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
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
16198b6c-e326-49b5-9d61-aaedec5fc486 | isometric-transformation-invariant-and | 2005.06316 | null | https://arxiv.org/abs/2005.06316v4 | https://arxiv.org/pdf/2005.06316v4.pdf | Isometric Transformation Invariant and Equivariant Graph Convolutional Networks | Graphs are one of the most important data structures for representing pairwise relations between objects. Specifically, a graph embedded in a Euclidean space is essential to solving real problems, such as physical simulations. A crucial requirement for applying graphs in Euclidean spaces to physical simulations is learning and inferring the isometric transformation invariant and equivariant features in a computationally efficient manner. In this paper, we propose a set of transformation invariant and equivariant models based on graph convolutional networks, called IsoGCNs. We demonstrate that the proposed model has a competitive performance compared to state-of-the-art methods on tasks related to geometrical and physical simulation data. Moreover, the proposed model can scale up to graphs with 1M vertices and conduct an inference faster than a conventional finite element analysis, which the existing equivariant models cannot achieve. | ['Toshiaki Hishinuma', 'Yu Ihara', 'Naoto Mitsume', 'Masanobu Horie', 'Naoki Morita'] | 2020-05-13 | null | https://openreview.net/forum?id=FX0vR39SJ5q | https://openreview.net/pdf?id=FX0vR39SJ5q | iclr-2021-1 | ['physical-simulations'] | ['miscellaneous'] | [-7.58561352e-03 -7.21811503e-02 2.18190312e-01 -3.83705437e-01
-1.11944959e-01 -3.43191147e-01 5.87301373e-01 1.53292820e-01
-9.69378501e-02 4.67319787e-01 -1.84613705e-01 -5.42201698e-01
-5.77273726e-01 -1.14145923e+00 -1.11972034e+00 -6.98117375e-01
-2.09395856e-01 6.05530620e-01 2.74559498e-01 -3.89483780e-01
2.11066633e-01 9.67426121e-01 -1.20000470e+00 -3.12141240e-01
6.12351656e-01 6.84933662e-01 -1.26314491e-01 6.83621228e-01
1.25467569e-01 4.18640673e-01 7.25100413e-02 -1.75690457e-01
1.95545241e-01 -1.51248500e-01 -7.45768845e-01 -5.55304289e-02
6.55887902e-01 -1.48701340e-01 -1.10178947e+00 1.16994607e+00
2.09420756e-01 4.42425668e-01 8.14588726e-01 -1.41690230e+00
-8.34087372e-01 5.96660256e-01 -2.06236809e-01 -1.24094337e-02
1.31999269e-01 1.53654590e-01 8.16769958e-01 -7.84266829e-01
5.41349828e-01 1.43701077e+00 5.86248279e-01 1.75879434e-01
-1.09283876e+00 -6.28518283e-01 -4.47984599e-02 3.71064633e-01
-1.53054798e+00 -1.77602634e-01 9.25899148e-01 -3.36258292e-01
8.50244701e-01 2.63825774e-01 6.60641074e-01 6.70132935e-01
6.16488159e-01 2.02634096e-01 7.85994887e-01 -3.46565023e-02
2.01925874e-01 -6.22874141e-01 1.09072283e-01 9.98827934e-01
3.83227289e-01 -1.35560960e-01 -1.71436578e-01 -1.97656415e-02
1.26128435e+00 2.78870575e-02 -1.09603368e-01 -9.12019610e-01
-1.51186311e+00 6.41470313e-01 8.77713859e-01 1.89834684e-01
-2.62112673e-02 7.15656877e-01 2.19130561e-01 1.79672584e-01
1.44697785e-01 5.74308872e-01 -9.68020186e-02 1.59334630e-01
-3.86821777e-01 3.84189367e-01 5.11210382e-01 1.02489161e+00
6.76346898e-01 1.21747218e-01 2.30724663e-01 2.98572034e-01
4.43981171e-01 7.67362058e-01 -2.56523073e-01 -8.92516732e-01
3.44129264e-01 7.99276650e-01 -1.69940546e-01 -1.40322554e+00
-8.48476052e-01 -2.45755285e-01 -1.27563334e+00 1.87109426e-01
1.96661547e-01 3.23763639e-01 -7.44370759e-01 1.63607788e+00
5.00028729e-01 4.50827986e-01 -9.38435048e-02 8.45935464e-01
9.75987613e-01 5.41154027e-01 -3.03077370e-01 2.18060970e-01
1.07764888e+00 -5.73519945e-01 -5.14203370e-01 1.27532363e-01
5.49705923e-01 -5.23978472e-01 8.51398826e-01 1.16416914e-02
-9.20740902e-01 -5.88057697e-01 -1.42682672e+00 -1.87591031e-01
-4.03181940e-01 -1.38746694e-01 1.13190699e+00 3.39133412e-01
-8.52294922e-01 1.05467725e+00 -1.09119463e+00 -3.20186675e-01
2.57794619e-01 6.35397494e-01 -7.18252540e-01 -7.88676888e-02
-1.13067758e+00 6.98167503e-01 2.71892101e-01 2.55147457e-01
-7.54967809e-01 -7.40370393e-01 -1.19758308e+00 5.26724644e-02
3.73706341e-01 -6.63762271e-01 7.43800640e-01 -7.71986097e-02
-1.40496087e+00 4.78082031e-01 2.62107044e-01 -2.09638447e-01
3.59204680e-01 1.16794311e-01 -5.88167250e-01 5.94281778e-02
-2.75122255e-01 2.06630990e-01 6.34964705e-01 -9.83661652e-01
1.85087726e-01 -4.77528095e-01 6.14778996e-01 4.50908206e-02
-1.15418704e-02 -2.85790384e-01 -5.78467667e-01 -4.71249968e-01
4.42762554e-01 -1.25058532e+00 -4.31247890e-01 1.96086928e-01
-3.94904941e-01 -3.18877637e-01 8.64730179e-01 -4.12233442e-01
8.76234829e-01 -1.96303713e+00 5.34458756e-01 5.05925536e-01
8.07322800e-01 2.20905572e-01 -4.45534773e-02 3.55564982e-01
-1.22713849e-01 -1.51113778e-01 -3.50261927e-01 3.31277758e-01
3.07246268e-01 1.41887262e-01 -7.11905956e-02 9.64372993e-01
-1.52432416e-02 1.10314298e+00 -9.33425725e-01 -3.37344706e-01
7.25001216e-01 5.36029339e-01 -5.28234482e-01 9.77119058e-02
-2.45237738e-01 7.66278267e-01 -6.10696495e-01 9.75840539e-02
9.59881663e-01 -5.02612948e-01 3.02663743e-01 -4.84268993e-01
1.29869238e-01 1.78602964e-01 -1.37878704e+00 1.85901952e+00
-4.01429355e-01 4.54122156e-01 -6.96052164e-02 -1.35079134e+00
8.88497233e-01 5.45412190e-02 6.85648084e-01 -6.16454244e-01
3.25864822e-01 -8.43380988e-02 2.41267681e-01 -2.42564157e-01
4.82306391e-01 2.16899842e-01 -4.73265827e-01 4.16322410e-01
1.48552611e-01 -3.59170973e-01 8.48300457e-02 4.95554596e-01
1.32217991e+00 -2.07782891e-02 2.16294616e-01 -5.57561696e-01
8.54284644e-01 -4.67896551e-01 4.49461699e-01 4.94710028e-01
2.04838723e-01 2.99037874e-01 4.31717932e-01 -6.71974361e-01
-1.15140104e+00 -1.47998333e+00 -8.28436390e-02 4.01816756e-01
5.23244441e-01 -5.76949298e-01 -6.95286930e-01 -5.00108719e-01
2.50644863e-01 2.83572853e-01 -5.38621902e-01 -4.98835653e-01
-8.20337772e-01 -1.83712840e-01 4.66940403e-01 6.34269714e-01
7.14990973e-01 -7.17472434e-01 -7.71953687e-02 1.01964951e-01
1.58201426e-01 -1.35035741e+00 -3.94814193e-01 -2.91192770e-01
-7.27972031e-01 -1.26097119e+00 -2.77072079e-02 -6.06548429e-01
9.14728165e-01 3.80068004e-01 1.00413930e+00 3.15340728e-01
-4.56942618e-01 3.70842427e-01 3.33966687e-02 -3.90942469e-02
-3.85111600e-01 9.47290882e-02 3.58684212e-01 -8.79037678e-02
-2.04513967e-03 -9.40612197e-01 -5.75952172e-01 3.75506222e-01
-1.00044072e+00 3.13768715e-01 4.59612668e-01 4.68184471e-01
7.06117690e-01 2.36217067e-01 1.96502194e-01 -9.66898799e-01
1.97565287e-01 -2.79200703e-01 -8.79564285e-01 2.52471775e-01
-2.04484552e-01 4.85245526e-01 1.05865574e+00 -1.07626036e-01
-6.23259366e-01 8.47587436e-02 1.48020908e-02 -5.25384724e-01
8.99872743e-03 5.99958837e-01 -5.54969132e-01 -7.62379169e-01
3.44262719e-01 -8.02109167e-02 -1.44441172e-01 -2.33175367e-01
5.85109830e-01 1.47308499e-01 5.84621131e-01 -7.46557713e-01
1.21921051e+00 5.27188957e-01 1.00818384e+00 -9.61374640e-01
-4.11561012e-01 -2.29884341e-01 -1.08355904e+00 -4.12777960e-01
7.92799294e-01 -4.83497709e-01 -1.21710587e+00 5.19980252e-01
-9.34418678e-01 -1.99703634e-01 1.41988114e-01 7.84171224e-01
-8.00755143e-01 6.22418582e-01 -4.58964199e-01 -2.32459351e-01
-7.63789285e-03 -1.18900073e+00 9.27749634e-01 9.79179665e-02
1.43934414e-01 -1.23749888e+00 1.10130027e-01 1.23445168e-01
4.18902218e-01 4.76054370e-01 1.13587642e+00 -2.74177343e-01
-1.01382732e+00 -1.37619823e-01 -3.42549384e-01 -1.82788193e-01
1.42861038e-01 6.57467078e-03 -3.45046103e-01 -4.76282597e-01
-2.80805945e-01 -1.65913686e-01 5.60249507e-01 1.97782427e-01
1.60480428e+00 3.52442451e-02 -3.42117250e-01 9.76351917e-01
1.24678969e+00 -1.30182534e-01 5.93043804e-01 -3.19068134e-01
1.32969701e+00 1.25818878e-01 3.58226925e-01 1.38783097e-01
4.30424869e-01 6.95337117e-01 5.17256379e-01 -8.51843506e-02
-5.07479310e-02 -3.01301062e-01 -1.08443545e-02 1.48401499e+00
-2.08044931e-01 1.24190319e-02 -9.34091449e-01 2.32727379e-01
-2.09625602e+00 -6.91119194e-01 -3.26389283e-01 2.17572284e+00
1.17797568e-01 1.33700281e-01 -4.08419758e-01 -1.57068539e-02
7.99792290e-01 2.79583752e-01 -7.59828985e-01 -4.03911650e-01
1.37471959e-01 4.89264339e-01 6.60727024e-01 4.43192631e-01
-1.25277257e+00 8.59023452e-01 6.23871708e+00 6.65616989e-01
-9.82232928e-01 -2.47547239e-01 4.17485759e-02 3.81188482e-01
-3.97828192e-01 1.50740713e-01 -2.51932085e-01 1.15526259e-01
9.08334732e-01 -4.12148774e-01 7.99338937e-01 7.73077726e-01
-9.85404402e-02 3.65555942e-01 -1.48563886e+00 1.04972947e+00
1.40686870e-01 -1.56282747e+00 2.47704655e-01 1.57030985e-01
6.62205398e-01 -1.94944322e-01 2.73217801e-02 1.44810289e-01
5.21762252e-01 -1.35305274e+00 3.92581880e-01 6.56435013e-01
7.62948751e-01 -9.03004825e-01 4.86213982e-01 6.45774230e-02
-1.86800504e+00 5.29666543e-01 -7.70315707e-01 -8.98046717e-02
1.93417743e-01 4.98046339e-01 -4.68190581e-01 1.13911021e+00
4.18625027e-01 9.53190804e-01 -5.56051672e-01 7.84373105e-01
-1.79967552e-01 3.71101052e-01 -2.68788755e-01 3.07179093e-02
1.79332644e-01 -5.76223671e-01 4.60620970e-01 6.60082221e-01
3.32813799e-01 3.44298780e-01 2.46517465e-01 9.40769553e-01
-4.65859592e-01 -4.59634177e-02 -1.03537560e+00 -1.88534856e-01
3.09254736e-01 1.34476376e+00 -1.08451450e+00 -4.94993217e-02
-5.29783845e-01 7.48128116e-01 5.05756557e-01 1.95068419e-01
-1.26883030e+00 -5.74991643e-01 6.89548850e-01 -6.18079081e-02
2.77638972e-01 -8.36229801e-01 2.18339749e-02 -1.16977811e+00
5.24842665e-02 -6.08048320e-01 1.19318448e-01 -7.17705250e-01
-1.35796404e+00 2.17026189e-01 1.08697295e-01 -1.15080452e+00
1.10948004e-01 -1.04041934e+00 -6.40384793e-01 6.61682248e-01
-9.64007378e-01 -1.35747111e+00 -6.55788779e-01 9.54769254e-01
-1.18360877e-01 5.38450703e-02 6.77986324e-01 3.07494462e-01
-2.84337461e-01 5.13622165e-01 1.32114679e-01 4.32011366e-01
4.00623322e-01 -1.21860659e+00 1.08768237e+00 8.44539642e-01
2.96922773e-01 7.60816157e-01 4.32829231e-01 -6.24195039e-01
-2.32265139e+00 -1.32732618e+00 3.21200669e-01 -3.17533106e-01
8.13073516e-01 -6.23811126e-01 -1.04096103e+00 8.31428885e-01
-2.43336588e-01 4.84426498e-01 3.62146139e-01 3.15378746e-03
-4.85441387e-01 -1.17938079e-01 -9.00492311e-01 8.60450745e-01
1.66863763e+00 -6.57034755e-01 -3.46104801e-01 4.78729248e-01
8.44524741e-01 -7.27810979e-01 -1.27113390e+00 6.18150651e-01
5.43346882e-01 -5.55861771e-01 1.32540905e+00 -9.05299187e-01
2.89812148e-01 -4.61836994e-01 -1.63461015e-01 -1.32076597e+00
-6.47626102e-01 -5.00218153e-01 -6.47613853e-02 8.50859523e-01
3.82687077e-02 -6.68125927e-01 6.51495576e-01 4.62486267e-01
-1.79302678e-01 -4.70963776e-01 -8.15210640e-01 -1.01937890e+00
1.69368342e-01 -5.14151216e-01 1.02884293e+00 1.16330051e+00
-1.16552904e-01 2.66744882e-01 -2.16288731e-01 5.03741384e-01
9.14683521e-01 3.41054112e-01 1.09281802e+00 -1.42011809e+00
-4.23299596e-02 -3.01279545e-01 -1.22923815e+00 -9.07115579e-01
6.92304015e-01 -1.34783494e+00 -1.15107596e-01 -1.61130548e+00
1.65178031e-01 -5.02592266e-01 -4.10822093e-01 4.53342572e-02
-3.60874971e-03 1.23845056e-01 -2.37955414e-02 -2.35998482e-01
-7.98141599e-01 7.79627681e-01 1.40680313e+00 -3.01540673e-01
3.00694644e-01 -2.01262876e-01 -3.33543599e-01 8.14475179e-01
8.20697904e-01 -2.30904907e-01 -5.54427922e-01 -4.04840857e-01
4.20421690e-01 2.10704446e-01 5.45322239e-01 -1.13818681e+00
4.33010101e-01 -4.10694689e-01 2.42979094e-01 -5.77578187e-01
2.28977397e-01 -7.37457871e-01 5.37987769e-01 3.66753042e-01
-1.95426464e-01 3.25977802e-01 1.00943998e-01 6.92747474e-01
9.44376141e-02 2.25625783e-01 8.01631153e-01 1.16681997e-02
-6.37752891e-01 9.15137768e-01 3.37452143e-02 8.98669362e-02
1.06108439e+00 2.30483860e-01 -3.88860464e-01 -3.80706847e-01
-4.86025333e-01 1.05046928e-01 7.27543473e-01 5.03902137e-01
7.65512228e-01 -1.76532340e+00 -5.51056683e-01 9.88202468e-02
2.36820549e-01 8.92324671e-02 5.81656814e-01 6.18317127e-01
-9.68198180e-01 5.73731780e-01 -5.80660582e-01 -7.34026790e-01
-1.11898291e+00 7.34096825e-01 1.96612209e-01 -3.24458808e-01
-9.65343833e-01 3.93800914e-01 4.45447356e-01 -1.03674626e+00
-1.53215304e-01 -4.65905666e-01 7.99119025e-02 -8.75756443e-01
1.16806731e-01 5.17463386e-01 2.88554251e-01 -8.84300828e-01
-5.71898639e-01 8.99984539e-01 2.25298122e-01 1.74130797e-01
1.27268696e+00 2.15413421e-01 -4.62989420e-01 3.63645047e-01
1.20039117e+00 -8.82205516e-02 -1.05771339e+00 -3.46267313e-01
-4.24744934e-01 -4.65959787e-01 -6.75130263e-02 1.33426219e-01
-1.14485145e+00 9.90942001e-01 1.87683955e-01 -6.54276228e-03
5.54945827e-01 7.43979663e-02 7.88545966e-01 9.43898380e-01
6.79073989e-01 -8.22870076e-01 -2.39729304e-02 6.29604459e-01
9.01538968e-01 -1.00941360e+00 3.69980127e-01 -7.32353568e-01
-9.76006538e-02 1.17846704e+00 6.88083887e-01 -6.36727631e-01
1.08007121e+00 1.12616584e-01 -5.69157243e-01 -5.01687765e-01
-4.09008741e-01 6.44564703e-02 6.04285359e-01 5.28771758e-01
1.75222442e-01 3.85253936e-01 -1.13666795e-01 2.95570046e-01
-4.35298443e-01 -4.49479252e-01 4.37969506e-01 6.86520219e-01
-1.59631982e-01 -9.84748244e-01 -9.64335129e-02 5.15912652e-01
-3.64427119e-02 2.18473107e-01 -3.87380987e-01 9.11243737e-01
-2.24400848e-01 5.91437638e-01 2.23224878e-01 -6.49590671e-01
3.53062898e-01 -3.34883928e-01 9.10762012e-01 -3.79438549e-01
-1.04431912e-01 -5.74416459e-01 -2.59742469e-01 -9.01594937e-01
-5.71683407e-01 -3.97314042e-01 -1.58959198e+00 -9.26107764e-01
-1.68377534e-03 4.26042899e-02 5.86815119e-01 1.06862712e+00
3.45250130e-01 9.37645078e-01 6.61168694e-01 -9.37443852e-01
-9.23060924e-02 -6.19649410e-01 -6.30786002e-01 6.09460294e-01
5.50905354e-02 -1.06965101e+00 -9.30323824e-02 -3.34686130e-01] | [6.841278076171875, 6.071304798126221] |
40ae1eca-6755-4eee-9f51-4defff252633 | does-the-explanation-satisfy-your-needs-a | 2211.05667 | null | https://arxiv.org/abs/2211.05667v2 | https://arxiv.org/pdf/2211.05667v2.pdf | What Makes a Good Explanation?: A Harmonized View of Properties of Explanations | Interpretability provides a means for humans to verify aspects of machine learning (ML) models and empower human+ML teaming in situations where the task cannot be fully automated. Different contexts require explanations with different properties. For example, the kind of explanation required to determine if an early cardiac arrest warning system is ready to be integrated into a care setting is very different from the type of explanation required for a loan applicant to help determine the actions they might need to take to make their application successful. Unfortunately, there is a lack of standardization when it comes to properties of explanations: different papers may use the same term to mean different quantities, and different terms to mean the same quantity. This lack of a standardized terminology and categorization of the properties of ML explanations prevents us from both rigorously comparing interpretable machine learning methods and identifying what properties are needed in what contexts. In this work, we survey properties defined in interpretable machine learning papers, synthesize them based on what they actually measure, and describe the trade-offs between different formulations of these properties. In doing so, we enable more informed selection of task-appropriate formulations of explanation properties as well as standardization for future work in interpretable machine learning. | ['Finale Doshi-Velez', 'Weiwei Pan', 'Marton Havasi', 'Varshini Subhash', 'Zixi Chen'] | 2022-11-10 | null | null | null | null | ['interpretable-machine-learning'] | ['methodology'] | [ 5.24756849e-01 5.96396208e-01 -4.86687779e-01 -8.07740033e-01
-3.14525604e-01 -7.29147613e-01 5.40026128e-01 4.66754854e-01
-3.86530417e-03 7.65092254e-01 3.05870473e-01 -1.13884819e+00
-5.75164139e-01 -3.40599507e-01 -4.29089040e-01 -4.75861669e-01
3.48451465e-01 6.59854352e-01 -3.65351796e-01 2.52768219e-01
3.26690286e-01 6.07129335e-01 -1.45432973e+00 5.14822185e-01
7.35718906e-01 6.49001837e-01 -2.34992370e-01 4.72131819e-01
-2.43365318e-01 8.73006463e-01 -6.67929232e-01 -2.70497531e-01
1.43374592e-01 -7.76399732e-01 -1.10180235e+00 3.08837056e-01
3.80778491e-01 -8.92469883e-02 4.45625782e-01 6.02053583e-01
-6.02363087e-02 -3.67533356e-01 1.11573660e+00 -1.61613250e+00
-6.15070999e-01 7.05026269e-01 1.48106232e-01 1.14752464e-01
4.94332314e-01 2.35060692e-01 1.11340511e+00 -3.22265625e-01
2.74558216e-01 1.18408275e+00 3.96301329e-01 7.56498098e-01
-1.18024051e+00 -2.72784740e-01 3.92980218e-01 1.43178761e-01
-8.01271796e-01 -3.96541148e-01 3.18651199e-01 -7.11629927e-01
8.72564495e-01 7.36750603e-01 5.57008088e-01 8.04726720e-01
3.02938968e-01 3.90810966e-01 1.26656520e+00 -7.21328378e-01
3.31521958e-01 4.39005196e-01 7.06906617e-01 7.69643784e-01
8.41615021e-01 -8.41520652e-02 -4.98577893e-01 -2.73028314e-01
6.82887435e-01 4.77771088e-02 -4.72603887e-01 -1.79367930e-01
-1.33010697e+00 8.73025298e-01 -9.80418473e-02 3.95186245e-01
-9.93249118e-02 1.81713745e-01 3.31103504e-01 5.37968457e-01
8.44971910e-02 9.95331705e-01 -7.05145240e-01 -1.29326358e-01
-4.30925310e-01 2.04611838e-01 1.01926196e+00 7.97529578e-01
6.89443588e-01 -7.60298455e-04 -2.47339699e-02 2.24083707e-01
4.61126715e-01 1.17530257e-01 2.77166873e-01 -1.18098295e+00
1.56558722e-01 9.73418951e-01 2.36400783e-01 -6.72721326e-01
-4.95245039e-01 -3.44299823e-01 -5.45001090e-01 4.05503780e-01
7.15672314e-01 -1.43204674e-01 -5.02147317e-01 1.60043585e+00
-9.37953219e-02 -1.06242590e-01 1.04652084e-01 9.01638269e-01
7.06284225e-01 1.21732503e-01 2.71925807e-01 -3.21705818e-01
1.44364405e+00 -6.41427219e-01 -6.20692432e-01 -5.28180003e-01
1.09440637e+00 -5.59410334e-01 1.32956684e+00 4.05396432e-01
-8.81442487e-01 -3.05120826e-01 -1.15017343e+00 -1.11759966e-02
-2.92641073e-01 -1.40550330e-01 1.09672248e+00 6.85456216e-01
-5.50036848e-01 5.82474172e-01 -5.91342032e-01 -2.31589183e-01
7.57202357e-02 4.45902556e-01 -3.64391148e-01 7.55453408e-02
-9.26004052e-01 1.30206883e+00 1.71300188e-01 1.25314549e-01
-2.93605179e-01 -6.79335654e-01 -9.13519561e-01 2.38313198e-01
2.90772647e-01 -9.66292202e-01 1.37202096e+00 -1.34606242e+00
-7.83396006e-01 8.82828891e-01 -3.62283528e-01 -4.96511996e-01
4.83656496e-01 -1.54139519e-01 -2.76536375e-01 -2.62570083e-01
1.47232696e-01 3.15310925e-01 5.09141743e-01 -9.55739558e-01
-7.29811847e-01 -3.51235002e-01 5.39511383e-01 1.49154410e-01
3.68471928e-02 6.91713300e-03 -2.04223394e-02 -2.96827883e-01
2.86603123e-01 -9.71256137e-01 -1.55051142e-01 7.16908425e-02
-3.44977885e-01 -2.82392085e-01 7.05409765e-01 -2.02830374e-01
1.15390515e+00 -1.84768260e+00 7.38292113e-02 9.43577290e-02
5.60815871e-01 -1.16970107e-01 9.22263339e-02 2.83732682e-01
-3.72608662e-01 6.30126715e-01 -1.72462627e-01 -2.16247998e-02
2.70355850e-01 5.09437740e-01 -4.37323302e-01 3.69763374e-01
3.37438107e-01 6.52681947e-01 -7.74420917e-01 -4.92937088e-01
3.08124721e-01 2.93968588e-01 -2.62773007e-01 7.13630393e-02
-1.42250001e-01 4.46485341e-01 -6.10861123e-01 4.58743572e-01
3.43479365e-02 -4.23692584e-01 2.58057535e-01 1.59619436e-01
1.30689234e-01 6.81769967e-01 -1.25867951e+00 7.89643705e-01
-6.39151931e-01 8.70648980e-01 -3.62723202e-01 -9.19733942e-01
6.08293772e-01 5.26860297e-01 1.15095377e-02 -1.60126209e-01
-2.53385175e-02 3.89031410e-01 4.05477345e-01 -6.91519797e-01
-1.11306868e-01 -6.12203062e-01 7.63532594e-02 8.87477338e-01
-4.82525229e-01 -2.06553221e-01 1.01033352e-01 -1.81180723e-02
1.12286985e+00 -1.98088452e-01 7.11479962e-01 -3.23329598e-01
5.14901042e-01 2.64795404e-02 5.93846560e-01 9.87104535e-01
1.07577927e-01 5.60342669e-01 8.50356698e-01 -8.96117926e-01
-7.62340307e-01 -8.10570061e-01 -2.53563702e-01 6.65925324e-01
-1.52465552e-01 -4.36688989e-01 -5.63654184e-01 -1.06191683e+00
3.18778083e-02 1.12137175e+00 -6.96748316e-01 -6.05182312e-02
-2.95379102e-01 -4.25562471e-01 3.05553734e-01 6.38419569e-01
-9.65885669e-02 -9.59580421e-01 -1.18908226e+00 -6.71104863e-02
-1.62494346e-01 -1.00732982e+00 -2.61325747e-01 4.06934917e-01
-1.09503973e+00 -1.46865535e+00 -1.97576475e-03 -3.57283294e-01
9.44665253e-01 1.78613231e-01 1.30620372e+00 8.83236706e-01
1.46694154e-01 6.77281141e-01 -4.24860477e-01 -7.93941736e-01
-7.63107598e-01 2.27600127e-03 6.75598010e-02 -3.82889807e-01
5.64831197e-01 -1.42872334e-01 -2.49709800e-01 3.60823601e-01
-1.08885324e+00 2.84444511e-01 5.97916663e-01 8.00668061e-01
3.26454133e-01 -1.68489322e-01 4.61284608e-01 -1.19735503e+00
7.35030115e-01 -2.71728903e-01 -2.55328298e-01 6.21823490e-01
-1.06202662e+00 5.01324475e-01 5.58078408e-01 -4.29173559e-01
-5.91471136e-01 2.39131376e-02 3.97640496e-01 2.49678195e-01
-4.01879460e-01 5.77021658e-01 -2.02561185e-01 2.19270065e-01
6.11558735e-01 -3.13641101e-01 -2.29247719e-01 -2.13034943e-01
2.08887145e-01 7.48738587e-01 -3.37650217e-02 -5.51710904e-01
8.20605338e-01 1.18543424e-01 1.01472482e-01 -4.00518298e-01
-1.00713861e+00 -2.01962858e-01 -6.93838835e-01 -2.04417538e-02
7.85631776e-01 -3.51715207e-01 -5.78704000e-01 -3.37167531e-01
-1.23154819e+00 -2.49668315e-01 -3.84575695e-01 5.11181414e-01
-6.87652171e-01 2.16202125e-01 -1.76113158e-01 -7.81894803e-01
-1.22966737e-01 -1.34264863e+00 9.18229699e-01 -2.04895176e-02
-1.12350321e+00 -1.18636501e+00 -3.37394804e-01 6.40552163e-01
1.51597053e-01 1.95466608e-01 1.49455369e+00 -1.02584851e+00
-4.13469315e-01 -3.01205128e-01 3.19275148e-02 3.09399456e-01
5.22271156e-01 9.80149060e-02 -8.32531631e-01 1.91989362e-01
2.38272101e-01 8.36347491e-02 2.27419794e-01 3.64725888e-01
1.03860617e+00 -7.20199585e-01 -2.90221751e-01 1.72405675e-01
9.02866304e-01 2.82102436e-01 3.69611233e-01 4.59586173e-01
4.30744052e-01 9.46032465e-01 6.41738176e-01 9.29043218e-02
1.84386924e-01 7.23863959e-01 2.20659122e-01 -2.82983780e-01
1.87244728e-01 1.12663969e-01 2.35885397e-01 1.92378491e-01
-1.14230551e-01 -6.15287237e-02 -1.02044666e+00 1.84129074e-01
-1.98670220e+00 -9.76810634e-01 -5.33111393e-01 2.40829420e+00
6.34870827e-01 3.57544124e-01 -1.53483704e-01 3.85806769e-01
3.25802624e-01 -2.31261209e-01 -4.14276689e-01 -8.11117828e-01
9.87425968e-02 -7.73051903e-02 8.41903538e-02 8.60887885e-01
-6.51187062e-01 4.87792045e-01 6.79996395e+00 2.32246846e-01
-9.32113647e-01 -1.19352698e-01 8.47668231e-01 2.99213797e-01
-6.68301523e-01 4.87468243e-01 -3.95003766e-01 1.42639220e-01
8.73354733e-01 -2.71013796e-01 1.65242195e-01 8.47772717e-01
4.95418102e-01 -7.08559528e-02 -1.90932584e+00 7.08066702e-01
-7.37576410e-02 -1.36846113e+00 6.92611933e-02 8.86848792e-02
1.51046142e-01 -6.23376250e-01 -1.47948593e-01 -6.20259047e-02
3.85236256e-02 -1.47783554e+00 7.77285993e-01 4.02743936e-01
3.39136690e-01 -3.12896788e-01 8.79589260e-01 4.27981943e-01
-7.42095768e-01 -1.62699893e-01 8.20263010e-03 -7.53769279e-01
-1.33322284e-01 4.86383319e-01 -1.32830548e+00 3.25259477e-01
2.75545210e-01 3.38591188e-01 -5.12894690e-01 7.48973429e-01
-7.28412688e-01 6.79657996e-01 3.74871977e-02 -1.39976457e-01
5.26464395e-02 -1.01539746e-01 3.31967533e-01 1.13826597e+00
8.94685909e-02 1.91672713e-01 -1.50391832e-01 9.03790057e-01
4.22594219e-01 -2.45252084e-02 -7.29678512e-01 -1.11538813e-01
4.07577246e-01 1.02406192e+00 -7.66326845e-01 -4.16497171e-01
-4.71747726e-01 6.95942342e-01 -9.33596194e-02 2.73209959e-01
-8.10089827e-01 1.87390845e-03 7.37453878e-01 4.99695390e-01
-5.17692745e-01 -2.07573950e-01 -7.14211047e-01 -1.07345879e+00
1.29232109e-01 -1.23242605e+00 4.90522772e-01 -9.68529522e-01
-1.10326004e+00 5.13976455e-01 2.70801336e-01 -1.28838158e+00
-4.98907298e-01 -7.37346530e-01 -5.76055467e-01 8.81609082e-01
-1.23110509e+00 -9.39392686e-01 -2.60134190e-01 -6.97442666e-02
6.05119109e-01 1.34489775e-01 1.01934052e+00 -3.05050254e-01
-3.95306408e-01 2.51228124e-01 -6.09034181e-01 -7.51331151e-02
5.83710849e-01 -1.47844720e+00 2.11980060e-01 6.52011335e-01
3.12004924e-01 1.06296682e+00 1.04974520e+00 -4.43839818e-01
-1.29067624e+00 -7.48647869e-01 1.28939199e+00 -1.01385033e+00
3.18278283e-01 3.55105735e-02 -8.93701911e-01 1.11604083e+00
-5.07718744e-03 -4.74927455e-01 1.01300645e+00 4.82616007e-01
-3.59994620e-01 1.39772370e-01 -9.70223129e-01 8.91339540e-01
8.26764405e-01 -3.90792459e-01 -8.49802196e-01 5.35513461e-01
6.76696539e-01 -1.27371326e-01 -7.28061557e-01 3.74200642e-01
5.49442410e-01 -1.06175363e+00 5.63040316e-01 -1.14593577e+00
1.70839280e-01 -5.79649091e-01 -8.80811438e-02 -8.96441579e-01
-3.00674170e-01 -4.43956673e-01 1.19455941e-01 9.42587912e-01
9.61384594e-01 -9.77551281e-01 5.80434144e-01 1.47096169e+00
-1.95505068e-01 -8.98560762e-01 -7.57480502e-01 -6.56204522e-01
-4.93202358e-02 -7.29073346e-01 8.36189926e-01 1.15487778e+00
3.91736805e-01 5.34169793e-01 -2.22624838e-02 3.36022943e-01
1.85520962e-01 2.00347289e-01 7.46176541e-01 -1.50573909e+00
-3.29275340e-01 -5.65059602e-01 -5.42582691e-01 -6.66968167e-01
2.66726196e-01 -7.90314734e-01 -2.85276473e-01 -1.79280853e+00
2.74306595e-01 -5.06536305e-01 -8.10672045e-02 1.01018441e+00
