01 4.70090769e-02 -5.83683252e-01
2.39626154e-01 -1.30981952e-01 5.79352617e-01 -5.09812415e-01
-4.75368679e-01 2.49140970e-02 -1.36252150e-01 -8.72073054e-01
-8.35952580e-01 -4.17575359e-01 -1.37349939e+00 -6.21736757e-02
-3.33914131e-01 -1.76485434e-01 3.55295718e-01 1.03806627e+00
4.66544777e-01 7.05830336e-01 8.50848675e-01 -6.50624752e-01
1.20248169e-01 -8.10641408e-01 -8.43762875e-01 5.98176718e-01
6.10364089e-03 -1.03449142e+00 -2.01689735e-01 -1.99624255e-01] | [13.644853591918945, 1.618565320968628] |
2dfa2798-c4b8-499a-93af-608102891dc0 | automated-mobile-attention-kpconv-networks-1 | null | null | https://openreview.net/group?id=ICLR.cc/2022/Conference | https://openreview.net/forum?id=VZC5Lzyl0le | Automated Mobile Attention KPConv Networks via a Wide and Deep Predictor | Kernel Point Convolution (KPConv) achieves cutting-edge performance on 3D point cloud applications. Unfortunately, the large size of KPConv network limits its usage in mobile scenarios. In addition, we observe that KPConv ignores the kernel relationship and treats each kernel point equally when formulating neighbor-kernel correlation via Euclidean distance. This leads to a weak representation power. To mitigate the above issues, we propose a module named Mobile Attention Kernel Point Convolution (MAKPConv) to improve the efficiency and quality of KPConv. MAKPConv employs a depthwise kernel to reduce resource consumption, and re-calibrates the contribution of kernel points towards each neighbor point via Neighbor-Kernel attention to improve representation power. Furthermore, we capitalize Inverted Residual Bottleneck (IRB) to craft a design space and employ a predictor-based Neural Architecture Search (NAS) approach to automate the design of efficient 3D networks based on MAKPConv. To fully exploit the immense design space via an accurate predictor, we identify the importance of carrying feature engineering on searchable features to improve neural architecture representations, and propose a Wide & Deep Predictor to unify dense and sparse neural architecture representations for lower error in performance prediction. Experimental evaluations show that our NAS-crafted MAKPConv network uses 96% fewer parameters on 3D point cloud classification and segmentation benchmarks with better performance. Compared with state-of-the-art NAScrafted model SPVNAS, our NAS-crafted MAKPConv network achieves 1% better mIOU with 83% fewer parameters and 52% fewer Multiply-Accumulates. | ['Anonymous'] | 2022-09-28 | null | null | null | international-conference-on-learning | ['point-cloud-classification'] | ['computer-vision'] | [-3.63566697e-01 -2.63823539e-01 -3.76825124e-01 -1.91751689e-01
-3.82406533e-01 -4.29503947e-01 1.33968338e-01 -2.58795619e-01
-1.97734594e-01 -4.38337997e-02 6.35533407e-02 -8.75229120e-01
-2.59539455e-01 -9.21853721e-01 -1.00197327e+00 -2.84911036e-01
1.11875117e-01 1.82579339e-01 2.97826678e-01 -1.86846793e-01
3.20081264e-01 1.03098691e+00 -1.38401604e+00 4.03880864e-01
7.22230554e-01 1.32321215e+00 3.36893499e-01 4.72720504e-01
-4.27305013e-01 4.80603278e-01 -4.09347028e-01 -2.81151712e-01
5.29940903e-01 3.41536552e-01 -6.83270216e-01 -4.26499009e-01
6.31939113e-01 -4.50521976e-01 -5.86183965e-01 6.40721142e-01
5.70432365e-01 9.89572704e-03 6.59193635e-01 -1.30954850e+00
-8.23212385e-01 5.57098746e-01 -7.10271597e-01 4.19291705e-01
-3.94991875e-01 3.67724359e-01 9.30735588e-01 -1.22447276e+00
-3.96063700e-02 1.01191342e+00 1.13869166e+00 3.16564441e-01
-1.07271659e+00 -8.95629227e-01 3.07827443e-01 -1.06050828e-02
-1.67181766e+00 -2.76621044e-01 7.32029557e-01 -2.34231949e-01
1.54169869e+00 4.34214413e-01 7.17296302e-01 8.57814074e-01
1.15110159e-01 8.44351232e-01 2.75278062e-01 -8.59409794e-02
2.09450111e-01 -6.77930936e-02 3.08427006e-01 8.04916680e-01
2.63187796e-01 5.90606406e-02 -2.82123625e-01 -1.62560701e-01
1.39194798e+00 4.29076552e-01 -2.64061213e-01 -5.17344654e-01
-1.03353560e+00 7.20098376e-01 9.35745299e-01 1.17145471e-01
-3.71547312e-01 6.03899300e-01 4.12584573e-01 1.89333782e-01
2.00933918e-01 5.73039055e-01 -8.28025401e-01 -2.59120047e-01
-8.81209075e-01 1.49248734e-01 5.49970567e-01 1.16927147e+00
7.87051260e-01 2.99388230e-01 -1.65941283e-01 9.24231768e-01
4.92323330e-03 5.59869111e-01 3.15108091e-01 -9.13088083e-01
5.87951005e-01 1.04221249e+00 -2.93107659e-01 -1.01722503e+00
-3.91078115e-01 -6.95747018e-01 -7.48800874e-01 2.66925365e-01
-1.00570068e-01 1.35973886e-01 -9.97237623e-01 1.19873619e+00
1.95766583e-01 6.79629207e-01 -1.98205382e-01 1.08100235e+00
8.51541877e-01 6.95528686e-01 -1.72844216e-01 4.93628293e-01
1.11320233e+00 -1.16091895e+00 2.17769980e-01 -9.16608647e-02
8.42200696e-01 -4.47258890e-01 1.35519385e+00 4.96819019e-02
-9.61111188e-01 -7.20123053e-01 -9.87562835e-01 -2.02787414e-01
-2.93643266e-01 3.25634778e-01 1.06349409e+00 6.07753575e-01
-1.09204805e+00 7.44208217e-01 -1.07834494e+00 -1.85947912e-03
6.92222595e-01 9.75743413e-01 -9.69807133e-02 3.48033756e-02
-5.49335241e-01 4.89438146e-01 3.12202483e-01 9.05974731e-02
-5.51696658e-01 -1.38604271e+00 -6.82323635e-01 3.96066427e-01
1.55016959e-01 -7.75250971e-01 1.02065182e+00 -6.97569251e-01
-1.50248635e+00 3.27605426e-01 -4.01632115e-02 -4.56449479e-01
-6.17610440e-02 -4.27129269e-01 -2.60944337e-01 -2.87990216e-02
-2.53205776e-01 9.48772132e-01 7.74278879e-01 -8.41132879e-01
-6.38076067e-01 -1.60946608e-01 1.04806744e-01 8.69894102e-02
-5.79280257e-01 -3.38057995e-01 -1.04112327e+00 -6.99623942e-01
9.84869301e-02 -1.12347043e+00 -2.25343049e-01 5.64999180e-03
-1.96666136e-01 -3.32067728e-01 1.10142779e+00 -1.41898587e-01
1.31252944e+00 -2.44192982e+00 -2.08379924e-01 3.81565243e-01
4.20242071e-01 6.29961073e-01 -3.70220542e-01 -3.35601121e-02
-1.64330080e-01 1.43467143e-01 1.81043908e-01 -3.55565399e-01
-1.47199094e-01 3.00644130e-01 -5.26967049e-01 3.06523442e-01
3.75754029e-01 1.24044716e+00 -4.15603399e-01 -1.77157700e-01
3.83640587e-01 7.61241078e-01 -1.10649872e+00 -1.56197280e-01
-1.20747179e-01 -2.58259028e-01 -6.24113977e-01 9.63247538e-01
8.04550707e-01 -6.72907352e-01 -3.83821160e-01 -5.76472700e-01
-8.05094242e-02 1.90068170e-01 -8.80693972e-01 1.77875066e+00
-7.24017859e-01 5.01150310e-01 -1.78565592e-01 -7.24521816e-01
1.06298661e+00 -1.45040944e-01 4.30180848e-01 -5.22837937e-01
2.90899545e-01 2.13671491e-01 -2.43592232e-01 -8.75315536e-03
6.27891123e-01 5.50086796e-01 6.79548755e-02 9.54318643e-02
-1.74801722e-02 1.99522167e-01 -5.55375934e-01 6.09639101e-02
1.29422069e+00 -2.77884621e-02 -2.20756918e-01 -2.64163196e-01
3.12766314e-01 8.84053335e-02 3.31863731e-01 7.01006770e-01
-2.62337271e-02 7.06032455e-01 3.26764673e-01 -8.48750293e-01
-9.30511355e-01 -9.30020511e-01 -2.67534047e-01 1.21171641e+00
2.35828981e-01 -5.44527531e-01 -5.91567099e-01 -7.57447183e-01
4.58181143e-01 7.54544914e-01 -6.27931118e-01 -2.59148598e-01
-7.91580915e-01 -5.54039419e-01 7.49629200e-01 1.08444405e+00
4.13154900e-01 -7.32216716e-01 -5.18696606e-01 4.77194637e-02
5.19732714e-01 -9.32558179e-01 -6.54167354e-01 4.77204025e-01
-1.07942283e+00 -6.58167899e-01 -4.72394139e-01 -7.62632191e-01
6.18431807e-01 5.95102251e-01 1.02639079e+00 1.67807743e-01
-5.13226315e-02 2.56142378e-01 -3.27831924e-01 -2.98978001e-01
1.32891521e-01 5.66296995e-01 6.89112246e-02 -4.41077650e-01
6.39938653e-01 -8.20733368e-01 -8.40016901e-01 4.89148378e-01
-5.71518004e-01 1.08486444e-01 9.30973649e-01 8.06436419e-01
8.16237509e-01 -1.92383185e-01 1.99686199e-01 -4.98860002e-01
3.99858743e-01 -5.36826074e-01 -6.94745362e-01 -6.04998097e-02
-8.44750702e-01 2.50523448e-01 5.98971128e-01 -6.48777902e-01
-3.58577073e-01 1.72620356e-01 -1.58941448e-01 -1.40942919e+00
7.21116513e-02 2.49721929e-01 2.86302138e-02 -5.24489939e-01
6.78313494e-01 1.99129850e-01 1.43030882e-01 -6.48002446e-01
4.31909621e-01 5.32238245e-01 5.18986821e-01 -6.40831888e-01
6.80256844e-01 2.47080460e-01 5.20464852e-02 -6.24507844e-01
-4.18065250e-01 -4.01064128e-01 -3.77440214e-01 3.17718089e-01
7.15978086e-01 -1.13989103e+00 -1.00025749e+00 7.87947476e-02
-1.09843755e+00 -3.69717300e-01 -2.91253418e-01 1.96119010e-01
-1.71646863e-01 -8.87512416e-02 -5.98753214e-01 -4.02176321e-01
-8.05872262e-01 -1.37503541e+00 1.27093458e+00 1.20960638e-01
-1.91536993e-01 -6.05204582e-01 -3.49132121e-01 2.24842057e-01
7.11004138e-01 -1.82227954e-01 1.03351831e+00 -6.63548470e-01
-9.31847513e-01 -1.95506990e-01 -7.33749688e-01 3.55866224e-01
-1.19148590e-01 -1.55729622e-01 -8.55794013e-01 -2.37152725e-01
-3.39334130e-01 1.78181782e-01 8.25742483e-01 4.35699731e-01
1.66589618e+00 -9.42979530e-02 -4.81160611e-01 1.35068142e+00
1.39420819e+00 -4.85744094e-03 3.85508597e-01 3.66682976e-01
1.19345474e+00 -1.77325308e-01 1.33572698e-01 2.50367224e-01
3.60768080e-01 6.18592203e-01 7.44316399e-01 -1.32130966e-01
-2.39874572e-01 -2.69697458e-01 1.76924512e-01 8.68867278e-01
-1.47837386e-01 7.12899417e-02 -1.00864089e+00 4.72612768e-01
-1.72195327e+00 -4.40010369e-01 1.75543986e-02 1.79175723e+00
4.43258733e-01 3.13527197e-01 -9.38725695e-02 -1.81868747e-01
3.38678420e-01 4.79576178e-02 -7.65733123e-01 -4.35753495e-01
9.10014585e-02 4.53991741e-01 9.29922998e-01 9.07087177e-02
-1.04864907e+00 1.11357629e+00 5.68732166e+00 1.19787669e+00
-1.61307836e+00 3.06264870e-03 6.60270452e-01 -4.70316440e-01
-2.57775426e-01 -2.85836697e-01 -1.18419325e+00 4.10712898e-01
9.59308386e-01 3.26208025e-01 4.31357861e-01 1.56502330e+00
-6.75013661e-02 4.91371274e-01 -1.17335129e+00 1.41915154e+00
-1.98922977e-01 -2.00038862e+00 2.69500077e-01 2.64019966e-01
4.40027982e-01 5.87190926e-01 2.74617791e-01 5.61564863e-01
5.04677631e-02 -1.29264319e+00 7.01791823e-01 1.53608575e-01
9.73431885e-01 -1.11139774e+00 6.94576740e-01 8.98733214e-02
-1.33258522e+00 -2.18822286e-01 -6.74334109e-01 -7.71996528e-02
-2.58959562e-01 4.67204571e-01 -9.97992098e-01 1.98499650e-01
9.15822208e-01 6.18538618e-01 -6.06429458e-01 9.35969532e-01
2.93092519e-01 5.92213750e-01 -5.16813755e-01 -1.96061373e-01
4.92949098e-01 1.17256485e-01 3.31323504e-01 1.14049315e+00
4.25105035e-01 9.80352759e-02 -5.17834425e-02 1.00308692e+00
-1.79577351e-01 -8.31037015e-02 -4.56293076e-01 1.40958756e-01
8.24961483e-01 1.09288359e+00 -6.32029176e-01 -5.37189357e-02
-5.80980539e-01 1.10230577e+00 5.33405185e-01 2.62170255e-01
-1.03415632e+00 -2.79159993e-01 1.27103579e+00 2.51477003e-01