-2.26224422e-01 -2.47922346e-01 3.23195726e-01 2.44377479e-01
-2.53318876e-01 -1.20602727e-01 8.19573224e-01 -1.83931753e-01
-1.43563703e-01 2.86940575e-01 -1.21270025e+00 1.06610334e+00
7.73792207e-01 -6.35623038e-01 -6.19858921e-01 -4.50449407e-01
4.61072534e-01 8.66989978e-03 5.76708257e-01 -7.14464426e-01
-2.59706285e-02 -6.69675887e-01 3.40459764e-01 1.74137920e-01
-6.66407719e-02 -1.11711729e+00 4.78384078e-01 7.31123745e-01
-7.18507111e-01 2.62944430e-01 1.15825333e-01 1.74785003e-01
-1.44236675e-02 -7.24778891e-01 3.29555899e-01 -7.34216943e-02
-2.94952929e-01 -1.66706815e-01 -6.24761343e-01 -9.03854668e-02
1.01687026e+00 -5.06858051e-01 -3.33663732e-01 -6.53006434e-01
-7.48876810e-01 1.03827000e-01 7.06247091e-01 4.50074375e-01
4.71143782e-01 -8.81722271e-01 -5.16165137e-01 -7.57322982e-02
1.85598880e-01 -2.54626393e-01 -1.80052444e-01 8.83518159e-01
-3.83601844e-01 6.35140598e-01 9.42753255e-02 -3.19440812e-01
-1.54594839e+00 3.95010620e-01 5.59230208e-01 -1.53760985e-01
-5.91446459e-01 1.78992271e-01 3.33651185e-01 -2.18031719e-01
3.50130588e-01 -8.15239489e-01 -1.30966187e-01 -2.43514970e-01
5.66450357e-01 1.16507642e-01 1.38512835e-01 -2.30655953e-01
-5.82527220e-01 4.24541295e-01 1.66509449e-01 -1.08688943e-01
1.09035504e+00 1.50000438e-01 -8.67492333e-02 7.77947366e-01
5.28004885e-01 -8.00304785e-02 -8.99644494e-01 2.28067085e-01
1.49029002e-01 -2.88387805e-01 -3.25131953e-01 -1.06878006e+00
-6.79006636e-01 9.54769254e-01 3.24569255e-01 5.71096838e-01
1.00658762e+00 1.66237339e-01 1.04734488e-01 6.27653003e-01
3.41521710e-01 -7.60202527e-01 -3.07303041e-01 1.29912212e-01
1.14067304e+00 -1.21617973e+00 3.96345913e-01 -6.02114618e-01
-9.08094466e-01 1.47124791e+00 4.33503181e-01 4.43951309e-01
3.18084598e-01 1.83901340e-01 3.21742117e-01 -4.25435901e-01
-8.66211116e-01 7.82845840e-02 4.78035033e-01 4.51639205e-01
7.43848145e-01 2.74794638e-01 -5.24205625e-01 5.64681649e-01
-2.94705689e-01 -1.18706152e-01 7.27274895e-01 8.26579094e-01
-3.85599524e-01 -1.35862362e+00 -4.94212091e-01 6.93586588e-01
-2.45606616e-01 -1.01126432e-01 -8.75570655e-01 1.10998654e+00
2.65675128e-01 1.25960517e+00 -1.93451345e-01 -2.12682411e-01
2.84713179e-01 3.92784178e-01 4.32898134e-01 -1.01669741e+00
-5.96278846e-01 -3.84163916e-01 4.45534527e-01 -2.18285650e-01
-4.78449434e-01 -7.27679491e-01 -1.59069490e+00 -1.47211760e-01
-3.43126208e-01 4.04486924e-01 5.00736833e-01 1.60373282e+00
2.46078268e-01 3.68068695e-01 1.67785510e-01 -2.70922035e-01
-3.66046965e-01 -5.79453766e-01 -2.09240735e-01 3.59384269e-01
3.83380055e-01 -5.97140312e-01 -6.98224485e-01 2.51433194e-01] | [8.802586555480957, 5.907437801361084] |
38b9bf0c-08db-4c47-9914-4b7024f7a257 | sample-crop-track-self-supervised-mobile-3d | 2209.10471 | null | https://arxiv.org/abs/2209.10471v1 | https://arxiv.org/pdf/2209.10471v1.pdf | Sample, Crop, Track: Self-Supervised Mobile 3D Object Detection for Urban Driving LiDAR | Deep learning has led to great progress in the detection of mobile (i.e. movement-capable) objects in urban driving scenes in recent years. Supervised approaches typically require the annotation of large training sets; there has thus been great interest in leveraging weakly, semi- or self-supervised methods to avoid this, with much success. Whilst weakly and semi-supervised methods require some annotation, self-supervised methods have used cues such as motion to relieve the need for annotation altogether. However, a complete absence of annotation typically degrades their performance, and ambiguities that arise during motion grouping can inhibit their ability to find accurate object boundaries. In this paper, we propose a new self-supervised mobile object detection approach called SCT. This uses both motion cues and expected object sizes to improve detection performance, and predicts a dense grid of 3D oriented bounding boxes to improve object discovery. We significantly outperform the state-of-the-art self-supervised mobile object detection method TCR on the KITTI tracking benchmark, and achieve performance that is within 30% of the fully supervised PV-RCNN++ method for IoUs <= 0.5. | ['Niki Trigoni', 'Andrew Markham', 'Kaichen Zhou', 'Madhu Vankadari', 'Stuart Golodetz', 'Sangyun Shin'] | 2022-09-21 | null | null | null | null | ['object-discovery'] | ['computer-vision'] | [ 1.35122731e-01 -1.61068030e-02 -3.75447184e-01 -4.46158499e-01
-8.00901890e-01 -6.49890065e-01 6.54544413e-01 1.09059317e-02
-6.46470845e-01 5.89867711e-01 -4.07107621e-02 -3.63058180e-01
2.91840672e-01 -4.65232313e-01 -8.23092282e-01 -7.12137341e-01
-9.09989327e-02 5.61430752e-01 1.04449725e+00 -1.13586791e-01
1.06111906e-01 2.30800584e-01 -1.83317327e+00 6.45728707e-02
7.63145924e-01 9.59717572e-01 3.23953241e-01 7.90001988e-01
-1.02881648e-01 8.75677586e-01 -3.28561217e-01 -5.21999858e-02
3.56719643e-01 -1.05539180e-01 -6.18614674e-01 1.15187056e-01
8.18643689e-01 -1.36255264e-01 -3.56889755e-01 8.82943392e-01
4.52561468e-01 1.27848610e-01 6.46992505e-01 -1.40351522e+00
8.04706067e-02 2.98755914e-01 -7.92925775e-01 6.30333543e-01
1.20398297e-04 2.48954743e-01 9.29220676e-01 -8.60172272e-01
6.62996233e-01 9.11957979e-01 1.05846512e+00 5.19523799e-01
-1.10666525e+00 -5.99040389e-01 2.81359047e-01 7.54064396e-02
-1.50720215e+00 -6.54121935e-01 7.33008981e-01 -5.94304323e-01
9.82502580e-01 1.54864147e-01 6.75861239e-01 8.44428539e-01
-9.70161408e-02 1.19070983e+00 8.42124164e-01 -2.26739883e-01
2.53984720e-01 3.22141290e-01 1.02040403e-01 8.11082661e-01
4.45044577e-01 2.12314412e-01 -3.32613200e-01 1.90954320e-02
4.21181500e-01 -1.51811793e-01 1.56967625e-01 -8.96137416e-01
-1.18071270e+00 6.92128599e-01 6.85558438e-01 2.09286422e-01
2.12038122e-02 4.55750763e-01 4.89670515e-01 -1.44038960e-01
5.59072733e-01 1.14934631e-01 -3.68116111e-01 -3.91416609e-01
-1.28261781e+00 3.22842389e-01 4.57850367e-01 1.10794377e+00
7.53518462e-01 3.35254632e-02 -6.44134134e-02 5.21610498e-01
4.59109306e-01 3.71609092e-01 4.22696680e-01 -8.42733443e-01
4.94500905e-01 5.84485531e-01 4.44801152e-01 -8.61512125e-01
-5.70261478e-01 -5.34166694e-01 -3.70126724e-01 4.65420902e-01
6.30432725e-01 -4.86131236e-02 -1.14576006e+00 1.42654788e+00
5.73559463e-01 2.21630141e-01 1.08005688e-01 9.49801743e-01
8.16390455e-01 4.92259979e-01 7.92079493e-02 5.39678037e-02
9.80904937e-01 -1.18601465e+00 -3.39827716e-01 -7.46601522e-01
1.07110989e+00 -5.65211177e-01 7.88227499e-01 6.27107918e-02
-7.38637090e-01 -6.32689118e-01 -1.07937574e+00 5.37793711e-02
-3.37892592e-01 2.04266340e-01 7.48093545e-01 9.86868083e-01
-8.45034361e-01 4.34652567e-01 -1.06612861e+00 -4.58524197e-01
9.59326148e-01 4.62207705e-01 -1.75657108e-01 -4.57412470e-03
-6.96805954e-01 7.73604929e-01 4.00200576e-01 1.92410499e-01
-7.90626466e-01 -4.80254859e-01 -9.13522601e-01 -3.71840686e-01
6.07341349e-01 -4.24865842e-01 1.27675366e+00 -9.99897599e-01
-7.97860920e-01 1.09494233e+00 -3.00809532e-01 -8.84837270e-01
9.23013449e-01 -2.27155745e-01 -2.01269612e-01 -1.80912957e-01
4.32941377e-01 1.41147149e+00 9.10244763e-01 -1.40282023e+00
-1.35676312e+00 -1.97235271e-01 -1.02791719e-01 1.78520560e-01
3.38192447e-03 -1.40583709e-01 -7.15019405e-01 -3.65380377e-01
5.67069128e-02 -1.32447529e+00 -2.99178421e-01 6.27833083e-02
-2.24249080e-01 -4.28087205e-01 1.34297943e+00 -2.40524128e-01
9.04717565e-01 -2.08645463e+00 -3.53409588e-01 1.93728204e-03
3.59187692e-01 4.70726103e-01 1.95429474e-01 -2.22630143e-01
4.27291304e-01 -1.01862907e-01 -2.70534873e-01 -6.10305369e-01
-1.36593640e-01 4.38412935e-01 -5.72887883e-02 6.30958438e-01
1.74824685e-01 1.09291625e+00 -1.05382633e+00 -7.69905210e-01
3.88903707e-01 3.50589693e-01 -5.08335650e-01 -1.77037731e-01
-2.99621284e-01 3.33593279e-01 -2.82197893e-01 8.28347445e-01
4.97854561e-01 -3.24425340e-01 -1.58959687e-01 1.62566617e-01
-3.04706842e-01 2.63792187e-01 -1.18832612e+00 1.36187708e+00
-1.41804054e-01 1.22164381e+00 2.28881910e-02 -1.00503230e+00
7.20313907e-01 -1.28088117e-01 6.72464192e-01 -4.48321044e-01
6.41573742e-02 2.64459729e-01 1.37488514e-01 -4.30305302e-01
6.48689091e-01 2.19848156e-01 1.17107160e-01 5.89502305e-02
-3.20477694e-01 7.79685825e-02 2.54676729e-01 1.76182181e-01
1.13916159e+00 4.61313188e-01 1.33660072e-02 -2.38548666e-01
3.42164338e-01 5.86672187e-01 6.58954442e-01 1.08796787e+00
-6.62854970e-01 6.91476762e-01 8.64410922e-02 -5.38968384e-01
-9.59764957e-01 -6.67833507e-01 -1.55763656e-01 1.22431946e+00
5.68132222e-01 -2.66245097e-01 -6.19373083e-01 -1.01391351e+00
1.96788996e-01 3.13575149e-01 -6.30847871e-01 4.56932001e-02
-8.35098624e-01 -8.54495466e-01 5.52671671e-01 1.07755387e+00
5.34313619e-01 -9.44854915e-01 -8.93213451e-01 3.36518675e-01
-1.74447492e-01 -1.42801714e+00 -3.97024691e-01 4.98791665e-01
-7.76348293e-01 -9.90749359e-01 -6.92077219e-01 -9.06147659e-01
6.94681883e-01 6.75051212e-01 9.31101084e-01 6.47727698e-02
-4.10936028e-01 2.77877778e-01 -3.67126852e-01 -5.28068125e-01
-2.97271192e-01 2.11811706e-01 1.12101771e-01 -6.74218312e-02
4.68177557e-01 -7.18541220e-02 -7.26052344e-01 6.99337125e-01
-4.26826984e-01 -2.60635931e-02 5.77365339e-01 6.36277020e-01
5.94078660e-01 -3.60067636e-02 4.10376340e-01 -7.20699847e-01
-1.42176852e-01 -2.46055573e-01 -7.87751019e-01 -2.68100947e-01
-6.22409940e-01 -2.02991962e-02 1.31285697e-01 -5.71509540e-01
-8.12984228e-01 5.76493859e-01 -1.30479008e-01 -4.07538831e-01
-4.09106344e-01 6.33069724e-02 -9.06522274e-02 -4.06614512e-01
8.07010770e-01 1.34615144e-02 -1.79274678e-01 -1.53051436e-01
2.72872686e-01 6.03570282e-01 5.98805368e-01 -6.38916045e-02
8.50299060e-01 9.35827076e-01 -8.72032791e-02 -8.70313406e-01
-9.66290057e-01 -1.10885441e+00 -8.16414535e-01 -4.05909836e-01
8.13144863e-01 -1.09958625e+00 -1.48862615e-01 3.26674134e-01
-8.18815410e-01 -6.43732905e-01 -2.37454146e-01 4.31605160e-01
-5.62550902e-01 3.25820655e-01 -1.16772272e-01 -1.02022314e+00
-5.82163334e-02 -1.06720126e+00 1.26568782e+00 1.14941284e-01
-2.87139714e-01 -8.12500238e-01 -1.79490983e-01 6.00195706e-01
3.60500634e-01 2.61159986e-01 5.39936610e-02 -5.81475616e-01
-8.91798556e-01 -3.60221326e-01 -2.81781286e-01 -1.00350520e-02
-4.47505675e-02 -2.52427369e-01 -1.09052503e+00 -2.09399328e-01
-6.09567702e-01 -2.22270682e-01 1.29003263e+00 5.47081769e-01
8.33709180e-01 1.03928514e-01 -9.51467037e-01 5.29436886e-01
9.78245437e-01 3.30482386e-02 4.17842299e-01 6.93488955e-01
9.54983115e-01 4.70896810e-01 9.99652386e-01 4.42580283e-02
5.44062674e-01 8.94693553e-01 6.66819930e-01 -2.26057261e-01
-2.53375232e-01 -1.56684980e-01 2.80952483e-01 -2.03104559e-02
-8.42264593e-02 -1.54265016e-01 -1.04887891e+00 7.88908184e-01
-2.20075321e+00 -9.72918570e-01 -5.60146451e-01 2.05428600e+00
5.25218308e-01 9.13684607e-01 5.81122935e-01 2.42086425e-01
6.30237460e-01 1.47062793e-01 -6.75507784e-01 2.56059170e-01
-9.01459977e-02 -4.54355627e-01 1.01388645e+00 4.12863314e-01
-1.67331874e+00 1.29873109e+00 5.92799282e+00 7.79858768e-01
-1.04009628e+00 1.09622948e-01 4.83443230e-01 6.01144582e-02
2.10485950e-01 -3.01007205e-03 -1.37408686e+00 3.54739636e-01
5.83494782e-01 3.95008296e-01 -1.54081017e-01 1.29431152e+00
2.21503600e-01 -4.95964617e-01 -8.82320285e-01 1.01556993e+00
1.24265037e-01 -1.32818019e+00 -6.18078053e-01 8.21421817e-02
8.43406618e-01 5.03961802e-01 -3.81929800e-02 4.44027871e-01
1.14942029e-01 -7.48260081e-01 8.67170870e-01 5.10434881e-02
4.46955174e-01 -6.10852718e-01 8.38918746e-01 7.06680477e-01
-1.51579571e+00 -1.52180985e-01 -3.91671598e-01 -1.55742913e-01
1.54349595e-01 3.90521109e-01 -1.05110741e+00 -4.71593440e-02
9.35548544e-01 7.91973233e-01 -8.51358473e-01 1.46295929e+00
5.90100549e-02 7.31153846e-01 -6.01655543e-01 -2.35415325e-01
6.23157680e-01 3.27378899e-01 6.83059037e-01 1.35279095e+00
-5.04514836e-02 -2.38373533e-01 4.27242190e-01 4.62447554e-01
1.71687126e-01 -1.32488325e-01 -5.97216070e-01 2.91388601e-01
1.79469571e-01 1.27243185e+00 -1.23797464e+00 -4.12791669e-01
-3.66687506e-01 6.98047459e-01 1.86170697e-01 1.37148157e-01
-9.86057520e-01 -1.96149703e-02 6.18649125e-01 6.06798708e-01
8.31523240e-01 -4.56610978e-01 -2.45633438e-01 -8.26670408e-01
5.18174842e-02 -2.40970477e-01 3.04185271e-01 -5.70944965e-01
-8.00428987e-01 3.62944186e-01 -5.82052916e-02 -1.40342784e+00
-2.51214206e-01 -6.53884351e-01 -4.92934734e-01 1.92692623e-01
-1.55812883e+00 -1.20390344e+00 -5.32919943e-01 2.60493964e-01
7.14576244e-01 -3.58584076e-02 1.20409548e-01 3.68697494e-01
-3.22048068e-01 5.32454908e-01 4.79134694e-02 2.78198421e-01
6.77976072e-01 -1.29754364e+00 5.34296870e-01 8.48291218e-01
4.29909706e-01 1.18859924e-01 8.50320995e-01 -7.29700029e-01
-1.17018378e+00 -1.49153745e+00 7.25277781e-01 -9.49698031e-01
4.58774149e-01 -5.29663861e-01 -7.58199215e-01 5.96976578e-01
-3.01505715e-01 5.93465984e-01 2.84529060e-01 -4.77947704e-02
-1.07105404e-01 -1.42004967e-01 -8.75487626e-01 4.79273975e-01
1.44525933e+00 -5.48133925e-02 -3.53017837e-01 3.77184927e-01
4.67668980e-01 -6.46160781e-01 -1.90977424e-01 6.18558943e-01
3.90128374e-01 -7.89903283e-01 1.06374729e+00 -2.73229241e-01
-1.69032127e-01 -7.34225154e-01 1.71072453e-01 -6.75074995e-01
-1.63474813e-01 -5.82888126e-01 -3.53934526e-01 9.09569561e-01
4.77665395e-01 -2.36121789e-01 1.49931300e+00 3.51618588e-01
-3.50685984e-01 -6.43079221e-01 -1.01399243e+00 -9.71480191e-01
-2.81478584e-01 -7.23358512e-01 1.82832927e-02 6.99452460e-01
-2.50419021e-01 2.84132659e-01 -3.54222596e-01 2.08187103e-01
7.34249890e-01 4.89404276e-02 1.24531579e+00 -1.31572974e+00
4.97595780e-02 -6.12578094e-01 -8.77953768e-01 -1.55705774e+00
5.93887493e-02 -7.30374694e-01 4.84305918e-01 -1.58665073e+00
6.42452575e-03 -8.36859345e-01 1.11082487e-01 5.10252714e-01
-1.41771480e-01 7.53097057e-01 -1.15549318e-01 4.09673035e-01
-1.20304644e+00 2.87339836e-01 8.16394031e-01 -1.76730424e-01
-5.34248173e-01 2.91541994e-01 -4.41025615e-01 8.59729648e-01
8.05092990e-01 -5.97038746e-01 -1.12727888e-01 -1.14363953e-01
1.22452654e-01 -4.52024460e-01 6.98606074e-01 -1.43515408e+00
4.59544271e-01 1.75110474e-01 4.44584399e-01 -1.06984127e+00
3.64961743e-01 -7.47737348e-01 -2.64893293e-01 3.88500869e-01
-1.16768308e-01 -2.18387902e-01 3.54082227e-01 8.66186142e-01
-6.81499466e-02 -1.08328395e-01 7.70865262e-01 -1.75565518e-02
-1.20661688e+00 2.23185882e-01 -5.04004240e-01 1.20218014e-02
1.10672355e+00 -7.27011144e-01 -3.15675102e-02 -2.54139632e-01
-7.30417490e-01 4.63555574e-01 4.20973361e-01 6.03306413e-01
3.29826057e-01 -1.05538642e+00 -4.20203894e-01 -2.67085694e-02
3.66231501e-01 3.11908990e-01 -1.31455585e-01 1.06157351e+00
-3.95006239e-01 5.50628781e-01 1.76550239e-01 -1.30092824e+00
-1.57792544e+00 3.87962520e-01 3.80200416e-01 5.03488667e-02
-8.20640504e-01 9.47504044e-01 1.17830373e-01 -2.36947253e-01
4.91105765e-01 -4.71689850e-01 -2.00874493e-01 -4.56843451e-02
3.50449800e-01 4.00514722e-01 7.19604492e-02 -9.92989063e-01
-5.74497640e-01 5.50984323e-01 -1.63193405e-01 -1.72815789e-02
9.42767322e-01 -1.92716032e-01 5.61906755e-01 3.17950279e-01
1.03240991e+00 -6.56070039e-02 -1.76780570e+00 -2.43708879e-01
3.29907507e-01 -4.11491305e-01 2.18423367e-01 -3.94079864e-01
-7.64927685e-01 6.80304170e-01 7.86940038e-01 1.98117942e-01
5.14256537e-01 3.20412129e-01 7.40214169e-01 4.86610860e-01
3.86991560e-01 -1.17761970e+00 9.55050513e-02 5.76326370e-01
2.22964406e-01 -1.89086962e+00 -2.46870685e-02 -4.70423490e-01
-6.19431973e-01 7.49012351e-01 8.12996268e-01 2.04534810e-02
5.34667432e-01 3.70262206e-01 1.37436807e-01 -1.00276023e-01
-3.54965329e-01 -7.79187143e-01 5.04053652e-01 7.62685180e-01
1.37012362e-01 -2.09062457e-01 1.02480717e-01 3.82992476e-02
-3.17295827e-02 -1.08516499e-01 1.57671899e-01 1.31548452e+00
-8.83607745e-01 -6.83560073e-01 -5.27966022e-01 5.42041063e-01
-2.53495932e-01 2.16642842e-01 -3.49272162e-01 1.01382709e+00
3.60651761e-01 9.67166960e-01 2.26982579e-01 -2.36830816e-01
1.57293618e-01 -1.17791079e-01 2.62332708e-01 -5.19090474e-01
-2.47694805e-01 2.58251369e-01 2.42010444e-01 -4.61464792e-01
-7.71136701e-01 -9.31772411e-01 -1.38872492e+00 1.00244187e-01
-6.90889418e-01 -2.47543249e-02 6.04234755e-01 1.06984556e+00
2.22260848e-01 1.49607167e-01 1.76932231e-01 -1.22899175e+00
-7.36842677e-02 -7.78886318e-01 -9.03851464e-02 1.60871223e-01
5.66781402e-01 -9.48720217e-01 -4.36704040e-01 1.50305927e-01] | [8.001323699951172, -1.3578487634658813] |
b91a4f33-0a01-4c46-9804-3bd23a6c01c6 | forward-and-inverse-approximation-theory-for | 2305.18478 | null | https://arxiv.org/abs/2305.18478v1 | https://arxiv.org/pdf/2305.18478v1.pdf | Forward and Inverse Approximation Theory for Linear Temporal Convolutional Networks | We present a theoretical analysis of the approximation properties of convolutional architectures when applied to the modeling of temporal sequences. Specifically, we prove an approximation rate estimate (Jackson-type result) and an inverse approximation theorem (Bernstein-type result), which together provide a comprehensive characterization of the types of sequential relationships that can be efficiently captured by a temporal convolutional architecture. The rate estimate improves upon a previous result via the introduction of a refined complexity measure, whereas the inverse approximation theorem is new. | ['Qianxiao Li', 'Haotian Jiang'] | 2023-05-29 | null | null | null | null | ['temporal-sequences'] | ['reasoning'] | [ 8.52679014e-02 6.29117712e-02 -2.33538643e-01 -1.74824908e-01
-1.30253300e-01 -8.45172405e-01 4.99997824e-01 2.74745613e-01
-3.91718417e-01 3.18286508e-01 3.49185288e-01 -7.29692936e-01
-4.24218386e-01 -4.59209591e-01 -1.05000222e+00 -2.84209102e-01
-7.33783424e-01 7.51348510e-02 4.55647767e-01 -4.16782379e-01
3.30640450e-02 8.93366575e-01 -1.34429252e+00 2.68375903e-01
2.28598386e-01 1.52463579e+00 -2.23250955e-01 1.18995166e+00
-3.28135863e-02 1.35776317e+00 -1.72865108e-01 -3.04685146e-01
2.47163057e-01 -5.89638174e-01 -1.07344723e+00 -3.83298993e-01
1.79680958e-01 -4.98271167e-01 -7.58657038e-01 8.52845490e-01
-1.55160531e-01 4.45495248e-01 5.50362945e-01 -1.31101584e+00
-5.06187260e-01 8.31645906e-01 9.36386082e-03 8.04104149e-01
2.07456082e-01 -2.01987401e-01 1.12488031e+00 -4.77842987e-01
3.91790181e-01 1.01320457e+00 1.29503584e+00 4.66257989e-01
-1.12111139e+00 -2.52619714e-01 2.52890915e-01 2.30247051e-01
-1.31808281e+00 -4.70367581e-01 3.65785867e-01 -4.01630670e-01
1.22801018e+00 1.80562198e-01 7.51124740e-01 5.94112813e-01
6.02019280e-02 1.04635668e+00 5.19759417e-01 -3.61396015e-01
1.23476543e-01 -3.25999320e-01 4.61457849e-01 9.54825103e-01
2.35113502e-01 2.70021200e-01 -5.01397967e-01 -2.02799857e-01
1.09796572e+00 -1.65523723e-01 -3.17016214e-01 -1.75662369e-01
-9.71061110e-01 4.45156246e-01 4.55886066e-01 5.77009022e-01
-3.61799479e-01 8.49782586e-01 6.53031468e-01 5.21302819e-01
2.05795929e-01 9.45803076e-02 -4.74771351e-01 -3.79400551e-01
-8.36626351e-01 4.24479127e-01 1.04167378e+00 1.25615609e+00
3.41977715e-01 2.15668872e-01 -8.90480950e-02 2.01010630e-01
3.38399187e-02 -6.91979378e-02 2.31261611e-01 -1.20963073e+00
1.69884607e-01 2.29923233e-01 2.46217802e-01 -8.87620270e-01
-5.37356615e-01 -7.32156277e-01 -9.53096509e-01 -5.18535018e-01
7.25508213e-01 9.56033021e-02 -2.83592016e-01 2.27767611e+00
-1.90059006e-01 3.83298457e-01 -6.58015981e-02 4.15459037e-01
1.76501766e-01 7.31735229e-01 -1.93998024e-01 -5.19828081e-01
1.01267374e+00 -7.46448755e-01 -9.12417889e-01 2.01895609e-01
7.54534364e-01 -2.16634676e-01 5.22753239e-01 2.16171190e-01
-1.30004966e+00 -6.07034266e-01 -1.14216733e+00 -7.52148554e-02
-4.72257473e-02 -2.27918223e-01 9.63001311e-01 4.59689111e-01
-1.47841537e+00 1.05133057e+00 -9.54046667e-01 -3.27527195e-01
3.86853516e-01 3.24461222e-01 -1.79510102e-01 2.68201709e-01
-9.99884844e-01 6.61535144e-01 4.13461685e-01 3.25528860e-01
-9.41482782e-01 -9.91412580e-01 -6.47857428e-01 3.95315170e-01
2.43544698e-01 -5.88955164e-01 1.79989219e+00 -1.06637490e+00
-1.23597527e+00 4.53017265e-01 -3.97412390e-01 -1.07598960e+00
6.04455888e-01 -2.41740853e-01 -2.49055043e-01 4.46243823e-01
-4.62493241e-01 4.55056876e-02 2.60121346e-01 -5.90276122e-01
-6.48948193e-01 6.03523515e-02 4.97921348e-01 -3.45260262e-01
-3.79110873e-01 6.18730858e-02 -2.01128080e-01 -7.50916958e-01
-2.14221627e-01 -6.18699431e-01 -5.55384398e-01 3.21683496e-01
1.13315219e-02 -3.36261600e-01 1.00852586e-01 -5.19195378e-01
1.62884629e+00 -2.16170669e+00 2.11411148e-01 1.50461212e-01
4.30622846e-01 4.06609513e-02 -1.77073851e-01 6.72855794e-01
-5.11124253e-01 2.24626392e-01 -1.52516380e-01 -1.34556577e-01
1.19237162e-01 1.94652632e-01 -5.35820186e-01 4.34624732e-01
1.82295814e-01 1.06773686e+00 -9.59156871e-01 -2.52275258e-01
-1.02680594e-01 -1.75368190e-02 -6.86168492e-01 1.61624223e-01
-3.29660535e-01 -2.05387115e-01 -1.64953977e-01 9.22330320e-02
3.78090769e-01 -4.66927201e-01 3.68187845e-01 -3.15071344e-01
-2.08572686e-01 5.12281001e-01 -9.86959279e-01 1.42313898e+00
-3.78430843e-01 9.99512017e-01 -1.42915815e-01 -1.10611773e+00
5.91663122e-01 6.14456713e-01 5.59441626e-01 -1.81279749e-01
3.36385012e-01 2.81741291e-01 3.92506905e-02 -2.46818423e-01
5.62724471e-01 -3.47404599e-01 -4.49038334e-02 8.44128907e-01
-8.69885273e-03 4.17256713e-01 4.06336308e-01 3.97848547e-01
1.30266678e+00 1.56978548e-01 3.47186953e-01 -7.21566796e-01
4.64471042e-01 -1.25581920e-01 2.30464980e-01 1.04937923e+00
-2.64861792e-01 5.53409522e-03 8.99268329e-01 -9.86844420e-01
-1.49231303e+00 -9.50762272e-01 3.28636058e-02 8.60224426e-01
-2.91012824e-01 -7.92176127e-01 -6.15780890e-01 -2.90293753e-01
-1.97476283e-01 1.70086682e-01 -1.06960535e+00 -4.10491467e-01
-7.57049561e-01 -2.19920754e-01 1.11539924e+00 1.34621835e+00
2.51057327e-01 -5.34019530e-01 -7.40696371e-01 3.85357827e-01
-3.12940001e-01 -1.26024508e+00 -4.72075880e-01 4.31782067e-01
-1.11623621e+00 -1.00231516e+00 -4.39550787e-01 -5.91610670e-01
5.80341183e-02 -2.71074306e-02 1.06140995e+00 2.61631370e-01
1.14118703e-01 4.70112234e-01 -1.49080291e-01 1.92166660e-02
-6.12323225e-01 7.85572603e-02 2.47288257e-01 -3.68762799e-02
2.71713078e-01 -9.19763088e-01 -4.61851209e-01 1.24266200e-01
-1.11275947e+00 -1.34938747e-01 2.80942112e-01 6.44146562e-01
1.28235385e-01 1.76680386e-01 3.01747620e-01 -3.79619181e-01
5.33634365e-01 -5.70457697e-01 -8.50306451e-01 8.96514803e-02
-2.56032288e-01 5.63981533e-01 9.34001982e-01 -8.35103035e-01
-5.94026506e-01 4.41899300e-02 -4.73156273e-02 -7.03406811e-01
2.25066990e-01 7.71682322e-01 3.49290699e-01 8.17854777e-02
4.86700684e-01 4.82615769e-01 -8.92480612e-02 -4.25388336e-01
3.60924929e-01 1.33596957e-01 9.06440437e-01 -7.38859057e-01
5.58767796e-01 5.72233319e-01 4.68714118e-01 -7.25968838e-01
-9.51640546e-01 -2.05789864e-01 -6.55837834e-01 -3.47837836e-01
4.07728314e-01 -5.89248359e-01 -1.08396614e+00 1.99326962e-01
-1.43753266e+00 -5.16333103e-01 -2.39237651e-01 4.76572931e-01
-9.49590087e-01 5.51616848e-01 -1.01544404e+00 -1.23386812e+00
1.09957188e-01 -5.96244574e-01 3.76678586e-01 -2.39125878e-01
-4.39346433e-01 -9.42632616e-01 1.40569136e-01 -4.21013534e-01
4.42673892e-01 2.77756453e-01 1.25920987e+00 -7.53093779e-01
-5.97482204e-01 -2.57079929e-01 -3.32529575e-01 3.90678018e-01
-3.58323693e-01 1.74376182e-02 -5.93800783e-01 3.98077033e-02
3.06313820e-02 1.53362513e-01 8.38949502e-01 4.61976260e-01
1.18425548e+00 -5.87370455e-01 -2.14291021e-01 5.30592799e-01
1.59053409e+00 1.67062402e-01 6.13112748e-01 1.70288473e-01
3.68990570e-01 2.51451075e-01 -9.33359712e-02 7.07797229e-01
-1.83780994e-02 5.18474519e-01 2.75782198e-01 5.66685736e-01
1.03235967e-01 -2.79273123e-01 2.58560538e-01 9.54497993e-01
-5.92227221e-01 -1.39767051e-01 -7.82054424e-01 8.78498495e-01
-2.15070152e+00 -1.10722101e+00 -2.23424450e-01 2.08985996e+00
8.11802745e-01 2.51504689e-01 7.21008599e-01 4.97737586e-01
5.80326557e-01 3.51737626e-02 -4.71428514e-01 -7.68006146e-01
-3.46646048e-02 5.19243538e-01 8.40444744e-01 5.49869418e-01
-7.96041787e-01 7.14345634e-01 9.24619579e+00 6.77008331e-01
-3.29640865e-01 3.69967930e-02 1.63795009e-01 1.97966490e-03
-2.93041825e-01 -1.33406639e-01 -5.78103960e-01 -9.82317049e-03
1.63022971e+00 -5.98621547e-01 6.15962744e-01 8.26652229e-01
5.77728786e-02 4.38455790e-01 -1.63968360e+00 7.38887191e-01
-3.18183124e-01 -1.74510098e+00 1.01954907e-01 1.55011728e-01
4.74193156e-01 -3.50801468e-01 2.73199957e-02 -5.19279838e-02
3.58243465e-01 -9.25194025e-01 1.13664281e+00 6.45559907e-01
5.56575000e-01 -9.64235604e-01 4.52859998e-01 1.84915543e-01
-1.80338144e+00 -4.95937347e-01 -2.37077475e-01 -5.41511476e-01
1.21900633e-01 3.07503551e-01 -1.99921340e-01 5.97440839e-01
5.72380483e-01 7.76013613e-01 -2.12875798e-01 1.03727603e+00
4.06847382e-03 4.70490366e-01 -5.38250208e-01 -2.44949937e-01
5.84680200e-01 2.39088297e-01 5.51020384e-01 1.37405360e+00
4.59056161e-02 2.76116312e-01 -4.48901385e-01 9.99053180e-01
2.95132659e-02 -4.03196305e-01 -5.79911113e-01 -4.64697659e-01
4.79712188e-01 6.21861994e-01 -5.89116633e-01 -3.96369219e-01
-4.37187284e-01 7.20318854e-01 4.49660391e-01 2.14215070e-01
-8.36680174e-01 -2.85033256e-01 8.64605725e-01 3.88959385e-02
7.62238801e-01 -6.64891124e-01 -2.24142343e-01 -9.36412394e-01
2.25379094e-01 -6.96915269e-01 3.93482685e-01 -6.18652999e-01
-8.56532753e-01 6.11369848e-01 3.44615579e-01 -1.08311927e+00
-5.61896384e-01 -8.26898634e-01 -4.12867904e-01 6.49277449e-01
-1.06901026e+00 -6.00946903e-01 1.72739148e-01 5.16231120e-01
2.06856012e-01 2.27535516e-01 7.83464372e-01 3.06537092e-01
-3.83846015e-01 7.95421422e-01 -6.80641904e-02 2.65961200e-01
-3.79852206e-01 -1.09715688e+00 4.69731927e-01 1.22153854e+00
-1.17654286e-01 9.00137007e-01 8.80567610e-01 -2.34179899e-01
-1.51296222e+00 -9.68727946e-01 9.46934760e-01 -2.40488708e-01
1.14220214e+00 -1.24735329e-02 -8.86911690e-01 1.19898498e+00
4.49810140e-02 -1.35146931e-01 5.61944544e-01 1.19599655e-01
-8.18656504e-01 -2.21428089e-02 -4.94062722e-01 6.89955890e-01
1.43720233e+00 -8.30163121e-01 -4.89268094e-01 -7.69051490e-03
1.16462266e+00 -2.31762946e-01 -1.07732737e+00 2.86892474e-01
1.05966842e+00 -9.33142543e-01 8.16210985e-01 -1.27778816e+00
6.80480957e-01 -3.32436502e-01 -5.24581373e-01 -4.03778166e-01
-6.98780775e-01 -1.10558844e+00 -7.91220188e-01 6.13205433e-01
3.84836793e-01 -4.24343735e-01 5.78363478e-01 5.27292073e-01
-2.97073126e-01 -8.15619588e-01 -9.24421668e-01 -1.48908627e+00
1.65145755e-01 -9.33942854e-01 5.71935594e-01 6.80444658e-01
4.02524292e-01 -1.13586329e-01 -5.34119070e-01 1.14083380e-01
4.22997683e-01 -1.56216606e-01 2.69694507e-01 -1.00780392e+00
-5.18718064e-01 -1.05787349e+00 -7.21523106e-01 -1.75464582e+00
1.84221059e-01 -6.07653737e-01 -5.81034236e-02 -1.08513451e+00
-2.59113256e-02 -4.43937667e-02 -5.87556124e-01 2.79327273e-01
5.20262182e-01 5.44999167e-02 -3.86976707e-03 1.15499772e-01
-4.55649614e-01 2.58448154e-01 6.51792049e-01 5.09058908e-02
2.13848665e-01 1.35265753e-01 -5.27311802e-01 5.86513042e-01
6.58119023e-01 -3.36415738e-01 -4.50977296e-01 -4.61621106e-01
9.52944517e-01 2.07074419e-01 4.20911431e-01 -9.92786646e-01
4.04572785e-01 1.96206309e-02 -3.57879363e-02 -7.08011627e-01
1.88710794e-01 -7.15787768e-01 2.34106064e-01 1.00741160e+00
-9.61872518e-01 5.39701045e-01 5.12908518e-01 8.23751330e-01
7.79478103e-02 -2.27832109e-01 7.93394983e-01 2.32603084e-02
-6.36822820e-01 4.57181305e-01 -8.23176742e-01 -1.84550107e-01
6.77105367e-01 -1.11015312e-01 -5.06377593e-02 -6.94902241e-01
-8.10797095e-01 -3.76718561e-03 1.00101400e-02 5.38983457e-02
5.14444947e-01 -1.63728893e+00 -3.95787954e-01 -9.36673209e-02
-8.82223696e-02 -3.94235104e-01 4.45316210e-02 1.09967923e+00
-7.21237361e-01 6.68627858e-01 -7.66482949e-02 -2.70121545e-01
-7.20362604e-01 8.98783922e-01 6.59625292e-01 -2.58862406e-01
-6.14820123e-01 7.10210085e-01 -1.56492845e-03 3.58557343e-01
4.89998877e-01 -7.29779124e-01 1.57264918e-01 -2.98329681e-01
9.30691302e-01 4.36314374e-01 -1.24625616e-01 -2.23723739e-01
-5.24064422e-01 3.23324442e-01 1.75753802e-01 -3.12388450e-01
1.39339960e+00 -8.90897438e-02 -2.86714166e-01 8.63726735e-01
1.26428747e+00 -7.12833166e-01 -9.45207238e-01 -5.93656123e-01
1.01402275e-01 -6.94883466e-02 -3.43639493e-01 -1.93811819e-01
-7.41149902e-01 8.25259686e-01 1.48315191e-01 7.09330797e-01
1.34316194e+00 1.10658020e-01 6.34879947e-01 5.96239805e-01
1.46561228e-02 -7.57284582e-01 -6.94891587e-02 9.95715022e-01
6.80186689e-01 -7.71907568e-02 -3.07708502e-01 -3.41743112e-01
4.45773117e-02 1.48531699e+00 1.49464846e-01 -3.16263109e-01
9.80615020e-01 5.40641665e-01 -6.91855133e-01 -1.03409458e-02
-1.35525298e+00 -2.80832440e-01 9.02274773e-02 5.27415216e-01
5.33556163e-01 -2.25962192e-01 -2.07610965e-01 5.89130998e-01
-3.26560587e-01 1.85194388e-01 4.93205220e-01 7.61723757e-01
-2.55545706e-01 -6.46573424e-01 2.11671203e-01 3.11318133e-02
-4.95582342e-01 -3.14801365e-01 -1.65754274e-01 7.98223019e-01
-4.10169840e-01 7.52045691e-01 5.02779067e-01 -5.85941315e-01
1.04552016e-01 1.71361431e-01 9.18834150e-01 -1.66487675e-02
-3.59956384e-01 -1.47979781e-01 1.92442685e-01 -7.98267424e-01
-7.25634098e-01 -3.85024041e-01 -9.14021492e-01 -8.37615907e-01
-1.13267496e-01 1.63717896e-01 2.73297340e-01 1.20093834e+00
8.81501511e-02 6.23066664e-01 5.45560718e-01 -4.51295942e-01
-4.41432655e-01 -8.28163385e-01 -4.56329197e-01 -1.02614649e-01
7.00393081e-01 -2.06228048e-01 -4.17044610e-01 1.70720041e-01] | [7.666269779205322, 3.5000860691070557] |
49806351-763c-4a04-9f5b-777d0aabdf5e | video-object-segmentation-in-panoptic-wild | 2305.04470 | null | https://arxiv.org/abs/2305.04470v2 | https://arxiv.org/pdf/2305.04470v2.pdf | Video Object Segmentation in Panoptic Wild Scenes | In this paper, we introduce semi-supervised video object segmentation (VOS) to panoptic wild scenes and present a large-scale benchmark as well as a baseline method for it. Previous benchmarks for VOS with sparse annotations are not sufficient to train or evaluate a model that needs to process all possible objects in real-world scenarios. Our new benchmark (VIPOSeg) contains exhaustive object annotations and covers various real-world object categories which are carefully divided into subsets of thing/stuff and seen/unseen classes for comprehensive evaluation. Considering the challenges in panoptic VOS, we propose a strong baseline method named panoptic object association with transformers (PAOT), which uses panoptic identification to associate objects with a pyramid architecture on multiple scales. Experimental results show that VIPOSeg can not only boost the performance of VOS models by panoptic training but also evaluate them comprehensively in panoptic scenes. Previous methods for classic VOS still need to improve in performance and efficiency when dealing with panoptic scenes, while our PAOT achieves SOTA performance with good efficiency on VIPOSeg and previous VOS benchmarks. PAOT also ranks 1st in the VOT2022 challenge. Our dataset is available at https://github.com/yoxu515/VIPOSeg-Benchmark. | ['Yi Yang', 'Zongxin Yang', 'Yuanyou Xu'] | 2023-05-08 | null | null | null | null | ['semi-supervised-video-object-segmentation', 'video-object-segmentation', 'video-semantic-segmentation'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 1.02499267e-02 -5.78708649e-01 -4.67208982e-01 -2.32120335e-01
-6.63587630e-01 -7.25170672e-01 4.60995644e-01 -2.34099850e-01
-2.13181287e-01 3.84207070e-01 -7.00581223e-02 -9.43000987e-02
1.98368058e-01 -6.72426701e-01 -8.96856785e-01 -6.10882103e-01
-8.79190862e-02 6.26784503e-01 8.91872108e-01 -1.07502356e-01
-2.12017417e-01 1.53912544e-01 -1.64708531e+00 5.64077020e-01
7.10993886e-01 1.34243870e+00 4.24091309e-01 8.52754354e-01
2.41890848e-02 6.30573571e-01 -6.33189261e-01 -2.28932053e-01
9.55872774e-01 -1.09800898e-01 -9.66759920e-01 4.02309477e-01
1.41959357e+00 -3.71762693e-01 -2.38707200e-01 1.14863789e+00
1.72798410e-01 6.04259968e-02 5.86040616e-01 -1.43319190e+00
-3.76533717e-01 5.10659099e-01 -7.42978334e-01 8.09879780e-01
-2.60444164e-01 7.92385101e-01 1.55236268e+00 -8.70233834e-01
7.54181683e-01 1.34577513e+00 7.04277694e-01 3.77151191e-01
-1.01896501e+00 -7.62862146e-01 4.46403623e-01 2.41751298e-01
-1.27580976e+00 -2.81659931e-01 2.51874059e-01 -5.83407164e-01
8.87621820e-01 5.52233338e-01 1.03046751e+00 8.90292048e-01
-2.17035279e-01 1.45147526e+00 1.01815546e+00 2.92483270e-01
-1.20540790e-01 1.13438517e-02 2.30203047e-01 4.76744622e-01
2.80087113e-01 -4.26858515e-02 -1.46831319e-01 2.39680022e-01
7.17796743e-01 -8.88496786e-02 -3.56936216e-01 -2.35297814e-01
-1.45435703e+00 6.20710015e-01 7.22627997e-01 1.12534054e-01
-3.19950193e-01 2.69704074e-01 4.68676180e-01 7.67612830e-02
5.61795354e-01 6.27236187e-01 -8.69959831e-01 2.30699971e-01
-1.26757240e+00 3.14273238e-01 7.33123362e-01 1.15480947e+00
6.16858482e-01 2.08414748e-01 -4.01393294e-01 9.87970710e-01
2.12876216e-01 6.05501235e-01 8.43843296e-02 -1.10896170e+00
7.97878385e-01 5.08439779e-01 6.47984967e-02 -5.55658400e-01
-3.06249946e-01 -3.68900716e-01 -7.03429699e-01 7.97245800e-02
5.50264716e-01 -4.91847657e-02 -1.56833303e+00 1.24819100e+00
4.82381493e-01 5.56035757e-01 -3.48410010e-02 1.14274848e+00
1.41410148e+00 1.17617643e+00 2.90354103e-01 -2.73795016e-02
1.57208800e+00 -1.71871293e+00 -4.13932502e-01 -4.93450016e-01
2.11083978e-01 -8.77847612e-01 1.25801516e+00 2.54415870e-01
-7.92649209e-01 -7.25799322e-01 -8.58267963e-01 1.02905735e-01
-4.13992643e-01 9.25791711e-02 7.46878624e-01 3.69778842e-01
-1.14149487e+00 3.06123763e-01 -5.64765453e-01 -5.92229724e-01
7.44819582e-01 2.47254044e-01 -4.74462807e-02 6.21451698e-02
-1.15341413e+00 3.56196463e-01 8.48829806e-01 8.13630526e-04
-1.45488203e+00 -9.06703770e-01 -7.93881357e-01 2.27759778e-02
8.92859101e-01 -7.00744629e-01 1.26769650e+00 -8.64764273e-01
-9.90386188e-01 9.22295928e-01 2.81324983e-01 -7.87834823e-01
4.45022702e-01 -4.20699209e-01 -3.88945222e-01 4.96031463e-01
3.62972826e-01 1.39327824e+00 9.42801774e-01 -1.25936675e+00
-1.11397123e+00 9.74280164e-02 2.09477305e-01 3.06844890e-01
5.44771440e-02 1.64661065e-01 -1.15377581e+00 -8.26423883e-01
-6.11583814e-02 -1.03992772e+00 -4.29986417e-01 -1.18686378e-01
-6.14659727e-01 -3.61588389e-01 1.04924572e+00 -3.46961498e-01
9.75568175e-01 -1.89930689e+00 -1.89589709e-01 -2.00242713e-01
2.30962217e-01 6.13505244e-01 -3.43209237e-01 -1.56553686e-01
9.40521508e-02 2.46635213e-01 -4.55760181e-01 -2.80996770e-01
-1.77049935e-01 3.70684177e-01 -6.07451618e-01 1.10402420e-01
2.69806534e-01 1.14574647e+00 -8.65267158e-01 -7.93526828e-01
3.81767899e-01 1.92389376e-02 -6.09777570e-01 2.63729930e-01
-8.29310060e-01 4.30466652e-01 -3.63674194e-01 1.32352448e+00
7.54798651e-01 -5.33510566e-01 -2.51958400e-01 -3.59292209e-01
-1.17222279e-01 1.74358740e-01 -1.11424708e+00 1.27098465e+00
-3.80847268e-02 7.57651806e-01 1.38360308e-02 -7.07501054e-01
4.75158393e-01 1.59695297e-01 5.94832897e-01 -5.03961563e-01
2.59173028e-02 2.94897735e-01 -2.26731956e-01 -6.50654137e-01
8.14611256e-01 2.36327171e-01 1.23666460e-02 -1.72552675e-01
4.26470578e-01 -5.70714712e-01 7.63224304e-01 2.58991361e-01
5.96181154e-01 2.52547175e-01 2.47610196e-01 -4.51308966e-01
2.54274249e-01 4.92043525e-01 8.31948340e-01 9.43216562e-01
-4.45475698e-01 9.75400448e-01 2.87746519e-01 -4.77901042e-01
-9.61611569e-01 -1.03938138e+00 -3.84268016e-01 1.11212337e+00
6.58498585e-01 -5.09333432e-01 -6.51585460e-01 -1.10724950e+00
-3.25370543e-02 3.67781341e-01 -6.18852317e-01 5.98327458e-01
-6.30099118e-01 -1.37421310e+00 5.34085572e-01 7.02801108e-01
8.62019837e-01 -1.35765100e+00 -3.99715960e-01 1.36252671e-01
-4.98899758e-01 -1.70776808e+00 -7.44444966e-01 5.81501536e-02
-7.73681402e-01 -1.27386081e+00 -8.50196958e-01 -7.21263230e-01
1.46328747e-01 5.09385467e-01 1.36108494e+00 -1.33004189e-01
-3.32015634e-01 3.63046169e-01 -3.40622872e-01 -4.08232659e-01
-1.29567266e-01 1.49038434e-01 -6.69754595e-02 1.67475104e-01
1.66333452e-01 -3.30168724e-01 -9.34581161e-01 7.29901195e-01
-9.20953870e-01 1.73031032e-01 5.39279342e-01 3.58421266e-01
9.15462911e-01 -2.14449167e-01 2.56862789e-01 -9.19185340e-01
-4.45524931e-01 -7.04656303e-01 -9.10839677e-01 2.39317745e-01
-3.71337265e-01 -4.79595959e-01 2.65009850e-01 -4.73722875e-01
-8.64847541e-01 9.70813259e-03 -3.03580791e-01 -6.44066811e-01
-3.17492813e-01 8.67790878e-02 -1.98147058e-01 5.86470142e-02
4.40676868e-01 1.40842691e-01 -7.56995082e-01 -7.55288601e-01
3.11729819e-01 4.18705225e-01 7.23864317e-01 -4.56258595e-01
8.53840172e-01 7.34310389e-01 -2.92934388e-01 -8.37744772e-01
-1.41210115e+00 -1.14245260e+00 -5.39637506e-01 -2.26879850e-01
1.15182078e+00 -1.34659994e+00 -3.71501476e-01 5.81389368e-01
-9.00977552e-01 -7.34101474e-01 -5.01505315e-01 3.75724167e-01
-4.61288154e-01 2.92353064e-01 -6.93436801e-01 -4.34842795e-01
-3.95176977e-01 -1.25072670e+00 1.24293828e+00 2.63836294e-01
2.08934456e-01 -6.95649862e-01 -2.80610216e-03 7.27069795e-01
1.30424276e-01 2.09608629e-01 1.89701259e-01 -6.19112968e-01
-1.22706819e+00 3.37239385e-01 -6.01682603e-01 6.78849041e-01
-1.21904701e-01 7.92383179e-02 -9.52111363e-01 -1.08900793e-01
-1.48920104e-01 -4.66054708e-01 1.71556377e+00 5.47935784e-01
1.16285110e+00 -3.60053301e-01 -3.40279877e-01 1.18718243e+00
1.43240368e+00 2.39498243e-01 5.40687859e-01 5.24148345e-01
1.29532433e+00 3.84112090e-01 1.02210307e+00 1.21757714e-02
4.39948887e-01 7.37223089e-01 7.03729510e-01 -1.89361021e-01
-6.35210752e-01 -7.97552615e-02 4.54422593e-01 6.37129724e-01
-2.10535303e-01 -6.99799836e-01 -9.16648507e-01 1.05415499e+00
-1.84116316e+00 -8.61331224e-01 -4.99318808e-01 1.72289479e+00
8.38596046e-01 1.44023716e-01 2.55314022e-01 -2.02879712e-01
6.73136890e-01 6.22030377e-01 -4.41230029e-01 2.39041626e-01
-5.20128846e-01 -1.60064045e-02 7.96069860e-01 1.35808706e-01
-2.00507474e+00 1.41495371e+00 5.79640579e+00 1.19318104e+00
-1.13060427e+00 3.27170074e-01 8.88852358e-01 -9.13823768e-02
2.76243448e-01 6.63122311e-02 -1.44022524e+00 4.34047431e-01
5.94235361e-01 2.12707058e-01 -4.69817556e-02 1.05169606e+00
-1.16713926e-01 -7.86656588e-02 -8.43261123e-01 9.02167201e-01
-4.93311062e-02 -1.48386836e+00 4.18163419e-01 -2.55751073e-01
1.37406313e+00 8.16161215e-01 8.02787319e-02 4.85549003e-01
3.05477232e-01 -7.80964434e-01 8.71699750e-01 -1.04437232e-01
6.28795087e-01 -7.55462945e-02 6.78815007e-01 -4.73528467e-02
-1.67777979e+00 -2.73614936e-02 -3.97056013e-01 3.93268973e-01
2.74174184e-01 1.45146474e-01 -7.46918559e-01 5.46077967e-01
1.29328728e+00 1.08799314e+00 -8.00267458e-01 1.71669686e+00
-1.75761640e-01 1.20600438e+00 -8.01649749e-01 4.06732023e-01
8.97245228e-01 -1.69778585e-01 9.85271394e-01 1.57156658e+00
1.30870566e-01 -3.50118876e-02 5.45813084e-01 4.85084295e-01
-2.68884510e-01 2.91078657e-01 -2.45603174e-01 1.12234257e-01
-2.05366910e-02 1.41366160e+00 -1.10442698e+00 -8.84627044e-01
-3.33506197e-01 6.36700988e-01 -2.98512459e-01 4.52277839e-01
-1.02229679e+00 1.36611223e-01 8.12035918e-01 3.31465989e-01
8.24841857e-01 9.71086398e-02 -9.65379402e-02 -1.43723571e+00
-1.14898711e-01 -7.73294091e-01 9.29850161e-01 -7.95874596e-01
-1.28728974e+00 8.87608826e-01 3.72055292e-01 -1.55093455e+00
4.37533736e-01 -5.64595640e-01 -5.87462246e-01 1.05635963e-01
-1.82068837e+00 -1.38027751e+00 -5.24006009e-01 5.64558506e-01
1.25203633e+00 -4.26476961e-03 8.02093223e-02 5.55140793e-01
-5.62432289e-01 1.50411800e-01 -9.36173946e-02 2.38130823e-01
6.28535092e-01 -1.48362124e+00 6.67812943e-01 9.52708721e-01
6.35732114e-01 -5.33720404e-02 6.11067414e-01 -6.56767666e-01
-7.68786371e-01 -1.84619784e+00 3.70529354e-01 -7.74263561e-01
8.87253404e-01 -3.22604567e-01 -1.00808239e+00 9.35155511e-01
2.94526070e-01 4.53118265e-01 6.43515289e-02 -2.62519568e-01
-3.91124338e-01 -2.27395654e-01 -6.81214690e-01 5.20702839e-01
1.26240623e+00 -3.55013236e-02 -6.25238836e-01 7.87783325e-01
1.17445600e+00 -5.20151079e-01 -6.76927030e-01 9.21328545e-01
7.96506032e-02 -6.99787080e-01 1.22537446e+00 -4.26025152e-01
2.74222881e-01 -7.19672740e-01 -2.16522560e-01 -8.77473176e-01
8.94425884e-02 -7.33039021e-01 -2.98544653e-02 1.19916558e+00
3.65074962e-01 -3.41626406e-01 7.09357083e-01 -1.81092054e-01
-3.68843019e-01 -6.11213386e-01 -7.12246180e-01 -1.17331004e+00
-2.39255667e-01 -7.20635295e-01 4.66242611e-01 9.39693809e-01
-9.08776879e-01 3.46831471e-01 -6.11828804e-01 4.14391488e-01
7.84636080e-01 5.08667886e-01 9.01145160e-01 -1.03812361e+00
-3.72215211e-01 -5.34375429e-01 -3.87726456e-01 -1.61532199e+00
-3.67120355e-01 -9.46332633e-01 2.92541325e-01 -1.63647544e+00
3.74604315e-01 -5.83673298e-01 -4.99523878e-01 5.31615376e-01
-4.26985711e-01 1.07778430e+00 5.35804093e-01 6.70661092e-01
-1.24142158e+00 5.56390524e-01 1.43071508e+00 -4.92705941e-01
-3.32324445e-01 2.61279881e-01 -3.28865409e-01 1.06669188e+00
7.67780840e-01 -5.69123030e-01 -2.55126059e-01 -4.49569494e-01
-1.03146829e-01 -3.07743728e-01 6.75939202e-01 -1.18969381e+00
-1.04622003e-02 -2.17762589e-01 1.25892997e-01 -1.03698862e+00
4.44144547e-01 -6.07820034e-01 -6.12403825e-02 2.29564518e-01
2.02524915e-01 -2.38533542e-01 4.46060359e-01 5.37957609e-01
-5.30767202e-01 -5.24148569e-02 8.52545917e-01 -1.64324388e-01
-1.48536146e+00 9.65729475e-01 -1.01168998e-01 4.82046038e-01
1.05781376e+00 -6.66687414e-02 -4.85818624e-01 7.40797147e-02
-8.75868857e-01 7.63648808e-01 1.50447264e-01 7.24676609e-01
4.14349943e-01 -9.43266213e-01 -7.92747319e-01 -2.02835545e-01
3.49561930e-01 3.35341841e-01 4.69526112e-01 1.02385283e+00
-9.36180770e-01 5.58336616e-01 -2.07325980e-01 -1.04211426e+00
-1.46405768e+00 6.08311713e-01 2.83584237e-01 -3.16995174e-01
-8.06180000e-01 1.05190277e+00 1.06276214e+00 -2.37499595e-01
2.45482057e-01 -7.22985864e-01 -4.84857500e-01 2.86597818e-01
4.21861261e-01 -2.89229658e-02 -3.27586263e-01 -6.91192150e-01
-2.77319610e-01 7.19329357e-01 1.71219930e-02 3.00455123e-01
1.38459492e+00 -1.05545893e-01 9.07570422e-02 4.69853044e-01
8.91206861e-01 -1.53178334e-01 -1.72611547e+00 -5.04878342e-01
-4.43116613e-02 -4.46540236e-01 -1.67666957e-01 -7.21827686e-01
-1.56802654e+00 8.53286028e-01 5.16898036e-01 1.55657172e-01
1.08787274e+00 4.58406597e-01 1.05002594e+00 4.00334924e-01
6.00156710e-02 -8.76306117e-01 2.01219261e-01 5.49849927e-01
9.04555798e-01 -1.58931732e+00 3.38983536e-02 -8.46096873e-01
-9.68859673e-01 7.08935976e-01 1.01913357e+00 -5.98811619e-02
5.95581114e-01 -1.34302711e-03 2.64419109e-01 -3.76035541e-01
-7.63235748e-01 -7.75054455e-01 7.37423778e-01 1.99008629e-01
-3.05463105e-01 8.86882283e-03 1.65439889e-01 2.40433797e-01
-9.48027000e-02 -2.82211244e-01 1.01857826e-01 4.08847064e-01
-4.96817440e-01 -5.73714674e-01 -3.43923628e-01 4.70263302e-01
-6.32985294e-01 -4.12615955e-01 -1.08600765e-01 1.10152507e+00
5.39403498e-01 5.78612328e-01 2.69980997e-01 -1.39663622e-01
1.82875067e-01 -3.03592145e-01 2.06107855e-01 -6.88884556e-01
-7.76177526e-01 3.88316095e-01 2.37682119e-01 -5.82374513e-01
-7.83005893e-01 -6.53978705e-01 -9.29681420e-01 1.46332383e-01
-1.62378386e-01 4.06598598e-02 3.17664534e-01 7.61572123e-01
2.98680067e-02 5.23863077e-01 2.60737449e-01 -8.99518788e-01
8.92972574e-03 -9.61222887e-01 -3.00216168e-01 5.09493232e-01
2.77328849e-01 -5.46806276e-01 -2.38354787e-01 3.71362597e-01] | [9.289186477661133, 0.10923925787210464] |
efded397-3645-4c39-9973-1ed1fc0700c3 | class-interference-regularization | 2009.02396 | null | https://arxiv.org/abs/2009.02396v1 | https://arxiv.org/pdf/2009.02396v1.pdf | Class Interference Regularization | Contrastive losses yield state-of-the-art performance for person re-identification, face verification and few shot learning. They have recently outperformed the cross-entropy loss on classification at the ImageNet scale and outperformed all self-supervision prior results by a large margin (SimCLR). Simple and effective regularization techniques such as label smoothing and self-distillation do not apply anymore, because they act on multinomial label distributions, adopted in cross-entropy losses, and not on tuple comparative terms, which characterize the contrastive losses. Here we propose a novel, simple and effective regularization technique, the Class Interference Regularization (CIR), which applies to cross-entropy losses but is especially effective on contrastive losses. CIR perturbs the output features by randomly moving them towards the average embeddings of the negative classes. To the best of our knowledge, CIR is the first regularization technique to act on the output features. In experimental evaluation, the combination of CIR and a plain Siamese-net with triplet loss yields best few-shot learning performance on the challenging tieredImageNet. CIR also improves the state-of-the-art technique in person re-identification on the Market-1501 dataset, based on triplet loss, and the state-of-the-art technique in person search on the CUHK-SYSU dataset, based on a cross-entropy loss. Finally, on the task of classification CIR performs on par with the popular label smoothing, as demonstrated for CIFAR-10 and -100. | ['Sikandar Amin', 'Fabio Galasso', 'Bharti Munjal'] | 2020-09-04 | null | null | null | null | ['person-search'] | ['computer-vision'] | [-1.74398944e-02 -8.12158734e-02 -1.55041501e-01 -2.70783573e-01
-5.77544749e-01 -1.67627886e-01 7.95095086e-01 7.64103830e-02
-9.07579243e-01 8.13806057e-01 7.14974999e-02 2.41529897e-01
-3.17281634e-01 -6.27749801e-01 -6.39227688e-01 -7.01899052e-01
-1.45262182e-01 6.50375783e-01 -5.02835289e-02 -2.02733099e-01
-2.33883426e-01 2.80434966e-01 -1.69388771e+00 -1.55464783e-01
7.37658143e-01 1.19007289e+00 -4.48350221e-01 1.91518202e-01
1.97407931e-01 7.11652935e-01 -3.61423552e-01 -1.20085156e+00
3.93178672e-01 -2.56333262e-01 -6.43724799e-01 -1.81616127e-01
8.06586742e-01 -1.66874439e-01 -4.13771987e-01 1.12652445e+00
1.04523337e+00 3.25271964e-01 9.21366692e-01 -1.13492715e+00
-7.66166866e-01 5.77590704e-01 -5.85296631e-01 9.77123603e-02
3.11885655e-01 1.79116517e-01 9.73390341e-01 -9.18431401e-01
5.43925405e-01 1.38196266e+00 1.08849382e+00 9.29455161e-01
-1.48549092e+00 -8.62949312e-01 -1.13011347e-02 4.02506143e-01
-1.53256357e+00 -4.25125182e-01 4.93688077e-01 -4.62853432e-01
9.30877566e-01 2.20379114e-01 2.54572332e-01 1.37713599e+00
-4.54003423e-01 6.83732569e-01 1.07516122e+00 -5.76343000e-01
1.10650264e-01 6.09245062e-01 4.85671014e-01 6.77576005e-01
1.59115866e-01 4.17367488e-01 -5.58284819e-01 -1.93744212e-01
2.10592017e-01 -3.43700908e-02 -3.45022142e-01 -4.92362469e-01
-6.37765646e-01 8.62758636e-01 3.84085983e-01 4.28324372e-01
-1.07289821e-01 1.05038539e-01 7.05945253e-01 3.22459072e-01
6.33481383e-01 1.71811685e-01 -2.88293809e-01 -1.02549279e-02
-9.96030569e-01 2.45746925e-01 8.46382856e-01 5.09026706e-01
7.50881076e-01 5.47793470e-02 -8.02010715e-01 1.15954947e+00
1.43969692e-02 4.16506857e-01 5.36252499e-01 -5.85007429e-01
3.39002311e-01 3.16068918e-01 -2.13027269e-01 -5.36779642e-01
-3.31470162e-01 -8.16375315e-01 -1.00924838e+00 4.15103227e-01
5.90718567e-01 -1.71572909e-01 -1.01323092e+00 2.05125237e+00
1.40553787e-01 6.06925368e-01 -9.91790593e-02 7.90255904e-01
7.10471690e-01 1.00254685e-01 3.23603094e-01 -1.20918706e-01
1.46389031e+00 -8.28251302e-01 -5.85673153e-01 9.67402905e-02
6.66122317e-01 -2.61076927e-01 8.69998455e-01 2.57930011e-01
-1.00350511e+00 -6.86496377e-01 -9.56307471e-01 1.52698413e-01
-6.80372834e-01 6.19218834e-02 4.31993335e-01 1.17698014e+00
-1.10463035e+00 1.05054653e+00 -3.73652428e-01 -4.62832034e-01
7.73198724e-01 4.98617291e-01 -5.73716581e-01 -2.08788160e-02
-1.45925283e+00 1.09455681e+00 2.22868800e-01 -2.08454520e-01
-6.08562171e-01 -1.31070948e+00 -1.03938937e+00 3.80716980e-01
3.21633518e-01 -7.66198933e-01 8.35274875e-01 -8.52730036e-01
-1.57318747e+00 1.17544687e+00 1.41586572e-01 -1.19440985e+00
9.63235617e-01 -5.47343567e-02 -4.81220514e-01 2.89193150e-02
-2.90458165e-02 7.37246811e-01 8.16860259e-01 -9.84564364e-01
-3.72326940e-01 -4.86091346e-01 -2.95840591e-01 -6.81482479e-02
-6.92173779e-01 -3.48690115e-02 -3.00964504e-01 -6.40444934e-01
-7.84238279e-01 -8.89880598e-01 1.40560448e-01 -6.58300370e-02
-2.35104844e-01 -4.61966306e-01 6.90312922e-01 -7.41927922e-01
1.19982016e+00 -2.20784259e+00 6.73068911e-02 3.33815217e-01
-2.64089014e-02 7.43538320e-01 -9.46740434e-02 6.26038462e-02
-4.21647489e-01 4.26311232e-02 -3.98319751e-01 -9.56142724e-01
4.53109801e-01 -1.84333652e-01 -2.05531299e-01 7.00829804e-01
-4.60489877e-02 1.03622389e+00 -6.99084520e-01 -3.71557891e-01
4.39105511e-01 8.12643588e-01 -5.50001442e-01 -9.67102945e-02
1.81097716e-01 9.83915180e-02 1.99494556e-01 2.72521675e-01
7.56470144e-01 -4.39418294e-02 -1.82878882e-01 -2.62684077e-01
1.98985428e-01 -3.67634714e-01 -1.03253996e+00 1.57407963e+00
-3.11672270e-01 3.40358883e-01 -2.90374756e-01 -1.03047693e+00
7.28979170e-01 3.07498187e-01 4.78727579e-01 -7.55302906e-01
1.03088796e-01 1.49741188e-01 -1.42411873e-01 -7.42383897e-02
1.18010446e-01 -1.60170257e-01 -1.58827677e-02 2.29570180e-01
6.18469298e-01 6.64259315e-01 3.50417167e-01 1.65221184e-01
8.42536926e-01 -7.13681653e-02 5.21756150e-02 -2.67376959e-01
8.26745510e-01 -7.28932500e-01 4.83734220e-01 8.83109212e-01
-3.63914311e-01 5.92440367e-01 3.86308938e-01 -2.01080620e-01
-9.80368257e-01 -1.00234520e+00 -4.57944125e-01 1.08164811e+00
-3.40564325e-02 -3.59631270e-01 -9.27706003e-01 -8.86003375e-01
4.75728899e-01 8.48064780e-01 -8.59954953e-01 -3.06974202e-01
-2.67673105e-01 -9.36268687e-01 7.46067047e-01 3.59113872e-01
7.50317335e-01 -8.46415818e-01 4.55729887e-02 -2.53058970e-02
1.55466303e-01 -1.20174551e+00 -5.60435176e-01 1.40631765e-01
-2.49307141e-01 -9.65367734e-01 -1.24539685e+00 -6.44828677e-01