9.61553454e-01 -2.32953236e-01 -6.81571782e-01 -9.03504312e-01
-1.32232830e-01 -5.72737932e-01 -4.35597301e-02 -6.84623420e-01
-1.13887393e+00 8.10026228e-01 -1.85456872e-01 -1.13351977e+00
2.16055065e-01 -8.10900390e-01 -6.71129167e-01 1.00240028e+00
-1.42126727e+00 -1.26175559e+00 -3.29431981e-01 6.52873099e-01
5.67983747e-01 -4.20742810e-01 6.87125206e-01 4.50005233e-01
-6.68427646e-01 9.75657463e-01 -1.76083222e-01 1.28097728e-01
1.75673962e-01 -9.67251956e-01 1.00167584e+00 3.51942509e-01
8.05159882e-02 9.99125600e-01 -2.63573192e-02 -4.15922701e-01
-1.99993122e+00 -1.54293466e+00 1.81133926e-01 -7.23778009e-01
4.89688843e-01 -2.17237264e-01 -9.12050664e-01 5.85830390e-01
-3.49324107e-01 5.34801483e-01 6.68754578e-01 3.02982241e-01
-4.89733875e-01 -2.97156006e-01 -7.61860549e-01 7.05223083e-01
1.30954754e+00 -4.70262557e-01 -3.50645483e-01 -7.96876550e-02
1.44281888e+00 -5.62565804e-01 -8.60565841e-01 7.15080380e-01
4.51035798e-01 -8.81422341e-01 1.38666952e+00 -4.69632208e-01
3.48352134e-01 -3.77107143e-01 -3.00357521e-01 -8.51808012e-01
-7.36089587e-01 -2.83364385e-01 -4.56242502e-01 8.82994056e-01
5.53038657e-01 -4.46370304e-01 1.34426892e+00 6.36852860e-01
-6.21905804e-01 -1.60130525e+00 -8.82407486e-01 -7.53925264e-01
6.41016886e-02 -8.64804804e-01 1.05521870e+00 9.48171735e-01
-5.64867437e-01 2.29672417e-01 -8.22775736e-02 4.50064391e-01
1.40228227e-01 2.15052947e-01 9.11078215e-01 -9.83268082e-01
-3.88378918e-01 -6.90085828e-01 -5.78962922e-01 -1.50380838e+00
1.53668197e-02 -9.60341156e-01 -4.78077501e-01 -1.01361823e+00
-1.49809733e-01 -9.67168808e-01 -4.72394913e-01 6.75199091e-01
2.62670554e-02 2.76862532e-01 2.14534372e-01 3.61735344e-01
-4.82388914e-01 5.36051750e-01 1.08553040e+00 3.26011926e-02
-5.32137215e-01 -6.98267967e-02 -9.11905825e-01 7.89179802e-01
5.60478151e-01 -2.23818123e-01 -5.64816952e-01 -9.58196461e-01
1.31797656e-01 -2.66972601e-01 4.72711444e-01 -1.21530867e+00
4.25121009e-01 -2.60853698e-03 6.58992887e-01 -9.27519202e-01
4.99713361e-01 -9.57880855e-01 1.03204027e-01 2.11013123e-01
1.42218903e-01 4.03816700e-01 5.64269722e-01 3.92259628e-01
6.42012209e-02 5.42572327e-02 4.59137052e-01 8.52037743e-02
-9.72418725e-01 6.19868100e-01 2.32818380e-01 -3.99815500e-01
9.06084597e-01 -5.48854887e-01 -2.75398403e-01 2.43190795e-01
-2.47825041e-01 1.41801000e-01 5.71872294e-01 3.99013191e-01
9.68044698e-01 -1.38685632e+00 -1.67484760e-01 5.28194726e-01
-5.96258566e-02 4.44678456e-01 2.47958452e-01 6.52762830e-01
-1.07538676e+00 5.74529171e-01 -6.36201501e-02 -7.96113908e-01
-9.20854449e-01 6.17614269e-01 3.04843694e-01 2.99678044e-03
-9.37226355e-01 1.28035188e+00 2.19048128e-01 -4.91646171e-01
3.76161933e-01 -7.31926799e-01 1.69555783e-01 -5.05119681e-01
4.19485122e-01 3.62163663e-01 3.02095592e-01 -3.01708430e-01
-5.14352739e-01 6.38924778e-01 -2.77998865e-01 5.89287400e-01
1.45135427e+00 3.90095055e-01 8.58146697e-02 -1.25104114e-01
1.46602392e+00 3.69008407e-02 -1.37498963e+00 -9.33987275e-02
-3.68027061e-01 -4.09442484e-01 6.04436338e-01 -6.75804257e-01
-1.45739102e+00 7.01913297e-01 6.60952806e-01 -2.29799390e-01
1.11995602e+00 4.00979668e-02 1.20635819e+00 5.80162048e-01
2.87723631e-01 -6.55740678e-01 -7.25555867e-02 7.92528987e-01
7.10188627e-01 -1.01711655e+00 -5.57303019e-02 -6.30181253e-01
-4.57029581e-01 1.06160545e+00 9.82719123e-01 -4.90511715e-01
9.19304609e-01 4.43920553e-01 -2.45391771e-01 -3.63068610e-01
-7.19759405e-01 1.50026292e-01 4.84569460e-01 4.34223831e-01
8.02818462e-02 -1.10797100e-02 4.80991185e-01 1.01579690e+00
-3.65617722e-01 -7.76187479e-02 -9.38970596e-02 8.07433426e-01
-2.38753885e-01 -8.62470925e-01 -1.93680689e-01 6.68461263e-01
-1.66565374e-01 -5.33389986e-01 -1.77075546e-02 7.47470915e-01
2.00730264e-01 2.56941646e-01 4.41460729e-01 -9.70544934e-01
4.93062705e-01 -2.33958408e-01 1.04898803e-01 -4.37116504e-01
-8.97751451e-01 -6.53116629e-02 -2.80433655e-01 -8.87189865e-01
2.92986542e-01 -2.09749267e-01 -1.29701293e+00 -6.62477612e-01
-4.01932538e-01 -1.25728443e-01 8.53082836e-01 5.76450527e-01
1.29079199e+00 7.37454355e-01 4.18085277e-01 -1.01647234e+00
-5.27937531e-01 -5.37615299e-01 -2.72794485e-01 -6.11499473e-02
2.16684073e-01 -6.02504492e-01 -1.43086731e-01 -5.37499070e-01] | [7.88428258895874, -3.5417747497558594] |
52dd90e8-b533-4479-b5e8-b42955c1a412 | qmul-sds-diacr-ita2020-evaluating | 2011.02935 | null | https://arxiv.org/abs/2011.02935v2 | https://arxiv.org/pdf/2011.02935v2.pdf | QMUL-SDS @ DIACR-Ita: Evaluating Unsupervised Diachronic Lexical Semantics Classification in Italian | In this paper, we present the results and main findings of our system for the DIACR-ITA 2020 Task. Our system focuses on using variations of training sets and different semantic detection methods. The task involves training, aligning and predicting a word's vector change from two diachronic Italian corpora. We demonstrate that using Temporal Word Embeddings with a Compass C-BOW model is more effective compared to different approaches including Logistic Regression and a Feed Forward Neural Network using accuracy. Our model ranked 3rd with an accuracy of 83.3%. | ['Maria Liakata', 'Arkaitz Zubiaga', 'Adam Tsakalidis', 'Rabab Alkhalifa'] | 2020-11-05 | null | null | null | null | ['diachronic-word-embeddings'] | ['natural-language-processing'] | [-1.13780782e-01 -1.59272105e-01 -3.68391067e-01 -4.68987614e-01
-8.38293254e-01 -6.17672741e-01 9.58260000e-01 2.92036027e-01
-9.11268890e-01 5.05337656e-01 4.18690950e-01 -2.92084634e-01
1.28607497e-01 -5.77207685e-01 -1.49538293e-01 -4.89648134e-01
-1.63521960e-01 6.97904468e-01 3.45879287e-01 -4.45320040e-01
3.79790694e-01 2.02096254e-01 -1.15334821e+00 3.99479806e-01
2.88699210e-01 1.04614198e+00 -5.70164248e-02 5.85553765e-01
-1.85874477e-01 4.89888966e-01 -8.51585686e-01 -5.20896614e-01
2.31785774e-01 1.51455833e-03 -1.14837623e+00 -7.63241768e-01
1.77758723e-01 2.59227246e-01 -4.18418080e-01 9.75490689e-01
5.88027954e-01 2.57366151e-01 8.20858955e-01 -1.11946809e+00
-8.04210603e-01 7.05765426e-01 -3.43907624e-01 7.76337981e-01
4.16957229e-01 -3.80636930e-01 1.17500305e+00 -9.83797550e-01
8.34116518e-01 1.29550993e+00 9.91644442e-01 5.23351490e-01
-9.62321699e-01 -7.58106530e-01 2.72946376e-02 6.98630273e-01
-1.42281556e+00 -2.04669595e-01 6.77934229e-01 -4.79622781e-01
1.72823226e+00 2.67354101e-01 4.29895967e-01 1.29857266e+00
3.21198851e-01 6.49816811e-01 8.42717826e-01 -8.06289554e-01
9.61931944e-02 1.62090734e-01 6.92063987e-01 3.12350243e-01
1.21557914e-01 -7.21267611e-02 -4.91251290e-01 -1.64502963e-01
-1.15350615e-02 -4.77590948e-01 3.87371927e-01 1.67232100e-02
-9.37500000e-01 1.10088134e+00 3.41997355e-01 1.02297103e+00
-2.52787530e-01 4.74659175e-01 7.89105237e-01 4.04846519e-01
1.00735116e+00 7.07433701e-01 -9.75941956e-01 -5.01575112e-01
-8.99419785e-01 1.13098517e-01 4.40643787e-01 6.36486351e-01
2.16038182e-01 -8.68232846e-02 -4.94195484e-02 9.65632617e-01
7.84401745e-02 1.75013825e-01 1.21577084e+00 -5.71823418e-01
5.69719672e-01 2.27944240e-01 -1.13452331e-03 -1.02008343e+00
-6.26320779e-01 -2.27662288e-02 -2.09459230e-01 -1.96751997e-01
2.28434056e-01 -3.76888007e-01 -8.58252704e-01 1.71217060e+00
8.18565860e-02 1.13763595e-02 3.69589701e-02 4.51146662e-01
5.12260854e-01 7.75603592e-01 3.67674559e-01 -5.23877926e-02
1.44562364e+00 -1.09667814e+00 -1.03148580e+00 -2.29165584e-01
8.55474532e-01 -8.76107514e-01 6.61568940e-01 4.28184241e-01
-9.09214675e-01 -7.55319417e-01 -1.06584680e+00 -3.91712822e-02
-9.47913170e-01 7.82354474e-02 4.64068413e-01 6.24387205e-01
-1.01359308e+00 7.38939226e-01 -9.10020530e-01 -8.93821537e-01
1.99753135e-01 2.13000625e-01 -4.73296225e-01 3.56252521e-01
-1.55342078e+00 1.41039121e+00 6.67272687e-01 -3.47021580e-01
-9.19626281e-02 -6.87161982e-01 -8.68767202e-01 -1.88418552e-01
-3.78293455e-01 2.05607563e-02 1.29361165e+00 -9.64288712e-01
-1.36148858e+00 9.82251048e-01 -7.50456378e-02 -8.13680828e-01
1.64572477e-01 -6.41754270e-01 -8.55599880e-01 -1.45633053e-02
1.47706136e-01 8.65144849e-01 3.93895626e-01 -4.87518162e-01
-6.62845671e-01 -3.45306128e-01 -4.72554058e-01 -1.02575555e-01
-7.38110363e-01 6.46535575e-01 -3.65370482e-01 -9.38035011e-01
-1.40154853e-01 -9.17886734e-01 -2.46613175e-01 -5.51420331e-01
-2.06613657e-03 -8.41246188e-01 7.43423998e-01 -1.15807474e+00
1.72623229e+00 -2.04589248e+00 9.63273197e-02 -1.98695380e-02
-2.35481188e-01 4.10488397e-01 -1.58557341e-01 4.30453718e-01
-6.48994029e-01 2.28199720e-01 8.08239207e-02 -1.80350915e-01
1.60337724e-02 3.43285617e-03 -1.60552979e-01 3.26346874e-01
4.08901393e-01 9.28976417e-01 -1.03616989e+00 -3.76294583e-01
3.46225470e-01 3.84527922e-01 -4.49643046e-01 2.00895369e-02
6.23366721e-02 -1.87902793e-01 -3.25735956e-01 3.85340422e-01
3.74728620e-01 1.91395387e-01 4.35311109e-01 -3.72031957e-01
-2.52211004e-01 5.28147757e-01 -5.95587134e-01 1.99015868e+00
-5.26028514e-01 1.08501422e+00 -8.18467975e-01 -1.35407090e+00
1.11049652e+00 4.68831807e-01 5.56769013e-01 -1.08811569e+00
1.84187204e-01 1.79291293e-01 -6.85333833e-02 -5.29582024e-01
6.27195895e-01 2.97686290e-02 -6.71757042e-01 1.10424109e-01
4.15651292e-01 1.00166202e-01 5.16684130e-02 -5.55595756e-02
1.39010394e+00 2.38705114e-01 4.88687336e-01 -5.31724095e-01
5.08125424e-01 2.17172265e-01 5.57483912e-01 3.68724763e-01
-5.00115812e-01 5.37303984e-01 3.94128650e-01 -7.90903986e-01
-1.05780804e+00 -1.02994251e+00 -3.07972819e-01 9.84743655e-01
-2.13707536e-01 -6.65267408e-01 -5.98936975e-01 -9.73966599e-01
2.12505329e-02 1.31162274e+00 -1.03030574e+00 -2.83507973e-01
-1.03694987e+00 -8.94443035e-01 6.13287926e-01 7.34103322e-01
7.79903904e-02 -1.01548922e+00 -4.53016460e-01 4.63835090e-01
-3.15076739e-01 -1.03496265e+00 -4.36093301e-01 5.48099279e-01
-6.90011680e-01 -8.91084373e-01 -3.51011515e-01 -1.03686178e+00
1.09762773e-01 -2.51778871e-01 1.10970819e+00 -6.11098826e-01
-6.19879663e-01 1.00971855e-01 -4.47762251e-01 -4.37904894e-01
-2.34535381e-01 3.41709554e-01 8.41608047e-02 -3.64575237e-01
1.08511031e+00 -2.06684217e-01 -4.52187866e-01 -4.80374433e-02
-5.48898339e-01 -5.56546390e-01 2.57770270e-01 7.80703425e-01
9.86296460e-02 -2.13843957e-01 6.13849938e-01 -7.88302302e-01
4.99752134e-01 -6.58660591e-01 -3.69003117e-01 1.77282304e-01
-1.02018535e+00 1.59899816e-01 2.04428509e-01 -4.98140514e-01
-9.96820927e-01 -1.54294474e-02 -4.37782288e-01 -1.30265191e-01