3.34467769e-01 -3.27579737e-01 9.99356210e-01 -9.31375772e-02
-5.53074956e-01 5.67222893e-01 -6.34536147e-02 -2.15263054e-01
-2.21420452e-03 1.43339157e-01 3.36727977e-01 4.44257200e-01
6.03187561e-01 -6.00460708e-01 -6.61055267e-01 2.84452498e-01
-4.76625264e-01 -5.92972755e-01 2.21337959e-01 1.17028069e+00
3.16822201e-01 -2.76734978e-02 5.97298503e-01 -9.16965187e-01
4.76094633e-01 -1.95114896e-01 -3.30948740e-01 4.99457538e-01
-8.97758067e-01 1.30557064e-02 6.25915349e-01 -4.79177088e-01
-1.08562136e+00 -1.99729621e-01 -2.01357737e-01 -7.72057891e-01
-1.50178418e-01 -1.35090470e-01 1.45633677e-02 -4.28037107e-01
7.37731636e-01 7.51859471e-02 2.75225677e-02 -6.43140197e-01
3.48378628e-01 6.40360177e-01 5.06647050e-01 -1.92625746e-01
7.24618793e-01 5.62226295e-01 1.44883066e-01 -7.53125429e-01
-8.32389891e-01 -6.61169946e-01 -3.33502114e-01 6.18957691e-02
9.22661185e-01 -9.96002913e-01 -1.08545470e+00 8.39372933e-01
-7.02602983e-01 -1.04718104e-01 -7.01566815e-01 3.67403477e-01
-4.18089688e-01 5.88523328e-01 -6.33530080e-01 -8.96071970e-01
-6.48400009e-01 -7.74666131e-01 8.75077546e-01 3.66156608e-01
-1.58765465e-02 -1.00713968e+00 2.50717308e-02 3.39049667e-01
4.22733814e-01 3.25357318e-02 5.30479252e-01 -8.79551411e-01
2.40105148e-02 -3.56562585e-01 -4.46565390e-01 5.60371935e-01
-1.51846737e-01 -6.11127675e-01 -1.37212765e+00 -6.51643455e-01
-3.09335500e-01 -4.66118813e-01 1.54964900e+00 3.58600616e-01
1.17820346e+00 -2.54169535e-02 -2.24379122e-01 7.61821687e-01
1.43602061e+00 -3.81617844e-01 7.44498670e-01 1.66849777e-01
7.23686516e-01 5.96189618e-01 7.73947984e-02 5.15389621e-01
2.93676198e-01 1.07595909e+00 1.96227416e-01 -7.33497068e-02
-4.41695541e-01 -2.82993138e-01 1.81749269e-01 1.02093339e-01
-2.03788012e-01 -4.95172292e-02 -6.28189266e-01 3.19818467e-01
-1.96402717e+00 -1.39269972e+00 2.51172811e-01 2.52819490e+00
6.60939276e-01 4.66580652e-02 4.51939046e-01 8.07417929e-02
8.81947815e-01 1.82872176e-01 -4.54617918e-01 -2.29660422e-01
-3.05114746e-01 4.61308330e-01 7.93068945e-01 5.75253904e-01
-1.52375448e+00 9.71223414e-01 5.09798813e+00 1.47464943e+00
-8.80868971e-01 5.98600388e-01 7.59267330e-01 -3.46241862e-01
1.81533039e-01 -3.41758400e-01 -1.32884312e+00 9.26292717e-01
9.48965192e-01 -2.33821366e-02 4.54928964e-01 7.86470234e-01
-2.14564070e-01 1.14786059e-01 -1.18203759e+00 1.57388020e+00
6.06121778e-01 -1.10600567e+00 -1.31752357e-01 1.51934654e-01
7.78689802e-01 -7.37848952e-02 4.11610305e-01 7.24653840e-01
-4.35491465e-02 -1.07560420e+00 4.46813464e-01 6.05078816e-01
9.72430468e-01 -8.63544047e-01 9.27195668e-01 4.38008867e-02
-9.88977194e-01 -4.47822839e-01 -4.53761458e-01 2.76575804e-01
3.28170270e-01 5.59573233e-01 -3.58093739e-01 5.56560636e-01
7.06445038e-01 7.78396726e-01 -7.02256262e-01 1.15923774e+00
1.04843363e-01 3.96348655e-01 -2.42958143e-01 2.61654586e-01
1.92597322e-02 -2.03177303e-01 5.91234386e-01 1.42109334e+00
2.87370719e-02 -2.69754529e-01 -4.46720719e-02 8.26263249e-01
-3.68044406e-01 9.79439542e-02 -3.17562282e-01 3.05336326e-01
1.37644662e-02 1.15076160e+00 -1.47163704e-01 -4.13047433e-01
-2.68094212e-01 1.16271889e+00 3.94176215e-01 4.93242323e-01
-8.65023613e-01 -4.13967520e-01 6.71911061e-01 2.13642821e-01
6.26876593e-01 5.01016676e-01 -3.99967097e-02 -1.25462592e+00
5.72962761e-02 -6.87654376e-01 6.34003043e-01 -8.60563386e-03
-1.88754737e+00 4.66377378e-01 -9.21240300e-02 -1.21619236e+00
-3.06695282e-01 -6.45463645e-01 -4.77104008e-01 9.71018910e-01
-1.81274140e+00 -1.28110695e+00 -1.04255408e-01 7.49903321e-01
1.19918391e-01 -7.05275476e-01 8.88640642e-01 7.79409766e-01
-7.71940887e-01 1.32081068e+00 1.71921682e-02 2.71235824e-01
8.72199833e-01 -1.24159110e+00 1.84000954e-01 6.23452127e-01
6.81551695e-02 3.33233148e-01 5.21954298e-01 -4.42489386e-01
-7.06951857e-01 -9.50090885e-01 8.31675708e-01 -2.33795270e-01
5.39245486e-01 -5.12274086e-01 -7.61508942e-01 3.51360679e-01
9.42399651e-02 3.41008484e-01 6.38562500e-01 2.60713160e-01
-6.41835392e-01 -4.05810446e-01 -1.47558343e+00 3.44601989e-01
1.23735845e+00 -8.09363067e-01 -3.35084468e-01 4.39010024e-01
3.68949831e-01 8.58339667e-02 -7.81165898e-01 6.59285367e-01
5.87604463e-01 -1.03090107e+00 1.32622850e+00 -6.36631548e-01
-3.40398885e-02 5.29773608e-02 -2.89833434e-02 -1.10917056e+00
-4.38021362e-01 -5.93529999e-01 -3.11282992e-01 1.72715831e+00
2.77174979e-01 -7.80752480e-01 9.68411684e-01 7.25942016e-01
4.00707245e-01 -7.58553386e-01 -1.18370640e+00 -1.11665726e+00
5.94333559e-02 -1.77012533e-01 3.24460000e-01 8.51686537e-01
-2.41240695e-01 1.73209757e-01 -7.58739591e-01 -3.02734047e-01
1.24286258e+00 -4.61958647e-01 4.76283669e-01 -1.46544158e+00
-6.04259551e-01 -6.15043938e-01 -8.02341819e-01 -5.10666072e-01
7.47638583e-01 -1.09798658e+00 -5.12126565e-01 -9.07192409e-01
3.97789568e-01 -2.76907325e-01 -6.67532742e-01 4.32499141e-01
-7.79589713e-02 5.48727572e-01 5.01221895e-01 -1.47433402e-02
-6.08794272e-01 8.15764606e-01 5.89993298e-01 -5.00544429e-01
-6.21335320e-02 1.14382952e-01 -4.72722530e-01 5.77803552e-01
4.59195286e-01 -2.64694661e-01 -1.83067471e-01 1.01897165e-01
-2.26917028e-01 -4.99792278e-01 6.57742381e-01 -1.29213452e+00
3.54381174e-01 8.06669593e-01 5.26061535e-01 -1.98147282e-01
7.34738767e-01 -7.01762676e-01 4.07135393e-03 4.74496543e-01
-3.54358494e-01 -5.20632446e-01 3.65046063e-03 6.13828599e-01
-1.34874925e-01 -2.75020033e-01 1.11990213e+00 -4.79748771e-02
-6.91401124e-01 5.76287568e-01 4.01298732e-01 2.04372317e-01
1.04924572e+00 -3.69830877e-01 -3.72667015e-01 -2.08751857e-01
-8.87663543e-01 2.51054347e-01 2.54176646e-01 3.75115991e-01
2.56072938e-01 -1.41891825e+00 -8.97405803e-01 3.44441205e-01
4.37451191e-02 -8.88112545e-01 6.07944429e-01 9.50425506e-01
5.71317412e-02 4.42031503e-01 -2.25177780e-01 -5.07834792e-01
-1.39068127e+00 7.04108894e-01 6.95706248e-01 -5.78532517e-01
-5.01943290e-01 1.10248113e+00 2.79900849e-01 -4.03104037e-01
7.57748544e-01 3.95713598e-01 -3.35288852e-01 4.41671669e-01
6.73161864e-01 6.86283290e-01 2.30352521e-01 -7.43591905e-01
-4.73861098e-01 8.32210124e-01 -2.68560022e-01 -9.28677097e-02
1.20125365e+00 1.38631929e-02 1.13670647e-01 1.51346624e-01
1.55940056e+00 -3.21184188e-01 -1.18172812e+00 -3.91865611e-01
-9.13422033e-02 -3.69692147e-01 -6.79645734e-03 -9.35618699e-01
-1.05793250e+00 7.52811849e-01 1.27436388e+00 2.87976442e-03
8.32458019e-01 -1.19317621e-01 8.03241670e-01 1.99394356e-02
3.43410909e-01 -1.30398583e+00 -6.14816099e-02 2.17288345e-01
4.57717448e-01 -1.45918608e+00 -9.02274475e-02 -2.74639726e-01
-6.47268236e-01 5.10977507e-01 4.72461462e-01 -1.00228332e-01
8.99295032e-01 7.12179393e-03 -3.51773202e-01 2.17946038e-01
-3.72100264e-01 -5.72476149e-01 4.76022512e-01 6.56328380e-01
1.46651149e-01 5.64629398e-02 -3.13987911e-01 5.72767317e-01
7.70446435e-02 1.01790942e-01 -2.26936460e-01 1.14858128e-01
-7.17830285e-03 -1.20970225e+00 -1.17849171e-01 5.39329052e-01
-6.71756983e-01 -1.88844517e-01 -1.31349400e-01 5.67195833e-01
5.06213903e-01 7.44682014e-01 1.31506249e-01 -2.92381853e-01
4.23790187e-01 2.46734917e-01 5.40896475e-01 -2.98805177e-01
-9.40953672e-01 -3.82702440e-01 7.51361698e-02 -4.09360230e-01
-2.78053105e-01 -6.13336504e-01 -6.61010504e-01 -4.76879448e-01
-3.57785165e-01 1.22745529e-01 4.73997593e-01 8.75536680e-01
2.87860334e-01 1.68591216e-01 5.64934731e-01 -8.18805277e-01
-1.08617866e+00 -9.35750067e-01 -8.44975889e-01 9.49345946e-01
2.16834843e-01 -7.75952578e-01 -6.14741266e-01 -4.60599303e-01] | [14.74284553527832, 1.051376223564148] |
f9a14fb2-0828-44f7-b158-8ff274bb6f32 | few-shot-referring-relationships-in-videos | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Kumar_Few-Shot_Referring_Relationships_in_Videos_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Kumar_Few-Shot_Referring_Relationships_in_Videos_CVPR_2023_paper.pdf | Few-Shot Referring Relationships in Videos | Interpreting visual relationships is a core aspect of comprehensive video understanding. Given a query visual relationship as <subject, predicate, object> and a test video, our objective is to localize the subject and object that are connected via the predicate. Given modern visio-lingual understanding capabilities, solving this problem is achievable, provided that there are large-scale annotated training examples available. However, annotating for every combination of subject, object, and predicate is cumbersome, expensive, and possibly infeasible. Therefore, there is a need for models that can learn to spatially and temporally localize subjects and objects that are connected via an unseen predicate using only a few support set videos sharing the common predicate. We address this challenging problem, referred to as few-shot referring relationships in videos for the first time. To this end, we pose the problem as a minimization of an objective function defined over a T-partite random field. Here, the vertices of the random field correspond to candidate bounding boxes for the subject and object, and T represents the number of frames in the test video. This objective function is composed of frame level and visual relationship similarity potentials. To learn these potentials, we use a relation network that takes query-conditioned translational relationship embedding as inputs and is meta-trained using support set videos in an episodic manner. Further, the objective function is minimized using a belief propagation-based message passing on the random field to obtain the spatiotemporal localization or subject and object trajectories. We perform extensive experiments using two public benchmarks, namely ImageNet-VidVRD and VidOR, and compare the proposed approach with competitive baselines to assess its efficacy. | ['Anand Mishra', 'Yogesh Kumar'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['video-understanding'] | ['computer-vision'] | [ 2.08376661e-01 -1.10772841e-01 -3.98620099e-01 -2.98471540e-01
-6.38918817e-01 -5.39976180e-01 5.28404236e-01 2.59813607e-01
-4.73187447e-01 4.33770835e-01 9.17102247e-02 4.95873988e-02
-5.47690690e-02 -5.66439450e-01 -1.14957821e+00 -5.78000367e-01
-2.20632300e-01 3.54147136e-01 5.09814203e-01 1.50619626e-01
-3.85986008e-02 1.59751058e-01 -1.45533061e+00 4.55805361e-01
5.42000651e-01 1.14330256e+00 5.23801684e-01 3.25385273e-01
2.17640623e-01 8.79192233e-01 -3.90905231e-01 -5.05069017e-01
1.85819745e-01 -3.58146399e-01 -1.01086128e+00 3.19950163e-01
6.78214371e-01 -1.92852870e-01 -6.00208759e-01 1.11412680e+00
7.80147240e-02 7.12750673e-01 4.35605139e-01 -1.53175557e+00
-6.61624789e-01 3.26541305e-01 -6.20387137e-01 5.64532399e-01
6.50405645e-01 2.23937392e-01 1.25121999e+00 -8.51373851e-01
1.03259206e+00 1.20514417e+00 1.31212994e-01 3.33504081e-01
-1.18392670e+00 -3.73866498e-01 4.86506492e-01 6.87028944e-01
-1.53279281e+00 -2.03207344e-01 6.83772802e-01 -5.04438519e-01
9.07643259e-01 3.72699127e-02 8.96511436e-01 1.12963068e+00
-6.93696812e-02 8.83245111e-01 5.81443429e-01 -6.61573485e-02
2.01231837e-01 2.31076345e-01 -5.43735251e-02 7.97534049e-01
-7.39116743e-02 -5.50099909e-02 -8.12799633e-01 6.41139224e-02
5.04007757e-01 1.79720402e-01 -6.44345343e-01 -6.31286621e-01
-1.39024317e+00 6.91601157e-01 6.67285860e-01 2.51750946e-01
-4.03722048e-01 2.89640605e-01 2.72873670e-01 9.71502662e-02
3.13734889e-01 2.62903959e-01 -3.71601582e-01 1.54943550e-02
-7.51980424e-01 1.86423987e-01 5.93305707e-01 1.02075863e+00
8.08498144e-01 -4.41962481e-01 -2.59317875e-01 4.41291928e-01
2.87777603e-01 1.83292776e-01 1.82540163e-01 -9.52998102e-01
6.96775913e-01 5.73367655e-01 1.34286448e-01 -1.34133434e+00
-4.91284160e-03 -2.31643215e-01 -3.94617528e-01 -2.59057194e-01
3.04960191e-01 2.78672904e-01 -8.11287642e-01 1.90692902e+00
6.12615407e-01 9.46063638e-01 1.58862337e-01 1.20094883e+00
9.14219141e-01 8.63092601e-01 1.75470665e-01 -2.73265898e-01
1.43987811e+00 -1.07773638e+00 -6.16692960e-01 -3.39765847e-01
4.81636763e-01 -4.75482643e-01 9.35766876e-01 -1.08216256e-01
-1.01848626e+00 -5.87241948e-01 -7.81970084e-01 -2.38867432e-01
-2.99651712e-01 -6.29023612e-02 3.44033837e-01 -4.64502983e-02
-7.19292521e-01 5.22341907e-01 -8.79183948e-01 -4.20102656e-01
5.22406638e-01 2.34134749e-01 -5.93280017e-01 -2.42054626e-01
-1.08014929e+00 7.17390001e-01 7.26074517e-01 2.02115327e-01
-1.21901214e+00 -6.49983644e-01 -1.17751038e+00 2.45379657e-02
7.19610453e-01 -6.28006518e-01 8.37126851e-01 -1.03991461e+00
-8.70330334e-01 9.21355546e-01 -3.16288948e-01 -6.09868288e-01
2.14138642e-01 -2.99140155e-01 -3.71719629e-01 4.98176008e-01
2.80980885e-01 8.89075398e-01 8.02566111e-01 -1.32196152e+00
-8.26107979e-01 -1.76471904e-01 4.82424974e-01 4.52319026e-01
-1.50673047e-01 7.78132826e-02 -1.14291835e+00 -5.42018354e-01
1.87361285e-01 -8.17635238e-01 1.52373657e-01 1.86474457e-01
-3.87210131e-01 -3.71464103e-01 1.04682040e+00 -8.34721088e-01
9.61378634e-01 -2.30217099e+00 4.81902421e-01 7.49568567e-02
1.22439906e-01 -4.88456152e-02 -1.64340958e-01 4.78312299e-02
-8.98144245e-02 -1.77875510e-03 -1.96694836e-01 -1.91740826e-01
-2.43765533e-01 3.89775604e-01 -3.15851748e-01 6.04858279e-01
2.49985814e-01 9.88962352e-01 -1.17725754e+00 -5.53239107e-01
2.49772325e-01 5.84844232e-01 -5.30431747e-01 3.97758514e-01
-5.86912870e-01 5.48566282e-01 -4.61780399e-01 6.37719989e-01
2.25986972e-01 -5.87360203e-01 4.94698482e-03 -7.16539383e-01
6.96001574e-02 -1.90560613e-02 -1.09970868e+00 1.95595264e+00
-1.11545496e-01 7.19359994e-01 -3.76400828e-01 -1.28304493e+00
5.82263768e-01 2.80718178e-01 7.79241204e-01 -5.18004060e-01
6.33535162e-02 -3.30597311e-02 -1.20053582e-01 -9.67979670e-01
2.68684506e-01 1.10083073e-01 5.51975146e-02 6.22754507e-02
3.42594981e-01 2.11699396e-01 3.53371948e-01 4.11499947e-01
1.02325630e+00 4.52371508e-01 1.09390222e-01 2.01168492e-01
6.33403659e-01 -9.99461021e-03 6.50077283e-01 4.28071171e-01
-2.72478253e-01 4.65233475e-01 6.40696108e-01 -3.54257882e-01
-7.12573886e-01 -9.75381494e-01 1.36506066e-01 8.35254312e-01
8.31224501e-01 -4.44913834e-01 -2.92900980e-01 -8.20939898e-01
-6.60739914e-02 5.40525198e-01 -6.73204482e-01 -1.42142594e-01
-7.19448447e-01 -2.05062315e-01 -8.85797851e-03 4.37418789e-01
5.41758239e-01 -1.01994014e+00 -6.96478367e-01 6.15705699e-02
-5.61351478e-01 -1.65703213e+00 -6.97688520e-01 -8.61290917e-02
-4.98059690e-01 -1.27741420e+00 -5.20281971e-01 -1.08375239e+00
7.43945539e-01 2.40630358e-01 1.08934975e+00 9.06476751e-02
-1.96300432e-01 5.36378741e-01 -3.81933808e-01 1.71789780e-01
7.71205500e-02 -3.73708397e-01 -1.37715787e-01 4.24488097e-01
2.22019121e-01 -1.88733384e-01 -7.61945724e-01 4.08040345e-01
-7.76027858e-01 5.75095415e-02 2.12557524e-01 6.96955740e-01
1.12656248e+00 1.23946890e-01 1.99521497e-01 -4.25747395e-01
-6.08743541e-02 -6.37913406e-01 -5.54189205e-01 5.48858523e-01
1.23098195e-01 -9.87829342e-02 4.06379402e-01 -6.62563205e-01
-7.92987883e-01 3.49981785e-01 3.40689242e-01 -1.02713168e+00
5.76393269e-02 6.23793185e-01 -2.92426944e-01 5.56094982e-02
3.28132272e-01 2.50843644e-01 -2.53056347e-01 -1.71609893e-01
5.63630521e-01 4.81028706e-02 7.97848880e-01 -4.00528431e-01
8.10241699e-01 7.92085052e-01 -1.54333472e-01 -5.41787624e-01
-1.08766818e+00 -7.91891992e-01 -5.69099188e-01 -3.91179502e-01
1.31776464e+00 -9.43229139e-01 -8.25439453e-01 -1.86033458e-01
-1.21327281e+00 -1.24617741e-01 -2.47700378e-01 7.40248740e-01
-6.88784361e-01 2.64285326e-01 -1.17980033e-01 -4.16430771e-01
6.83870018e-02 -1.26092494e+00 1.04791927e+00 2.22229540e-01
-3.51661816e-02 -9.66827095e-01 -1.74879491e-01 4.21319753e-01
-2.23416150e-01 4.26194489e-01 7.99816012e-01 -6.38339341e-01
-1.04666078e+00 -1.64545432e-01 -2.75069594e-01 2.07488209e-01
6.09539449e-02 -1.13323227e-01 -5.86522281e-01 -3.16276401e-01
-4.98211868e-02 -3.79911542e-01 7.49963045e-01 3.28217864e-01
1.31476963e+00 -3.94046217e-01 -6.32562160e-01 6.90415084e-01
1.38410425e+00 1.96363449e-01 2.51333803e-01 1.45651937e-01
8.12101126e-01 6.42452896e-01 8.58298838e-01 1.87442318e-01
5.35950482e-01 8.95154774e-01 5.79083204e-01 1.50796473e-01
-1.14827655e-01 -3.01070392e-01 2.99003720e-01 4.02209997e-01
-8.81566629e-02 -4.26134884e-01 -7.81228483e-01 6.79321945e-01
-2.06731629e+00 -1.27454376e+00 1.45281136e-01 2.27242541e+00
7.57550180e-01 -1.11324573e-02 -5.29633388e-02 -3.40397745e-01
7.37235963e-01 2.58928210e-01 -6.42334342e-01 3.47194910e-01
-1.09161936e-01 -2.00430185e-01 1.89319417e-01 3.05409223e-01
-1.24118602e+00 1.07743990e+00 4.92041969e+00 4.84714687e-01
-1.11558259e+00 1.78656951e-01 5.62907934e-01 -3.14031810e-01
-6.07935190e-02 1.90529108e-01 -6.94768846e-01 4.49997276e-01
5.29784977e-01 -1.25634447e-01 5.67838252e-01 5.97308755e-01
2.86317438e-01 -2.29077458e-01 -1.52585554e+00 1.06912565e+00
4.05006289e-01 -1.35661364e+00 1.77875776e-02 -2.98629940e-01
6.78384244e-01 -3.40030305e-02 5.23044430e-02 3.49642485e-01
-2.09867954e-01 -8.85402501e-01 8.67068231e-01 3.92248124e-01
6.01830900e-01 -4.29848999e-01 4.59866941e-01 2.05882013e-01
-1.62912440e+00 2.30680946e-02 -2.35055432e-01 3.83191049e-01
4.43668067e-01 -1.02456100e-02 -4.78883088e-01 5.47634065e-01
9.64166701e-01 1.19253576e+00 -2.64437884e-01 1.17505443e+00
-3.28395635e-01 2.73470342e-01 -3.61206204e-01 2.44290173e-01
2.58865565e-01 -1.70861498e-01 5.91565907e-01 7.17829347e-01
1.50482893e-01 4.13056433e-01 5.15800238e-01 9.87539530e-01
-2.37832487e-01 -2.19752789e-02 -5.67238927e-01 3.99702527e-02
4.15899873e-01 1.10057020e+00 -7.92247832e-01 -4.05412287e-01
-7.06779301e-01 1.15416276e+00 4.76794392e-01 6.79833412e-01
-1.22704983e+00 6.45597726e-02 5.59273779e-01 -6.57134354e-02
4.99251425e-01 -1.94677293e-01 2.84210116e-01 -1.27708018e+00
2.77359486e-01 -7.84773529e-01 6.08376265e-01 -8.13925803e-01
-9.93549824e-01 5.11950672e-01 2.90201515e-01 -1.33197188e+00
6.06262311e-02 -2.78964013e-01 -4.81904984e-01 5.80703437e-01
-1.44980752e+00 -1.08110976e+00 -5.24076581e-01 8.81764114e-01
7.69305110e-01 6.06788583e-02 3.40220571e-01 3.07452619e-01
-6.07358158e-01 3.44515532e-01 -3.73557568e-01 1.35987133e-01
4.17570919e-01 -9.08978283e-01 -5.25196791e-02 8.32552552e-01
5.29263496e-01 4.27144051e-01 6.73438191e-01 -7.72528946e-01
-1.39797854e+00 -1.39698875e+00 8.10816407e-01 -5.12613177e-01
7.50819504e-01 -3.66507798e-01 -1.05442560e+00 9.26976800e-01
7.33863488e-02 6.93632245e-01 3.67719322e-01 -2.90454239e-01
-3.21393639e-01 4.27604914e-02 -7.96331227e-01 5.14553070e-01
1.27257228e+00 -7.13820159e-01 -6.42559707e-01 7.21912861e-01
8.46327007e-01 -6.23631775e-01 -8.82968903e-01 3.95216495e-01
3.02740395e-01 -6.22166753e-01 1.23736799e+00 -8.15152407e-01
5.12595832e-01 -5.62619448e-01 -2.86363006e-01 -1.03222632e+00
7.46459439e-02 -4.68262136e-01 -4.55339789e-01 1.03552485e+00
3.68879735e-01 -2.56323200e-02 7.24952936e-01 6.52283967e-01
-2.78447755e-03 -9.96326506e-01 -9.33322310e-01 -9.04162586e-01
-4.87935543e-01 -3.28658640e-01 2.92514324e-01 8.59409690e-01
-1.51992053e-01 3.29264015e-01 -4.24484581e-01 5.99778175e-01
5.41506171e-01 2.46338606e-01 5.34895480e-01 -7.48323381e-01
-3.53246242e-01 -5.18168807e-02 -6.78192794e-01 -1.20012462e+00
5.00939071e-01 -9.26774681e-01 2.52072871e-01 -1.58080411e+00
3.57624829e-01 -3.36449087e-01 -2.49824598e-01 6.45375967e-01
-2.82546759e-01 1.37261242e-01 2.87300259e-01 4.19514120e-01
-1.01034129e+00 7.15126932e-01 1.19894755e+00 -3.87043923e-01
-2.45115280e-01 -2.27045581e-01 -1.08506881e-01 7.42737830e-01
3.95217478e-01 -4.92058009e-01 -6.65551484e-01 -6.10884905e-01
1.73696935e-01 3.19711179e-01 8.26470017e-01 -8.12399268e-01
4.14913595e-01 -2.09168196e-01 1.57055244e-01 -5.94981968e-01
7.14707196e-01 -1.03489172e+00 1.57397449e-01 2.37392977e-01
-4.23440129e-01 -1.47137195e-02 -9.18842703e-02 9.43387210e-01
-4.85502243e-01 -2.70060562e-02 5.76311529e-01 6.53505027e-02
-1.18144655e+00 7.81541049e-01 2.83801317e-01 2.07691893e-01
1.56152046e+00 -1.82130024e-01 -2.96398342e-01 -2.54868120e-01
-8.52269292e-01 4.59807605e-01 3.63509327e-01 7.24804163e-01
9.23647404e-01 -1.20945108e+00 -4.59137857e-01 -4.12658565e-02
4.38915402e-01 1.13229364e-01 2.76148915e-01 9.95037675e-01
-1.82909489e-01 1.62972182e-01 -2.66511440e-02 -8.81001830e-01
-1.22317505e+00 8.31310868e-01 4.40130770e-01 9.46014673e-02
-8.55534434e-01 1.13908935e+00 4.16779131e-01 1.07335746e-01
4.55644131e-01 -2.64584839e-01 -4.86608624e-01 -2.69101392e-02
3.74928564e-01 -7.45345354e-02 -3.40717912e-01 -1.15735900e+00
-5.47894239e-01 6.06690705e-01 4.96937670e-02 -2.82523893e-02
1.06084979e+00 -1.90016493e-01 -3.53116691e-02 3.76275301e-01
1.60655475e+00 -5.03261328e-01 -1.42407763e+00 -5.60512662e-01
-1.17586352e-01 -7.37834811e-01 -1.03054568e-01 -3.98361266e-01
-1.29559553e+00 5.51106274e-01 5.21544039e-01 -7.08441213e-02
9.73075688e-01 5.56567729e-01 6.87275171e-01 2.36279398e-01
2.45096460e-01 -7.65312314e-01 4.61565971e-01 2.58217931e-01
9.12567854e-01 -1.42133272e+00 -7.58219808e-02 -5.98673165e-01
-7.32432425e-01 7.65795171e-01 9.04219210e-01 -4.52997275e-02
5.77468157e-01 -2.65087008e-01 -4.12012339e-01 -4.19020504e-01
-8.98362279e-01 -3.20654303e-01 6.87366545e-01 3.89041394e-01
1.71100795e-01 -1.95570737e-01 -7.63237625e-02 2.01507807e-01
1.70493394e-01 -2.78202482e-02 3.00898738e-02 8.94409478e-01
-2.31903031e-01 -6.93054378e-01 -1.13028340e-01 3.71079743e-01
-2.79493213e-01 1.14803179e-03 -2.40257774e-02 6.70130968e-01
3.52646440e-01 8.28427672e-01 2.36978129e-01 -3.49246979e-01
2.62670904e-01 -2.30187267e-01 3.74185979e-01 -6.81596637e-01
-3.07134718e-01 -7.44941086e-02 -1.39438406e-01 -8.25796783e-01
-7.93382466e-01 -6.47003829e-01 -1.50287080e+00 1.67992651e-01
-3.23119700e-01 2.91649718e-02 3.35452825e-01 1.12766600e+00
2.29778856e-01 3.99413884e-01 5.01223087e-01 -6.68821454e-01
-1.60463259e-01 -3.96688074e-01 -1.59115821e-01 7.99360454e-01
2.97650754e-01 -9.30848837e-01 -1.32807940e-01 4.99154747e-01] | [9.620452880859375, 0.7579352259635925] |
541e6242-1e9e-42c4-9d24-2d4c49aa1468 | augmenting-control-over-exploration-space-in | 2306.14705 | null | https://arxiv.org/abs/2306.14705v1 | https://arxiv.org/pdf/2306.14705v1.pdf | Augmenting Control over Exploration Space in Molecular Dynamics Simulators to Streamline De Novo Analysis through Generative Control Policies | This study introduces the P5 model - a foundational method that utilizes reinforcement learning (RL) to augment control, effectiveness, and scalability in molecular dynamics simulations (MD). Our innovative strategy optimizes the sampling of target polymer chain conformations, marking an efficiency improvement of over 37.1%. The RL-induced control policies function as an inductive bias, modulating Brownian forces to steer the system towards the preferred state, thereby expanding the exploration of the configuration space beyond what traditional MD allows. This broadened exploration generates a more varied set of conformations and targets specific properties, a feature pivotal for progress in polymer development, drug discovery, and material design. Our technique offers significant advantages when investigating new systems with limited prior knowledge, opening up new methodologies for tackling complex simulation problems with generative techniques. | ['Gregory Rutledge', 'Neil Malur', 'Luis Martinez', 'Andrew Emmel', 'Paloma Gonzalez-Rojas'] | 2023-06-26 | null | null | null | null | ['drug-discovery'] | ['medical'] | [ 3.91712368e-01 -8.33464414e-02 -6.94699585e-01 3.01345527e-01
-6.12011373e-01 -9.30210710e-01 6.41610146e-01 2.34737605e-01
-4.70683038e-01 1.42628884e+00 1.22801643e-02 -7.59542763e-01
2.16327123e-02 -8.65620255e-01 -7.80838847e-01 -1.30556130e+00
-3.85906965e-01 5.42989194e-01 -3.49710211e-02 -3.28598261e-01
5.34052014e-01 9.41985905e-01 -1.00906467e+00 -1.30984277e-01
9.34219003e-01 2.32900664e-01 3.31870079e-01 6.03564382e-01
2.58937448e-01 1.87179834e-01 -2.77529508e-01 1.42454952e-01
3.35941911e-02 -4.20483351e-01 -4.49112713e-01 -2.44276240e-01
-3.34606580e-02 6.56711459e-02 -2.13019654e-01 7.47290492e-01
8.78987789e-01 4.73023891e-01 6.25083447e-01 -3.74497354e-01
-5.17062306e-01 3.72505546e-01 -2.00187132e-01 2.72051543e-01
4.83756393e-01 9.56858814e-01 6.76417470e-01 -3.88772845e-01
9.14955318e-01 1.27005124e+00 3.76580715e-01 7.03178048e-01
-1.79822230e+00 -6.10738933e-01 2.51918942e-01 -1.25777274e-01
-8.83765340e-01 -1.33000106e-01 8.41204464e-01 -2.90908158e-01
1.24575830e+00 1.22831449e-01 1.16346681e+00 1.33465004e+00
7.35783696e-01 3.53234977e-01 1.37204838e+00 -3.25726360e-01
8.36475194e-01 -3.09956968e-01 -2.99952328e-01 3.35053116e-01
3.42284471e-01 7.45098412e-01 -1.43072903e-01 -5.79629123e-01
9.13617492e-01 1.13818094e-01 -1.88112184e-01 -7.78168678e-01
-1.11649895e+00 9.84452844e-01 2.20809847e-01 -1.38927102e-01
-6.75854623e-01 1.03729941e-01 1.71631247e-01 -5.15802614e-02
-1.98854357e-01 1.43922722e+00 -5.28271794e-01 -4.16009367e-01
-5.08205414e-01 8.75050366e-01 8.03449750e-01 4.49572653e-01
4.16822046e-01 2.05070585e-01 2.55349465e-02 3.01813424e-01
3.64572465e-01 7.48508453e-01 1.26235828e-01 -1.03656673e+00
1.57465309e-01 3.70181650e-01 3.45180333e-01 -2.16496825e-01
-3.77092659e-01 -5.77990174e-01 -4.50770140e-01 5.10227859e-01
3.17875803e-01 -4.60482180e-01 -9.30824637e-01 1.77800333e+00
5.57748914e-01 -5.50853789e-01 4.52444665e-02 4.94756877e-01
4.29547504e-02 8.13671589e-01 5.43651998e-01 -6.24548554e-01
8.35040867e-01 -6.76160216e-01 -4.89846230e-01 7.78264627e-02
2.50948757e-01 -6.15266263e-01 9.93495166e-01 6.19502425e-01
-1.26824069e+00 -3.30851167e-01 -1.16796517e+00 6.41190350e-01
-4.10182089e-01 -5.46369970e-01 9.84191895e-01 7.80547917e-01
-7.82407165e-01 1.31011391e+00 -9.69120145e-01 -1.61024690e-01
4.10157591e-01 7.58453667e-01 5.34090102e-02 1.02589414e-01
-1.09181631e+00 9.66188610e-01 6.27987683e-01 -2.96186686e-01
-1.27721465e+00 -8.09737682e-01 -4.90428627e-01 -2.29655355e-01
3.62730026e-01 -8.23800445e-01 9.86701787e-01 -2.97748923e-01
-2.19201112e+00 1.08730398e-01 -3.89939249e-02 -4.42463875e-01
4.99489009e-01 9.29740071e-02 -1.49444699e-01 -1.01509258e-01
-2.63518184e-01 8.97584140e-01 7.80982852e-01 -1.10799038e+00
1.02395967e-01 -1.86627850e-01 -2.13260025e-01 2.73703992e-01
3.39651465e-01 -3.66567880e-01 1.93146139e-01 -6.09418213e-01
-2.85140336e-01 -1.33984876e+00 -1.08202946e+00 -5.87847412e-01
-4.30217832e-01 -9.95409340e-02 5.00316501e-01 -4.86458242e-02
1.03523803e+00 -1.45606840e+00 5.82061648e-01 5.91923594e-01
5.82962744e-02 5.31916797e-01 -9.55480933e-02 1.14475060e+00
-2.09882170e-01 1.33804530e-01 -1.07442215e-01 4.81578201e-01
-2.22551346e-01 3.67840528e-02 -4.41169828e-01 9.13540125e-02
8.90874788e-02 1.03127789e+00 -1.13364029e+00 -6.14351220e-02
3.56416941e-01 4.65837657e-01 -7.85887599e-01 1.41480565e-01
-7.80536830e-01 1.04879212e+00 -7.61623561e-01 6.89762115e-01
6.14517331e-01 -2.47070357e-01 6.90083027e-01 1.26211241e-01
-4.31735128e-01 2.56997079e-01 -7.10251331e-01 1.36224020e+00
-1.45581126e-01 -6.93015615e-03 -4.47611809e-01 -3.60733330e-01
1.07563996e+00 -1.71303943e-01 7.28178203e-01 -7.17736840e-01
8.96304250e-02 -3.48309688e-02 5.34017742e-01 -2.96583325e-01
1.65326923e-01 -4.44899738e-01 4.22176495e-02 4.41381127e-01
-2.98588008e-01 -4.43078220e-01 3.19377899e-01 -5.18218167e-02
9.28340137e-01 4.15143430e-01 2.70942032e-01 -4.23435390e-01
3.18211794e-01 1.60982326e-01 3.04141670e-01 8.86660218e-01
-1.67500511e-01 -1.39273897e-01 5.42413354e-01 -5.42285442e-01
-1.46272731e+00 -1.28809631e+00 -2.15929314e-01 8.81882012e-01
4.20809574e-02 -4.17112261e-01 -5.10681808e-01 -3.41465652e-01
3.89548719e-01 5.61240137e-01 -6.07455850e-01 -3.37355107e-01
-9.39117134e-01 -1.12409759e+00 7.15882778e-02 2.77022511e-01
-6.22238144e-02 -1.31188071e+00 -5.02881050e-01 8.29550266e-01
5.33385932e-01 -5.57907343e-01 -3.74473870e-01 5.09317517e-01
-1.27035356e+00 -1.01963258e+00 -5.52540839e-01 -3.92873824e-01
4.62876022e-01 -3.34582567e-01 9.31045711e-01 -4.04325873e-01
-4.96767223e-01 7.19275847e-02 -3.16580348e-02 -9.70303565e-02
-8.42399895e-01 1.66172743e-01 1.15960456e-01 -9.12376285e-01
9.27586760e-03 -7.75010586e-01 -1.02126348e+00 1.71544358e-01
-4.93921041e-01 -2.13441581e-01 6.78947568e-01 8.76222432e-01
8.77961278e-01 -2.27297351e-01 7.67475128e-01 -6.10052228e-01
8.84811997e-01 -2.85430849e-01 -7.12620020e-01 1.08521702e-02
-9.02010262e-01 6.57485485e-01 7.61376381e-01 -9.00345981e-01
-9.69813466e-01 4.18956093e-02 -2.76301324e-01 1.16595805e-01
-1.09142661e-02 2.00529814e-01 -2.37487227e-01 -3.61176550e-01
6.08931243e-01 3.84087563e-01 4.29150492e-01 -3.95039320e-01
4.94703770e-01 -1.30467221e-01 -2.40590256e-02 -1.18165469e+00
2.94125378e-01 5.31858616e-02 5.04021764e-01 -7.19941020e-01
-1.08421065e-01 7.46442527e-02 -4.11844254e-01 5.10187559e-02
5.44169366e-01 -4.73507106e-01 -1.44323528e+00 -2.17702575e-02
-5.67634702e-01 -7.04367042e-01 -2.87676513e-01 4.66167063e-01
-6.75784349e-01 3.35191637e-01 -5.77778161e-01 -6.89282298e-01
-4.30341631e-01 -1.64825726e+00 7.00201452e-01 2.96827018e-01
-5.96702456e-01 -8.77829552e-01 6.80095792e-01 2.64979191e-02
4.60916698e-01 5.75891197e-01 1.22458720e+00 -3.36285532e-01
-7.81814516e-01 1.33724496e-01 5.17690539e-01 -1.59693271e-01
-7.08350562e-04 1.77244872e-01 -3.35566759e-01 -6.92794204e-01
-5.85563183e-01 -2.25863978e-01 7.47616172e-01 6.20072901e-01
1.15777338e+00 -2.80172974e-01 -7.03765273e-01 3.82461160e-01
1.14389038e+00 9.92071986e-01 6.76810622e-01 5.42356133e-01
3.70902151e-01 5.57812266e-02 5.17304897e-01 5.42618155e-01
-8.60846341e-02 6.10905528e-01 3.31341982e-01 8.41176510e-02
9.09918174e-02 -5.67784905e-01 3.53382677e-01 2.70628303e-01
-2.70804256e-01 -1.33567736e-01 -6.79340482e-01 -5.68987727e-02
-1.43818033e+00 -1.06177366e+00 5.29336452e-01 2.23296165e+00
1.33271241e+00 3.05249214e-01 5.84620655e-01 -3.29664528e-01
5.34303188e-01 1.74568802e-01 -1.36358547e+00 -7.18384147e-01
1.23235583e-01 6.94152951e-01 6.48827791e-01 6.20409012e-01
-9.22477901e-01 1.00274527e+00 8.42402935e+00 6.57640815e-01
-1.28998911e+00 -5.62828839e-01 5.16598225e-01 -2.51553535e-01
-4.38152134e-01 1.26318127e-01 -1.03605378e+00 6.44587159e-01
1.10507882e+00 -9.90368873e-02 6.49196506e-01 7.28244305e-01
6.35881901e-01 -2.02810578e-02 -8.54811668e-01 3.71955842e-01
-6.30576193e-01 -1.80712545e+00 2.74339080e-01 5.26698709e-01
8.15697074e-01 -2.72757351e-01 3.44398469e-01 2.63687134e-01
6.73762143e-01 -1.12430823e+00 2.85659969e-01 4.56891626e-01
6.35737121e-01 -1.15052247e+00 7.08129928e-02 4.81917262e-02
-6.20809913e-01 -1.20898977e-01 -9.03003216e-02 -4.30353470e-02
3.69330168e-01 3.23238999e-01 -1.04423356e+00 -1.96348503e-01
3.62994596e-02 3.35148960e-01 -2.97882594e-02 7.50953138e-01
-8.18872005e-02 5.76027274e-01 -2.75295913e-01 -6.16925061e-01
3.29901636e-01 -6.83295965e-01 6.75086200e-01 8.62171531e-01
-1.13106251e-01 6.94559216e-02 3.57093394e-01 9.71558392e-01
-2.55013257e-03 -1.50469258e-01 -5.67260444e-01 -3.98090750e-01
7.29775548e-01 6.26356006e-01 -7.69683063e-01 -2.07755655e-01
5.59687436e-01 4.05347824e-01 1.21743932e-01 5.06565750e-01
-8.67910028e-01 -2.36276954e-01 8.76036763e-01 7.49077797e-02
4.59289074e-01 -6.39236927e-01 1.80489775e-02 -6.64044499e-01
-5.77288628e-01 -1.21738183e+00 -8.39895234e-02 -4.39255014e-02
-8.97388101e-01 1.03361532e-01 2.22574174e-02 -6.25324726e-01
-3.24338377e-01 -6.00705266e-01 -5.50405383e-01 6.68409050e-01
-1.31545603e+00 -5.87491632e-01 4.09331501e-01 -1.78888105e-02
4.07228857e-01 -9.62288976e-02 6.58218980e-01 -2.11709782e-01
-6.78799272e-01 3.77934307e-01 6.78513467e-01 -7.89303720e-01
5.94104767e-01 -1.20961535e+00 4.28122133e-01 2.69816160e-01
-5.40435791e-01 1.02848268e+00 1.11738348e+00 -1.02229118e+00
-1.92408180e+00 -6.80170357e-01 -1.19827434e-01 -4.47128475e-01
6.40914798e-01 -1.49335057e-01 -6.61894023e-01 1.09314725e-01
6.69705719e-02 -5.41356683e-01 8.39204133e-01 8.75408724e-02
-4.07171957e-02 1.91002324e-01 -1.13089728e+00 9.25835609e-01
9.74679649e-01 -2.12874964e-01 -6.49939030e-02 4.54079717e-01
7.69256651e-01 -5.31833708e-01 -1.32804108e+00 3.53576422e-01
8.49881351e-01 -5.50426364e-01 1.49278581e+00 -1.01643109e+00
2.35892922e-01 -2.08695441e-01 1.20741509e-01 -1.40863597e+00
-4.54361260e-01 -1.25550377e+00 -4.46565449e-01 5.58210313e-01
6.08931482e-01 -7.81504035e-01 9.08473194e-01 5.50896645e-01
-1.04897274e-02 -1.26008916e+00 -5.70632339e-01 -7.86550879e-01
5.98609805e-01 1.95114553e-01 7.64975727e-01 5.16093433e-01
3.41045946e-01 5.66858836e-02 -1.22196577e-01 -2.67064154e-01
6.97616398e-01 2.56053388e-01 4.80053127e-01 -6.53952658e-01
-5.71131289e-01 -6.41496360e-01 8.47930461e-02 -1.10233784e+00
-6.49546832e-02 -6.08594239e-01 -2.90101707e-01 -9.96099830e-01
2.06387594e-01 -6.57155156e-01 -3.95248160e-02 7.77588189e-02
-1.05763718e-01 -3.55893880e-01 2.42134765e-01 7.51913115e-02
-4.42840397e-01 8.70154679e-01 1.88700271e+00 1.19665628e-02
-5.81650972e-01 7.60777220e-02 -7.51509368e-01 2.71031529e-01
9.56033349e-01 -2.47146696e-01 -3.79076064e-01 4.39273924e-01
2.28874117e-01 2.34924719e-01 5.21783456e-02 -7.44161129e-01
-2.81748116e-01 -6.08538508e-01 6.15728140e-01 -6.08419001e-01
2.85201728e-01 -4.54731882e-01 2.64005899e-01 1.02373958e+00
-4.90478337e-01 1.43114135e-01 3.85039866e-01 8.20960641e-01
4.05158699e-01 1.60979182e-01 9.49929416e-01 -2.18055993e-01
-2.54173338e-01 4.93278265e-01 -7.84975410e-01 -3.62109169e-02
1.19349086e+00 -1.69867799e-01 -2.49094531e-01 5.85710593e-02
-1.15219855e+00 3.60730216e-02 8.37600112e-01 8.09803829e-02
3.08826119e-01 -1.13529205e+00 -1.70109011e-02 2.93409795e-01
-2.80076295e-01 -1.54496640e-01 1.13015369e-01 2.79035330e-01
-6.18527949e-01 6.20170236e-01 -4.23490345e-01 -5.07729828e-01
-8.62436116e-01 7.43642569e-01 3.18009704e-01 -5.30372918e-01
-4.35676903e-01 3.78268123e-01 -3.22601020e-01 -2.73613185e-01
-2.57859051e-01 -1.88334316e-01 7.71314278e-02 -2.20244408e-01
1.93615764e-01 5.46342015e-01 -1.55565426e-01 8.13014358e-02
-2.83107549e-01 5.11838078e-01 -3.43882263e-01 1.42067224e-01
1.46953225e+00 2.88553029e-01 1.62584335e-01 -1.86399892e-01
7.28003085e-01 -2.75426991e-02 -1.88453364e+00 2.25528046e-01
-1.16860792e-01 -2.60697067e-01 1.10942246e-02 -1.02137053e+00
-1.89856723e-01 5.65850914e-01 6.07936323e-01 -1.38828129e-01
4.31917757e-01 -8.36731866e-02 7.91507185e-01 6.80823207e-01
4.09716278e-01 -8.95198166e-01 4.15009081e-01 5.78929424e-01
6.71215236e-01 -1.03448868e+00 3.24340045e-01 6.82636201e-02
-4.92996126e-01 1.28911972e+00 4.84881788e-01 -2.38684252e-01
3.84275675e-01 3.68303001e-01 -4.25307274e-01 -1.99384987e-01
-8.21035981e-01 1.20054141e-01 -1.31243154e-01 7.91490972e-01
4.17296499e-01 2.78619200e-01 -3.57775867e-01 -6.00504503e-02
7.06232935e-02 -2.34283164e-01 2.99641520e-01 1.19443965e+00
-6.02452576e-01 -1.80009365e+00 -3.85492146e-01 1.22768499e-01
-2.09684810e-03 -3.49443965e-02 -1.83653921e-01 8.93390059e-01
-5.87752238e-02 5.03921926e-01 -2.09309965e-01 1.38790563e-01
1.99661911e-01 1.32552702e-02 9.86926675e-01 -2.92254597e-01
-5.31810522e-01 1.84710026e-01 2.36848034e-02 -4.90036905e-01
2.33790036e-02 -9.07373250e-01 -1.22211361e+00 -5.51512182e-01
-3.53743821e-01 3.82496893e-01 6.00600123e-01 6.20031834e-01
7.89481223e-01 5.48957288e-01 7.76213527e-01 -1.19142127e+00
-7.15516269e-01 -5.16911209e-01 -1.97776154e-01 -3.49607058e-02
2.91577280e-01 -8.95530343e-01 3.50899026e-02 -3.25325370e-01] | [4.888193607330322, 5.540378093719482] |
9fee7939-d2a0-475b-ba0b-f6e54efb3a77 | impact-of-feature-selection-on-micro-text | 1708.08123 | null | http://arxiv.org/abs/1708.08123v1 | http://arxiv.org/pdf/1708.08123v1.pdf | Impact of Feature Selection on Micro-Text Classification | Social media datasets, especially Twitter tweets, are popular in the field of
text classification. Tweets are a valuable source of micro-text (sometimes
referred to as "micro-blogs"), and have been studied in domains such as
sentiment analysis, recommendation systems, spam detection, clustering, among
others. Tweets often include keywords referred to as "Hashtags" that can be
used as labels for the tweet. Using tweets encompassing 50 labels, we studied
the impact of word versus character-level feature selection and extraction on
different learners to solve a multi-class classification task. We show that
feature extraction of simple character-level groups performs better than simple
word groups and pre-processing methods like normalizing using Porter's Stemming
and Part-of-Speech ("POS")-Lemmatization. | ['Ankit Vadehra', 'Gordon V. Cormack', 'Maura R. Grossman'] | 2017-08-27 | null | null | null | null | ['spam-detection'] | ['natural-language-processing'] | [ 7.56783336e-02 -2.28548422e-01 -3.74784470e-01 -5.45838833e-01
-5.52643836e-01 -6.04413569e-01 7.51749516e-01 1.36023211e+00
-1.00122094e+00 4.54843491e-01 5.26916265e-01 -4.40523416e-01
1.43237367e-01 -1.11643422e+00 -2.25492433e-01 -6.49288476e-01
3.17146406e-02 1.18298009e-01 1.35552391e-01 -3.39747161e-01
7.46592581e-01 3.50573927e-01 -1.83883929e+00 4.45439368e-01
8.28054905e-01 8.36295784e-01 -2.99449880e-02 2.57495373e-01
-1.06986129e+00 5.07083833e-01 -8.20070922e-01 -4.67799991e-01
-1.06662929e-01 -2.10042268e-01 -6.56690300e-01 3.79216932e-02
2.27382835e-02 2.89984703e-01 1.53590113e-01 1.39888859e+00
3.73867691e-01 2.54930437e-01 7.44858801e-01 -9.57860887e-01
-7.01152235e-02 9.19121444e-01 -7.10396171e-01 2.42359564e-01
4.20995831e-01 -5.38834870e-01 1.20060337e+00 -8.18495750e-01
5.87714493e-01 1.28067708e+00 4.94740367e-01 1.82912666e-02
-7.84514129e-01 -8.53349268e-01 7.28851631e-02 8.44376627e-03
-1.35425031e+00 1.96361989e-02 5.10794818e-01 -6.51319325e-01
6.69886768e-01 4.95804399e-01 5.20951569e-01 7.83528030e-01
4.69689250e-01 8.27014685e-01 1.03044510e+00 -5.11320770e-01
1.81012213e-01 4.56856340e-01 6.33877039e-01 3.14620078e-01
2.20380291e-01 -8.49182665e-01 -4.04964209e-01 -4.22379762e-01
-2.82980740e-01 1.66000977e-01 2.03567788e-01 4.93991792e-01
-1.00348258e+00 1.42333877e+00 5.30534312e-02 5.41529238e-01
-3.14975411e-01 -2.08009467e-01 7.57629752e-01 4.15346593e-01
1.03571403e+00 6.57893956e-01 -3.13073725e-01 3.46362218e-02
-8.25427592e-01 1.19327486e-01 7.47029006e-01 7.77808368e-01
1.07456052e+00 -3.21547657e-01 -2.12959096e-01 9.75241125e-01
1.67904377e-01 4.88035887e-01 1.17497683e+00 -8.54121819e-02
6.37285233e-01 8.63289833e-01 -7.80845508e-02 -1.30940890e+00
-7.86052763e-01 -5.50124228e-01 -7.04607129e-01 -6.83818460e-01
2.72045761e-01 -5.23064435e-01 -5.82444370e-01 1.16280854e+00
5.16699970e-01 -5.70585392e-02 -1.99249879e-01 2.33338326e-01
8.96454990e-01 8.88185561e-01 2.73617506e-01 -5.12821615e-01
1.61143720e+00 -5.72091639e-01 -6.98553145e-01 -8.46618041e-02
9.73923564e-01 -1.08182013e+00 1.01168132e+00 3.83471400e-01
-5.72837114e-01 -2.44118109e-01 -4.38298345e-01 2.76911795e-01
-1.24210131e+00 -3.26571912e-01 5.31262696e-01 9.09422159e-01
-4.08618867e-01 6.47670686e-01 -3.70743364e-01 -4.34880018e-01
2.55790412e-01 2.29685634e-01 -2.04957306e-01 4.75847363e-01
-1.28510046e+00 6.63652182e-01 1.99037597e-01 -5.96747875e-01
-3.93578671e-02 -4.25634682e-01 -6.70251906e-01 -1.17356792e-01
2.91081965e-01 -1.07111216e-01 1.00664580e+00 -9.67743099e-01
-1.12241471e+00 1.05121493e+00 -1.54413760e-01 -3.83884370e-01
1.00835957e-01 -1.87636524e-01 -7.00660110e-01 2.97724642e-02
2.86662638e-01 7.95012116e-02 1.09920299e+00 -6.43101990e-01
-1.15899634e+00 -2.51602888e-01 -3.25637013e-01 -2.44552940e-02
-9.72224295e-01 7.66072571e-01 -9.74475220e-02 -8.51980686e-01
2.88534045e-01 -9.97887671e-01 -2.44026840e-01 -9.06141698e-01
-7.69907355e-01 -6.48247600e-01 7.93907166e-01 -3.55742961e-01
1.72403479e+00 -2.19512081e+00 -3.96596551e-01 7.07368255e-01
2.46685386e-01 1.70584083e-01 3.05161685e-01 6.72906518e-01
-1.12852370e-02 7.57031739e-01 2.51449525e-01 -2.57737100e-01
8.53285368e-04 -2.03070447e-01 -3.24420661e-01 4.99210179e-01
-9.82531756e-02 4.37102526e-01 -9.60629523e-01 -8.17483425e-01
5.57427220e-02 1.32598549e-01 -1.68934226e-01 -2.04887986e-01
-1.16489649e-01 1.13049753e-01 -7.71865189e-01 2.80383885e-01
2.96292037e-01 -3.35597023e-02 -2.03527316e-01 -1.38230234e-01
-4.80795979e-01 5.70088625e-01 -1.08065212e+00 6.25937462e-01
-5.70526659e-01 7.97549009e-01 -2.58242667e-01 -6.88665748e-01
9.00176883e-01 -1.42766595e-01 5.88505566e-01 -3.54401469e-01
8.57230127e-01 1.15555942e-01 -1.69880558e-02 -3.68592829e-01
9.67635155e-01 1.04257721e-03 -4.56931353e-01 6.02105558e-01
-1.92993194e-01 -1.46158129e-01 7.28296459e-01 3.43290269e-01
7.94329047e-01 -7.93336630e-01 3.29227418e-01 -3.38833809e-01
6.12065256e-01 4.01703455e-02 1.53451040e-01 5.97459018e-01
2.39611804e-01 5.07293582e-01 5.83635211e-01 9.01659951e-02
-6.53771400e-01 -2.73281902e-01 -3.19353551e-01 1.57808542e+00
-7.96209276e-02 -9.87279713e-01 -7.96301186e-01 -6.53555572e-01
2.45076641e-01 7.11470306e-01 -5.55894434e-01 3.62799540e-02
-5.05006552e-01 -1.12099195e+00 5.03870666e-01 7.74059966e-02
6.21854588e-02 -1.00833559e+00 -5.62916286e-02 3.77652287e-01
-6.17073327e-02 -9.40787971e-01 -4.41400081e-01 4.75347966e-01
-5.14929235e-01 -8.71864498e-01 -4.88483816e-01 -6.94461465e-01
8.10618758e-01 4.26136732e-01 7.57894278e-01 7.14260712e-02
3.16149034e-02 1.33042201e-01 -1.01291931e+00 -7.17388511e-01
-4.91605222e-01 5.78249633e-01 8.26530233e-02 5.21907985e-01
5.15265167e-01 -2.84685105e-01 -2.65670061e-01 3.52975786e-01
-1.17073739e+00 -2.46460378e-01 2.61212319e-01 4.34860706e-01
3.84028196e-01 2.38685399e-01 6.04380190e-01 -1.69677734e+00
1.10709560e+00 -6.44408166e-01 -2.85337538e-01 -9.92044657e-02
-5.63835144e-01 -9.51900333e-02 9.29785907e-01 -6.07595682e-01
-6.29799962e-01 -2.90259480e-01 -4.00058180e-01 4.49920565e-01
-6.09273426e-02 9.70083356e-01 3.25886495e-02 -1.68931603e-01
5.31261325e-01 1.72711074e-01 -1.60444021e-01 -5.26277125e-01
1.27754524e-01 1.30045795e+00 -2.03746110e-01 -1.72030836e-01
8.37274611e-01 2.94962883e-01 -2.73516536e-01 -1.23109758e+00
-1.08417261e+00 -1.11483288e+00 -5.22368848e-01 -1.33223087e-01
7.99556196e-01 -6.13512039e-01 -5.76970041e-01 6.74169421e-01
-7.37253964e-01 2.49515399e-01 -1.43425748e-01 4.91116375e-01
2.31511712e-01 3.31116527e-01 -6.23526573e-01 -5.25243104e-01
-4.84491050e-01 -7.96269536e-01 7.28150427e-01 4.45749253e-01
-5.70476234e-01 -1.09971094e+00 -3.20132226e-01 1.92441300e-01
3.57106805e-01 6.69392571e-02 1.13466525e+00 -1.45368207e+00
4.42490250e-01 -7.13426173e-01 -3.81695740e-02 4.31130290e-01
1.78203955e-01 2.09667802e-01 -7.34804153e-01 -6.81681186e-02
-1.65361211e-01 4.38520834e-02 9.44033206e-01 6.89364644e-03
1.29764903e+00 -5.82440436e-01 -4.63383764e-01 2.71994919e-01
9.46862698e-01 -1.52885541e-03 4.44358915e-01 7.24518478e-01
8.03098679e-01 7.18979478e-01 7.62322605e-01 9.00584817e-01
3.27930570e-01 9.61894244e-02 1.19323939e-01 1.49196714e-01
4.83850092e-01 -8.36306661e-02 3.91844749e-01 1.29186809e+00
4.60255057e-01 -4.10321563e-01 -9.85273838e-01 3.24732780e-01
-1.35105455e+00 -7.93131888e-01 -7.18502164e-01 1.99242330e+00
9.25722480e-01 4.17952448e-01 3.16513985e-01 5.55325329e-01
1.06217992e+00 3.70956004e-01 5.78230396e-02 -4.11465585e-01
-2.10403442e-01 3.42998892e-01 9.25338924e-01 3.14805686e-01
-1.30830109e+00 9.18444395e-01 5.07144976e+00 1.22276306e+00
-1.31014287e+00 9.44006629e-03 3.50778967e-01 4.55149531e-01
-2.35468164e-01 -1.38199180e-01 -1.33879924e+00 8.68851244e-01
9.10424948e-01 -3.82523984e-01 -2.95163810e-01 7.18837380e-01
4.33157891e-01 -5.56595147e-01 -5.90385199e-01 7.88609743e-01
1.54075816e-01 -1.02392864e+00 1.79385677e-01 -2.51220092e-02
8.11440527e-01 4.33326475e-02 2.86559854e-03 2.44903475e-01
1.05640739e-02 -5.24929583e-01 6.51268780e-01 -1.20577251e-03
2.50685096e-01 -8.39681923e-01 9.17609155e-01 3.71927321e-01
-1.04127705e+00 -5.75987548e-02 -2.27179274e-01 3.61986086e-02
6.02872036e-02 1.02391529e+00 -7.10870743e-01 2.67629236e-01
3.33766460e-01 7.30416477e-01 -9.24662530e-01 1.16294861e+00
3.36102247e-02 9.88808930e-01 -4.35444534e-01 -8.24530542e-01
5.01081228e-01 -2.96919405e-01 3.57257813e-01 1.68015277e+00
1.76606894e-01 -2.18058318e-01 2.97020823e-01 -2.17477098e-01
-2.73484260e-01 1.07496464e+00 3.77032794e-02 -2.95702159e-01
4.13369209e-01 1.41741586e+00 -1.47215760e+00 -5.41634738e-01
-3.42076093e-01 3.09024513e-01 -2.42854863e-01 -5.09650633e-02
-5.34895420e-01 -1.00318801e+00 3.97574484e-01 5.11494219e-01
-2.01101322e-02 -3.26642483e-01 -6.77397549e-02 -9.72377241e-01
-4.12981153e-01 -7.46099591e-01 3.41341853e-01 -2.95324117e-01
-1.39026260e+00 5.20426631e-01 3.83385760e-03 -1.35425711e+00
-1.18284747e-01 -5.38252056e-01 -6.49181485e-01 3.58042657e-01
-1.38846064e+00 -6.30681992e-01 -2.08184466e-01 6.26430392e-01
4.22992438e-01 -2.69632161e-01 6.72464371e-01 6.89570487e-01
-6.80960238e-01 5.66902280e-01 6.35127485e-01 4.17671472e-01
8.76539230e-01 -1.11441243e+00 1.72588959e-01 3.91645610e-01
8.80180001e-02 6.47383869e-01 8.29924405e-01 -7.77063847e-01
-1.29070497e+00 -1.30881274e+00 1.39520490e+00 7.68678933e-02
1.15024853e+00 -3.57221603e-01 -6.52534604e-01 3.14599484e-01
2.42944937e-02 -3.04014504e-01 1.04576945e+00 1.29594669e-01
-1.18200794e-01 -1.16122961e-01 -9.90479469e-01 5.25712311e-01
3.87753218e-01 -4.22221243e-01 -4.05860901e-01 7.72565365e-01
4.50687081e-01 -4.31344025e-02 -5.60395122e-01 -2.71067351e-01
3.43144685e-01 -5.19965291e-01 5.49794436e-01 -5.79066992e-01
1.62531033e-01 1.29246237e-02 1.64291799e-01 -1.43552864e+00
-2.60022897e-02 -7.56110132e-01 5.29380918e-01 1.75258231e+00
5.64377367e-01 -9.55165625e-01 6.34231091e-01 1.35412589e-01
-2.62623485e-02 -4.38077748e-01 -4.35228318e-01 -5.01281083e-01
7.98161794e-03 -6.20162368e-01 4.64076042e-01 1.07578278e+00
2.25381374e-01 5.76369107e-01 -1.73043981e-01 -3.93769413e-01
2.08666056e-01 -6.55921474e-02 7.12093294e-01 -1.56019115e+00
4.42529172e-01 -8.59139442e-01 -3.97529691e-01 -7.65229583e-01
3.58254522e-01 -1.10669839e+00 -2.16327578e-01 -1.20544982e+00
-2.61032015e-01 -5.15921116e-01 -5.78477932e-03 1.68962359e-01
-6.68966398e-02 2.07837373e-01 7.32140020e-02 1.41231656e-01
-7.04860210e-01 1.20780148e-01 8.66682053e-01 -2.27791533e-01
-3.43659371e-01 5.33618808e-01 -6.27494335e-01 1.09200370e+00
7.87745416e-01 -9.69293177e-01 1.08495586e-01 1.76544219e-01
9.70339715e-01 -5.09061575e-01 -4.22625691e-01 -6.48853540e-01
4.15620208e-01 -2.05301613e-01 -1.41751230e-01 -6.90526783e-01
-4.02311087e-02 -6.86120689e-01 -3.13336462e-01 2.19784066e-01
-5.35130620e-01 1.14157878e-01 -1.27064973e-01 3.12022507e-01
-4.90583628e-01 -7.74618149e-01 6.38195992e-01 -2.93792374e-02
-3.50275248e-01 4.91585396e-02 -1.13518417e+00 4.00726080e-01
8.81179512e-01 1.51359728e-02 -2.66305864e-01 -5.02773464e-01
-4.42912579e-01 2.21496388e-01 1.19370416e-01 3.83594841e-01
2.38591939e-01 -8.23731780e-01 -7.05508232e-01 -6.98136836e-02
1.21229291e-01 -1.93691671e-01 -2.49243930e-01 9.41359699e-01
-5.97578347e-01 1.30395457e-01 2.14368165e-01 -2.97953755e-01
-1.43856072e+00 2.12417230e-01 -2.71182805e-01 -2.19084546e-01
-2.44659394e-01 8.19653332e-01 -2.35166878e-01 -1.52494878e-01
3.69436234e-01 -3.19376469e-01 -7.72599757e-01 1.36997604e+00
6.25015199e-01 5.00446260e-01 3.94719481e-01 -8.76861274e-01
-2.91556031e-01 5.77706397e-01 -1.48248032e-01 1.29483521e-01
1.31527233e+00 -2.39846975e-01 -4.88884807e-01 7.78332174e-01
1.54185963e+00 5.19955218e-01 8.40881001e-03 -3.30457389e-01
5.62455237e-01 -3.22520882e-01 3.89695540e-03 -4.08985233e-03
-7.98415124e-01 7.30754256e-01 1.01492397e-01 9.81735587e-01
6.15791857e-01 -1.05444446e-01 9.97482300e-01 5.49133003e-01
1.39738321e-01 -1.53503764e+00 -1.41173244e-01 1.03522110e+00
2.81363130e-01 -1.09726024e+00 1.95554689e-01 -5.66977382e-01
-5.94641387e-01 1.01428115e+00 4.13110033e-02 -1.48907036e-01
1.23494697e+00 2.59988099e-01 2.75116265e-02 2.15535715e-01
-3.31825495e-01 -6.96446002e-01 1.96969986e-01 -8.33753794e-02
5.13012528e-01 -1.32688303e-02 -9.22714412e-01 7.70026147e-01
-6.49114609e-01 -6.81715906e-01 6.87820971e-01 1.05038166e+00
-8.45108569e-01 -1.12761533e+00 -5.42806387e-01 1.28276849e+00
-1.13852453e+00 -3.13471317e-01 -3.57458711e-01 4.88315910e-01
1.73431233e-01 1.18842864e+00 1.65395752e-01 -7.09629118e-01
6.10314347e-02 2.49573514e-01 -1.54526278e-01 -1.00077784e+00
-1.44008982e+00 7.39542469e-02 2.99717963e-01 2.20788702e-01
-4.19016331e-01 -9.05628741e-01 -1.34360480e+00 -4.19282317e-01
-7.93046951e-01 7.56312549e-01 1.14944816e+00 1.18520641e+00
6.71489686e-02 1.13397107e-01 1.08442891e+00 -6.77553058e-01
-2.10059747e-01 -1.12768579e+00 -6.12755537e-01 7.26440489e-01
1.37916729e-01 -4.23983783e-01 -7.39379048e-01 -8.56700540e-03] | [10.798110961914062, 7.16273832321167] |
bd43d382-7758-4480-973e-394556b6314c | unsupervised-hyperalignment-for-multilingual | 1811.01124 | null | https://arxiv.org/abs/1811.01124v3 | https://arxiv.org/pdf/1811.01124v3.pdf | Unsupervised Hyperalignment for Multilingual Word Embeddings | We consider the problem of aligning continuous word representations, learned in multiple languages, to a common space. It was recently shown that, in the case of two languages, it is possible to learn such a mapping without supervision. This paper extends this line of work to the problem of aligning multiple languages to a common space. A solution is to independently map all languages to a pivot language. Unfortunately, this degrades the quality of indirect word translation. We thus propose a novel formulation that ensures composable mappings, leading to better alignments. We evaluate our method by jointly aligning word vectors in eleven languages, showing consistent improvement with indirect mappings while maintaining competitive performance on direct word translation. | ['Marco Cuturi', 'Edouard Grave', 'Armand Joulin', 'Jean Alaux'] | 2018-11-02 | null | null | null | null | ['multilingual-word-embeddings'] | ['methodology'] | [ 3.72669280e-01 6.79876655e-02 -5.22075593e-01 -3.59894127e-01