1.62625104e-01 1.80347010e-01 1.99276254e-01 4.47059780e-01
7.77791023e-01 -2.55218655e-01 -2.10733220e-01 -9.04572606e-01
7.05613613e-01 1.05979300e+00 5.10340273e-01 -2.54558891e-01
4.77024615e-01 2.90785521e-01 -5.65001786e-01 -5.92865527e-01
-8.86379898e-01 -9.50711787e-01 -9.66641545e-01 -1.23457327e-01
1.17724812e+00 -8.08484197e-01 -2.82646678e-02 4.34385568e-01
-1.34478486e+00 -5.24820164e-02 -3.08765978e-01 5.83816826e-01
-3.87645781e-01 2.52047420e-01 -6.72620952e-01 -3.23076993e-01
-4.98709679e-01 -6.86665952e-01 9.47242200e-01 -4.35962938e-02
-9.73476827e-01 -1.30723596e+00 6.80865645e-01 2.03596741e-01
6.20221138e-01 4.06728797e-02 7.37807333e-01 -1.25798607e+00
5.42827666e-01 -3.96069825e-01 -5.06071523e-02 4.11146164e-01
3.47625107e-01 -1.36456072e-01 -1.02986658e+00 -1.98201343e-01
-3.09432238e-01 -1.78482220e-01 8.42524409e-01 3.22955072e-01
1.21155751e+00 -1.26987889e-01 -6.13042235e-01 1.71968743e-01
1.45654655e+00 4.39914614e-01 6.49238050e-01 9.10001755e-01
2.93004364e-01 5.53994119e-01 8.11144292e-01 2.31728762e-01
2.07302332e-01 1.24511468e+00 -9.20398440e-03 2.14735791e-01
-4.63443488e-01 -1.07689865e-01 5.22074878e-01 1.18590987e+00
2.18638927e-01 -2.07761183e-01 -1.16583848e+00 1.03664708e+00
-1.71578157e+00 -9.10474837e-01 -2.49101475e-01 1.92555070e+00
9.90101933e-01 2.46319875e-01 2.22452525e-02 6.80012032e-02
7.64915109e-01 2.52935648e-01 9.52838138e-02 -1.06047630e+00
-8.83238837e-02 1.02109647e+00 6.38298154e-01 4.33866352e-01
-1.45907843e+00 1.36923492e+00 7.67060852e+00 1.02046120e+00
-8.97232175e-01 7.28546977e-01 4.02035177e-01 -7.23035336e-02
-1.16980756e-02 -2.74268717e-01 -7.85350919e-01 5.55253148e-01
1.67665946e+00 -2.72134393e-01 -1.91707775e-01 6.90269351e-01
-8.66227522e-02 1.71911135e-01 -8.36070538e-01 7.76859939e-01
6.40684247e-01 -1.17599642e+00 -2.32440978e-01 -5.27429402e-01
7.22585559e-01 3.18823189e-01 -1.36215851e-01 6.15944505e-01
5.39418161e-01 -1.01180840e+00 6.47810400e-01 3.64160061e-01
6.12620294e-01 -6.11653209e-01 1.03336954e+00 -5.24985969e-01
-1.06086111e+00 -2.33195182e-02 -1.38699397e-01 4.28740792e-02
1.66273803e-01 3.98879051e-01 -6.74555779e-01 6.26276076e-01
9.31513369e-01 1.08120656e+00 -7.25042939e-01 8.40249062e-01
-1.02199979e-01 8.28887343e-01 -6.42238483e-02 -2.94533342e-01
4.54535335e-01 1.24384418e-01 3.04905891e-01 1.76473212e+00
3.36550534e-01 -2.20758915e-01 -2.51257718e-01 2.35632330e-01
5.81471846e-02 4.51910347e-01 -5.30058086e-01 5.18498756e-02
3.71644557e-01 1.07815135e+00 -5.67462981e-01 -4.74525154e-01
-4.89769667e-01 1.21586823e+00 4.40737933e-01 4.67652269e-02
-1.06124794e+00 -6.66744709e-01 8.37579072e-01 -3.14124912e-01
5.96396446e-01 -1.98408276e-01 -3.96296352e-01 -9.09157455e-01
-2.03212947e-01 -2.91325301e-01 6.35295331e-01 -8.17860246e-01
-1.33128822e+00 6.74310803e-01 4.52273563e-02 -9.67852056e-01
-2.41779208e-01 -9.37742889e-01 -5.79800248e-01 6.36970520e-01
-1.33944762e+00 -9.02819157e-01 8.46665129e-02 1.48200303e-01
7.71906972e-01 -4.77759570e-01 1.10088003e+00 4.17975634e-01
-5.41204512e-01 6.72642827e-01 3.67676884e-01 2.92268068e-01
1.01280868e+00 -1.45132291e+00 7.76147962e-01 6.33757472e-01
2.88565308e-01 1.87715366e-01 7.22875655e-01 -4.38320696e-01
-7.31878936e-01 -1.15793395e+00 1.74305689e+00 -5.48494160e-01
1.37265229e+00 -1.87123463e-01 -6.09899104e-01 8.27308655e-01
5.71020186e-01 -1.04560263e-01 6.79486811e-01 3.37844282e-01
-5.52079797e-01 -2.24655867e-01 -9.20743942e-01 3.48659426e-01
8.56276631e-01 -4.86423194e-01 -1.32963574e+00 4.78855312e-01
8.75662386e-01 8.36029202e-02 -1.02367234e+00 3.57095778e-01
4.93653804e-01 -2.06521735e-01 1.08504891e+00 -1.13404965e+00
2.26253644e-01 1.33915290e-01 -4.42160666e-01 -1.42561698e+00
-5.49970329e-01 -2.93100089e-01 1.84058636e-01 1.25806785e+00
6.36908054e-01 -5.04148662e-01 5.57942331e-01 8.40893462e-02
1.13856653e-02 -2.00255588e-01 -1.33679092e+00 -9.70010102e-01
7.09403396e-01 -7.89609253e-01 1.99276075e-01 1.14035511e+00
3.04544300e-01 4.59533870e-01 -1.23601235e-01 -1.75960168e-01
-2.79403497e-02 -3.12091351e-01 1.89512521e-01 -1.17030203e+00
1.27836615e-01 -6.46796227e-01 -9.33100045e-01 -4.91977036e-01
3.94143671e-01 -1.17028821e+00 -1.62144259e-01 -1.24095130e+00
3.44343372e-02 -3.74173641e-01 -8.82384896e-01 5.59240758e-01
-1.54058293e-01 3.31701428e-01 -1.45320877e-01 -3.94269861e-02
-7.28287041e-01 3.55171472e-01 1.25307128e-01 -4.41750437e-01
7.57608637e-02 -2.39483461e-01 -3.18428278e-01 5.64762414e-01
1.07656169e+00 -5.38747132e-01 8.79486650e-02 -4.40232903e-01
1.30380884e-01 -6.55983031e-01 -4.01823334e-02 -7.91866124e-01
-6.49142498e-03 3.35523605e-01 2.43449166e-01 -5.71153164e-01
1.96196333e-01 -6.19885623e-01 -7.27041857e-03 9.19700742e-01
-3.57690275e-01 7.46828020e-01 6.03214800e-01 3.10848147e-01
-3.76042187e-01 -3.84473264e-01 6.72049224e-01 1.98249936e-01
-1.15689445e+00 -2.34590933e-01 -6.15474164e-01 1.02426611e-01
1.28420365e+00 3.55073988e-01 -2.25183040e-01 1.76261634e-01
-9.48704541e-01 4.33795042e-02 -1.48752615e-01 1.20259225e+00
2.17887536e-01 -1.75707626e+00 -6.62448466e-01 3.77426185e-02
4.35229331e-01 -9.82375681e-01 -1.94097549e-01 6.02539718e-01
-6.48831069e-01 7.92827427e-01 -1.24476351e-01 -4.52848583e-01
-1.45838797e+00 5.59593797e-01 3.20334107e-01 -4.41378355e-01
-4.56108660e-01 9.49377120e-01 -4.30134803e-01 -6.20765090e-01
1.67617038e-01 -1.93304881e-01 -6.29299700e-01 5.64406574e-01
4.71748024e-01 3.76315266e-01 3.40371996e-01 -5.55086911e-01
-8.09421122e-01 6.08214021e-01 -1.12900354e-01 -2.60544956e-01
1.50075471e+00 1.01678729e-01 -6.44380823e-02 7.98510969e-01
1.60143018e+00 -1.23243295e-01 -3.47464323e-01 -3.40317875e-01
7.24569857e-01 -2.02324361e-01 2.29323760e-01 -9.09027636e-01
-8.86550486e-01 7.95962036e-01 1.07252300e+00 5.01550771e-02
7.57874370e-01 2.18503401e-01 7.95742035e-01 1.78186819e-01
2.01522037e-01 -1.68454278e+00 7.24211484e-02 4.39430773e-01
7.20662951e-01 -1.24739766e+00 -1.47628680e-01 -1.17049560e-01
-4.11095828e-01 1.34364104e+00 3.44353288e-01 -2.39963546e-01
1.13182580e+00 1.68436497e-01 2.08842754e-01 -2.49831393e-01
-8.63333404e-01 -1.90303698e-01 3.98471981e-01 4.84302461e-01
5.99883199e-01 2.04406619e-01 -1.00319791e+00 7.03276992e-01
-4.32126045e-01 -3.65026534e-01 1.46378621e-01 7.25665390e-01
-2.84804642e-01 -1.45400727e+00 9.47771315e-03 2.10540742e-01
-7.45239258e-01 -3.02884340e-01 -2.76091248e-01 1.01445448e+00
1.93154767e-01 1.00459051e+00 6.21739805e-01 -5.15995562e-01
4.28484976e-01 5.87921798e-01 4.11080033e-01 -5.94797730e-01
-7.18518794e-01 -2.22663090e-01 5.30597270e-01 -7.30366349e-01
-4.28814322e-01 -1.27674401e+00 -1.02957547e+00 -9.33336467e-02
-2.53119171e-01 2.41456583e-01 1.02376390e+00 9.03008878e-01
1.22695908e-01 6.24409854e-01 8.10083389e-01 -2.65335977e-01
-4.45708334e-01 -1.43144608e+00 -2.06866339e-01 6.03960454e-01
-3.01046103e-01 -5.49625218e-01 -4.27178234e-01 3.39613147e-02] | [10.148763656616211, 9.040098190307617] |
84256bae-8f19-46aa-90f8-bf96bf3a52b5 | video-text-retrieval-by-supervised-multi | 2302.09473 | null | https://arxiv.org/abs/2302.09473v1 | https://arxiv.org/pdf/2302.09473v1.pdf | Video-Text Retrieval by Supervised Multi-Space Multi-Grained Alignment | While recent progress in video-text retrieval has been advanced by the exploration of better representation learning, in this paper, we present a novel multi-space multi-grained supervised learning framework, SUMA, to learn an aligned representation space shared between the video and the text for video-text retrieval. The shared aligned space is initialized with a finite number of concept clusters, each of which refers to a number of basic concepts (words). With the text data at hand, we are able to update the shared aligned space in a supervised manner using the proposed similarity and alignment losses. Moreover, to enable multi-grained alignment, we incorporate frame representations for better modeling the video modality and calculating fine-grained and coarse-grained similarity. Benefiting from learned shared aligned space and multi-grained similarity, extensive experiments on several video-text retrieval benchmarks demonstrate the superiority of SUMA over existing methods. | ['Peng Shi', 'Yimu Wang'] | 2023-02-19 | null | null | null | null | ['video-text-retrieval'] | ['computer-vision'] | [ 3.38711619e-01 -6.12328053e-01 -5.46721637e-01 -2.98319101e-01
-1.02483034e+00 -4.22176123e-01 1.05236638e+00 3.68450701e-01
-2.56118208e-01 2.49758780e-01 5.97396791e-01 2.90018857e-01
-4.17578191e-01 -4.15820748e-01 -4.88237619e-01 -7.05210984e-01
3.06509584e-02 5.13808250e-01 2.86140130e-03 1.86711568e-02
2.83844948e-01 -1.01891132e-02 -1.97642016e+00 7.33542740e-01
6.00764513e-01 1.10115814e+00 3.56971711e-01 6.46362126e-01
-2.79299706e-01 7.70033658e-01 -3.70500267e-01 -1.31476596e-01
3.86799812e-01 -4.64181393e-01 -9.84870553e-01 2.01248437e-01
8.33025217e-01 -4.75276619e-01 -5.19932568e-01 8.22209537e-01
2.12559164e-01 6.61673605e-01 8.56770992e-01 -1.29590738e+00
-4.44865435e-01 4.93828952e-01 -4.29901123e-01 1.69806123e-01
6.11381888e-01 -3.52487236e-01 1.61392212e+00 -1.04829633e+00
7.33580232e-01 1.56253147e+00 2.34033838e-01 2.77846843e-01
-9.60459411e-01 -5.71198940e-01 4.87210482e-01 5.61085641e-01
-1.56785071e+00 -2.00180471e-01 6.01963937e-01 -3.61526906e-01
8.42979789e-01 1.70031771e-01 5.67341745e-01 1.14215267e+00
1.04206376e-01 1.18958926e+00 6.66402817e-01 -6.49396718e-01
2.15938196e-01 -1.46213144e-01 2.52391368e-01 6.15364254e-01
7.13670403e-02 -1.83155268e-01 -8.79374027e-01 -3.52872193e-01
6.78341210e-01 4.14690018e-01 -2.31711060e-01 -8.94535959e-01
-1.38140059e+00 9.51524973e-01 1.52505443e-01 6.95759475e-01
-2.24432483e-01 2.08581164e-01 7.54680216e-01 3.06696266e-01
5.53615332e-01 1.43113703e-01 -1.09958723e-01 -2.47890845e-01
-1.40693998e+00 3.38889271e-01 6.30629420e-01 1.08705080e+00
1.01484263e+00 -2.56949872e-01 -5.44445097e-01 8.78130198e-01
3.67104411e-01 5.56877971e-01 9.85392451e-01 -7.61844277e-01
5.00318050e-01 5.30616701e-01 -1.40106991e-01 -1.13924050e+00
4.21877243e-02 -3.10453847e-02 -7.30946362e-01 -3.26032639e-01
-6.82526315e-03 5.00953019e-01 -1.00912344e+00 1.57640588e+00
8.48284736e-02 6.62664056e-01 -2.48884074e-02 8.55119467e-01
5.16199291e-01 7.52528131e-01 6.68646917e-02 -1.51398674e-01
1.09720051e+00 -1.38029361e+00 -8.37673247e-01 1.46428749e-01