-1.22044885e+00 -9.59940493e-01 7.43924320e-01 -3.83546464e-02
-6.74564838e-01 9.36650455e-01 4.55128610e-01 -5.96925855e-01
2.63386369e-01 -7.24975824e-01 -7.21221149e-01 -3.69678319e-01
4.16496933e-01 7.40708292e-01 -1.49376601e-01 -5.38291574e-01
4.40912470e-02 4.11100209e-01 -8.25409889e-01 2.79805452e-01
7.48937786e-01 -6.95796087e-02 3.19273680e-01 5.55368662e-01
-5.83094597e-01 -5.76808341e-02 -4.49649692e-01 -5.62533975e-01
5.08494496e-01 -6.03589118e-01 -1.04582679e+00 5.67940958e-02
8.19431901e-01 2.33850613e-01 6.27868921e-02 1.03215098e+00
1.08397044e-01 -1.10460617e-01 5.00492692e-01 -1.15203881e+00
-8.99637043e-01 6.75576031e-01 -3.83452088e-01 1.05460502e-01
2.40818858e-01 -3.57924700e-01 1.50726092e+00 -8.41502845e-01
7.03969121e-01 1.08112121e+00 5.03742397e-01 5.80127537e-01
-1.66646731e+00 -6.42808437e-01 2.69007683e-01 -2.95959860e-01
-1.44973791e+00 -5.24740219e-01 3.97132099e-01 -4.27361459e-01
1.36833549e+00 1.04601957e-01 2.48628587e-01 1.07577753e+00
3.99516732e-01 3.08255851e-01 1.01052153e+00 -7.29985058e-01
-1.94480032e-01 2.68034548e-01 9.22923535e-02 5.60037136e-01
5.78005612e-01 -6.31110668e-02 -5.09700596e-01 -2.56760687e-01
7.92165518e-01 -2.69595891e-01 -1.48231030e-01 -4.82396126e-01
-1.48880601e+00 9.38489854e-01 4.08839971e-01 6.00288749e-01
-6.21028990e-02 1.55521229e-01 1.45053834e-01 6.64765000e-01
5.23989499e-01 8.73649776e-01 -4.17414337e-01 1.88563630e-01
-6.81452215e-01 2.35098302e-01 8.59796822e-01 1.05011463e+00
8.85142922e-01 7.85296485e-02 1.25780970e-01 6.30974054e-01
2.10130975e-01 7.51553118e-01 5.15920699e-01 -5.27118564e-01
9.52335060e-01 2.65502900e-01 2.38184169e-01 -7.15489984e-01
-2.62840003e-01 -5.51828086e-01 -5.00373065e-01 -4.02882621e-02
2.82745361e-01 -1.57646403e-01 -6.37558341e-01 2.24380350e+00
-4.11180891e-02 8.12249109e-02 3.38645428e-01 7.82060623e-01
1.95614576e-01 8.18377912e-01 -8.50600377e-02 -6.61642104e-02
1.10777020e+00 -1.05078661e+00 -6.44427180e-01 -5.85825741e-01
7.98558533e-01 -1.20839167e+00 1.03736794e+00 7.82877430e-02
-1.03975606e+00 -5.45907259e-01 -1.14375794e+00 -1.58396050e-01
-1.29709214e-01 -9.02846605e-02 6.23611212e-01 5.42758584e-01
-1.25512719e+00 5.18801868e-01 -8.63545835e-01 -7.00125515e-01
-2.12879285e-01 5.61066985e-01 -7.43052661e-01 9.60560814e-02
-1.24666607e+00 1.44207990e+00 2.22101837e-01 -3.36547673e-01
-3.44081044e-01 -3.69216621e-01 -9.22837436e-01 -1.43904150e-01
7.18439929e-03 -8.06390703e-01 1.27894568e+00 -1.14452899e+00
-1.43283761e+00 9.10391271e-01 -4.88041639e-01 -3.69761646e-01
1.52397215e-01 -3.03087801e-01 -3.94241929e-01 -4.79978532e-01
3.90268177e-01 6.44495964e-01 4.67805386e-01 -1.05751705e+00
-4.86410320e-01 -1.86421126e-01 1.08155496e-01 4.47383761e-01
-6.00123525e-01 1.47131354e-01 -5.04011989e-01 -6.90855861e-01
6.99678436e-02 -1.22396243e+00 -4.32424188e-01 -2.30833322e-01
-2.57933021e-01 -8.89245942e-02 -2.87140720e-03 -6.96786940e-01
1.01342285e+00 -1.97213590e+00 9.78337824e-01 2.72686601e-01
-6.91025108e-02 6.23063818e-02 -7.67898917e-01 6.43730581e-01
-2.63834357e-01 1.64062724e-01 -4.24598843e-01 -5.76434731e-01
-6.60180077e-02 6.84038997e-01 -6.96228623e-01 6.21746778e-01
3.74449730e-01 1.10629654e+00 -8.08920681e-01 -2.70483911e-01
-2.53725588e-01 2.91105628e-01 -6.43843889e-01 3.38104069e-01
9.46313664e-02 2.14560330e-01 -1.55541018e-01 1.55051902e-01
4.07613009e-01 -9.14568156e-02 6.79393053e-01 4.57642078e-02
-4.22587730e-02 6.72560096e-01 -1.11620367e+00 1.96842885e+00
-8.95329475e-01 5.86694121e-01 -1.32322013e-01 -1.00022912e+00
1.14044702e+00 4.94447947e-01 3.46250117e-01 -4.66506243e-01
-1.80984065e-01 4.68265593e-01 1.65656861e-02 1.29666612e-01
6.35292113e-01 -5.57278752e-01 -3.29334497e-01 7.09021628e-01
3.30495030e-01 -4.31402296e-01 -1.10890809e-02 -8.66877586e-02
6.52671695e-01 2.38576978e-01 5.14588356e-01 -4.73768383e-01
2.86094636e-01 8.22030231e-02 3.42379600e-01 4.98009384e-01
3.29874843e-01 6.08662724e-01 1.87068447e-01 -3.77823055e-01
-1.33186233e+00 -1.43292987e+00 6.22871779e-02 1.03307307e+00
3.55979204e-02 -5.98723471e-01 -6.90549552e-01 -5.06983161e-01
-6.46880344e-02 6.96044147e-01 -4.46943104e-01 -3.62987556e-02
-1.12101638e+00 -5.44988990e-01 6.92721665e-01 4.92134213e-01
-2.01565415e-01 -7.33490109e-01 -8.09198320e-02 4.34506446e-01
-2.30161190e-01 -1.08509421e+00 -9.04151678e-01 3.88219923e-01
-1.06054044e+00 -6.71573639e-01 -6.95037127e-01 -1.08761275e+00
7.93076813e-01 3.86453718e-01 1.22568917e+00 2.14584526e-02
2.44580120e-01 7.14377090e-02 7.12802336e-02 3.43141705e-03
-7.08370268e-01 7.31132627e-01 4.39699948e-01 -1.32716805e-01
3.01664710e-01 -4.31727797e-01 2.13047311e-01 1.64550170e-01
-1.00359750e+00 2.80275524e-01 5.12489438e-01 9.16698098e-01
5.95858693e-01 -8.12035620e-01 2.86315620e-01 -8.26978624e-01
7.56692350e-01 -5.32119572e-01 -7.42048383e-01 4.49084848e-01
-7.01737583e-01 5.41976929e-01 8.13118160e-01 -4.38146442e-01
-5.75185776e-01 1.10810280e-01 -9.91608799e-02 -6.11675307e-02
1.39654294e-01 4.44642872e-01 -8.24289694e-02 -7.63365701e-02
7.58794010e-01 1.25848159e-01 -2.16452545e-03 -5.68911791e-01
7.88131952e-01 6.06389105e-01 5.04648209e-01 -7.87336051e-01
1.18981493e+00 1.26268864e-01 -8.71759281e-02 -5.60076296e-01
-7.14516342e-01 -4.12076741e-01 -8.35337996e-01 5.67510247e-01
7.24370539e-01 -1.01053786e+00 2.79740155e-01 -1.89037681e-01
-1.59367573e+00 -2.33318418e-01 -2.53418088e-01 5.14657021e-01
-6.46742761e-01 2.09304914e-01 -3.02906245e-01 -7.23493770e-02
-2.21161410e-01 -1.12370384e+00 1.05325162e+00 -2.19389185e-01
-8.50783885e-01 -1.29390419e+00 7.20872104e-01 2.71689203e-02
3.65037680e-01 -3.27592283e-01 8.81823242e-01 -8.32087517e-01
-3.54875892e-01 6.54779896e-02 7.09719062e-02 2.75510937e-01
6.68438673e-01 -3.71174484e-01 -4.84704494e-01 -5.71957767e-01
-2.63360679e-01 -4.06476595e-02 6.09801590e-01 -1.13268055e-01
2.98432142e-01 -3.49158406e-01 -2.80776322e-01 7.74845719e-01
1.55833876e+00 -1.40115306e-01 2.72253603e-01 4.97732967e-01
7.72326589e-01 4.74679798e-01 2.53564656e-01 -2.82331616e-01
2.12714374e-01 1.26266456e+00 -9.48287323e-02 -3.40250254e-01
-2.09737808e-01 -2.88565397e-01 5.78021646e-01 1.45670378e+00
4.93417978e-02 -2.69556493e-01 -1.04130626e+00 8.77343059e-01
-1.76812077e+00 -5.45735538e-01 9.19345617e-02 2.15121818e+00
1.10118401e+00 -9.37873200e-02 -1.37617603e-01 -4.39584553e-01
5.58828354e-01 8.04742053e-02 -2.09675692e-02 -8.89246762e-01
-2.71938473e-01 7.24127531e-01 7.67510891e-01 1.27937675e+00
-8.33913326e-01 1.28359675e+00 7.36823750e+00 2.69628704e-01
-1.11123991e+00 4.48701411e-01 1.24834836e-01 -9.47138965e-02
-9.75202382e-01 2.19244555e-01 -1.00744808e+00 -4.24328074e-02
1.01798809e+00 -5.31203747e-01 6.62428200e-01 5.54438949e-01
-1.77797586e-01 5.97304165e-01 -1.32777011e+00 7.39595175e-01
2.64856160e-01 -1.30754781e+00 4.88705784e-01 -1.11924978e-02
1.09712732e+00 2.48635337e-01 -6.32035825e-03 1.08007111e-01
5.94892442e-01 -1.20755994e+00 4.57999974e-01 1.81891650e-01
9.02132034e-01 -7.41255164e-01 6.02497339e-01 1.55018657e-01
-1.12244391e+00 4.65697914e-01 -4.25103188e-01 -2.12737128e-01
3.63449752e-01 8.74786377e-02 -9.04078066e-01 7.92640686e-01
9.04511884e-02 6.48330808e-01 -4.60695833e-01 6.34612024e-01
-3.52154672e-01 4.17596936e-01 -2.08061114e-01 1.00962877e-01
3.67314190e-01 -4.81367260e-01 4.70183432e-01 1.33633912e+00
6.83578134e-01 -2.57401675e-01 4.48624492e-01 6.89631879e-01
-1.28469989e-01 4.31339085e-01 -1.03898001e+00 -1.42031103e-01
2.84273565e-01 9.44241703e-01 -4.12737966e-01 -3.14133078e-01
-7.02706397e-01 1.19785535e+00 7.13967800e-01 3.89175683e-01
-5.97307086e-01 -2.76461214e-01 1.21303248e+00 -1.89382099e-02
1.02793597e-01 -7.47698843e-01 -4.09799546e-01 -1.35436475e+00
6.20867731e-03 -9.86633539e-01 1.97383463e-01 -4.10122484e-01
-1.23749697e+00 9.14556563e-01 3.98955084e-02 -1.01880395e+00
-4.38204825e-01 -4.71859694e-01 -3.42326164e-01 1.48469853e+00
-1.30185556e+00 -1.06963181e+00 3.39569777e-01 3.73295605e-01
6.20288491e-01 -5.34368098e-01 1.20366514e+00 1.98019058e-01
-1.23723321e-01 7.79789507e-01 7.95027241e-02 -8.14430267e-02
9.86304224e-01 -1.27326739e+00 1.19384754e+00 9.60403025e-01
1.01422131e+00 9.77469087e-01 6.71282589e-01 -6.46142900e-01
-1.43249333e+00 -1.19101393e+00 1.59943318e+00 -6.19791687e-01
9.18930471e-01 -4.72082376e-01 -1.03291643e+00 1.07321239e+00
6.46786213e-01 -2.82542616e-01 8.31520319e-01 2.88679689e-01
-5.77290952e-01 4.75122184e-02 -5.53777814e-01 8.36253941e-01
1.05693007e+00 -8.66924882e-01 -9.90326345e-01 4.04211491e-01
9.97002423e-01 -3.33396405e-01 -6.64267361e-01 -4.76544052e-02
2.70714492e-01 -2.86178496e-02 9.30403352e-01 -1.22285306e+00
3.15957159e-01 -5.06063581e-01 -5.52156031e-01 -1.66487205e+00
-4.03258532e-01 -6.55316532e-01 2.62201399e-01 1.03151512e+00
7.69137263e-01 -8.20202112e-01 4.04063195e-01 1.54595524e-01
1.61447562e-02 -5.15446126e-01 -1.19323266e+00 -1.17753470e+00
6.81734562e-01 -2.54341632e-01 6.61515415e-01 1.07606220e+00
-2.90311757e-04 6.98042810e-01 -5.12060583e-01 2.78866082e-01
2.57618010e-01 2.91375250e-01 8.75113249e-01 -8.91422272e-01
-5.37403584e-01 -4.17092472e-01 -3.62024546e-01 -1.16752815e+00
5.58209896e-01 -1.50436270e+00 1.06183216e-01 -1.50339103e+00
1.36554062e-01 -4.43675339e-01 -4.75741625e-01 6.08664572e-01
-2.15819523e-01 4.83504981e-01 3.00977468e-01 1.88193351e-01
-5.20645306e-02 2.12490022e-01 9.29774940e-01 -2.11729288e-01
-7.71711394e-02 -3.55893552e-01 -9.52173889e-01 2.25894257e-01
9.72491860e-01 -7.29682565e-01 -3.74022007e-01 -9.85579550e-01
3.86699378e-01 -2.34080136e-01 -1.07880466e-01 -6.61391854e-01
8.05121288e-02 -4.84701782e-01 -8.29285532e-02 5.44907562e-02
2.47032553e-01 -5.86159527e-01 3.09115320e-01 3.98085117e-01
-5.48184991e-01 6.41224444e-01 3.45210552e-01 2.24392712e-01
-2.80322820e-01 -3.43146831e-01 6.53567433e-01 3.18792015e-02
-5.02688706e-01 9.55918133e-02 -1.31168723e-01 1.39644608e-01
8.11276615e-01 1.03503592e-01 4.45532501e-02 -1.56050220e-01
-5.61748981e-01 -1.19903728e-01 7.62428105e-01 7.89557219e-01
2.63506711e-01 -1.68605042e+00 -1.03638387e+00 3.21880519e-01
1.59542590e-01 -5.61930478e-01 -6.75558925e-01 4.61314231e-01
-4.20213938e-01 6.28664434e-01 -3.22362334e-01 -4.63667035e-01
-1.33060586e+00 3.59521896e-01 3.12854648e-01 -2.63890535e-01
-4.39526111e-01 6.66667998e-01 2.11066529e-01 -8.36840868e-01
-2.36351296e-01 -1.64601415e-01 4.18484583e-02 -9.53753367e-02
3.15713704e-01 -1.07173629e-01 1.38786986e-01 -9.61405754e-01
-5.10949314e-01 1.04615366e+00 -1.44591093e-01 -6.50809705e-01
1.31231010e+00 -6.95427507e-03 -3.12855810e-01 5.56183636e-01
1.33130586e+00 4.16209549e-01 -6.92593873e-01 -5.53598404e-01
2.64722735e-01 -4.14831579e-01 -4.25993115e-01 -1.82610169e-01
-8.36666286e-01 7.64239311e-01 4.06638116e-01 -1.83773950e-01
7.27837801e-01 6.44244850e-02 6.52653456e-01 7.38554955e-01
4.43381488e-01 -7.41058290e-01 -1.80512950e-01 7.59433866e-01
8.40027213e-01 -1.03255641e+00 5.48678823e-02 -1.44000575e-01
-5.49859524e-01 1.14518666e+00 4.69893903e-01 -3.47682804e-01
1.49518788e-01 3.54289502e-01 3.82932186e-01 2.63489097e-01
-7.73407340e-01 -2.37618610e-02 5.22715390e-01 5.32410681e-01
7.31016099e-01 3.24720204e-01 -3.40569258e-01 2.32553616e-01
-4.03127849e-01 -3.06788623e-01 5.25446653e-01 7.22073853e-01
-2.29791075e-01 -2.04724884e+00 -3.80692035e-01 -7.90414773e-03
-2.91476309e-01 -3.79986048e-01 -5.23568511e-01 7.56738305e-01
-2.92519834e-02 7.04853833e-01 2.58977532e-01 -1.00401573e-01
3.04112583e-01 2.98207402e-01 6.02264643e-01 -1.16638720e+00
-5.90185285e-01 -8.07572976e-02 2.24008888e-01 -4.57670510e-01
-2.35493422e-01 -5.55094779e-01 -1.03474927e+00 -1.44931689e-01
-2.23304078e-01 4.00949508e-01 6.87134802e-01 9.93889630e-01
1.61217704e-01 2.94558525e-01 4.71893817e-01 -5.65223515e-01
-6.93093598e-01 -5.44907391e-01 -2.15032101e-01 5.21577775e-01
5.11610150e-01 -3.37393492e-01 1.76682249e-02 8.05337876e-02] | [11.169116020202637, 10.145537376403809] |
c0568466-bb17-4cd1-9a45-8b9fb65891a9 | enriched-long-term-recurrent-convolutional | 1805.08417 | null | http://arxiv.org/abs/1805.08417v1 | http://arxiv.org/pdf/1805.08417v1.pdf | Enriched Long-term Recurrent Convolutional Network for Facial Micro-Expression Recognition | Facial micro-expression (ME) recognition has posed a huge challenge to
researchers for its subtlety in motion and limited databases. Recently,
handcrafted techniques have achieved superior performance in micro-expression
recognition but at the cost of domain specificity and cumbersome parametric
tunings. In this paper, we propose an Enriched Long-term Recurrent
Convolutional Network (ELRCN) that first encodes each micro-expression frame
into a feature vector through CNN module(s), then predicts the micro-expression
by passing the feature vector through a Long Short-term Memory (LSTM) module.
The framework contains two different network variants: (1) Channel-wise
stacking of input data for spatial enrichment, (2) Feature-wise stacking of
features for temporal enrichment. We demonstrate that the proposed approach is
able to achieve reasonably good performance, without data augmentation. In
addition, we also present ablation studies conducted on the framework and
visualizations of what CNN "sees" when predicting the micro-expression classes. | ['Raphael C. -W. Phan', 'Weiyao Lin', 'Huai-Qian Khor', 'John See'] | 2018-05-22 | null | null | null | null | ['micro-expression-recognition'] | ['computer-vision'] | [ 1.99378401e-01 -2.16117740e-01 -9.64071453e-02 -7.89011240e-01
-5.80096543e-01 -1.65615216e-01 6.25525594e-01 -4.66600567e-01
-4.41595703e-01 5.59243023e-01 2.40848795e-01 -5.68752848e-02
1.49564788e-01 -4.62229401e-01 -6.03026688e-01 -9.24649656e-01
-4.06498015e-01 -3.10465902e-01 -3.50338906e-01 -2.76739895e-01
-7.42706880e-02 9.47118938e-01 -1.78250575e+00 6.14746749e-01
1.84094906e-01 1.39359927e+00 -5.43473102e-02 6.56174779e-01
-2.25212678e-01 1.33627975e+00 -5.21415412e-01 -2.33970225e-01
-1.41635254e-01 -4.01815534e-01 -7.62830138e-01 -7.88245723e-02
5.67491591e-01 -2.15264961e-01 -2.45808885e-01 7.66940773e-01
5.20259440e-01 6.87746331e-02 3.19552660e-01 -1.24834168e+00
-3.86103541e-01 5.81313930e-02 -7.16712713e-01 4.67889607e-02
1.83751389e-01 2.97063570e-02 7.30311155e-01 -1.01259065e+00
8.56397331e-01 1.03002918e+00 7.28025734e-01 6.22511506e-01
-1.04285240e+00 -7.81213880e-01 -1.84470881e-02 2.78338313e-01
-1.43950963e+00 -6.38007939e-01 7.69505918e-01 -4.93170768e-01
1.33544290e+00 1.46254331e-01 7.25563765e-01 1.41196108e+00
1.42329380e-01 8.76257300e-01 9.96446550e-01 -1.79917544e-01
6.82491735e-02 1.19193256e-01 1.58041809e-02 8.96084845e-01
-6.53266013e-01 8.85680690e-02 -7.24241674e-01 -9.63296089e-03
8.17099273e-01 -1.90076716e-02 -6.33925050e-02 -1.34357452e-01
-7.87678838e-01 6.25618219e-01 5.53774416e-01 6.72065914e-01
-3.98706377e-01 5.11591077e-01 5.98749101e-01 3.54782671e-01
6.60172164e-01 2.53547698e-01 -6.20416999e-01 -5.06327212e-01
-1.23409569e+00 2.49215394e-01 4.10713613e-01 5.93847394e-01
9.32696462e-01 3.73376727e-01 -3.35337460e-01 1.05824316e+00
6.70745969e-03 1.19338326e-01 8.00798178e-01 -7.72270679e-01
6.57113940e-02 5.03770769e-01 4.03385684e-02 -1.20755100e+00
-6.78956389e-01 -3.61390531e-01 -8.65364194e-01 4.75153029e-01
1.88990220e-01 -4.34741169e-01 -9.30612445e-01 2.08789587e+00
-2.94108987e-02 3.72649848e-01 -6.98634097e-03 9.99853194e-01
8.40129793e-01 7.75777936e-01 4.35303718e-01 -1.07552297e-01
1.21197808e+00 -9.93846059e-01 -8.68870080e-01 -9.57845598e-02
1.04775655e+00 -3.24794054e-01 1.05729163e+00 7.38745704e-02
-9.12091017e-01 -5.29301941e-01 -8.95894527e-01 -1.64778829e-01
-6.16972864e-01 5.52275002e-01 7.49417007e-01 5.02844691e-01
-1.23093343e+00 7.72540867e-01 -6.71596408e-01 -3.21251214e-01
5.48375845e-01 6.36509359e-01 -8.34612787e-01 5.02208591e-01
-1.08854842e+00 1.01863372e+00 -2.08483145e-01 5.13917148e-01
-6.32907569e-01 -7.86656141e-01 -1.04799819e+00 2.76361927e-02
-9.43387449e-02 -5.26218057e-01 1.18966305e+00 -1.85246289e+00
-2.01252532e+00 1.02663898e+00 -5.40214837e-01 -3.31433773e-01
2.77288437e-01 -3.60674053e-01 -5.19124746e-01 -1.37135191e-02
-3.60017449e-01 8.35435331e-01 9.72290277e-01 -6.85413361e-01
-5.03832638e-01 -5.73420584e-01 -6.51684403e-02 -9.70519800e-03
-3.97764862e-01 4.04750645e-01 -3.10029626e-01 -7.34291017e-01
-2.42600888e-01 -8.20641577e-01 -2.75468707e-01 1.41448185e-01
-1.22426115e-01 -1.35407597e-01 1.21162450e+00 -5.22372782e-01
1.28563631e+00 -2.36268282e+00 1.01516306e-01 1.50793195e-01
7.48613030e-02 3.68389666e-01 -2.96939582e-01 -2.14847028e-02
-4.73780870e-01 3.45093794e-02 -6.03666827e-02 -5.46591640e-01
-1.18030466e-01 1.47230521e-01 -2.89495260e-01 4.96711075e-01
4.08747405e-01 1.20159674e+00 -5.94767451e-01 -2.30796412e-01
9.34133753e-02 8.67818296e-01 -2.84721881e-01 2.66280472e-01
-1.42589912e-01 3.66467386e-01 -1.84520319e-01 7.60158479e-01
4.90101367e-01 -1.40277639e-01 1.15299225e-01 -5.18859148e-01
-2.35785186e-01 -2.25528002e-01 -6.98903263e-01 1.67925191e+00
-6.90990865e-01 1.10943878e+00 2.90833652e-01 -9.03999984e-01
9.43386257e-01 3.65150154e-01 4.01055396e-01 -7.91453958e-01
2.65449762e-01 5.92505075e-02 -3.81570488e-01 -8.32469106e-01
5.34457982e-01 -2.92961836e-01 2.74360981e-02 2.35266462e-01
3.42851102e-01 5.89459598e-01 -2.12273329e-01 -2.57471025e-01
9.77336228e-01 4.70146626e-01 1.58469260e-01 -1.38109341e-01
4.34287190e-01 -3.56093824e-01 5.18029034e-01 2.01223224e-01
-1.83560967e-01 4.14722800e-01 6.75391376e-01 -7.16157436e-01
-8.36709261e-01 -4.89724159e-01 8.56301039e-02 1.48248863e+00
-3.93103451e-01 -4.52452481e-01 -6.35272980e-01 -4.99929130e-01
-1.67144299e-01 3.08899045e-01 -1.20540273e+00 -9.55989137e-02
-5.64943254e-01 -6.08074486e-01 1.10773277e+00 7.31694400e-01
4.92069602e-01 -1.16277361e+00 -9.80211675e-01 9.62683186e-02
1.15411520e-01 -9.97951984e-01 -1.28176764e-01 2.91911751e-01
-7.16054797e-01 -5.11791110e-01 -9.12498653e-01 -6.58223152e-01
5.22501945e-01 -1.03394620e-01 7.79662251e-01 -2.25441024e-01
-4.85706985e-01 1.72427259e-02 -2.14702442e-01 -1.56073421e-01
3.12155485e-01 -8.80306289e-02 -1.38270095e-01 5.87294221e-01
4.63010103e-01 -7.61292100e-01 -5.18693924e-01 6.51459992e-02
-6.96855009e-01 -6.16855435e-02 6.43892527e-01 8.46513152e-01
5.64796805e-01 -4.96873707e-01 3.64950269e-01 -8.63423526e-01
6.07031524e-01 -3.69251043e-01 -4.00263101e-01 1.77094951e-01
-1.15305439e-01 5.95988669e-02 5.41699708e-01 -4.54319447e-01
-1.18226755e+00 4.12955552e-01 -3.82083446e-01 -8.92500997e-01
-2.80174822e-01 6.16557896e-01 -7.32130036e-02 -3.13860267e-01
4.84747201e-01 2.00373784e-01 1.91220865e-01 -5.41158974e-01
4.06593978e-01 5.08098245e-01 5.77005625e-01 -3.31774890e-01
8.29115883e-02 4.85702515e-01 6.75271899e-02 -1.14318001e+00
-6.22689486e-01 -1.57412320e-01 -8.15619767e-01 -2.91985720e-01
8.90914917e-01 -9.07460213e-01 -7.22504437e-01 6.28903687e-01
-1.10383236e+00 -5.76185822e-01 -1.21724010e-01 1.80042744e-01
-4.72448021e-01 -7.89435580e-02 -6.69660568e-01 -7.65513420e-01
-2.69043416e-01 -1.01714313e+00 1.09209108e+00 2.24117219e-01
-4.95177776e-01 -9.69880223e-01 1.23227648e-01 -1.20661952e-01
6.01747036e-01 6.09020352e-01 6.52302504e-01 -1.51833326e-01
-5.82319386e-02 -3.36591244e-01 -3.29998404e-01 2.49628574e-01
-8.74593332e-02 4.36628044e-01 -1.50398636e+00 1.97079126e-02
-2.95261413e-01 -5.38863778e-01 7.97835171e-01 6.23494864e-01
1.53905141e+00 -3.46161604e-01 -2.76058286e-01 1.06428576e+00
1.06945682e+00 7.31634125e-02 7.95534492e-01 2.97947884e-01
6.74803793e-01 7.91734278e-01 4.43437248e-01 5.48078775e-01
2.41666455e-02 1.02126253e+00 1.47123560e-01 -5.53833902e-01
9.68044251e-02 -1.56048402e-01 5.02858460e-01 1.74950808e-01
-3.13772112e-01 2.29029983e-01 -6.16719723e-01 3.59743834e-01
-1.91128719e+00 -1.11202908e+00 1.60800949e-01 1.53974926e+00
5.05719304e-01 -4.21341509e-01 2.18174621e-01 -1.20280698e-01
2.65710145e-01 4.04582769e-01 -4.19831246e-01 -9.36312377e-01
-3.93221617e-01 4.71756727e-01 9.14984047e-02 3.02485794e-01
-1.13374817e+00 1.11789751e+00 6.54702187e+00 7.42483556e-01
-1.92595172e+00 6.15085252e-02 8.73606384e-01 -3.52685124e-01
1.02669872e-01 -3.78505319e-01 -5.87677240e-01 1.58587918e-01
1.07549763e+00 1.39462128e-01 -8.53617638e-02 1.02644873e+00
2.92470425e-01 6.07568249e-02 -9.45252240e-01 1.33989203e+00
8.10627788e-02 -1.52422035e+00 8.99639577e-02 -7.45162293e-02
3.24231237e-01 -1.78512514e-01 3.51724148e-01 3.09154153e-01
-1.02272786e-01 -1.49593890e+00 5.66101134e-01 6.57457471e-01
1.11387193e+00 -8.52474868e-01 7.35392690e-01 -2.66236924e-02
-1.13744438e+00 -1.45848140e-01 -2.37849638e-01 -2.90276289e-01
7.68969432e-02 3.28502834e-01 -5.71287990e-01 1.88179910e-01
8.15617204e-01 9.17390764e-01 -4.99847859e-01 5.05375326e-01
-1.22948289e-01 4.90146995e-01 -1.53225526e-01 -1.10313855e-01
6.07151031e-01 -1.17263887e-02 1.26248106e-01 1.81664765e+00
4.26938325e-01 1.51146024e-01 -3.93491536e-01 7.90869474e-01
-2.68486664e-02 1.76898673e-01 -7.41824865e-01 -1.13841243e-01
-1.41278073e-01 1.64038420e+00 -2.04165041e-01 -1.88273609e-01
-3.33086073e-01 1.27758825e+00 6.15663230e-01 6.19551599e-01
-6.87646508e-01 -3.99387121e-01 1.11128950e+00 7.13986997e-03
2.00912938e-01 -8.90429318e-02 -1.04619779e-01 -1.00982404e+00
-1.77723482e-01 -5.29458702e-01 3.89028102e-01 -9.86509562e-01
-9.08981621e-01 1.00878036e+00 -3.60188872e-01 -9.08764184e-01
-6.17927074e-01 -7.56120980e-01 -5.52181661e-01 9.42276061e-01
-1.26620257e+00 -1.43910432e+00 -4.12733912e-01 8.32254410e-01
3.02627474e-01 -1.97830617e-01 1.26301134e+00 4.56306487e-01
-8.23254645e-01 9.88754213e-01 -2.07066327e-01 3.55396986e-01
5.87419808e-01 -8.22577357e-01 -6.03312105e-02 3.60550851e-01
2.00416341e-01 5.88009298e-01 5.67391336e-01 -4.19324823e-02
-1.26336014e+00 -1.32057440e+00 9.68662798e-01 -1.17720976e-01
5.47204912e-01 -4.71447349e-01 -8.04470837e-01 8.78848851e-01
5.32646850e-02 4.24699277e-01 9.66011822e-01 2.66767085e-01
-5.44043481e-01 -3.59882802e-01 -1.11793327e+00 5.74021459e-01
7.14462221e-01 -7.94188440e-01 -1.23792641e-01 -1.36322111e-01
2.80250162e-01 -2.89107442e-01 -7.00580359e-01 5.36919653e-01
1.06945574e+00 -1.07886291e+00 7.32796907e-01 -7.98848450e-01
6.02494359e-01 -1.90257445e-01 -1.73682764e-01 -1.10529149e+00
-3.64460886e-01 -5.05945444e-01 -1.10859722e-01 8.99549723e-01
5.28659105e-01 -2.87916511e-01 9.07066166e-01 6.69455528e-01
8.07368457e-02 -1.21363783e+00 -1.03958750e+00 -4.00863200e-01
-2.11530358e-01 -5.24047554e-01 4.00473684e-01 1.01234567e+00
2.64578164e-01 3.94586235e-01 -7.34606147e-01 -1.47900462e-01
-5.96715175e-02 2.80308515e-01 8.26045692e-01 -5.54542124e-01
-1.12767540e-01 -5.95212758e-01 -7.46735692e-01 -9.69438672e-01
4.58667785e-01 -5.63946307e-01 -1.17286764e-01 -9.10963237e-01
1.04288459e-01 -7.21590072e-02 -3.61160457e-01 8.45429540e-01
2.95024276e-01 5.48268557e-01 -1.24134207e-02 -1.65017426e-01
-4.72341895e-01 5.91059566e-01 8.78434360e-01 3.98757197e-02
-3.25678408e-01 -2.60649353e-01 -3.18522811e-01 7.72647023e-01
6.72824562e-01 -3.91026400e-03 -2.77306259e-01 -4.01019156e-01
1.15708441e-01 1.33608386e-01 4.25675362e-01 -8.59038413e-01
1.74924523e-01 4.49131168e-02 6.98952913e-01 -3.91925991e-01
6.80844903e-01 -5.37636280e-01 2.68586636e-01 1.65071800e-01
-4.59885716e-01 7.64065906e-02 6.38539314e-01 7.26069808e-02
-6.37799323e-01 2.67884105e-01 9.34561968e-01 5.39901033e-02
-1.07707596e+00 4.13006693e-01 -7.14572608e-01 -6.79670513e-01
9.83469546e-01 -3.93366456e-01 1.08873934e-01 -5.97413003e-01
-1.01962852e+00 -1.50369480e-01 1.30357832e-01 4.67362136e-01
6.86209917e-01 -1.39870322e+00 -4.48327810e-01 3.63226295e-01
1.13859832e-01 -6.17034018e-01 4.70209897e-01 8.22367072e-01
-3.05236399e-01 3.83035988e-01 -5.44249713e-01 -5.31244516e-01
-1.71389985e+00 1.93554044e-01 7.13817835e-01 -1.80209070e-01
-4.46288645e-01 1.19530928e+00 3.16324443e-01 -3.75345260e-01
2.67388761e-01 -1.67200774e-01 -4.85271394e-01 3.28989208e-01
8.03482533e-01 3.21331695e-02 -3.11499368e-02 -1.09500408e+00
-3.86741370e-01 7.77279556e-01 3.71928699e-02 -1.35952964e-01
1.47255969e+00 7.66075402e-02 -2.72563756e-01 5.27662754e-01
1.75472867e+00 -3.32414091e-01 -1.23748326e+00 -1.92532334e-02
1.33535773e-01 -3.32047731e-01 1.12551674e-01 -7.99448192e-01
-1.29094028e+00 1.21913314e+00 7.67999768e-01 -4.95391011e-01
1.40816474e+00 -1.36385217e-01 5.32768309e-01 3.58057916e-01
2.10012845e-03 -9.14257348e-01 -2.51499955e-02 5.81736803e-01
8.78406525e-01 -8.42567265e-01 -4.30149972e-01 4.06600647e-02
-6.86083257e-01 1.39468181e+00 6.75324380e-01 -1.09141111e-01