7.84710944e-01 -7.08624661e-01 9.41184342e-01 1.56692117e-01
-1.19657493e+00 -6.06804013e-01 -9.22995865e-01 -1.99461028e-01
-4.96680290e-01 4.33486328e-03 4.39251810e-01 4.38901395e-01
-1.14272499e+00 5.55484772e-01 -8.00190985e-01 -6.29578531e-01
3.98120373e-01 1.32272378e-01 -4.09972399e-01 -4.88617241e-01
-1.19312656e+00 4.95098233e-01 5.89925587e-01 -3.80000949e-01
-7.80028462e-01 -5.51648736e-01 -9.82354283e-01 3.76807600e-01
4.82748419e-01 -9.58077610e-01 9.73369658e-01 -1.19513655e+00
-1.35659981e+00 7.94893086e-01 -1.32633626e-01 -5.58063745e-01
1.95126995e-01 -5.59371948e-01 -2.07699239e-01 8.26133668e-01
1.85309783e-01 6.45290494e-01 1.58952785e+00 -1.08832300e+00
-6.93434596e-01 -2.74921417e-01 5.21707833e-02 6.06518924e-01
-8.35416555e-01 1.76042676e-01 -9.65944052e-01 -1.17617321e+00
-1.02886863e-01 -8.80915582e-01 1.34796739e-01 1.78127348e-01
4.43162657e-02 -3.31348121e-01 9.36186254e-01 -3.93494844e-01
1.45986736e+00 -2.23398018e+00 7.00750411e-01 2.60125548e-01
2.34428018e-01 2.53106207e-01 -4.79556501e-01 5.48808634e-01
-1.30887300e-01 -3.05592548e-02 1.00682214e-01 -7.94862151e-01
1.61382765e-01 3.05146873e-01 -3.87260437e-01 4.04717326e-01
-2.82209702e-02 8.40203822e-01 -1.10253644e+00 -7.19103754e-01
4.62766021e-01 5.60661316e-01 -6.95872366e-01 3.72148007e-01
-1.00226767e-01 1.26061246e-01 -7.70579517e-01 6.13563895e-01
4.93050247e-01 -5.38935781e-01 6.25653714e-02 -2.65869379e-01
3.28210533e-01 -3.79077941e-02 -1.22832882e+00 2.39334011e+00
-3.00184727e-01 5.74749589e-01 -1.20041825e-01 -1.14050794e+00
5.26958942e-01 3.13313216e-01 1.00876927e+00 -6.30832255e-01
1.20174177e-01 -4.04222868e-02 -7.10924387e-01 -1.68047071e-01
9.38428640e-01 1.37207866e-01 -2.53856797e-02 8.16979885e-01
4.99521047e-01 1.05872549e-01 2.31143013e-01 7.83601701e-01
8.51316094e-01 1.55627877e-01 3.33996475e-01 -1.51218429e-01
7.19621241e-01 -1.93385750e-01 -5.83646558e-02 9.33487833e-01
-6.15454130e-02 8.08844507e-01 -1.43497363e-01 -3.82881999e-01
-8.97726059e-01 -8.85249317e-01 -3.06893270e-02 1.70018208e+00
3.83087814e-01 -1.07077193e+00 -5.56794465e-01 -5.59106827e-01
6.45174040e-03 6.26590326e-02 -5.38888216e-01 -3.78635854e-01
-4.86576736e-01 -2.54942179e-01 2.91285127e-01 4.81966764e-01
1.88047290e-01 -7.56148458e-01 -3.39324385e-01 1.13825798e-01
-4.11714226e-01 -1.17795634e+00 -9.56498623e-01 -1.11449592e-01
-7.20817924e-01 -1.03758502e+00 -9.81985688e-01 -7.22978413e-01
3.97071511e-01 9.25435424e-01 1.02506268e+00 4.24600095e-01
-5.23407400e-01 1.26966846e+00 -9.89572465e-01 5.54020815e-02
-1.01056330e-01 -2.22778674e-02 1.07044779e-01 2.94591576e-01
2.24651709e-01 -1.86904430e-01 -5.39106786e-01 1.80240959e-01
-1.38773024e+00 -9.52416733e-02 3.20802361e-01 1.03018951e+00
6.15507245e-01 -8.14577937e-02 2.06800342e-01 -3.98141861e-01
4.51999784e-01 -5.93156695e-01 -2.32110247e-01 4.75935370e-01
-4.85521108e-01 1.10163026e-01 2.51570523e-01 -5.07178664e-01
-8.80155802e-01 -2.58431137e-01 4.82585400e-01 -9.70562339e-01
5.46320640e-02 6.20768785e-01 1.13835812e-01 -1.82646699e-02
1.76411629e-01 3.25008154e-01 -4.65668272e-03 -4.26503479e-01
7.88487613e-01 5.69001675e-01 3.73074621e-01 -8.70909393e-01
8.33113313e-01 6.16145730e-01 -1.64988413e-01 -6.70393229e-01
-8.85669529e-01 -1.13191485e+00 -8.18244755e-01 -1.26539826e-01
8.49953532e-01 -1.32980013e+00 -1.13808289e-01 3.86623293e-01
-7.24537551e-01 -1.18222572e-01 -4.29236621e-01 4.79413986e-01
-8.37283909e-01 8.99845123e-01 -4.92142051e-01 -3.92591983e-01
-3.64203006e-01 -1.07321262e+00 1.68094838e+00 -1.87645048e-01
-5.89763597e-02 -1.15723932e+00 2.41266847e-01 4.29375142e-01
2.61896223e-01 -2.22098619e-01 8.07708383e-01 -9.62092221e-01
-5.83245218e-01 -1.47578761e-01 -1.74203455e-01 2.96924889e-01
2.58067966e-01 -1.39402643e-01 -6.45196021e-01 -8.73978019e-01
-9.57810581e-02 -5.70503771e-01 1.08794248e+00 2.73090124e-01
1.26875067e+00 -1.52666271e-01 -4.05230641e-01 4.96732712e-01
1.28917873e+00 -1.95389971e-01 3.30293089e-01 4.72465336e-01
7.45975435e-01 3.71521503e-01 9.99483347e-01 7.61997879e-01
4.34213817e-01 9.56769347e-01 2.69541115e-01 1.95096806e-01
-2.66563952e-01 -6.19281419e-02 3.62592787e-01 8.79212320e-01
1.31314784e-01 -3.97076160e-01 -7.10803568e-01 5.59615731e-01
-2.19485140e+00 -1.17070305e+00 4.34825480e-01 2.18137670e+00
6.67097330e-01 -3.92840087e-01 1.67422041e-01 1.00263782e-01
8.35178196e-01 4.11489040e-01 -2.28223681e-01 6.84706196e-02
-2.46500984e-01 1.82375237e-01 -9.06969160e-02 3.34567428e-01
-1.28734243e+00 9.55946565e-01 6.27479410e+00 1.15306652e+00
-9.70822871e-01 -5.29901423e-02 2.06993967e-01 -3.91048431e-01
-2.69987613e-01 -9.54009965e-02 -6.90401793e-01 3.55833232e-01
8.24081898e-01 -4.16860580e-01 4.16883796e-01 6.31189764e-01
-3.15744072e-01 3.93737964e-02 -1.39833665e+00 1.43338466e+00
6.14079952e-01 -1.47870433e+00 7.54037261e-01 -1.83085635e-01
8.99762273e-01 -3.36385407e-02 9.24581438e-02 3.91992539e-01
-2.39760801e-02 -6.90652966e-01 6.67665660e-01 5.25510967e-01
6.46828830e-01 -5.31973660e-01 4.55708265e-01 -1.81236528e-02
-1.57867253e+00 -1.26965895e-01 -3.21588248e-01 3.35086852e-01
-1.02262765e-01 2.01272920e-01 -3.76461595e-01 9.34944928e-01
7.73629129e-01 1.28579509e+00 -5.65645695e-01 8.58500421e-01
4.42097545e-01 2.29648486e-01 -8.67507234e-02 2.61970788e-01
5.05813777e-01 -1.99707225e-01 5.20304501e-01 1.49907947e+00
2.71792442e-01 9.91626643e-03 6.18787169e-01 2.58398831e-01
-8.64488855e-02 3.69276464e-01 -3.59514743e-01 -1.76291317e-01
4.85130191e-01 1.20204365e+00 -4.15391624e-01 -6.43981278e-01
-6.04719639e-01 1.23846805e+00 1.84756279e-01 5.38800955e-01
-5.78707278e-01 -6.46156296e-02 6.66118205e-01 -2.52885491e-01
6.44585431e-01 -2.01822281e-01 4.22614872e-01 -1.58067620e+00
-3.74588482e-02 -1.01912868e+00 9.06621873e-01 -8.03818762e-01
-1.36183119e+00 4.58166033e-01 3.59974414e-01 -1.44503713e+00
-4.95416135e-01 -4.73914444e-01 -1.11458302e-01 5.35508513e-01
-1.70401597e+00 -1.38502359e+00 -3.50514323e-01 1.25842381e+00
9.04311001e-01 -4.73635226e-01 9.04640019e-01 3.20020050e-01
-2.37818554e-01 7.06603825e-01 5.64689279e-01 -7.41776302e-02
1.18618071e+00 -1.00966942e+00 -3.22626941e-02 3.51995707e-01
5.37858248e-01 5.54738998e-01 3.03861618e-01 -4.67883646e-01
-1.49342310e+00 -1.05209553e+00 5.10885417e-01 -3.44807357e-01
9.55111623e-01 -2.83682734e-01 -9.22050714e-01 8.10027480e-01
1.99794441e-01 3.01585823e-01 8.98320138e-01 -5.63600287e-02
-6.75462127e-01 -6.59717321e-02 -8.61105561e-01 3.84902030e-01
8.51254642e-01 -1.03765702e+00 -7.31646657e-01 4.96275365e-01
6.20791256e-01 -2.79673338e-01 -1.02301598e+00 1.86429113e-01
6.16733134e-01 -7.00377584e-01 1.18243420e+00 -6.81111097e-01
2.39444733e-01 -3.15620825e-02 -5.39639950e-01 -1.14158237e+00
-3.70938420e-01 -4.47088659e-01 -4.87317115e-01 9.83766139e-01
-3.90746295e-01 -1.41739830e-01 6.81156754e-01 2.56890804e-01
-3.36403847e-02 -5.29088199e-01 -8.90862346e-01 -8.41862142e-01
7.50973150e-02 -3.62026840e-01 4.42853928e-01 9.72540975e-01
8.25809687e-02 2.39767402e-01 -5.37868917e-01 -6.54961541e-02
6.30959153e-01 4.46672410e-01 5.84233105e-01 -1.29598081e+00
-2.31107250e-01 -5.02587974e-01 -6.99036479e-01 -1.35081100e+00
5.39793253e-01 -9.74983811e-01 -8.35640654e-02 -1.19229436e+00
6.91417873e-01 -1.76181585e-01 -8.25940609e-01 3.13440055e-01
-3.99937838e-01 1.94684863e-01 4.96465862e-01 6.04523957e-01
-1.45224595e+00 8.43023241e-01 8.29534471e-01 -3.02987844e-01
1.55907080e-01 -4.28474396e-01 -3.20284933e-01 4.21178162e-01
1.97469458e-01 -2.91525841e-01 -5.35041809e-01 -6.13401115e-01
-4.77079451e-02 -5.91605762e-03 2.14020893e-01 -9.10358608e-01
4.52357918e-01 1.71226896e-02 6.38020858e-02 -6.87579691e-01
5.57394981e-01 -1.07796061e+00 -2.11213425e-01 5.92654683e-02
-6.88073277e-01 -1.38401836e-01 1.83684155e-02 1.02967954e+00
-6.22434318e-01 -2.08777621e-01 4.44112688e-01 5.86497411e-02
-8.70509624e-01 5.44154108e-01 -2.47794569e-01 -4.48708534e-02
9.32043016e-01 -3.28464329e-01 2.86089862e-03 -4.99822378e-01
-7.49688089e-01 3.80397677e-01 4.27461803e-01 8.31439376e-01
6.68907225e-01 -1.65056062e+00 -5.96250534e-01 1.20914645e-01
6.30741477e-01 -2.78435171e-01 4.51155216e-01 3.84668976e-01
6.03582747e-02 7.36277699e-01 -2.81922519e-01 -8.49047303e-01
-1.46617758e+00 6.28280997e-01 -5.05647920e-02 -5.26670098e-01
-7.15560257e-01 5.82352757e-01 4.35912907e-01 7.00028194e-03
4.69587088e-01 -1.63263410e-01 -2.39806592e-01 2.09613368e-01
8.68134558e-01 4.46181506e-01 1.04809992e-01 -9.03588891e-01
-2.07577169e-01 8.58107746e-01 -6.05861723e-01 1.23736322e-01
1.32049322e+00 -6.26974523e-01 -1.13430470e-01 5.06978333e-01
1.46251345e+00 -4.15794909e-01 -1.18852687e+00 -7.88822830e-01
7.05462992e-02 -7.38691092e-01 2.60243505e-01 -2.89831489e-01
-8.96373928e-01 5.65755188e-01 6.69938564e-01 -7.89775103e-02
1.25682485e+00 1.20469406e-01 7.11590648e-01 7.91876912e-01
4.66564715e-01 -1.16741991e+00 8.14375401e-01 5.74746072e-01
6.44235015e-01 -1.26744962e+00 1.59506619e-01 -1.76893666e-01
-4.80057269e-01 1.16802764e+00 2.01329499e-01 -2.41632298e-01
6.00699067e-01 -2.28147089e-01 -2.34992191e-01 -2.22202718e-01
-8.36408377e-01 -1.93552554e-01 6.93277299e-01 4.94933248e-01
2.87345946e-01 -2.78119475e-01 -9.28952768e-02 3.16245109e-01
3.48270088e-01 -1.71320677e-01 1.66550949e-01 1.08087087e+00
-5.15935123e-01 -1.32010889e+00 -4.37673450e-01 3.86319131e-01
-2.53834039e-01 -2.31939003e-01 -1.90415636e-01 6.58465087e-01
-3.41854870e-01 7.88226068e-01 3.59245330e-01 -9.05209333e-02
-6.44664019e-02 9.75445881e-02 6.04506552e-01 -8.34764242e-01
-4.22958940e-01 3.25819880e-01 -3.64496738e-01 -8.64437461e-01
-9.26384807e-01 -7.47664273e-01 -1.11116278e+00 4.38680872e-02
-3.28335136e-01 1.91555485e-01 4.51365858e-01 1.13237917e+00
3.84450942e-01 2.28262320e-01 6.88753068e-01 -1.03131092e+00
-5.92272818e-01 -7.32222676e-01 -7.60898411e-01 8.24930668e-01
4.41758722e-01 -7.61998534e-01 -2.62831628e-01 3.15007716e-01] | [10.364459991455078, 0.9976592659950256] |