7.90713847e-01 5.04253924e-01 3.54185343e-01 -4.48436677e-01
-9.36036766e-01 -7.33374730e-02 2.01442212e-01 4.03553903e-01
8.11977267e-01 -4.57055122e-02 8.01131204e-02 8.05808604e-01
-1.85841992e-01 4.44538504e-01 -3.94512340e-02 8.14705312e-01
-1.57794163e-01 -7.78938174e-01 8.76928866e-02 2.68812627e-01
-7.27476358e-01 1.22618318e-01 -3.77263457e-01 7.16419220e-01
2.18090639e-01 3.70932728e-01 3.28586251e-01 -7.16559410e-01
2.83051103e-01 1.81168631e-01 4.60239291e-01 -1.67647749e-01
-6.64401174e-01 3.96720506e-02 2.18767881e-01 -1.14809287e+00
-5.47539175e-01 -3.47101271e-01 -1.13291645e+00 -1.61292896e-01
1.93088874e-01 -1.71580713e-03 6.04668021e-01 8.99994493e-01
7.25477815e-01 3.03929687e-01 4.76177812e-01 -1.13011122e+00
6.26303703e-02 -9.72740650e-01 -6.58363163e-01 4.72492516e-01
5.38492739e-01 -6.26219392e-01 -1.00104153e-01 1.79582581e-01] | [13.595952987670898, 1.777076005935669] |
db4cb614-1ae2-4edc-a2f1-33baf0bfc52f | unsupervised-inference-of-data-driven | 2210.09559 | null | https://arxiv.org/abs/2210.09559v1 | https://arxiv.org/pdf/2210.09559v1.pdf | Unsupervised Inference of Data-Driven Discourse Structures using a Tree Auto-Encoder | With a growing need for robust and general discourse structures in many downstream tasks and real-world applications, the current lack of high-quality, high-quantity discourse trees poses a severe shortcoming. In order the alleviate this limitation, we propose a new strategy to generate tree structures in a task-agnostic, unsupervised fashion by extending a latent tree induction framework with an auto-encoding objective. The proposed approach can be applied to any tree-structured objective, such as syntactic parsing, discourse parsing and others. However, due to the especially difficult annotation process to generate discourse trees, we initially develop such method to complement task-specific models in generating much larger and more diverse discourse treebanks. | ['Giuseppe Carenini', 'Patrick Huber'] | 2022-10-18 | null | null | null | null | ['discourse-parsing'] | ['natural-language-processing'] | [ 6.00597799e-01 7.98862875e-01 -6.22063987e-02 -4.30020481e-01
-9.11518395e-01 -6.75948143e-01 7.59377658e-01 2.78568000e-01
-1.69159457e-01 1.18244231e+00 6.59775138e-01 -5.22712648e-01
7.15104165e-03 -9.90519285e-01 -2.16810361e-01 -4.06252623e-01
2.30302736e-01 4.66956109e-01 3.00299317e-01 -1.77940696e-01
2.24642649e-01 -2.62040514e-02 -1.41264951e+00 5.00464737e-01
1.21061170e+00 3.47498119e-01 5.78207076e-01 4.99382168e-01
-4.89367396e-01 1.04271460e+00 -8.81998599e-01 -4.92178828e-01
-2.94579804e-01 -7.26595283e-01 -1.26937807e+00 3.61881733e-01
-6.36674464e-02 -1.68278337e-01 2.70426303e-01 9.63475227e-01
4.31171864e-01 2.11395815e-01 6.64082289e-01 -6.91659093e-01
-5.42040825e-01 1.08271682e+00 -4.74946722e-02 2.46959962e-02
4.00173604e-01 -5.76132722e-02 1.12019789e+00 -3.44919711e-01
9.74717736e-01 1.30821013e+00 2.05120847e-01 7.90790856e-01
-1.15567267e+00 -1.14885136e-01 4.19168204e-01 -1.18390415e-02
-6.02635741e-01 -4.39757586e-01 1.06492019e+00 -5.05364120e-01
8.81012380e-01 2.77524889e-01 1.81330964e-01 1.51913261e+00
-2.56511450e-01 7.89171934e-01 1.16148746e+00 -7.72397637e-01
1.44323677e-01 3.95572521e-02 2.23102346e-01 3.35546196e-01
4.45938520e-02 -4.24395084e-01 -3.32595378e-01 3.82713377e-02
5.25577128e-01 -8.43783677e-01 -2.54573226e-01 1.26801729e-01
-1.13447595e+00 9.99380946e-01 -3.38378027e-02 7.72498786e-01
-2.05456585e-01 -3.27481359e-01 6.37524426e-01 2.55051464e-01
8.10643733e-01 7.48327971e-01 -3.78028780e-01 -3.94731075e-01
-6.67282879e-01 3.84768277e-01 1.02790082e+00 7.95011878e-01
4.32024807e-01 1.66795477e-01 -3.93864483e-01 8.56015027e-01
2.70478398e-01 -2.09720746e-01 5.71249068e-01 -1.06295717e+00
9.43149686e-01 7.75500834e-01 -2.32040882e-03 -9.06468391e-01
-3.66880208e-01 -1.01100899e-01 -7.05711901e-01 -5.09136468e-02
6.73378468e-01 -3.09067428e-01 -6.39385760e-01 1.90153766e+00
5.29950380e-01 -3.62058342e-01 4.16293949e-01 6.15594566e-01
1.06812096e+00 8.20684612e-01 3.56990308e-01 -7.25392401e-01
1.59102368e+00 -1.09214091e+00 -1.05333662e+00 -2.73109645e-01
9.16781306e-01 -8.40188682e-01 1.22455335e+00 2.48255864e-01
-1.19356453e+00 -4.93876755e-01 -8.48441780e-01 -3.01706731e-01
-9.73998904e-02 1.03870258e-01 7.31340587e-01 7.76481032e-01
-7.20694005e-01 3.97909462e-01 -6.09959304e-01 -1.89665362e-01
1.54340446e-01 -3.19928415e-02 -2.77619064e-01 2.00362384e-01
-1.23798084e+00 1.06873095e+00 8.94228935e-01 1.53946713e-01
-3.87624264e-01 -2.66413212e-01 -1.23133862e+00 5.96938543e-02
7.16749191e-01 -6.26366854e-01 1.48542929e+00 -9.62039649e-01
-2.09085512e+00 8.52293074e-01 -1.97102934e-01 -3.43130291e-01
5.66643655e-01 -2.32399881e-01 4.04200517e-02 -1.35483816e-01
1.30295113e-01 3.31893533e-01 5.60114622e-01 -1.07032824e+00
-3.63364071e-01 -2.37098396e-01 2.89664567e-01 4.24616694e-01
-4.90766376e-01 3.47244114e-01 -6.86608022e-04 -6.55370772e-01
6.17691316e-02 -5.94781935e-01 -4.98930782e-01 -7.92803824e-01
-5.13258338e-01 -5.02716422e-01 7.29915917e-01 -8.11763465e-01
1.39591444e+00 -1.82430041e+00 5.53146660e-01 -4.58050311e-01
7.66435564e-02 3.55167627e-01 2.49391631e-03 4.59058166e-01
1.83016390e-01 2.04931840e-01 -5.71203411e-01 -5.89017153e-01
-5.44714741e-02 3.91983062e-01 -2.99583316e-01 -2.23962784e-01
5.78495800e-01 7.79228568e-01 -1.20384324e+00 -8.64604175e-01
1.27667382e-01 7.63156191e-02 -3.78789246e-01 5.13781607e-01
-8.73206019e-01 9.50064659e-01 -8.27798367e-01 2.93311268e-01
1.96986765e-01 -2.79531717e-01 6.85098410e-01 3.31116319e-01
-3.90133739e-01 8.82228971e-01 -1.02835298e+00 1.65205956e+00
-5.61140120e-01 6.37645066e-01 7.43422136e-02 -1.24693322e+00
1.10295212e+00 6.82724237e-01 1.02313846e-01 -2.84811348e-01
2.15362847e-01 2.02776745e-01 2.13702887e-01 -7.31828868e-01
8.51554930e-01 -4.07332599e-01 -4.53531414e-01 3.99103314e-01
1.32562786e-01 -2.64901638e-01 4.97684270e-01 3.60402726e-02
1.05528104e+00 4.26203132e-01 5.44650078e-01 -8.57619420e-02
6.85589671e-01 2.12360561e-01 6.00596070e-01 3.05052936e-01
4.83798720e-02 6.99021935e-01 7.50297844e-01 -8.92117098e-02
-9.21618462e-01 -5.68682194e-01 -1.60059616e-01 1.21594393e+00
-3.28303874e-01 -5.10272503e-01 -1.02372503e+00 -8.64302993e-01
-4.52174813e-01 9.14823711e-01 -2.27468550e-01 4.39998478e-01
-1.17667961e+00 -8.53822947e-01 6.60966814e-01 3.48131061e-01
5.38032413e-01 -1.51333749e+00 -4.45499271e-01 7.22910643e-01
-5.85694671e-01 -1.31402779e+00 1.00442857e-01 -4.49896557e-03
-9.02316213e-01 -9.87615108e-01 -6.53054237e-01 -9.55518365e-01
4.36724901e-01 -2.43140623e-01 1.15603316e+00 -1.70179608e-03
3.72552365e-01 -2.58691907e-01 -6.83495522e-01 -2.94571400e-01
-1.13218498e+00 5.14795065e-01 -4.30210978e-01 -2.22283095e-01
-2.82052509e-03 -4.03411001e-01 -2.29201820e-02 -1.22709289e-01
-8.94204259e-01 3.08589131e-01 1.69689566e-01 9.03765261e-01
1.11674421e-01 3.73038054e-02 1.13820517e+00 -1.46997893e+00
1.12636173e+00 -4.12931323e-01 -6.37703717e-01 2.63652146e-01
-1.98514253e-01 1.52361974e-01 7.25567758e-01 -3.57525080e-01
-1.80239427e+00 -5.65300956e-02 -5.21788239e-01 4.40142363e-01
-3.08309257e-01 7.62459338e-01 -5.79083383e-01 7.04623461e-01
6.13748610e-01 -7.77362585e-02 -2.40551710e-01 -5.60338497e-01
6.09979510e-01 6.55941308e-01 4.20021713e-01 -8.09279323e-01
7.17554808e-01 -1.63863674e-01 -1.23582840e-01 -8.17822456e-01
-1.16935408e+00 -2.11232081e-01 -8.24708641e-01 -1.08639516e-01
1.17410517e+00 -6.61876678e-01 -3.54056478e-01 2.96389639e-01
-1.64011073e+00 -5.50355554e-01 -2.07318917e-01 3.00793350e-01
-6.26234770e-01 7.17858195e-01 -6.98834598e-01 -1.01212323e+00
-4.30421710e-01 -1.14781392e+00 1.03773463e+00 1.94775060e-01
-6.34240925e-01 -1.21491051e+00 2.79024124e-01 6.73633158e-01
1.43589139e-01 4.61597651e-01 1.24412560e+00 -7.62135625e-01
-6.65747643e-01 4.08824056e-01 -7.35147372e-02 3.30225110e-01
3.75962585e-01 6.70721084e-02 -9.53253627e-01 1.52449429e-01
2.95174748e-01 -6.00035548e-01 4.46313053e-01 1.28884181e-01
8.13989460e-01 -4.77799892e-01 -1.24277279e-01 3.61493900e-02
9.37284172e-01 6.57992065e-02 6.05666339e-01 5.89329123e-01
5.69495380e-01 1.11759710e+00 7.69955337e-01 6.90989047e-02
5.71098745e-01 7.09539592e-01 -7.56404772e-02 1.28334269e-01
-2.45330751e-01 -2.50878900e-01 4.30611461e-01 1.32577837e+00
-1.28215387e-01 -5.15767157e-01 -9.43475366e-01 8.25638771e-01
-2.03039122e+00 -8.60913813e-01 -4.46908951e-01 1.72321844e+00
1.43023229e+00 3.26192588e-01 -9.88407955e-02 2.59467781e-01
7.22024083e-01 4.79761571e-01 -2.97400411e-02 -5.69253325e-01
-1.54156446e-01 2.77702421e-01 -3.26734155e-01 6.26745164e-01
-1.09342909e+00 1.31130707e+00 6.16742897e+00 7.09495664e-01
-9.52858925e-01 1.19184695e-01 6.51474655e-01 5.75450540e-01
-5.07801294e-01 2.21326724e-01 -7.82407403e-01 3.72259140e-01
9.65661168e-01 -2.44563892e-01 -1.32964626e-01 8.47453296e-01
1.55742347e-01 5.14893001e-03 -1.13106954e+00 2.81055927e-01
-2.59681582e-01 -1.50760496e+00 7.29451841e-03 -3.33377533e-02
6.89934969e-01 -5.04851520e-01 -3.93506497e-01 4.79742527e-01
3.96482557e-01 -9.88386512e-01 4.68241990e-01 -5.50871389e-03
6.00896776e-01 -2.33643144e-01 7.34779477e-01 8.76141667e-01
-1.01947856e+00 1.69799611e-01 -1.89264536e-01 -5.30475855e-01
6.07029378e-01 6.59994245e-01 -1.09022713e+00 9.15430248e-01
1.52871966e-01 2.86730736e-01 -2.14292362e-01 5.51748097e-01
-7.55618453e-01 9.13900733e-01 -6.30548820e-02 -1.46639109e-01
2.17826605e-01 -2.20423460e-01 6.59906566e-01 1.30199015e+00
1.49859965e-01 1.87496349e-01 2.70532072e-01 7.80605555e-01
-1.08451940e-01 4.26140308e-01 -8.14176738e-01 -1.23963326e-01
5.31395912e-01 1.29650557e+00 -7.54639983e-01 -3.66323769e-01
-3.64378810e-01 6.76561415e-01 5.49277961e-01 1.23325892e-01
-4.94222790e-01 -3.67738307e-02 -2.13868287e-03 -1.37107387e-01
-2.69989707e-02 -4.47950214e-01 -5.62643945e-01 -1.19732404e+00
2.47759789e-01 -1.16743517e+00 3.60321045e-01 -3.66962910e-01
-1.31731093e+00 8.83334816e-01 1.91741884e-01 -9.50921416e-01
-9.40380991e-01 -4.99146372e-01 -8.47666383e-01 8.84016216e-01
-1.54070628e+00 -1.16478920e+00 -1.81398585e-01 2.10149828e-02
1.05483282e+00 -1.99189708e-01 1.13152683e+00 1.29818141e-01
-5.79256892e-01 2.33288363e-01 -2.62543648e-01 1.95375875e-01
5.51392376e-01 -1.42230844e+00 5.25705159e-01 8.00034881e-01
4.67800582e-03 5.23122728e-01 7.28786767e-01 -7.84924626e-01
-8.16488028e-01 -9.82743204e-01 1.25448000e+00 -6.18966460e-01
7.76130259e-01 -5.66140294e-01 -1.28246236e+00 6.89919472e-01
3.83360118e-01 -7.17743993e-01 6.18156731e-01 4.28445250e-01
1.71359226e-01 4.67851907e-01 -7.93807328e-01 6.99462354e-01
1.04039061e+00 -2.91572779e-01 -1.32070816e+00 4.09192562e-01
9.60040092e-01 -6.35117114e-01 -1.03504288e+00 6.79893851e-01
-5.63238340e-04 -6.76264465e-01 5.34670591e-01 -6.21412218e-01
7.34820008e-01 -4.84420359e-02 1.50938168e-01 -1.03981912e+00
7.69869536e-02 -1.00408864e+00 4.77707013e-02 1.87889838e+00
5.63800156e-01 -5.13277352e-01 8.49669576e-01 6.56971157e-01
-3.66492242e-01 -5.05689085e-01 -1.05158424e+00 -5.88762820e-01
3.49859834e-01 -1.79912448e-01 4.09541041e-01 1.06282473e+00
3.50516945e-01 9.96324599e-01 -2.97142118e-01 -1.67580634e-01
2.56044149e-01 3.63221139e-01 8.30046713e-01 -1.29672122e+00
-3.55767488e-01 -4.05311108e-01 1.41349316e-01 -1.12610579e+00
5.03716171e-01 -7.06600726e-01 3.77078861e-01 -1.66545260e+00
1.59648657e-02 -7.26978540e-01 2.90744513e-01 1.45379364e-01
-5.91770351e-01 -3.94148588e-01 2.84595549e-01 1.15780585e-01
-4.34343219e-01 9.08139944e-01 1.44616866e+00 1.69627368e-01
-3.94147485e-01 1.85120493e-01 -5.86792111e-01 9.72803414e-01
8.46268654e-01 -5.48243344e-01 -4.99953121e-01 -4.73662674e-01
5.01271337e-02 3.90945643e-01 -1.31036267e-01 -5.34862459e-01
-4.67824414e-02 -3.14621091e-01 -1.99469268e-01 -3.28837574e-01
2.84313381e-01 -1.95903376e-01 -1.15055725e-01 1.08893052e-01
-4.52227265e-01 -8.92182291e-02 5.38882166e-02 2.05054313e-01
-6.90294743e-01 -8.82061481e-01 2.96532571e-01 -4.33566362e-01
-4.66167837e-01 -2.72749096e-01 -5.40986061e-01 1.99219465e-01
9.10164177e-01 -2.62037188e-01 -6.00038707e-01 -2.31213376e-01
-6.55615449e-01 2.10295931e-01 3.51849824e-01 5.07763863e-01
2.40836963e-01 -1.01410913e+00 -9.57350373e-01 -1.59319073e-01
-1.17415227e-01 5.64025581e-01 8.32591727e-02 3.01057875e-01
-2.69235760e-01 5.44998884e-01 -1.26752362e-01 -3.75992417e-01
-1.23114789e+00 2.27645904e-01 -1.28919497e-01 -9.53769565e-01
-7.00822353e-01 4.73381251e-01 3.47285599e-01 -5.33498704e-01
1.28078818e-01 -1.94853261e-01 -7.50594437e-01 1.55316159e-01
2.59504795e-01 4.86027226e-02 -8.81824866e-02 -5.84622443e-01
8.70736092e-02 1.13410443e-01 2.09575772e-01 -2.82181054e-01
1.27457321e+00 -1.83421373e-01 -1.76315933e-01 5.46519220e-01
6.27846420e-01 2.60836899e-01 -1.13494146e+00 1.62652861e-02
6.93868279e-01 -2.45014876e-01 -2.82871842e-01 -3.38757694e-01
-3.47509116e-01 8.93408775e-01 -1.64026722e-01 7.13936567e-01
9.82149959e-01 1.89498797e-01 7.60838747e-01 3.98799181e-01
3.88339847e-01 -1.06706178e+00 1.17311120e-01 8.63622248e-01
8.50971162e-01 -1.20588815e+00 -1.64053485e-01 -9.18003380e-01
-7.18252778e-01 1.13165331e+00 7.97496378e-01 2.99443036e-01
1.05747700e-01 1.60438672e-01 2.11154744e-01 -9.11947191e-02
-8.81740451e-01 -3.34070355e-01 1.59617662e-01 6.51615858e-01
1.05209088e+00 -1.21909261e-01 -8.18141043e-01 5.80604672e-01
-2.41000444e-01 -5.75701855e-02 5.62691629e-01 7.72745252e-01
-4.38235015e-01 -1.92811978e+00 -2.32992008e-01 2.61985026e-02
-6.30484343e-01 -9.14715528e-02 -4.50030953e-01 8.28080118e-01
-1.44360766e-01 1.01215160e+00 -1.06083646e-01 8.21845904e-02
1.63950622e-01 3.32544327e-01 3.91026318e-01 -1.26060271e+00
-7.28165627e-01 1.15078926e-01 1.07989609e+00 -1.22039337e-02
-8.89538229e-01 -6.71201408e-01 -1.27169037e+00 1.59146115e-01
-5.20428419e-01 2.47007966e-01 3.72237384e-01 1.14364815e+00
3.49033885e-02 6.53784096e-01 3.45178872e-01 -5.89766562e-01
-5.67063570e-01 -1.40674341e+00 5.08293845e-02 6.24112546e-01
-1.30643487e-01 -6.18125796e-01 3.31029966e-02 2.26891235e-01] | [10.811488151550293, 9.334900856018066] |
51fe91f0-84f6-4189-89f8-ccf81496fb16 | real-time-universal-and-robust-adversarial | 2003.02301 | null | https://arxiv.org/abs/2003.02301v2 | https://arxiv.org/pdf/2003.02301v2.pdf | Real-time, Universal, and Robust Adversarial Attacks Against Speaker Recognition Systems | As the popularity of voice user interface (VUI) exploded in recent years, speaker recognition system has emerged as an important medium of identifying a speaker in many security-required applications and services. In this paper, we propose the first real-time, universal, and robust adversarial attack against the state-of-the-art deep neural network (DNN) based speaker recognition system. Through adding an audio-agnostic universal perturbation on arbitrary enrolled speaker's voice input, the DNN-based speaker recognition system would identify the speaker as any target (i.e., adversary-desired) speaker label. In addition, we improve the robustness of our attack by modeling the sound distortions caused by the physical over-the-air propagation through estimating room impulse response (RIR). Experiment using a public dataset of 109 English speakers demonstrates the effectiveness and robustness of our proposed attack with a high attack success rate of over 90%. The attack launching time also achieves a 100X speedup over contemporary non-universal attacks. | ['Zhuohang Li', 'Yi Xie', 'Yingying Chen', 'Cong Shi', 'Bo Yuan', 'Jian Liu'] | 2020-03-04 | null | null | null | null | ['room-impulse-response'] | ['audio'] | [ 1.88062936e-01 -1.19411074e-01 4.36762482e-01 -2.48738259e-01
-1.11282694e+00 -8.08816135e-01 3.97010863e-01 -3.90451312e-01
-1.91452250e-01 2.81247467e-01 2.30217457e-01 -7.42170990e-01
1.75688416e-01 -4.44099307e-01 -6.39495730e-01 -8.69493961e-01
-2.09456369e-01 3.78723145e-02 1.05314940e-01 -1.05303667e-01
-1.38087347e-01 8.54166567e-01 -1.19103074e+00 -2.14355692e-01
3.36221218e-01 1.01928127e+00 -3.60241443e-01 1.16074765e+00
2.43338704e-01 4.07006145e-01 -1.22872615e+00 -5.25461137e-02
2.46576816e-01 -1.81518689e-01 -7.01576352e-01 -5.11985660e-01
4.39462543e-01 -5.06390035e-01 -8.71913731e-01 1.05539131e+00
1.15699160e+00 1.56759381e-01 2.53908336e-01 -1.36770701e+00
-2.88995296e-01 9.43497956e-01 -3.47655296e-01 4.66804236e-01
3.23764414e-01 2.15995252e-01 6.99500680e-01 -5.39362431e-01
-6.95697218e-02 1.44828057e+00 3.94800812e-01 1.06713784e+00
-8.76372933e-01 -1.34487760e+00 5.24670072e-02 7.32476935e-02
-1.67827845e+00 -8.52958798e-01 9.92859483e-01 1.48842975e-01
6.45144880e-01 7.52434194e-01 -1.84282526e-01 1.47986042e+00
-1.97006941e-01 6.91454589e-01 6.83922589e-01 -3.40297073e-01
1.38405755e-01 1.68202370e-01 6.04582727e-02 2.59733975e-01
-2.73560524e-01 4.09438431e-01 -4.69222814e-01 -6.26418293e-01
4.27559853e-01 -3.26505959e-01 -4.68262583e-01 5.08866608e-01
-7.85891712e-01 6.08150184e-01 1.37545273e-01 1.68759584e-01
-2.51477361e-01 2.41770104e-01 5.84130406e-01 2.93159783e-01
1.54865324e-01 1.95387714e-02 -2.80219734e-01 -2.51826346e-01
-7.19323277e-01 3.79038975e-02 9.13135171e-01 7.30885923e-01
-2.47386191e-03 8.76279533e-01 3.98976281e-02 7.03901827e-01
5.74132681e-01 1.01079643e+00 4.83022213e-01 -4.64816988e-01
2.80555785e-01 -2.93940634e-01 -4.49604690e-02 -7.70936668e-01
-2.22230390e-01 -7.17285454e-01 -9.59890127e-01 2.29715809e-01
2.63225436e-01 -5.19207597e-01 -8.36355746e-01 1.93815970e+00
6.18206024e-01 8.00331533e-01 4.41816926e-01 8.68692040e-01
7.28407264e-01 8.59309196e-01 -1.78628623e-01 -3.13101076e-02
1.43756557e+00 -6.16599023e-01 -6.13208652e-01 -2.47087609e-02
1.46229371e-01 -9.48973000e-01 8.23584020e-01 4.88752276e-01
-6.51862025e-01 -3.69549304e-01 -1.10108054e+00 5.97053885e-01
3.59002943e-03 -4.00744706e-01 2.86619753e-01 1.51864243e+00
-9.19988394e-01 -1.56137168e-01 -6.32632852e-01 -1.61141809e-02
8.90073851e-02 6.98828757e-01 -1.88164666e-01 2.75770903e-01
-1.48200977e+00 3.56428146e-01 -3.55120271e-01 2.71345526e-01
-1.37334180e+00 -8.01526010e-01 -6.15977466e-01 7.93723539e-02
1.71150640e-01 -9.62209888e-03 1.45727658e+00 -8.83774459e-01
-1.90612650e+00 3.88522267e-01 -2.22335637e-01 -6.20300412e-01
2.95492262e-01 -7.14421347e-02 -1.18067050e+00 -4.05782126e-02
-5.17915964e-01 -1.15075253e-01 1.14217472e+00 -1.20765758e+00
-4.77767497e-01 -3.06704938e-01 -5.75712742e-03 -8.94312859e-02
-4.41787064e-01 6.58932626e-01 1.57280326e-01 -5.52548766e-01
6.71368316e-02 -9.68242943e-01 1.25034526e-01 -3.75713706e-01
-7.77071893e-01 -9.98828560e-02 1.47396374e+00 -7.53884077e-01
1.06395829e+00 -2.59105921e+00 -3.68416548e-01 5.85563302e-01
3.00227255e-02 6.24766827e-01 -1.25495940e-01 2.12311924e-01
-1.29215896e-01 -8.48373398e-03 9.88457948e-02 -4.29839432e-01
1.05846494e-01 4.38262261e-02 -7.25256085e-01 6.78673625e-01
-2.97766566e-01 2.08290920e-01 -4.17216063e-01 -3.66189852e-02
4.67514955e-02 8.06023359e-01 -3.55163783e-01 4.25125062e-01
3.94046545e-01 3.75884950e-01 -4.23275381e-01 6.79167390e-01
7.90696561e-01 5.76560974e-01 -8.66031349e-02 -1.45861292e-02
2.31931329e-01 4.51390207e-01 -1.43854427e+00 9.76951957e-01
-6.76690519e-01 8.12187910e-01 8.38001668e-01 -6.22871816e-01
9.28154826e-01 8.84263754e-01 -1.33548200e-01 -2.21726090e-01
2.93121070e-01 1.41153142e-01 3.99565488e-01 -9.69533399e-02
1.40460804e-01 -2.28456464e-02 -2.35235482e-01 6.20614648e-01
-4.38846238e-02 3.93137813e-01 -9.88133907e-01 1.82673350e-01
1.08135414e+00 -6.56531811e-01 -6.85711056e-02 -2.09950462e-01
9.30837154e-01 -9.32516098e-01 4.64378536e-01 9.74033892e-01
-6.90201342e-01 1.87242389e-01 -1.44616932e-01 -1.07435457e-01
-5.48638046e-01 -1.22709489e+00 -1.87096931e-02 1.12884736e+00
-1.72728017e-01 2.67402083e-01 -9.59768951e-01 -4.50121343e-01
-1.22453570e-01 7.84444273e-01 -2.28723407e-01 -2.10384294e-01
-7.25559056e-01 -2.69976348e-01 1.64094973e+00 4.62079167e-01
4.32727933e-01 -8.99089098e-01 -2.40694117e-02 1.74609095e-01
-1.27755046e-01 -1.20799530e+00 -8.57968330e-01 -5.04099056e-02
-2.50989825e-01 -4.24827963e-01 -4.89503711e-01 -8.62811863e-01
3.26688260e-01 2.85861731e-01 3.18940312e-01 -2.62268513e-01
-2.96351701e-01 4.77120429e-01 4.96778972e-02 -5.56506336e-01
-1.03289390e+00 -1.53116524e-01 8.47460806e-01 4.05392855e-01
2.03959197e-01 -7.62851417e-01 -5.07280111e-01 3.86331499e-01
-7.48404026e-01 -8.66997480e-01 3.38112488e-02 4.59149241e-01
-1.20458148e-01 2.74261177e-01 9.36837375e-01 -2.31165722e-01
6.82675183e-01 -2.43881375e-01 -6.28284097e-01 1.24051794e-02
4.00642976e-02 -7.33212456e-02 8.41155708e-01 -9.58927155e-01
-9.68443930e-01 -1.53541014e-01 -6.99442565e-01 -2.63810366e-01
-4.50074703e-01 4.94879037e-02 -7.19429135e-01 -4.51084584e-01
6.11451566e-01 4.74735826e-01 -1.66818112e-01 -3.77545506e-01
2.50617176e-01 1.26579499e+00 7.22754896e-01 -4.35807675e-01
1.23567247e+00 2.60473996e-01 -1.67410836e-01 -1.38036406e+00
-1.55883133e-01 -4.56286192e-01 1.04628168e-01 -2.93053806e-01
5.08068085e-01 -9.27244842e-01 -1.45949292e+00 1.04612803e+00
-1.25057137e+00 -1.05172424e-02 3.16840291e-01 6.64818585e-01
1.19445838e-01 4.69067305e-01 -6.67442918e-01 -1.45181799e+00
-8.36811781e-01 -1.28969598e+00 7.15097547e-01 1.11878447e-01
-1.05019808e-01 -7.23493755e-01 -1.78524584e-01 3.80297512e-01
8.75460565e-01 6.64004758e-02 5.67511916e-01 -1.23038363e+00
-4.48852718e-01 -4.84518975e-01 2.27116123e-01 5.59149683e-01
2.48519495e-01 -6.70790300e-02 -1.55165327e+00 -4.12131995e-01
3.26737285e-01 -1.44157320e-01 1.15175217e-01 2.41747543e-01
9.26360130e-01 -7.15407789e-01 -5.61190248e-02 6.57163143e-01
1.13865352e+00 3.91232789e-01 5.12737572e-01 -7.12390766e-02
7.22798705e-01 1.75844729e-01 -6.17093034e-02 3.85599375e-01
-1.47832558e-01 5.30687332e-01 5.04777253e-01 9.97787341e-02
2.00678959e-01 3.32931206e-02 7.06291080e-01 7.27451622e-01
2.30474517e-01 -5.16061425e-01 -8.15656304e-01 2.92775035e-01
-1.10247028e+00 -9.83528554e-01 5.95539398e-02 2.45498466e+00
8.12715530e-01 2.52668709e-01 2.15369016e-01 4.10843700e-01
8.22244465e-01 2.68216759e-01 -6.12126648e-01 -6.90285563e-01
1.41760081e-01 3.39981079e-01 5.86396813e-01 9.31912005e-01
-1.03840387e+00 8.40477586e-01 5.73480892e+00 7.47140229e-01
-1.61772120e+00 2.19069794e-01 4.42452639e-01 -1.82164684e-01
-3.69087700e-03 -6.83595955e-01 -8.95627201e-01 1.49730787e-01
1.54751134e+00 -1.61708668e-01 6.19220376e-01 9.02000785e-01
2.27638423e-01 6.62431359e-01 -8.83777440e-01 9.03911710e-01
2.57099271e-01 -8.68333161e-01 -1.85476273e-01 2.61017680e-01
-1.25663215e-02 1.73627138e-01 2.72869706e-01 2.63092250e-01
3.49071085e-01 -9.54169452e-01 6.25262916e-01 -6.46875501e-02
7.39845514e-01 -1.17297184e+00 6.44791186e-01 2.73630679e-01
-1.13479364e+00 -3.68018299e-02 -7.03072026e-02 2.49344051e-01
7.56802484e-02 1.58288896e-01 -9.60955679e-01 1.86078206e-01
5.57117581e-01 -5.85913360e-01 5.83527945e-02 6.61508739e-01
-1.47237005e-02 1.48087525e+00 -8.31221461e-01 -1.54957473e-01
4.02379245e-01 4.38014716e-01 1.02802134e+00 1.28331411e+00
2.47773200e-01 2.08090782e-01 -1.06227219e-01 4.01713550e-01
-3.15570891e-01 -7.51059800e-02 -6.16456270e-01 2.37736449e-01
8.85379255e-01 1.05295146e+00 -1.45678833e-01 1.75114170e-01
1.47137912e-02 1.14969766e+00 -3.42737824e-01 5.22961617e-01
-1.13229835e+00 -6.99049652e-01 1.28819454e+00 -2.69781917e-01
8.04546773e-02 -2.10456297e-01 5.35238907e-02 -6.14657342e-01
1.95628721e-02 -1.30400705e+00 1.18341753e-02 -2.50099868e-01
-9.18460667e-01 8.76254559e-01 -5.41140020e-01 -9.68575418e-01
-1.63213611e-01 -4.69703346e-01 -1.01873422e+00 1.22113025e+00
-1.19366169e+00 -1.08255529e+00 1.51250958e-01 9.13982749e-01
2.67338157e-01 -5.34567416e-01 1.19047976e+00 5.24991333e-01
-6.48675919e-01 1.55381024e+00 2.86309898e-01 6.69118047e-01
6.20173275e-01 -7.86745965e-01 7.94533908e-01 1.18989706e+00
2.13462785e-02 8.22072446e-01 9.98069048e-01 -1.00951932e-01
-1.55315232e+00 -1.02331829e+00 6.11699104e-01 -1.71580359e-01
7.86467254e-01 -8.51956069e-01 -9.78841126e-01 6.20678067e-01
2.16360286e-01 -1.32121488e-01 9.55315709e-01 -1.83797479e-01
-7.81303287e-01 -4.43651438e-01 -1.38246906e+00 7.25242496e-01
4.83259141e-01 -9.65702355e-01 -2.19523013e-01 1.86491460e-01
1.12242484e+00 -4.71380144e-01 -4.74912375e-01 -7.69148991e-02
6.11665189e-01 -4.60131913e-01 1.16066349e+00 -5.62812030e-01
-5.05941451e-01 -1.77733824e-01 -4.80917692e-01 -9.23561335e-01
-8.04192871e-02 -1.33736706e+00 -6.07369803e-02 1.65033400e+00
5.30301988e-01 -1.04975796e+00 6.84508681e-01 5.27373552e-01
2.92461310e-02 -2.53976196e-01 -1.42308831e+00 -8.54893446e-01
-2.31421724e-01 -8.21550369e-01 8.73136401e-01 7.57303894e-01
-2.19397858e-01 9.63230431e-02 -7.38563955e-01 1.35389042e+00
8.99088442e-01 -7.04413891e-01 8.04753363e-01 -9.21204388e-01