260ef057-23ca-498c-be5e-c2402121b26a | when-source-free-domain-adaptation-meets-1 | 2301.13381 | null | https://arxiv.org/abs/2301.13381v2 | https://arxiv.org/pdf/2301.13381v2.pdf | When Source-Free Domain Adaptation Meets Learning with Noisy Labels | Recent state-of-the-art source-free domain adaptation (SFDA) methods have focused on learning meaningful cluster structures in the feature space, which have succeeded in adapting the knowledge from source domain to unlabeled target domain without accessing the private source data. However, existing methods rely on the pseudo-labels generated by source models that can be noisy due to domain shift. In this paper, we study SFDA from the perspective of learning with label noise (LLN). Unlike the label noise in the conventional LLN scenario, we prove that the label noise in SFDA follows a different distribution assumption. We also prove that such a difference makes existing LLN methods that rely on their distribution assumptions unable to address the label noise in SFDA. Empirical evidence suggests that only marginal improvements are achieved when applying the existing LLN methods to solve the SFDA problem. On the other hand, although there exists a fundamental difference between the label noise in the two scenarios, we demonstrate theoretically that the early-time training phenomenon (ETP), which has been previously observed in conventional label noise settings, can also be observed in the SFDA problem. Extensive experiments demonstrate significant improvements to existing SFDA algorithms by leveraging ETP to address the label noise in SFDA. | ['Boyu Wang', 'A. Ian McLeod', 'Charles Ling', 'Ruizhi Pu', 'Jiaqi Li', 'Pengcheng Xu', 'Gezheng Xu', 'Li Yi'] | 2023-01-31 | null | null | null | null | ['learning-with-noisy-labels', 'source-free-domain-adaptation', 'learning-with-noisy-labels'] | ['computer-vision', 'computer-vision', 'natural-language-processing'] | [ 2.85897970e-01 -8.50185007e-02 -2.88739085e-01 -2.95760870e-01
-1.01184535e+00 -8.23219001e-01 6.42836690e-01 -1.87697560e-01
-1.79297775e-01 9.64101017e-01 -3.98650430e-02 -1.16724715e-01
-3.17743242e-01 -6.03948534e-01 -8.14432681e-01 -1.24097943e+00
2.83186436e-01 6.02797568e-01 1.92242116e-01 -8.69614910e-03
-1.26455083e-01 2.65494198e-01 -1.47879720e+00 -4.66798283e-02
7.55708098e-01 8.84012282e-01 1.08064272e-01 2.61824429e-01
-4.07040119e-01 5.55723667e-01 -9.46564913e-01 -1.98276639e-01
6.66123390e-01 -7.70656943e-01 -8.80219817e-01 2.59915024e-01
2.11535871e-01 -2.22477540e-02 -2.09932446e-01 1.40593958e+00
5.74472606e-01 2.44101912e-01 9.57577169e-01 -1.60304916e+00
-9.19319153e-01 4.62204158e-01 -5.62590420e-01 -3.92952077e-02
1.07691750e-01 -5.18366277e-01 6.52706087e-01 -7.64297903e-01
7.25452125e-01 1.16164196e+00 6.96352243e-01 9.49268222e-01
-1.42617130e+00 -8.21227968e-01 1.93061650e-01 1.45227328e-01
-1.34815001e+00 -2.43692979e-01 8.99364173e-01 -4.92028356e-01
2.90519297e-01 -1.31571397e-01 7.44532980e-03 1.68779624e+00
-3.18288177e-01 8.02313566e-01 1.46512783e+00 -8.80162060e-01
7.84060359e-01 3.09707344e-01 1.50207132e-01 8.57358649e-02
4.16967005e-01 2.33579203e-01 -4.89317477e-01 -5.27391553e-01
7.05091119e-01 -1.32393971e-01 -2.05509275e-01 -8.88301671e-01
-1.04127514e+00 9.43733752e-01 -4.32759710e-02 2.78749973e-01
-7.35783726e-02 -1.75115705e-01 3.35271031e-01 6.42387629e-01
6.82327688e-01 1.73899516e-01 -7.80890644e-01 1.17539443e-01
-6.30084038e-01 -6.62383586e-02 1.00654125e+00 1.29830718e+00
9.13936198e-01 -7.07889199e-02 -1.25131682e-02 8.55977476e-01
1.51382953e-01 7.42517114e-01 6.78682208e-01 -1.15218687e+00
3.20440024e-01 1.19663864e-01 4.00088698e-01 -4.89724696e-01
-2.49421000e-01 -6.08072221e-01 -6.50841177e-01 2.48022512e-01
8.62291336e-01 -2.14998603e-01 -9.70661700e-01 2.22532010e+00
6.21245503e-01 5.89988053e-01 4.59416002e-01 7.17916965e-01
2.34028414e-01 2.91933656e-01 5.00970781e-02 -4.71746176e-01
1.03375912e+00 -8.77226472e-01 -8.93240571e-01 -3.31986062e-02
9.41612542e-01 -6.48684800e-01 1.20580971e+00 4.28171754e-01
-4.90447700e-01 -3.51516128e-01 -1.00801408e+00 1.83139384e-01
-4.32592362e-01 -1.63254336e-01 3.48708987e-01 9.74246800e-01
-9.46529925e-01 2.18369290e-01 -4.97089803e-01 -7.47134089e-01
4.03408676e-01 2.12607592e-01 -3.28874767e-01 -4.35200065e-01
-1.30746901e+00 6.05608165e-01 2.26475820e-01 -3.71085703e-01
-8.83879006e-01 -7.00883985e-01 -5.49686909e-01 -1.39322355e-01
7.91146100e-01 -5.88783205e-01 1.53093743e+00 -1.45897913e+00
-1.57630551e+00 8.54384422e-01 -3.86751264e-01 -3.04574877e-01
4.29684997e-01 1.09423222e-02 -6.22508049e-01 7.27116019e-02
2.45593429e-01 1.98786974e-01 9.77665842e-01 -1.67836964e+00
-7.79824078e-01 -2.42607236e-01 -2.30435655e-01 9.62191671e-02
-3.65805864e-01 -1.26510948e-01 1.18738739e-02 -8.58349442e-01
-2.41820198e-02 -9.24512029e-01 8.20736587e-03 -1.09037369e-01
-1.20527722e-01 -5.65282524e-01 1.06479025e+00 -9.84438509e-02
1.04012990e+00 -2.44270635e+00 -1.81367233e-01 4.11608368e-01
5.30293621e-02 3.43951255e-01 -2.92066932e-01 3.81284148e-01
-1.70831576e-01 8.61736387e-02 -4.39415812e-01 -5.14304221e-01
2.38430887e-01 6.16710365e-01 -5.66110551e-01 4.88285303e-01
-9.65238288e-02 5.23357570e-01 -1.08243787e+00 -3.23860228e-01
-2.31629953e-01 2.01325521e-01 -2.98990220e-01 2.13496432e-01
-2.43115723e-01 5.74676633e-01 -3.68554175e-01 5.04385054e-01
9.84276235e-01 -3.66846710e-01 2.50650913e-01 2.02295452e-01
4.45423782e-01 3.22600678e-02 -1.34559047e+00 1.79757738e+00
-2.96614081e-01 3.06583494e-01 1.74866393e-01 -1.36856151e+00
7.77552068e-01 4.93854314e-01 5.20891547e-01 -4.25969213e-01
-3.51345167e-02 5.45914650e-01 -1.90542966e-01 -3.20497185e-01
7.96914771e-02 -5.08426964e-01 -1.82145372e-01 6.35474503e-01
3.01856399e-01 3.51175576e-01 -1.19664021e-01 7.73944929e-02
1.13743794e+00 9.18496698e-02 3.35315585e-01 -2.75300562e-01
3.63452643e-01 3.26555967e-02 7.83361375e-01 1.15939856e+00
-5.24712205e-01 7.03156710e-01 2.47743741e-01 7.30429515e-02
-9.85665202e-01 -1.16643989e+00 -3.85271817e-01 1.21359158e+00
2.52533585e-01 -3.03423293e-02 -9.37716365e-01 -1.20993495e+00
-1.86209768e-01 6.48041487e-01 -5.71942568e-01 -2.09505245e-01
-3.08465153e-01 -7.41932333e-01 6.49481356e-01 4.37452167e-01
5.01505792e-01 -5.15373766e-01 -2.23009847e-02 3.56670767e-01
-2.64279693e-01 -1.14924228e+00 -4.73895848e-01 4.85721022e-01
-6.03014469e-01 -9.29935336e-01 -9.33574259e-01 -8.12780321e-01
6.10542595e-01 4.00851667e-01 1.00039589e+00 -4.06978905e-01
1.78584635e-01 8.52969885e-01 -6.05028510e-01 -4.86932456e-01
-6.51379585e-01 2.03453809e-01 3.62174451e-01 1.40025765e-01
1.01833558e+00 -6.94755912e-01 -2.51519173e-01 4.96783793e-01
-1.19238818e+00 -5.24737120e-01 2.97076851e-01 1.04873967e+00
6.73420012e-01 4.23763424e-01 1.24203277e+00 -1.28977227e+00
4.73858804e-01 -1.04551995e+00 -4.80944812e-01 3.87347668e-01
-9.57054973e-01 1.07897416e-01 6.15588784e-01 -7.94418931e-01
-1.27862787e+00 1.46036983e-01 1.63794886e-02 -5.78869104e-01
-7.24539876e-01 3.08495611e-01 -4.39825296e-01 2.01710537e-02
8.54610920e-01 3.64162847e-02 -6.51472658e-02 -7.98968136e-01
3.07893485e-01 8.88607144e-01 4.67652649e-01 -8.84906828e-01
9.13292706e-01 7.91912556e-01 1.23569600e-01 -4.04960632e-01
-1.24830568e+00 -7.27787733e-01 -7.54877150e-01 3.20686400e-01
5.64175844e-01 -1.10836112e+00 1.29554775e-02 5.79416573e-01
-7.71466136e-01 -3.94176513e-01 -7.62715757e-01 4.86055404e-01
-7.91507483e-01 6.41226470e-01 -2.35353574e-01 -7.73730636e-01
3.29999268e-01 -8.57764542e-01 7.48311162e-01 1.23149104e-01
8.92704800e-02 -1.40450060e+00 1.01544872e-01 1.73539668e-02
3.02428007e-01 -7.30505362e-02 9.90349948e-01 -1.26486897e+00
-1.92938566e-01 6.93509579e-02 3.15492786e-03 5.65466702e-01
4.85128105e-01 -6.06074512e-01 -1.19186175e+00 -4.00282979e-01
4.50368106e-01 -3.27542514e-01 6.30992055e-01 3.79650623e-01
8.56115580e-01 -1.45715460e-01 -3.36591393e-01 3.46857369e-01
1.47820151e+00 2.33489007e-01 3.43184710e-01 4.27698821e-01
4.48418319e-01 4.80367571e-01 8.10441494e-01 4.49839771e-01
2.04651967e-01 7.90595114e-01 2.12246090e-01 -2.80220896e-01
-4.37890887e-01 -3.88570160e-01 5.25869548e-01 5.08342564e-01
5.71665227e-01 -4.13337648e-01 -7.66795576e-01 7.62059152e-01
-1.96047354e+00 -6.39594138e-01 -1.78161383e-01 2.41729736e+00
1.16084349e+00 -2.86238492e-01 2.58765340e-01 1.19070224e-01
8.68063271e-01 -3.47269833e-01 -7.27465212e-01 -3.97859141e-02
-5.12876511e-01 2.25786194e-01 6.84409320e-01 3.59996378e-01
-1.19809461e+00 6.64891481e-01 6.45351171e+00 1.18726623e+00
-8.28884661e-01 6.92814946e-01 2.58441150e-01 1.99112147e-02
-2.62590379e-01 -4.47692275e-02 -8.28493059e-01 4.69514996e-01
9.74422395e-01 -2.58864969e-01 4.27376390e-01 9.95112240e-01
-1.54497400e-01 -8.20125639e-02 -1.22921634e+00 9.02141094e-01
5.32306954e-02 -6.06363714e-01 -2.14920521e-01 2.31782630e-01
1.15588951e+00 -4.58900817e-02 1.74821749e-01 3.31535429e-01
8.28964949e-01 -5.69043636e-01 5.55086374e-01 1.84337944e-01
1.00840294e+00 -7.60521948e-01 8.43962193e-01 6.11795306e-01
-7.74215698e-01 -2.38146022e-01 -5.60204685e-01 8.04257765e-02
-1.29359022e-01 1.03009319e+00 -6.46862864e-01 7.04249442e-01
6.30970120e-01 5.79904854e-01 -3.14828813e-01 9.42570627e-01
-1.71796367e-01 1.09064043e+00 -5.00679374e-01 6.37699246e-01
5.68930060e-02 -1.19040065e-01 5.46208739e-01 1.00067186e+00
6.85034692e-01 -2.02769697e-01 2.19953641e-01 6.99077308e-01
-2.32315570e-01 -1.10730412e-03 -7.78973937e-01 2.54981518e-01
7.95015931e-01 6.43094063e-01 -4.95982796e-01 -2.66823679e-01
-6.73854768e-01 1.16710591e+00 1.18137151e-01 7.59528637e-01
-5.79986334e-01 -2.11438075e-01 6.94836855e-01 6.53743669e-02
4.66677576e-01 2.99389716e-02 -1.68784499e-01 -1.35859549e+00
4.91586886e-02 -8.30467880e-01 6.20509863e-01 -2.70986944e-01
-2.15803981e+00 1.51412547e-01 1.82072744e-01 -1.34351361e+00
-2.53185987e-01 -3.42817813e-01 -4.56295125e-02 8.95919263e-01
-1.71591246e+00 -9.96200442e-01 2.28768036e-01 1.06560802e+00
5.09726346e-01 -2.26489991e-01 9.69557643e-01 4.19011444e-01
-2.93070883e-01 8.96683991e-01 9.68084395e-01 -1.97803099e-02
1.27872121e+00 -1.34652436e+00 -3.78084518e-02 8.46513152e-01
1.66151896e-01 4.14676249e-01 5.04483461e-01 -5.87671280e-01
-8.04427862e-01 -1.17860067e+00 8.81411850e-01 -5.86583316e-01
5.37655056e-01 -6.13516808e-01 -1.00859225e+00 7.73581147e-01
-1.34667363e-02 3.63298208e-01 9.63640690e-01 6.51191622e-02
-7.56229818e-01 6.39264286e-02 -1.58758056e+00 7.95608237e-02
1.19438970e+00 -5.43929577e-01 -5.67281187e-01 3.44383508e-01
7.00835943e-01 -5.43180853e-02 -8.08496058e-01 2.93188393e-01
-4.19420861e-02 -6.35598481e-01 8.55034292e-01 -6.49651289e-01
-3.17281894e-02 -5.29685915e-01 -3.97402555e-01 -1.46259427e+00
-3.42677891e-01 -4.25030500e-01 -2.42344230e-01 1.61611664e+00
1.57607421e-01 -8.53980243e-01 6.31590664e-01 5.26758969e-01
1.47762388e-01 7.96937123e-02 -1.49002409e+00 -1.49529922e+00
6.54449344e-01 -3.85369509e-01 7.19642162e-01 1.36193156e+00
-2.20460460e-01 1.44899145e-01 -4.37592536e-01 4.02438313e-01
8.43993664e-01 -1.63291216e-01 5.49367726e-01 -1.53740978e+00
-4.10794228e-01 4.60358635e-02 1.22650862e-02 -1.21196556e+00
5.85330129e-01 -8.98103952e-01 1.98936552e-01 -1.07136619e+00
1.72465041e-01 -9.43409383e-01 -7.41848052e-01 4.66626495e-01
-1.41982764e-01 7.47780576e-02 1.05103828e-01 5.26956081e-01
-6.10768378e-01 3.76846313e-01 8.05501163e-01 1.35414690e-01
9.88820009e-03 1.99490517e-01 -8.41436505e-01 6.67933881e-01
7.40689993e-01 -1.01938069e+00 -7.23834753e-01 -2.31760487e-01
4.28005010e-02 -4.88666981e-01 3.08361739e-01 -8.94689143e-01
1.39112666e-01 -2.91879356e-01 -1.08695969e-01 -1.96994707e-01
2.38559041e-02 -1.41273582e+00 7.10800216e-02 3.58559340e-02
-3.18266630e-01 -5.18642485e-01 -1.31328896e-01 1.17543232e+00
-2.52165318e-01 -5.08465767e-01 7.85932660e-01 -8.92078280e-02
-7.08323598e-01 -2.08443999e-02 -6.00270450e-01 2.72191435e-01
1.08072567e+00 -1.82835143e-02 -4.13183510e-01 -4.43466038e-01
-7.62529433e-01 -1.13741443e-01 5.80127120e-01 1.62996903e-01
5.63536212e-02 -1.50054681e+00 -4.62539554e-01 2.25639015e-01
1.55442342e-01 1.85045063e-01 1.96304262e-01 7.11220562e-01
3.80165756e-01 8.53224322e-02 2.77029335e-01 -5.18532574e-01
-8.64687324e-01 9.23581362e-01 2.86402076e-01 -5.00588000e-01
-3.85026664e-01 7.86356151e-01 5.26744068e-01 -6.36181891e-01
3.86106074e-01 1.51810557e-01 2.09356412e-01 6.82310248e-03
2.97673404e-01 3.23468953e-01 6.66531846e-02 -4.31644887e-01
-1.50676593e-01 5.66013396e-01 -2.19280384e-02 -9.89440531e-02
1.06724536e+00 -7.31270134e-01 2.58982003e-01 5.93820572e-01
1.14406860e+00 1.76876724e-01 -1.42917657e+00 -9.79151785e-01
2.98020303e-01 -4.18036491e-01 -2.26270810e-01 -1.11541951e+00
-8.27937305e-01 8.12807798e-01 9.23724055e-01 3.08679253e-01
1.38858223e+00 2.81488806e-01 7.46155441e-01 3.15245688e-01
6.40559316e-01 -1.33105636e+00 3.16050835e-02 4.41716462e-01
3.64951849e-01 -1.32677937e+00 -4.39886987e-01 -5.59517562e-01
-5.79623103e-01 8.52808654e-01 3.63914251e-01 1.14514574e-01
9.58129764e-01 2.97999769e-01 3.04306805e-01 1.93976372e-01
-5.09708762e-01 -3.86954457e-01 -2.57107824e-01 1.40850592e+00
-2.38750532e-01 -2.11972632e-02 -1.40489101e-01 1.05475247e+00
2.33118668e-01 2.12875247e-01 7.04555094e-01 9.84202027e-01
-2.53233612e-01 -1.77785182e+00 -6.53417945e-01 1.14845924e-01
-5.52056015e-01 7.77858123e-02 -3.06608975e-01 7.29353547e-01
4.49860215e-01 1.13597858e+00 -1.47694260e-01 -1.40668765e-01
2.58858413e-01 5.52945614e-01 1.67926520e-01 -7.95382023e-01
-2.06826761e-01 1.97545245e-01 -1.33953661e-01 -1.27337188e-01
-8.74983609e-01 -7.95369983e-01 -1.11043739e+00 -1.08338542e-01
-4.44245368e-01 1.98635817e-01 5.16611755e-01 1.00115478e+00
5.30911148e-01 6.98483661e-02 7.38978446e-01 -2.54521310e-01
-9.78120208e-01 -8.70457292e-01 -1.11522686e+00 7.55807936e-01
4.22277570e-01 -9.80898380e-01 -8.24263930e-01 2.71765828e-01] | [10.39415454864502, 3.1836228370666504] |
e8c29ee5-69c5-42e3-ac53-77eb27e12746 | learning-mid-level-features-and-modeling | 1401.5535 | null | http://arxiv.org/abs/1401.5535v2 | http://arxiv.org/pdf/1401.5535v2.pdf | Learning Mid-Level Features and Modeling Neuron Selectivity for Image Classification | We now know that mid-level features can greatly enhance the performance of
image learning, but how to automatically learn the image features efficiently
and in an unsupervised manner is still an open question. In this paper, we
present a very efficient mid-level feature learning approach (MidFea), which
only involves simple operations such as $k$-means clustering, convolution,
pooling, vector quantization and random projection. We explain why this simple
method generates the desired features, and argue that there is no need to spend
much time in learning low-level feature extractors. Furthermore, to boost the
performance, we propose to model the neuron selectivity (NS) principle by
building an additional layer over the mid-level features before feeding the
features into the classifier. We show that the NS-layer learns
category-specific neurons with both bottom-up inference and top-down analysis,
and thus supports fast inference for a query image. We run extensive
experiments on several public databases to demonstrate that our approach can
achieve state-of-the-art performances for face recognition, gender
classification, age estimation and object categorization. In particular, we
demonstrate that our approach is more than an order of magnitude faster than
some recently proposed sparse coding based methods. | ['Shu Kong', 'Qiang Yang', 'Zhuolin Jiang'] | 2014-01-22 | null | null | null | null | ['object-categorization'] | ['computer-vision'] | [ 2.54475564e-01 2.30074325e-03 -1.44866139e-01 -8.55232835e-01
-5.10431826e-01 -1.81990549e-01 5.81473053e-01 3.21272463e-01
-6.56775951e-01 4.94854689e-01 -7.01170266e-02 -3.21685746e-02
-2.60788262e-01 -1.02808952e+00 -8.10369074e-01 -7.39599705e-01
-2.68916488e-01 2.57651627e-01 1.01991773e-01 1.60270661e-01
5.43755531e-01 7.77438641e-01 -2.31971741e+00 6.64197862e-01
6.54878020e-01 1.50990915e+00 -8.68923739e-02 5.49340487e-01
-1.82365164e-01 7.44618118e-01 -3.60173851e-01 -5.45937896e-01
-2.86202803e-02 -1.84104994e-01 -9.54443455e-01 4.97194119e-02
6.25668228e-01 -1.05948247e-01 -1.36440083e-01 9.52600539e-01
4.66274738e-01 2.39275265e-02 9.65721965e-01 -1.14036572e+00
-5.57210743e-01 5.35629570e-01 -5.42688906e-01 1.64251968e-01
2.41032556e-01 -2.02448279e-01 9.31062877e-01 -1.16959369e+00
4.58145142e-01 1.38338804e+00 7.40101516e-01 6.19262576e-01
-1.14805317e+00 -8.42689633e-01 2.13910282e-01 4.86778080e-01
-1.68796909e+00 -6.10817134e-01 5.75879157e-01 -3.61238837e-01
9.08483982e-01 1.71221063e-01 5.13006151e-01 8.03221583e-01
-6.04092516e-02 7.21585274e-01 1.21922398e+00 -5.43652773e-01
2.79300779e-01 2.64242589e-01 3.37413758e-01 1.27512586e+00
1.41394049e-01 -6.34889230e-02 -7.88865149e-01 -2.30788961e-01
5.29818594e-01 2.08819985e-01 1.27681330e-01 -3.36958528e-01
-9.18762505e-01 1.06579709e+00 5.28377414e-01 4.20283675e-01
-3.27395141e-01 1.77009150e-01 2.17839435e-01 1.38138533e-01
2.07653925e-01 1.40047848e-01 -3.22679907e-01 1.05932832e-01
-1.15328288e+00 1.09588787e-01 7.79203653e-01 5.40493309e-01
1.02756190e+00 -1.83255345e-01 -2.82548964e-01 1.15405834e+00
9.87086743e-02 4.75431830e-01 5.56713939e-01 -1.08497405e+00
2.24817153e-02 5.36566675e-01 -4.09391403e-01 -1.16389716e+00
-5.07676065e-01 -2.99311578e-01 -1.06358039e+00 1.62859187e-01
4.50619519e-01 6.16215244e-02 -9.05083954e-01 1.85487640e+00
8.00248012e-02 1.12348996e-01 -2.09661826e-01 4.70732898e-01
8.28057528e-01 4.34730619e-01 6.42586723e-02 -1.43063754e-01
1.52559280e+00 -7.40641057e-01 -3.96348983e-01 -1.03474729e-01
3.97980243e-01 -3.68253112e-01 7.19258189e-01 4.71343756e-01
-9.93330121e-01 -7.29324043e-01 -8.46070349e-01 -2.48697251e-02
-6.88140869e-01 3.82463545e-01 1.15504718e+00 7.73782551e-01
-1.02780890e+00 7.82518148e-01 -5.79970241e-01 -4.58915055e-01
1.00056779e+00 8.75914037e-01 -7.64781117e-01 -2.67348349e-01
-9.03909862e-01 5.53851962e-01 2.83200473e-01 -1.21967115e-01
-6.58750713e-01 -4.79361534e-01 -1.00475693e+00 3.71125251e-01
1.28326163e-01 -5.61385691e-01 7.87166357e-01 -7.83079803e-01
-1.34664154e+00 9.89272118e-01 -5.12302518e-01 -3.29098225e-01
-2.16514781e-01 1.43052816e-01 -8.34488273e-02 3.69069576e-01
-1.50720075e-01 1.11729503e+00 1.13353479e+00 -9.38176692e-01
-7.25390375e-01 -7.14104414e-01 -1.00102112e-01 -6.28541410e-02
-8.90779316e-01 1.03585415e-01 -4.18938398e-01 -3.76631647e-01
2.34089315e-01 -7.31143236e-01 -1.36521861e-01 2.71561533e-01
-1.32220566e-01 -6.39325857e-01 3.37902546e-01 -4.31182355e-01
1.09395778e+00 -2.31833363e+00 -2.83043515e-02 6.16270721e-01
2.53905535e-01 7.52929822e-02 2.73045022e-02 1.86489969e-02
-4.96201068e-02 -7.62856677e-02 -3.82435918e-01 -3.13999116e-01
-6.09502085e-02 1.17122948e-01 -4.86088358e-02 3.96279514e-01
2.39032641e-01 7.42586553e-01 -5.57345867e-01 -7.08232403e-01
-5.20778447e-02 6.00675762e-01 -1.06974292e+00 5.25683500e-02
3.25299382e-01 -6.40432760e-02 -1.68884352e-01 8.23973298e-01
6.63113236e-01 -2.60276794e-01 9.54710022e-02 -2.30796129e-01
-7.40593150e-02 6.79859817e-02 -1.31288421e+00 1.66466081e+00
-4.73917961e-01 3.33617598e-01 -1.16053708e-02 -1.60232830e+00
8.54337633e-01 1.55775556e-02 2.69902498e-01 -6.33387148e-01
2.05462128e-01 1.97591081e-01 -1.00890007e-02 -2.56944060e-01
1.12977713e-01 -3.17339152e-02 -7.79330060e-02 3.24739635e-01
7.53655732e-01 2.83328325e-01 2.24876553e-01 5.47946021e-02
8.68515968e-01 -3.29095244e-01 1.96867436e-01 -3.91795427e-01
8.07130635e-01 -3.79599601e-01 4.84643519e-01 8.55240643e-01
-4.30848189e-02 5.23545265e-01 5.81500769e-01 -3.96314979e-01
-5.79049170e-01 -9.06395912e-01 -3.29371512e-01 1.50718653e+00
-3.16136092e-01 -4.60055441e-01 -9.54472661e-01 -7.38924205e-01
2.49900416e-01 2.27331176e-01 -9.47285652e-01 -2.52727449e-01
-3.56978148e-01 -1.01703024e+00 4.84252006e-01 5.51146209e-01
5.74824274e-01 -1.18745720e+00 -3.36662441e-01 -1.14002600e-01
1.52668893e-01 -9.40089047e-01 -1.57988459e-01 5.15408158e-01
-9.74814892e-01 -8.52667749e-01 -6.24113619e-01 -1.03165138e+00
1.04927850e+00 -1.57949678e-03 8.18205893e-01 1.35097921e-01
-6.44463539e-01 1.37371331e-01 5.62589755e-03 -2.55035907e-01
1.90710023e-01 2.80382723e-01 1.88505113e-01 3.23843569e-01
6.28657758e-01 -7.09509552e-01 -5.72031736e-01 1.15909316e-02
-5.46414256e-01 -2.94065326e-01 8.65721166e-01 1.03068197e+00
5.16380847e-01 1.52075544e-01 6.18985176e-01 -1.13639283e+00
3.30861151e-01 -2.13348791e-01 -4.32427406e-01 4.20396142e-02
-5.53683043e-01 2.72961766e-01 6.70077443e-01 -3.01785737e-01
-8.09607089e-01 5.28810084e-01 -5.31565905e-01 -1.87547967e-01
-3.12707841e-01 2.83774674e-01 -1.35040373e-01 -3.47617269e-01
6.75376952e-01 5.16718268e-01 1.41165107e-01 -5.00947833e-01
4.24608022e-01 9.70375836e-01 5.11962593e-01 -6.18770063e-01
7.60793567e-01 6.41204357e-01 2.39967313e-02 -8.32392931e-01
-1.01352513e+00 -3.51814538e-01 -8.04345548e-01 1.39231801e-01
7.30842948e-01 -9.60566938e-01 -1.21684825e+00 4.95363802e-01
-7.81668782e-01 1.44291064e-03 -1.83136895e-01 3.12729418e-01
-6.12129450e-01 4.62164842e-02 -5.71225703e-01 -6.28419638e-01
-3.12809199e-01 -9.71473515e-01 8.64699900e-01 3.80937576e-01
3.57802846e-02 -5.47412217e-01 -4.34064925e-01 2.14945599e-01
4.90008354e-01 -3.05443257e-01 1.10719430e+00 -7.30124652e-01
-2.68593431e-01 -8.67375806e-02 -4.80549008e-01 3.95387620e-01
-2.12359056e-02 -3.14428389e-01 -1.31151974e+00 -2.95497000e-01
-1.76518306e-01 -6.48187637e-01 1.34782946e+00 3.35604161e-01
2.08416128e+00 -2.70663917e-01 -4.03726190e-01 9.19830203e-01
1.27042496e+00 -1.28916815e-01 6.00935698e-01 -1.36914372e-01
6.19656384e-01 7.89022624e-01 3.24845999e-01 4.67673928e-01
4.56129551e-01 4.92694676e-01 9.98982117e-02 -3.14552300e-02
-9.19599310e-02 -1.73757840e-02 1.06921919e-01 6.25999391e-01
-2.31256247e-01 4.63372111e-01 -5.68501353e-01 5.60323715e-01
-1.57027137e+00 -1.10701692e+00 4.71157312e-01 2.11219406e+00
9.88699973e-01 7.24431053e-02 1.90736040e-01 5.26275456e-01
5.02569556e-01 -4.01853211e-02 -3.26766193e-01 -6.97430968e-01
9.32576582e-02 7.22941279e-01 3.69963378e-01 3.53257477e-01
-1.27993417e+00 9.82456326e-01 6.78886080e+00 1.08014047e+00
-1.19888318e+00 5.41740889e-03 9.39241171e-01 5.70202656e-02
5.47695532e-02 -4.91976589e-01 -1.10775614e+00 2.48601899e-01
9.61270511e-01 1.94259524e-01 5.40580213e-01 1.03748536e+00
-4.53196913e-01 -5.69583885e-02 -1.18830383e+00 1.36439049e+00
3.95703465e-01 -1.41919887e+00 1.43821359e-01 5.29798828e-02
5.21519482e-01 -3.45976710e-01 2.33367980e-01 5.01866698e-01
-1.75161302e-01 -1.29133832e+00 3.84106249e-01 4.56976026e-01
8.44635725e-01 -1.10091484e+00 5.67888260e-01 2.26035357e-01
-1.04777443e+00 -4.84334767e-01 -7.64801025e-01 -5.70225343e-02
-6.61604583e-01 7.07552731e-01 -4.41427380e-01 1.13668904e-01
7.50586629e-01 5.54084420e-01 -9.70194638e-01 8.94191027e-01
7.46583752e-03 4.58880305e-01 -3.89990985e-01 -1.01274952e-01
1.57410726e-02 2.79180139e-01 -2.79028803e-01 1.36672688e+00
4.21544284e-01 1.58708528e-01 -7.84557238e-02 5.83562970e-01
-3.94798011e-01 2.25371182e-01 -5.30229628e-01 -1.08240684e-02
3.89621228e-01 1.44332802e+00 -8.53615582e-01 -3.95775437e-01
-5.96829593e-01 1.07771790e+00 6.55968785e-01 -9.10056289e-04
-3.63557369e-01 -8.40283096e-01 6.81541920e-01 9.76868868e-02
7.85075545e-01 2.69330889e-02 -2.56400824e-01 -1.09684682e+00
-1.55499801e-01 -7.99843729e-01 5.49773276e-01 -1.98170334e-01
-1.28047895e+00 4.62389499e-01 -2.31572971e-01 -6.77905142e-01
-3.93773556e-01 -9.60534155e-01 -2.47213483e-01 7.30501115e-01
-1.53858411e+00 -9.93478894e-01 -1.88350216e-01 9.68477190e-01
1.64911121e-01 -3.70826870e-01 1.14704847e+00 5.33268869e-01
-4.05912459e-01 9.81345475e-01 -6.69315010e-02 3.79700869e-01
5.20972371e-01 -1.17379308e+00 -1.08427286e-01 4.13435280e-01
3.92418444e-01 7.81153858e-01 2.35259026e-01 -4.69109081e-02
-1.38032568e+00 -8.84360135e-01 1.18810201e+00 -1.11857466e-02
2.63121009e-01 -8.07354629e-01 -7.42600501e-01 2.64605731e-01
-1.38022035e-01 5.26477218e-01 1.00860620e+00 5.38569391e-01
-6.43038869e-01 -5.30044317e-01 -1.34943295e+00 2.04098254e-01
1.20988417e+00 -6.76009476e-01 -6.00486875e-01 2.60706663e-01
2.16359898e-01 1.03247426e-01 -7.97902822e-01 3.17222893e-01
8.09450030e-01 -1.01247871e+00 1.29877150e+00 -5.69408476e-01
4.52406645e-01 -7.63048083e-02 -3.37062329e-01 -1.00387955e+00
-5.72317898e-01 -3.36534418e-02 -2.33641908e-01 1.22985101e+00
3.49207759e-01 -4.76239532e-01 9.28727746e-01 1.79838330e-01
3.11352193e-01 -1.03360128e+00 -8.62013698e-01 -6.01686180e-01
1.14135139e-01 -3.56987923e-01 4.98549938e-01 8.08895350e-01
1.50116354e-01 2.63056278e-01 -1.37426779e-01 6.94347769e-02
7.55153537e-01 4.78307128e-01 2.85919160e-01 -1.54489219e+00
-2.10369602e-01 -4.16224688e-01 -7.43978620e-01 -7.46167362e-01
3.66128325e-01 -1.11415184e+00 -2.75853090e-02 -1.27742898e+00
6.27916455e-01 -3.72296721e-01 -7.71980405e-01 8.15558553e-01
-1.84097201e-01 8.03357959e-01 3.08365356e-02 -1.08393095e-01
-7.27340400e-01 3.04868639e-01 9.27171826e-01 -1.51625574e-01
1.86736375e-01 -1.03241324e-01 -1.12680459e+00 8.21746051e-01
6.11848891e-01 -5.15264153e-01 -6.01484999e-02 -1.06528208e-01
-1.04528561e-01 -3.39579374e-01 2.14357942e-01 -1.31443608e+00
3.87417436e-01 4.62471172e-02 1.10579491e+00 -3.40080857e-01
4.94951010e-01 -5.19681871e-01 -6.16111279e-01 5.99958777e-01
-3.66988450e-01 -4.38328415e-01 5.30881882e-02 1.49233997e-01
-3.58460397e-01 -3.65734547e-01 1.01003706e+00 -1.54734626e-01
-6.37804806e-01 5.46439171e-01 -4.20001894e-01 -1.84832156e-01
6.36108875e-01 -4.11795527e-02 -4.10657302e-02 -7.96635896e-02
-9.41566885e-01 -1.21183984e-01 9.87560302e-02 1.99944347e-01
6.49506927e-01 -1.33803916e+00 -5.57879090e-01 6.30989850e-01
1.78353518e-01 -4.89729047e-01 2.07896158e-01 5.65639198e-01
-9.46646482e-02 5.50601900e-01 -4.66468483e-01 -5.67342877e-01
-1.34445918e+00 7.03762829e-01 4.12580222e-02 -9.37902182e-02
-3.01534861e-01 1.23954892e+00 2.01693729e-01 -5.87217867e-01
4.59778875e-01 -5.47134839e-02 -5.32586694e-01 4.61262435e-01
9.03978705e-01 6.96024895e-02 1.50556341e-01 -6.13788366e-01
-6.49902642e-01 8.33514988e-01 -1.68007448e-01 6.95080087e-02
1.54322183e+00 1.38370812e-01 -3.05287480e-01 2.85920769e-01
1.47682500e+00 -1.38926134e-01 -8.97793770e-01 -2.30219766e-01
3.68587747e-02 -4.54606771e-01 3.29180062e-02 -6.09130204e-01
-1.24819350e+00 1.25342631e+00 8.63940060e-01 -6.77268766e-03
1.35286999e+00 2.37720475e-01 3.82686585e-01 8.53059709e-01
3.99336100e-01 -1.17246485e+00 -3.25928442e-02 4.61582810e-01
5.97318828e-01 -1.21850979e+00 2.32211091e-02 -6.11019194e-01
-1.77933425e-01 1.05536020e+00 5.57284772e-01 -1.86041638e-01
9.78693366e-01 2.44143233e-01 -3.04584950e-01 -1.13441095e-01
-7.67859161e-01 -4.47969019e-01 3.16693544e-01 5.82722306e-01
4.80129540e-01 -3.89881581e-02 -2.91440159e-01 8.69976461e-01
-2.73882180e-01 5.90269268e-02 2.60397717e-02 8.72282922e-01
-7.21355379e-01 -1.22825027e+00 -3.01419377e-01 9.00071204e-01
-7.09523439e-01 -1.41514212e-01 -1.30703807e-01 3.54374528e-01
4.84112084e-01 6.20333612e-01 2.02143624e-01 -3.84486377e-01
6.71651438e-02 4.69557345e-01 8.19841743e-01 -7.06979871e-01
-4.66033131e-01 -3.85736823e-01 -2.00423613e-01 -7.44990826e-01
-2.70847231e-01 -7.46087968e-01 -1.26496983e+00 -2.65216947e-01
-3.41603048e-02 1.47594720e-01 8.04024756e-01 8.64026427e-01
3.69432896e-01 3.27921599e-01 7.42649078e-01 -7.37212718e-01
-3.86286676e-01 -8.90637994e-01 -7.38640487e-01 3.55013937e-01
2.44720262e-02 -8.46022487e-01 -2.97497332e-01 9.97909829e-02] | [9.472250938415527, 2.7609059810638428] |
b2c9630e-8ff2-48d9-bf94-4b666768e426 | few-shot-class-incremental-learning-via | 2006.15524 | null | https://arxiv.org/abs/2006.15524v3 | https://arxiv.org/pdf/2006.15524v3.pdf | MgSvF: Multi-Grained Slow vs. Fast Framework for Few-Shot Class-Incremental Learning | As a challenging problem, few-shot class-incremental learning (FSCIL) continually learns a sequence of tasks, confronting the dilemma between slow forgetting of old knowledge and fast adaptation to new knowledge. In this paper, we concentrate on this "slow vs. fast" (SvF) dilemma to determine which knowledge components to be updated in a slow fashion or a fast fashion, and thereby balance old-knowledge preservation and new-knowledge adaptation. We propose a multi-grained SvF learning strategy to cope with the SvF dilemma from two different grains: intra-space (within the same feature space) and inter-space (between two different feature spaces). The proposed strategy designs a novel frequency-aware regularization to boost the intra-space SvF capability, and meanwhile develops a new feature space composition operation to enhance the inter-space SvF learning performance. With the multi-grained SvF learning strategy, our method outperforms the state-of-the-art approaches by a large margin. | ['Fei Wu', 'Qi Tian', 'Mintong Kang', 'Xi Li', 'Yongjian Fu', 'Hanbin Zhao'] | 2020-06-28 | null | null | null | null | ['few-shot-class-incremental-learning'] | ['methodology'] | [ 1.85424253e-01 -4.51305270e-01 -2.28259951e-01 -2.24192426e-01
-4.50636864e-01 -2.82203048e-01 6.55938447e-01 1.87853470e-01
-6.36675358e-01 7.78597534e-01 1.45499825e-01 7.08616897e-02
-5.69632947e-01 -8.81018341e-01 -5.77237666e-01 -8.38434100e-01
2.38060996e-01 2.17423573e-01 8.88436496e-01 -2.10158780e-01
2.89110392e-01 2.74920046e-01 -1.80178630e+00 1.38439819e-01
1.08481503e+00 9.65584636e-01 6.03472233e-01 3.51648986e-01
-6.16341472e-01 7.94471383e-01 -3.42039824e-01 -1.72412977e-01
3.17262173e-01 -5.22957146e-01 -8.04139435e-01 -6.44562468e-02
2.18766540e-01 1.60285100e-01 -2.68256694e-01 1.06824696e+00
5.31920791e-01 6.69555366e-01 4.42092001e-01 -1.10073102e+00
-7.61924684e-01 4.96890396e-01 -5.09232700e-01 6.63531184e-01
-2.82790624e-02 2.41648033e-01 6.52437806e-01 -1.11583483e+00
6.41519368e-01 1.02440619e+00 6.59887493e-01 5.84433079e-01
-9.66776192e-01 -6.52859747e-01 7.44285643e-01 7.78886080e-01
-1.44363821e+00 -4.47799802e-01 9.55982685e-01 -4.84269053e-01
6.57653391e-01 3.49093489e-02 7.92107761e-01 9.81552422e-01
2.35822842e-01 7.88894951e-01 1.06398487e+00 -4.99152839e-01
6.41870201e-01 1.50445074e-01 3.81992310e-01 7.34703779e-01
2.40636393e-01 1.30588576e-01 -8.87406707e-01 3.36697139e-02
6.78670347e-01 2.21440479e-01 -2.77051568e-01 -6.53008759e-01
-1.02414453e+00 7.44556308e-01 2.51423508e-01 5.79572380e-01
-2.48758823e-01 -1.90425336e-01 5.68372965e-01 5.84348083e-01
3.82330120e-01 4.18008059e-01 -6.51721001e-01 -2.27963269e-01
-8.53833497e-01 9.30080004e-03 5.36730886e-01 7.24732161e-01
1.02868605e+00 1.98842570e-01 -3.53927433e-01 9.90186989e-01
4.46202606e-02 1.48309991e-01 1.09577322e+00 -3.80755812e-01
3.09907734e-01 5.64571321e-01 -3.95047553e-02 -8.39954257e-01
-3.96222442e-01 -8.13520432e-01 -6.59536421e-01 1.49273768e-01
3.42126429e-01 -1.28714845e-01 -9.77940381e-01 1.78543723e+00
6.61312580e-01 5.68948150e-01 6.67639226e-02 6.83905244e-01
6.28217459e-01 7.38777220e-01 1.35895818e-01 -7.33563602e-01
1.06707013e+00 -1.18591475e+00 -7.40067124e-01 -3.93906742e-01
2.57763594e-01 -4.77301121e-01 1.29139912e+00 1.56800851e-01
-7.63923049e-01 -9.56807613e-01 -1.18098044e+00 1.37333497e-01
-5.13549924e-01 -2.01959088e-01 5.01478314e-01 3.38963479e-01
-7.15587378e-01 6.59638703e-01 -6.15580738e-01 -2.72417426e-01
4.36462462e-01 8.90177190e-02 -1.63538307e-01 -1.52367741e-01
-1.43340456e+00 8.72543037e-01 8.55807602e-01 -6.97306842e-02
-8.05362046e-01 -9.20391738e-01 -6.17092550e-01 1.18844911e-01
8.81910682e-01 -6.71396911e-01 1.00978136e+00 -1.17096663e+00
-1.64613521e+00 3.90203595e-01 -3.84634919e-02 -2.39811122e-01
5.61087072e-01 -3.02518994e-01 -6.59537017e-01 -1.99895859e-01
8.90392736e-02 8.53508860e-02 1.30083048e+00 -1.06660724e+00
-9.43916380e-01 -4.03781921e-01 -1.82100236e-01 5.96586823e-01
-6.06806755e-01 -5.39604783e-01 -4.25015777e-01 -8.49496067e-01
1.03767894e-01 -5.42677283e-01 1.03020519e-02 -8.26098248e-02
1.38376608e-01 -2.60984212e-01 1.00792885e+00 -1.97166994e-01
1.54409981e+00 -2.08235312e+00 3.76534134e-01 -7.84237012e-02
1.78723902e-01 6.82011962e-01 -1.94141448e-01 1.19570680e-01
9.56365690e-02 -4.91347522e-01 -2.06370458e-01 -2.18604598e-02
-3.09342980e-01 1.83552638e-01 -1.98171690e-01 2.01562673e-01
-1.04042925e-01 9.63104665e-01 -1.27554941e+00 -4.75656867e-01
-4.68013957e-02 2.64574230e-01 -2.08004147e-01 2.13408813e-01
-5.27768731e-02 3.11428487e-01 -5.30350924e-01 7.13580072e-01
5.47640562e-01 -2.07854375e-01 4.23165858e-02 -1.27518475e-01
-3.40844393e-01 -4.51733142e-01 -1.40359640e+00 1.82775986e+00
-4.23274666e-01 2.53001034e-01 -2.03692198e-01 -1.01909196e+00
1.07449651e+00 -8.28360617e-02 3.12140048e-01 -7.76159704e-01
1.00553624e-01 2.24624813e-01 5.56021146e-02 -5.17727137e-01
4.52560604e-01 -5.06380975e-01 1.56852365e-01 3.30363065e-01
5.50573289e-01 2.48110816e-01 1.18362375e-01 -1.17504802e-02
1.03036726e+00 6.48280010e-02 5.70639133e-01 -4.42250520e-01
7.26487100e-01 -3.85816842e-01 9.50642586e-01 9.28404272e-01
-5.92937529e-01 2.60579795e-01 -9.45883337e-03 -6.82959974e-01
-7.14213550e-01 -1.12225449e+00 5.02264053e-02 1.59511364e+00
4.50085700e-01 -2.08929136e-01 -4.49968100e-01 -1.01728606e+00
3.02624315e-01 6.68753266e-01 -8.69208932e-01 -8.88064504e-01
-3.43076587e-01 -7.61558175e-01 2.54557598e-02 6.77890241e-01
6.50231063e-01 -1.04885840e+00 -6.41299665e-01 4.53061372e-01
3.54103386e-01 -3.76972973e-01 -7.43535459e-01 4.62212861e-01
-8.71115208e-01 -1.03712010e+00 -9.26117659e-01 -8.01760316e-01
4.51048076e-01 3.76568437e-01 6.64909005e-01 -2.70798922e-01
-3.22088063e-01 4.03825194e-01 -5.87866783e-01 -2.32924327e-01
5.61032258e-02 2.94714540e-01 3.40592295e-01 4.64364618e-01
3.57581139e-01 -5.88002980e-01 -5.22429466e-01 1.90400034e-01
-9.23890471e-01 -4.61071804e-02 5.66968739e-01 1.20255911e+00
6.57743633e-01 3.77361566e-01 1.03252745e+00 -1.01123226e+00
5.47053635e-01 -6.46503329e-01 -3.59027267e-01 5.44429719e-01
-1.09324956e+00 2.44122490e-01 7.97369361e-01 -9.35242593e-01
-1.41865194e+00 -5.58184236e-02 1.03551283e-01 -5.57987332e-01
3.09660822e-01 5.96012294e-01 -2.35393584e-01 -2.89247036e-01
6.86095774e-01 7.13218153e-01 -1.68016344e-01 -6.46101415e-01
5.33198953e-01 2.93237180e-01 5.40811241e-01 -6.01931274e-01
8.71846735e-01 2.56960303e-01 -3.25450778e-01 -5.36441922e-01
-1.11748827e+00 -4.31711018e-01 -9.05177712e-01 -2.71192163e-01
5.14397025e-01 -7.22124696e-01 -1.91989273e-01 8.84264469e-01
-5.37004888e-01 -4.27290678e-01 -9.71612871e-01 3.94881457e-01
-3.23087037e-01 2.20463723e-01 -3.47335368e-01 -7.64365911e-01
-3.07941943e-01 -8.55655551e-01 5.10830164e-01 6.55431271e-01
2.21159816e-01 -9.86700714e-01 3.92440587e-01 -3.92386429e-02
7.23897636e-01 -2.29850654e-02 8.02636981e-01 -6.84827328e-01
-1.93335652e-01 8.97069946e-02 -1.98583156e-01 2.28138417e-01
3.26193541e-01 -4.35916424e-01 -8.12675416e-01 -6.57125175e-01
2.33425692e-01 -3.01611423e-01 1.27057290e+00 1.13365479e-01
1.00043344e+00 -2.76995033e-01 -1.06752485e-01 7.38026321e-01
1.47535443e+00 5.34256399e-01 3.98276269e-01 6.69830024e-01
7.20279038e-01 2.90662646e-01 6.76161528e-01 5.62524796e-01
4.58334476e-01 4.89542842e-01 2.91223964e-03 3.38168085e-01
-4.06347960e-01 -3.97078902e-01 2.38745868e-01 1.10335028e+00
1.36986952e-02 1.48944408e-01 -7.74822295e-01 6.35186195e-01
-2.08195090e+00 -9.68352139e-01 6.13977373e-01 2.25079012e+00
1.03994524e+00 3.53722066e-01 -1.08285107e-01 -9.54496302e-03
7.77668238e-01 4.32358742e-01 -1.05601752e+00 5.20550720e-02
-1.98601097e-01 5.09554408e-02 2.07655489e-01 4.82232928e-01
-1.25786066e+00 9.58282053e-01 5.52716017e+00 1.28591895e+00
-1.34309995e+00 4.10984159e-01 3.48825395e-01 -1.05521843e-01
-1.92361861e-01 -2.80259605e-02 -8.97138357e-01 7.45181620e-01
6.61485672e-01 -5.61035454e-01 4.81043488e-01 9.67084110e-01
-4.17330772e-01 -6.49985820e-02 -6.89407408e-01 1.12962973e+00
3.22628826e-01 -1.30857897e+00 3.90622355e-02 -5.18340647e-01
9.33624208e-01 -1.48226246e-01 -1.68648399e-02 8.13860416e-01
-7.33922347e-02 -4.69635904e-01 7.74950802e-01 8.99835765e-01
7.76133537e-01 -7.82206833e-01 6.08254135e-01 3.78867835e-01
-1.68580639e+00 -5.27976453e-01 -3.38253915e-01 7.50549138e-02
1.54762454e-02 5.76258540e-01 -1.47115648e-01 6.12781286e-01
7.28228152e-01 8.32807541e-01 -8.45974803e-01 1.09966540e+00
-7.39174411e-02 3.65389258e-01 2.02583671e-01 -2.24990048e-03
1.23687811e-01 -1.12964502e-02 6.43822849e-01 1.02811551e+00
2.78014183e-01 2.06894323e-01 4.42628354e-01 5.17132878e-01
1.25028357e-01 2.28731260e-01 -1.80951089e-01 1.07316062e-01
6.71040297e-01 1.00173879e+00 -7.52781868e-01 -5.84511697e-01
-3.45761448e-01 1.09261572e+00 6.72217071e-01 3.26729596e-01
-5.13840318e-01 -6.56629086e-01 3.46809328e-01 7.73011148e-02
5.86132765e-01 -4.55252714e-02 -5.28554507e-02 -1.49660277e+00
-2.89805867e-02 -4.78161007e-01 6.98290110e-01 -3.63763243e-01
-1.56409383e+00 6.33490026e-01 8.94768746e-04 -1.24579358e+00
2.02801898e-01 -2.39952445e-01 -7.31143534e-01 5.14135718e-01
-1.80656624e+00 -1.00182378e+00 -3.93640906e-01 8.38617802e-01
8.94243896e-01 -6.23063743e-01 5.43585300e-01 3.54031235e-01
-5.43180048e-01 7.66877532e-01 5.88780403e-01 -4.09813523e-01
7.76477277e-01 -1.06012785e+00 1.11842319e-01 7.32673585e-01
-1.21713825e-01 5.50904512e-01 3.59318972e-01 -7.76552796e-01
-1.41778314e+00 -1.22573853e+00 6.07424676e-01 -6.25116080e-02
6.10856831e-01 -2.02429056e-01 -1.39328897e+00 2.66155273e-01
-3.94666284e-01 2.96304435e-01 5.46227038e-01 2.75259852e-01
-6.75952971e-01 -4.69032943e-01 -1.07245910e+00 3.35199386e-01
9.91562784e-01 -5.53449214e-01 -9.50228214e-01 3.72640148e-04
9.02298868e-01 -1.57159984e-01 -6.89292967e-01 4.80487645e-01
6.57678783e-01 -8.17437947e-01 8.96247506e-01 -6.40038669e-01
-5.31215742e-02 -5.79641104e-01 -1.19347677e-01 -1.50369680e+00
-8.70589674e-01 -5.41193664e-01 -7.12394416e-01 1.08116710e+00
-3.63685600e-02 -7.99731731e-01 7.06178248e-01 9.16111842e-03
-2.07356617e-01 -9.51488733e-01 -1.12792110e+00 -1.19873857e+00
1.08753115e-01 -1.28775045e-01 5.28930426e-01 1.19689417e+00
-6.43812045e-02 4.04772282e-01 -4.76020783e-01 -3.21131647e-01
5.18297136e-01 1.80680186e-01 3.36697578e-01 -1.48077571e+00
-3.20616126e-01 -5.49460173e-01 -2.97499031e-01 -6.54007554e-01
-4.76498455e-02 -7.97927201e-01 9.06757638e-02 -1.14363801e+00
4.66531664e-01 -3.32075775e-01 -1.01205361e+00 6.00633800e-01
-7.59281576e-01 -9.24057886e-02 4.24990028e-01 3.40501726e-01
-9.89998400e-01 9.65008318e-01 1.18428147e+00 -1.00399144e-01
-5.30758321e-01 -1.68408573e-01 -7.54978895e-01 6.15110934e-01
5.65487564e-01 -3.15827221e-01 -7.14336216e-01 -2.54981846e-01
-1.47662893e-01 -3.06805044e-01 -5.58585338e-02 -1.26525450e+00
6.02012753e-01 -2.92017788e-01 5.51842034e-01 -3.07805806e-01
1.05423473e-01 -6.45601451e-01 -3.12566571e-02 5.79029977e-01
-3.24190378e-01 -2.15482414e-01 2.08577961e-01 1.00733912e+00
-1.48478255e-01 -2.46408254e-01 1.14963150e+00 -3.36442202e-01
-1.40057600e+00 5.41214943e-01 -3.36009748e-02 3.45552385e-01
1.23471117e+00 -1.95676088e-01 -3.53040010e-01 7.88379908e-02
-8.39615166e-01 2.31964752e-01 3.19334447e-01 7.58603752e-01
7.21680701e-01 -1.51078129e+00 -4.50570196e-01 4.64543164e-01
3.24872911e-01 -2.42165834e-01 8.63469839e-01 6.37250423e-01
2.81404555e-02 -1.14815824e-01 -4.51048404e-01 -2.79325813e-01
-9.20644164e-01 8.66158664e-01 7.95157254e-02 -4.08178300e-01
-7.60781169e-01 1.31595361e+00 2.20331222e-01 -4.03153867e-01
3.42108965e-01 2.81032383e-01 -3.12110394e-01 5.26172101e-01
8.79364848e-01 5.81944942e-01 4.72682938e-02 -4.38964039e-01
-3.28814954e-01 7.80092776e-01 -4.84672248e-01 1.92062676e-01
1.26320684e+00 -3.99706692e-01 2.00853407e-01 9.74628210e-01
1.02533185e+00 -3.21222037e-01 -1.56037784e+00 -8.73368919e-01
2.32152194e-02 -6.78998649e-01 1.51982829e-01 -1.02106035e+00
-9.33502913e-01 5.42072773e-01 9.20637190e-01 3.72594707e-02
1.32099903e+00 -1.61621690e-01 8.42889309e-01 3.31026107e-01
5.18891692e-01 -1.62053823e+00 4.87648010e-01 8.24533284e-01
6.89081788e-01 -1.29346848e+00 2.13335425e-01 3.18120755e-02
-8.90958130e-01 1.11011887e+00 8.09443593e-01 -4.03988035e-03
8.58167768e-01 -8.76400992e-02 -2.01677606e-01 -1.32896870e-01
-8.69735122e-01 -3.59591037e-01 5.43555975e-01 4.97046947e-01
-3.63425654e-03 -1.12127230e-01 -2.92820930e-01 8.53336513e-01
2.26264983e-01 1.19921431e-01 1.02083385e-01 1.22752297e+00
-8.80331099e-01 -1.05897582e+00 -1.34377405e-01 6.44611299e-01
9.31016952e-02 1.37890894e-02 -2.21459102e-02 5.62833309e-01
5.46221435e-01 4.14057821e-01 -1.29248515e-01 -5.88306487e-01
4.06112969e-01 2.77056575e-01 4.20715988e-01 -7.21076548e-01
-4.19438899e-01 6.77031353e-02 -5.00314236e-01 -4.39938217e-01
-2.31804788e-01 -6.78451836e-01 -1.02644360e+00 1.63604558e-01
-4.25870806e-01 2.14173406e-01 2.26287246e-01 1.07667351e+00
3.80851895e-01 7.35334992e-01 7.67828465e-01 -5.71704984e-01
-7.43487895e-01 -7.25717425e-01 -8.82169902e-01 2.75239557e-01
3.91869605e-01 -1.01462734e+00 -3.37824106e-01 -9.92065072e-02] | [9.855484962463379, 3.3703994750976562] |
69dcd5d9-6a77-4fae-91f6-ec1913aff5f0 | handling-normalization-issues-for-part-of | null | null | https://aclanthology.org/L18-1014 | https://aclanthology.org/L18-1014.pdf | Handling Normalization Issues for Part-of-Speech Tagging of Online Conversational Text | null | ["Fr{\\'e}d{\\'e}ric B{\\'e}chet", "G{\\'e}raldine Damnati", 'Jeremy Auguste', 'Delphine Charlet', 'Alexis Nasr', 'Johannes Heinecke'] | 2018-05-01 | handling-normalization-issues-for-part-of-1 | https://aclanthology.org/L18-1014 | https://aclanthology.org/L18-1014.pdf | lrec-2018-5 | ['lexical-normalization'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.299404144287109, 3.713064193725586] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.