-4.32411879e-01 -2.03204289e-01 -6.70470893e-01 -9.96579528e-01
2.75677979e-01 -5.88131130e-01 2.49967292e-01 -7.29849696e-01
-4.42816228e-01 -2.53034741e-01 -5.92511475e-01 2.29039431e-01
-7.25121722e-02 -1.01203062e-01 7.56586939e-02 -2.81336010e-01
3.84015851e-02 2.98761785e-01 4.70103770e-01 -2.84119248e-01
-1.94947571e-01 5.90998888e-01 -5.76147258e-01 6.81840241e-01
1.01932824e+00 -5.29889941e-01 -3.72669667e-01 -1.31758049e-01
-5.36093593e-01 1.28522977e-01 3.31719220e-01 -1.17858231e+00
3.31941217e-01 1.53915748e-01 -2.20484078e-01 -3.03031743e-01
5.30750453e-01 -9.80939925e-01 -1.49002180e-01 4.82165664e-01
-3.75259250e-01 -3.07349622e-01 4.65929925e-01 5.76777399e-01
8.47002026e-03 3.29789072e-02 9.49071884e-01 5.92489719e-01
-1.85674220e-01 2.67485619e-01 -5.92107952e-01 -9.67328921e-02
6.57667994e-01 1.06712885e-01 -2.71331161e-01 -7.39417970e-01
-3.29598963e-01 -1.57453746e-01 -3.27136159e-01 5.10577142e-01
7.81174123e-01 -9.41563785e-01 -8.50324988e-01 4.61525947e-01
-2.01808587e-01 -3.37887019e-01 3.62468302e-01 2.03947663e-01
-3.30909759e-01 4.29443359e-01 3.62105459e-01 -4.58699346e-01
-1.98136115e+00 3.12460989e-01 6.32138908e-01 1.69023171e-01
-5.25633335e-01 1.25928521e+00 1.22834094e-01 -5.48864543e-01
9.54974532e-01 -9.05978531e-02 2.42743060e-01 -7.00051188e-01
9.87927616e-01 3.83547097e-01 -4.07114439e-03 -8.16387117e-01
-7.11667001e-01 1.28881000e-02 -1.02755658e-01 -4.50586855e-01
8.29429090e-01 -3.16135362e-02 1.16087973e-01 2.67820448e-01
1.32446277e+00 5.06549835e-01 -7.88607955e-01 -5.16015172e-01
-4.78036970e-01 -2.36325294e-01 3.20389897e-01 -7.32068419e-01
-9.31162000e-01 8.26486766e-01 9.92690265e-01 2.47472301e-01
8.99280190e-01 -2.98409134e-01 1.28949618e+00 4.91008043e-01
4.03622210e-01 -5.92262208e-01 -2.28002667e-01 3.91844541e-01
7.12323904e-01 -9.58690822e-01 -5.97911119e-01 -4.95972671e-02
-5.41566074e-01 8.32352757e-01 4.16648388e-01 1.74273118e-01
1.07619131e+00 4.89786834e-01 7.25534558e-01 3.57854575e-01
-2.96999246e-01 5.22287250e-01 9.28071290e-02 7.61358798e-01
1.11983173e-01 3.23209852e-01 5.66083312e-01 6.46878600e-01
-3.86825860e-01 -6.80921912e-01 4.76573318e-01 6.25840306e-01
-4.48203027e-01 -8.26756060e-01 -8.51794064e-01 -4.03479896e-02
-9.91584361e-01 -3.18819106e-01 -1.87434837e-01 3.76746744e-01
-5.63929200e-01 1.44127917e+00 -2.47488722e-01 -5.59961140e-01
2.42308870e-01 2.75836438e-01 -1.77743435e-01 -1.05601870e-01
-6.99944437e-01 3.86891216e-02 -3.24554592e-02 -3.39653403e-01
1.32243887e-01 -4.29280698e-01 -1.43002307e+00 -6.92563415e-01
-4.20683473e-01 2.82359689e-01 1.10855055e+00 9.15826023e-01
2.35641330e-01 6.57195449e-01 1.28698444e+00 -6.33110404e-01
-1.11388230e+00 -8.90337765e-01 -7.27677643e-01 -4.38307878e-03
9.27416623e-01 -3.99341211e-02 -9.05250967e-01 -4.14842635e-01] | [14.022130966186523, 5.843821048736572] |
d6fb0af8-713f-4c78-b38e-da0247a34ec7 | molecular-representation-learning-by | null | null | https://github.com/PaddlePaddle/PaddleHelix/blob/dev/competition/ogbg_molhiv/Molecule_Representation_Learning_by_Leveraging_Chemical_Information.pdf | https://github.com/PaddlePaddle/PaddleHelix/blob/dev/competition/ogbg_molhiv/Molecule_Representation_Learning_by_Leveraging_Chemical_Information.pdf | Molecular Representation Learning by Leveraging Chemical Information | Molecular property prediction is of great importance in AI drug design due to its high experimental efficiency compared with biological experiments. As graph neural networks have achieved great success in many domains, some studies apply graph neural networks to molecular property prediction and regard each molecule as a graph. A molecule’s atom is regarded as a node of the graph, while its bond is regarded as an edge of the graph. However, most existing methods simply apply general graph neural networks without considering the domain knowledge. As chemical information is highly related to molecular functions, it is critical for accurate property prediction. Thus, we leverage chemical information to learn molecular representation by integrating molecular fingerprints, i.e., the presence or absence of particular chemical substructures. We compare our proposed method to several strong baselines, and our proposed method significantly surpasses other methods. Up to now, our method ranks first in the Open Graph Benchmark(OGB) leaderboard for ogbg-molhiv. | ['Fan Wang', 'Shikun Feng', 'Xiaomin Fang', 'Jieqiong Lei', 'Zhengjie Huang', 'Lihang Liu', 'Shanzhuo Zhang', 'Weibin Li'] | 2021-03-15 | null | null | null | na-2021-3 | ['graph-property-prediction'] | ['graphs'] | [ 3.99985313e-01 4.29374762e-02 -9.46271420e-01 -1.39752731e-01
-7.65895322e-02 -4.61808383e-01 3.69500160e-01 9.86388505e-01
-1.22702055e-01 1.13804638e+00 1.35108069e-01 -6.84999764e-01
-1.05963551e-01 -1.18761849e+00 -9.11826193e-01 -7.77662277e-01
-1.48116380e-01 3.88865829e-01 2.17013329e-01 -2.73765206e-01
1.21693306e-01 7.21849263e-01 -6.73176706e-01 -4.38565798e-02
9.18689370e-01 9.06616688e-01 -1.02246784e-01 1.20918147e-01
1.45390302e-01 1.02730525e+00 -4.20835704e-01 -4.67200339e-01
-2.58090526e-01 -3.68660748e-01 -7.10796416e-01 -4.85521615e-01
4.57594246e-01 1.41278043e-01 -6.46865070e-01 1.17933893e+00
4.79571909e-01 2.85814136e-01 8.91664982e-01 -1.04616725e+00
-7.92645812e-01 4.80513304e-01 -4.62967455e-01 4.49873395e-02
3.52768809e-01 -1.26863923e-02 1.43340909e+00 -5.43289423e-01
7.32440293e-01 8.63255620e-01 4.53228027e-01 3.41575831e-01
-1.47437525e+00 -6.41337693e-01 2.45244324e-01 5.61611295e-01
-1.18864751e+00 -2.07203880e-01 9.50954616e-01 -3.44659239e-01
9.60510194e-01 1.17772613e-02 7.08077967e-01 9.52648699e-01
4.33496773e-01 6.57842398e-01 6.21891916e-01 6.35483563e-02
3.64494085e-01 -3.79209191e-01 2.96184450e-01 8.87839377e-01
5.27410984e-01 8.97502247e-03 -4.03797895e-01 -2.59068131e-01
6.08776569e-01 3.12735856e-01 -5.40058553e-01 -5.00798643e-01
-1.05402064e+00 9.01380539e-01 1.16375411e+00 2.44028401e-02
-5.21695256e-01 3.62748295e-01 3.05646807e-01 7.85247013e-02
3.64029557e-01 6.70187116e-01 -2.26714298e-01 3.42743397e-01
-5.57087064e-01 -3.20993550e-02 8.76680017e-01 6.61238194e-01
7.48727143e-01 3.32187712e-02 -1.58664619e-03 6.96441948e-01
2.42168352e-01 7.14303404e-02 1.90746307e-01 -1.70206755e-01
2.80447721e-01 9.15322125e-01 -1.70449957e-01 -1.42747641e+00
-6.19945705e-01 -4.65221614e-01 -1.15179574e+00 -1.26413375e-01
2.35251188e-01 1.22561768e-01 -9.18623626e-01 1.69309509e+00
3.81882817e-01 4.51631457e-01 -1.43372472e-02 6.53547466e-01
1.25161791e+00 7.11338699e-01 5.03097177e-01 -2.99562693e-01
1.11866868e+00 -8.42775643e-01 -5.84488928e-01 8.29581395e-02
6.49117589e-01 -2.63200611e-01 3.43539208e-01 4.25686657e-01
-4.74638999e-01 -2.65650928e-01 -1.38506985e+00 1.12155907e-01
-5.84709525e-01 -3.82288247e-01 1.17412138e+00 4.68532950e-01
-6.11075461e-01 1.22650063e+00 -7.57136762e-01 -1.87538750e-02
6.26803756e-01 7.04027176e-01 -7.30136812e-01 -3.12107980e-01
-1.34838510e+00 7.02563107e-01 6.99135721e-01 -1.46459613e-03
-7.36912191e-01 -6.11971617e-01 -1.04757869e+00 1.03794195e-01
5.74103892e-01 -6.79114282e-01 7.53514230e-01 -6.16954625e-01
-1.33799386e+00 2.80751795e-01 -4.28560302e-02 -4.57208157e-01
-8.98456480e-03 3.38999122e-01 -5.21378458e-01 1.46598369e-01
-3.06608319e-01 3.85641068e-01 4.97877538e-01 -8.97904515e-01
-2.12724239e-01 -4.38358516e-01 3.10196668e-01 -1.50253708e-02
-3.98910791e-01 -4.95152414e-01 -2.31646925e-01 -4.98757273e-01
4.31690961e-02 -8.58060062e-01 -4.53129023e-01 -2.60336876e-01
-9.04643893e-01 -3.85871559e-01 4.66951966e-01 -3.30224097e-01
1.36192405e+00 -1.84808564e+00 3.80506217e-01 4.31577235e-01
7.88651228e-01 1.71218485e-01 -2.27795526e-01 8.08493257e-01
-4.78980988e-01 1.73191145e-01 -1.75557837e-01 3.37559581e-01
-2.76075274e-01 -1.26605570e-01 -3.97530664e-03 6.41789019e-01
1.77460723e-02 9.98639107e-01 -1.22297513e+00 -3.04859430e-01
9.66876447e-02 6.72400475e-01 -4.83329624e-01 -1.30334171e-02
-6.83342993e-01 2.50513792e-01 -9.05688286e-01 7.42070794e-01
5.99717855e-01 -7.54482925e-01 7.10874975e-01 -4.78136957e-01
2.94047296e-01 4.10053164e-01 -6.00934267e-01 1.35830188e+00
-8.18765685e-02 3.28470141e-01 -5.33731759e-01 -1.17483544e+00
8.16957235e-01 2.80044585e-01 6.44873857e-01 -5.78592122e-01
-5.94404712e-02 1.55475318e-01 4.15255606e-01 -7.55166635e-02
1.16158493e-01 -1.00324377e-01 4.83346403e-01 4.23523523e-02
-2.98500266e-02 1.69206753e-01 2.44036376e-01 3.17490458e-01
1.13744688e+00 1.95062906e-01 7.24350452e-01 -4.71876711e-02
6.19450212e-01 -4.61642630e-02 6.65552258e-01 3.58670503e-01
1.47876233e-01 8.90633091e-02 9.46131587e-01 -6.02711618e-01
-6.06466174e-01 -7.35366642e-01 -9.98115540e-02 9.38817739e-01
3.25018466e-01 -7.98887432e-01 -4.77108836e-01 -7.85411716e-01
2.23483726e-01 3.50378603e-01 -7.69934952e-01 -5.23661673e-01
-2.33616307e-01 -7.99498856e-01 2.77996093e-01 3.62166971e-01
2.27507457e-01 -1.17197263e+00 2.17702553e-01 5.62319338e-01
2.24786744e-01 -7.86640108e-01 -5.25585592e-01 2.94225544e-01
-8.33941042e-01 -1.49321127e+00 -5.45710862e-01 -6.98401451e-01
5.55101275e-01 3.95196468e-01 1.20542920e+00 2.85432339e-01
-2.01405242e-01 -3.65611196e-01 -3.65512192e-01 -4.08468962e-01
-4.54690665e-01 3.23597550e-01 -4.52430174e-02 1.82674885e-01
2.14091823e-01 -8.16156149e-01 -9.41255927e-01 1.87123582e-01
-8.33641231e-01 -5.94784841e-02 6.97739720e-01 8.48534048e-01
9.45627570e-01 4.73068422e-03 6.70773387e-01 -1.37155211e+00
6.42541766e-01 -5.92422783e-01 -6.71049833e-01 3.52897763e-01
-7.21154988e-01 2.19936818e-01 9.74619627e-01 -1.36402026e-01
-3.23961437e-01 1.92302153e-01 -7.10604414e-02 -1.34715319e-01
9.17385891e-02 1.22229207e+00 -5.28622329e-01 -3.17973047e-01
5.76835096e-01 1.05467208e-01 -2.62439996e-01 -2.95218229e-01
1.58730030e-01 3.32462966e-01 5.63126840e-02 -5.36627054e-01
4.58795518e-01 4.84517477e-02 7.88223445e-01 -9.24675703e-01
-6.60826981e-01 -3.67929876e-01 -4.96385783e-01 1.23568565e-01
8.17112565e-01 -6.56760097e-01 -1.39019227e+00 1.57826766e-01
-1.03968811e+00 -6.52494803e-02 4.75645721e-01 4.50318456e-01
-1.72451496e-01 6.12921238e-01 -5.16068816e-01 -3.53503108e-01
-4.29738522e-01 -1.28536844e+00 6.69921398e-01 1.90463662e-01
9.08588544e-02 -1.30571902e+00 1.98264137e-01 2.39115551e-01
-1.31980227e-02 7.08987117e-01 1.47373760e+00 -9.24908817e-01
-6.56770349e-01 -4.33117926e-01 -2.29029387e-01 2.58173160e-02
5.15159428e-01 -1.01249013e-02 -5.73394358e-01 -3.33617330e-01
-7.75400698e-01 -2.84708828e-01 1.16921055e+00 4.54578400e-01
1.24250972e+00 -2.14966267e-01 -6.93064868e-01 7.08567679e-01
1.34757352e+00 3.34925532e-01 6.75111115e-01 1.03341438e-01
1.16547966e+00 2.96427220e-01 1.56099349e-01 1.83475688e-01
1.58258453e-01 7.87501097e-01 7.05975473e-01 -2.80914843e-01
8.07284340e-02 -5.65480590e-01 2.02592462e-01 5.17050266e-01
-4.57791984e-01 -7.09626198e-01 -7.73404121e-01 -6.60861060e-02
-1.87385643e+00 -8.26038897e-01 -2.70213604e-01 2.39232993e+00
9.13697064e-01 1.96611568e-01 2.31849626e-01 -2.09957421e-01
7.06556797e-01 4.04236197e-01 -1.06641614e+00 -4.87989485e-02
-5.70039786e-02 3.48036140e-01 5.07288039e-01 3.27974528e-01
-1.07849336e+00 1.06440639e+00 5.52111959e+00 8.83764267e-01
-1.11265922e+00 -3.63459438e-01 8.04833710e-01 3.59939665e-01
-2.35249966e-01 -5.48020899e-02 -4.70653981e-01 3.97107154e-01
7.27487445e-01 -1.85748786e-01 3.80960375e-01 7.63392746e-01
-5.03270552e-02 9.91556197e-02 -1.34214234e+00 7.84265161e-01
-9.19511169e-02 -1.73508036e+00 4.79569227e-01 3.48012894e-01
5.83135903e-01 2.73084417e-02 6.02588989e-02 3.15191858e-02
3.10974836e-01 -1.46799254e+00 -2.51610223e-02 3.10788661e-01
8.67323101e-01 -9.62604940e-01 6.63992047e-01 -6.38169236e-03
-1.31673956e+00 4.78777200e-01 -5.84579408e-01 1.44859776e-01
-2.19735250e-01 5.65779209e-01 -9.84927952e-01 9.46558774e-01
-1.35276735e-01 1.33994627e+00 -4.76298213e-01 1.21250737e+00
-5.26205301e-01 7.32153535e-01 4.88754548e-02 -2.66186893e-01
2.39845395e-01 -7.31719792e-01 2.03014240e-01 8.45982969e-01
-1.03066161e-01 5.90978451e-02 4.66809541e-01 5.58393955e-01
-8.68857622e-01 5.77650189e-01 -7.62188494e-01 -6.79875493e-01
2.49989465e-01 1.16502357e+00 -7.21229792e-01 -2.36897364e-01
-5.78763068e-01 8.33195806e-01 4.33788419e-01 4.67824489e-01
-7.52460122e-01 -6.31364226e-01 7.48226702e-01 -5.49509488e-02
-2.12507658e-02 8.87326896e-02 3.31131816e-01 -9.64145601e-01
-4.08603460e-01 -8.98095846e-01 5.46639681e-01 -5.07166624e-01
-1.33346999e+00 6.65651619e-01 -6.01351738e-01 -1.28530264e+00
2.13873908e-01 -9.41009343e-01 -4.17099774e-01 8.03602695e-01
-1.72331059e+00 -1.02075493e+00 -1.78928643e-01 4.09943879e-01
8.85364413e-03 -1.00776486e-01 9.60520089e-01 1.72285378e-01
-7.56669343e-01 5.59225202e-01 2.64629900e-01 1.73323900e-01
5.28642893e-01 -1.18546379e+00 4.15710062e-01 1.77917987e-01
4.04960185e-01 8.75086844e-01 4.69023556e-01 -1.00345206e+00
-1.64644265e+00 -1.14539838e+00 5.19055605e-01 -2.62460142e-01
9.16392207e-01 -2.94755369e-01 -1.02432442e+00 4.67326224e-01
-1.15313120e-01 1.27845496e-01 9.87705052e-01 4.22269404e-01
-6.11117005e-01 -3.21145207e-02 -6.52274847e-01 6.08357608e-01
1.07548523e+00 -5.04263759e-01 -3.84613089e-02 8.42726946e-01
9.31340456e-01 -2.28218675e-01 -1.09541523e+00 4.33077067e-01
4.57837105e-01 -6.52879477e-01 1.05428445e+00 -1.12807846e+00
6.10363901e-01 -3.43572408e-01 1.13223240e-01 -1.44884324e+00
-5.09069324e-01 -5.79067886e-01 -1.90356359e-01 5.22624850e-01
8.57028782e-01 -7.32376754e-01 9.83139038e-01 2.31494233e-01
-3.68816108e-02 -9.60348845e-01 -7.30177581e-01 -7.74551570e-01
-3.07779945e-02 1.73019350e-01 6.57646358e-01 1.14597332e+00
2.37609625e-01 9.74238813e-01 -6.24205887e-01 5.51861525e-02
4.86436665e-01 2.91336268e-01 5.24009049e-01 -1.58139277e+00
-5.30010164e-01 -6.67664647e-01 -8.39395165e-01 -9.10917163e-01
2.81748503e-01 -1.37395358e+00 -4.55149621e-01 -1.74360859e+00
4.75962073e-01 -1.95451438e-01 -8.64567995e-01 5.89184880e-01
-4.18167502e-01 1.49289727e-01 -7.82323033e-02 -6.97166054e-03
-6.39163673e-01 5.80235958e-01 1.57214272e+00 -6.95999682e-01
-2.71560699e-01 6.36061653e-02 -7.17175364e-01 3.74940783e-01
6.27364695e-01 -4.11722213e-01 -4.48928624e-01 7.19942302e-02
6.20524704e-01 2.92699963e-01 -3.79488356e-02 -8.49909425e-01
1.67708382e-01 -3.76439750e-01 3.80140901e-01 -2.06379414e-01
3.54570866e-01 -7.81837881e-01 2.84263253e-01 6.98541045e-01
-2.73718655e-01 -2.26539373e-01 -1.94202699e-02 1.13876450e+00
-3.69441628e-01 3.86117063e-02 6.34081542e-01 6.39198273e-02
-5.23519754e-01 1.14751804e+00 7.31858388e-02 -4.22539860e-01
7.20183253e-01 -1.83871672e-01 -4.04823035e-01 -3.14131320e-01
-5.70047259e-01 1.76057011e-01 4.53141451e-01 2.73496151e-01
6.36081398e-01 -1.19785821e+00 -4.40665364e-01 -2.12241650e-01
4.64174926e-01 -1.81207642e-01 4.15880641e-06 6.84864640e-01
-6.02784634e-01 4.39215034e-01 1.86117981e-02 -2.12560341e-01
-1.29294336e+00 9.76955116e-01 3.39356095e-01 -2.36688510e-01
-3.91870081e-01 6.68550134e-01 5.05429447e-01 -3.30693051e-02
1.79484457e-01 -2.16964781e-02 -5.20683825e-01 -3.64675634e-02
3.39500815e-01 7.01359510e-02 1.30233601e-01 -4.96863604e-01
-5.13063073e-01 4.55732435e-01 -3.73318970e-01 7.65882850e-01
1.54637134e+00 9.12799716e-01 -3.87827188e-01 1.96491554e-01
1.24183500e+00 2.07737237e-02 -7.94720829e-01 -2.27758214e-01
1.13588482e-01 -7.52984732e-02 1.54594734e-01 -8.38220417e-01
-1.07745624e+00 8.17994237e-01 -1.08407694e-03 -3.34795117e-02
7.96548486e-01 -1.08298823e-01 5.70522726e-01 9.11612928e-01
3.75288725e-01 -6.08051181e-01 8.44495744e-02 3.50173414e-01
6.10859871e-01 -1.29618573e+00 4.78288174e-01 -7.98448145e-01
-4.72110659e-01 1.21521890e+00 3.81089687e-01 8.44224244e-02
5.09093821e-01 -5.05601883e-01 -3.66749793e-01 -5.45445323e-01
-5.34487247e-01 -6.67943880e-02 5.55204272e-01 4.31980580e-01
7.37024367e-01 2.66805679e-01 -2.05923140e-01 2.36101136e-01
1.95097893e-01 -7.32251406e-02 1.07706875e-01 6.10045612e-01
-3.94749343e-01 -1.45425773e+00 1.82153255e-01 5.96298277e-01
-3.97648513e-01 -4.93809342e-01 -8.33456278e-01 7.29551494e-01
-2.72993475e-01 8.47251117e-01 -5.33773601e-01 -5.31598449e-01
3.15347999e-01 -1.70681879e-01 5.49751997e-01 -7.08655000e-01
-3.70065987e-01 -7.04097077e-02 1.03728086e-01 -5.33996820e-01
-3.40144366e-01 -1.55176148e-01 -1.38503623e+00 -4.20077771e-01
-4.27667767e-01 5.31644702e-01 4.13478643e-01 6.56547666e-01
4.78428960e-01 6.71185672e-01 6.72761261e-01 -3.64615202e-01
-4.23733145e-02 -7.48235345e-01 -7.24169731e-01 5.06019592e-01
2.19240516e-01 -6.98146760e-01 7.57290274e-02 -1.23283207e-01] | [5.209540367126465, 5.878997802734375] |
50704ecd-4b98-4248-88f8-e8562e12c12c | a-physics-informed-feature-engineering | 2003.01878 | null | https://arxiv.org/abs/2003.01878v3 | https://arxiv.org/pdf/2003.01878v3.pdf | Physics-informed machine learning for composition-process-property alloy design: shape memory alloy demonstration | Machine learning (ML) is shown to predict new alloys and their performances in a high dimensional, multiple-target-property design space that considers chemistry, multi-step processing routes, and characterization methodology variations. A physics-informed featured engineering approach is shown to enable otherwise poorly performing ML models to perform well with the same data. Specifically, previously engineered elemental features based on alloy chemistries are combined with newly engineered heat treatment process features. The new features result from first transforming the heat treatment parameter data as it was previously recorded using nonlinear mathematical relationships known to describe the thermodynamics and kinetics of phase transformations in alloys. The ability of the ML model to be used for predictive design is validated using blind predictions. Composition - process - property relationships for thermal hysteresis of shape memory alloys (SMAs) with complex microstructures created via multiple melting-homogenization-solutionization-precipitation processing stage variations are captured, in addition to the mean transformation temperatures of the SMAs. The quantitative models of hysteresis exhibited by such highly processed alloys demonstrate the ability for ML models to design for physical complexities that have challenged physics-based modeling approaches for decades. | ['Behnam Amin-ahmadi', 'Othmane Benafan', 'Sen Liu', 'Branden B. Kappes', 'Xiaoli Zhang', 'Aaron P. Stebner'] | 2020-03-04 | null | null | null | null | ['physics-informed-machine-learning'] | ['graphs'] | [ 1.94275528e-01 -2.55009979e-01 -2.17352241e-01 -3.34134877e-01
-8.50404441e-01 -1.91048980e-01 6.17651343e-01 3.63453746e-01
7.74273928e-03 5.89886427e-01 -7.17249587e-02 -1.86437890e-01
-8.49321663e-01 -6.59655392e-01 -6.62536204e-01 -1.01448476e+00
1.24557428e-01 9.29326713e-01 -6.07752055e-02 -6.20223463e-01
5.76342165e-01 8.61612082e-01 -1.65015650e+00 6.53317988e-01
3.08202088e-01 1.21700883e+00 4.06981707e-01 3.55030745e-01
2.10631073e-01 4.85904723e-01 -6.68782666e-02 3.98776591e-01
2.68046018e-02 -4.66487966e-02 -5.28177381e-01 -8.84175226e-02
-4.59252261e-02 1.93345681e-01 -4.06997740e-01 1.97951928e-01
6.76477924e-02 6.44367784e-02 1.19638121e+00 -1.01045549e+00
-9.07251775e-01 5.15950620e-01 -1.96764678e-01 -5.19544601e-01
4.70735997e-01 5.47584355e-01 1.17664301e+00 -9.64838743e-01
5.88452697e-01 1.07452500e+00 6.68686092e-01 3.25765699e-01
-1.49772203e+00 -4.33247864e-01 -4.29010093e-01 2.62540847e-01
-1.06742883e+00 -2.46691614e-01 1.10412896e+00 -9.77374911e-01
1.12837148e+00 5.25282264e-01 7.05893993e-01 8.03629220e-01
1.11600912e+00 2.18011767e-01 1.36931241e+00 -4.59233373e-01
5.59679151e-01 4.91828248e-02 1.15707457e-01 5.07416606e-01
1.17387079e-01 8.21899712e-01 -4.74049270e-01 -1.69824883e-01
2.88480163e-01 -4.29428136e-03 -3.75549048e-02 -3.84243459e-01
-7.86618650e-01 6.17542624e-01 2.61761963e-01 2.79256254e-01
-5.04312336e-01 1.94002584e-01 2.56765395e-01 5.75490296e-01
4.31919843e-01 1.49936557e+00 -1.19244218e+00 -1.40632922e-02
-1.18913531e+00 5.80553830e-01 7.08109617e-01 8.96616161e-01
1.23701274e+00 2.49583349e-01 7.02815428e-02 6.49119854e-01
7.50189722e-01 6.52387500e-01 5.01166642e-01 -8.16934526e-01
-1.30632922e-01 8.09243739e-01 -1.79579556e-01 -5.94537139e-01
-7.47858047e-01 1.55206382e-01 -2.42818534e-01 4.34454858e-01
-4.32570986e-02 3.54471058e-01 -1.12093318e+00 1.04517913e+00
1.86112329e-01 -7.60637939e-01 -1.74195886e-01 5.81737936e-01
5.94484568e-01 6.75121486e-01 2.43057385e-01 -3.76159430e-01
1.15709805e+00 -5.00068069e-01 -2.90541589e-01 5.39474050e-03
6.54114187e-01 -7.43359983e-01 1.05831158e+00 3.37720275e-01
-1.18051338e+00 -5.59325814e-01 -1.46389985e+00 3.89400035e-01
-4.90717113e-01 -2.88000315e-01 5.97516358e-01 5.19896805e-01
-6.28481686e-01 1.36624420e+00 -8.01769674e-01 -8.58668983e-03
8.15185308e-02 7.68324733e-01 -1.28340319e-01 -9.12255235e-03
-1.03855658e+00 1.13485861e+00 3.92908335e-01 1.36630639e-01
-8.65254223e-01 -1.58348751e+00 -7.14783609e-01 -4.32468176e-01
2.12186128e-01 -5.99182963e-01 1.15909290e+00 -2.19630450e-01
-1.84608448e+00 4.56205666e-01 -1.84515268e-02 2.49424100e-01
3.13011587e-01 -1.65180191e-01 -8.75819385e-01 6.97101802e-02
-8.46610889e-02 -1.37381941e-01 8.66901219e-01 -1.66971958e+00
-1.07268952e-01 -3.04323822e-01 -7.19605327e-01 -4.28195208e-01
2.41204187e-01 -1.94159344e-01 3.27523559e-01 -3.83628935e-01
5.36332071e-01 -9.69806552e-01 -3.43394518e-01 -4.78183091e-01
-5.13371706e-01 1.53483123e-01 1.01907492e+00 -6.94025517e-01
8.48076463e-01 -1.91599965e+00 1.56795368e-01 9.86248195e-01
2.25029647e-01 -4.62576061e-01 2.30566755e-01 1.00184822e+00
-3.96025091e-01 -2.55504828e-02 -4.71908629e-01 2.85072654e-01
2.07823306e-01 -1.49174072e-02 6.13259859e-02 5.80751002e-01
3.58296156e-01 1.11037505e+00 -4.31705326e-01 1.09820422e-02
4.28822517e-01 -5.36765642e-02 -4.65792269e-01 7.92172328e-02
-7.01546848e-01 4.90970850e-01 -5.16946554e-01 1.00441766e+00
3.94397646e-01 8.93998519e-02 -1.54650092e-01 -6.55920506e-01
-2.14751124e-01 -2.21407458e-01 -6.61461234e-01 1.26439703e+00
-5.06604493e-01 6.63876683e-02 2.20144372e-02 -7.14284360e-01
1.24291015e+00 3.42201263e-01 1.15438044e+00 -1.07827067e+00
2.61968285e-01 6.62883878e-01 1.42111510e-01 -6.63490355e-01
6.39559567e-01 -8.31240177e-01 -1.27750695e-01 3.74650002e-01
-5.84706590e-02 -1.06804967e+00 -2.48233765e-01 -2.81289101e-01
1.02747369e+00 2.61914402e-01 -1.55535668e-01 -7.72669852e-01
5.36467016e-01 4.54320490e-01 2.43799478e-01 4.39801633e-01
3.37975770e-01 5.86556137e-01 4.67330813e-02 -3.70268404e-01
-1.91275883e+00 -1.10849392e+00 -2.78206974e-01 6.87954843e-01
-2.21747868e-02 -3.69448781e-01 -1.86442822e-01 4.33552474e-01
5.48879921e-01 9.21824813e-01 -5.81037283e-01 -9.69606638e-01
-5.80464184e-01 -7.38420486e-01 -5.72467670e-02 3.90906125e-01
-3.02143693e-01 -7.87257791e-01 -2.95747459e-01 6.94390476e-01
8.45127285e-01 -7.57210076e-01 1.34714350e-01 4.22800899e-01
-1.05698490e+00 -1.30913591e+00 -8.14079717e-02 -5.11051416e-01
2.00614795e-01 -6.17695510e-01 9.70506132e-01 7.81057402e-02
-8.16921413e-01 7.13811278e-01 -1.25337750e-01 -3.32015008e-01
-1.44885087e+00 -1.37242720e-01 3.80043209e-01 -3.64143431e-01
6.41337559e-02 -6.25024438e-01 -5.81297457e-01 5.00378847e-01
-8.63258779e-01 -2.03507487e-02 7.95937598e-01 8.90438855e-01
7.33290553e-01 4.53311890e-01 5.05349159e-01 -7.90840209e-01
6.19687021e-01 -4.97471035e-01 -2.84991115e-01 6.54183984e-01
-1.46522141e+00 4.80282873e-01 4.29048508e-01 -3.39599460e-01
-1.12160432e+00 -4.29154746e-02 1.38600424e-01 -3.41436476e-01
-2.37524375e-01 6.85239673e-01 -1.60729691e-01 -1.84914067e-01
6.47890329e-01 2.64769420e-02 3.22755724e-01 -5.37280202e-01
8.98702145e-02 4.09118801e-01 4.81177479e-01 -1.49193740e+00
1.02432096e+00 -1.20806314e-01 5.73615134e-01 -8.60822499e-01
-9.02754217e-02 -1.28782675e-01 -7.22369432e-01 -4.64123279e-01
7.05293477e-01 -6.27118945e-01 -1.03210366e+00 4.87954915e-01
-4.38875526e-01 -2.25791857e-01 -4.97682363e-01 5.64666450e-01
-1.00012755e+00 -9.05332789e-02 -8.16878557e-01 -8.88773978e-01
-3.34857851e-01 -1.60045767e+00 1.00792170e+00 -1.47311866e-01
-7.80970633e-01 -9.58164334e-01 3.71164903e-02 4.75254655e-01
6.06258333e-01 5.58344543e-01 1.93365133e+00 -5.86079836e-01
-4.53808576e-01 -2.72352457e-01 4.93610531e-01 1.21157691e-01
4.98573959e-01 4.48237658e-01 -9.58585620e-01 -2.94493437e-01
3.33117634e-01 -7.25964755e-02 6.89488828e-01 4.41058636e-01
7.43852735e-01 5.52563742e-02 -3.22026104e-01 1.49027795e-01
1.66634512e+00 6.28760278e-01 3.57390642e-01 6.33911252e-01
8.78472805e-01 5.85408330e-01 4.10563380e-01 2.78162360e-01
-2.92037159e-01 5.28785229e-01 -2.04625428e-02 6.35686442e-02
1.08035222e-01 -8.37636590e-02 3.43363881e-01 9.13775504e-01
-2.52214521e-01 5.30716360e-01 -1.12304592e+00 -3.17299254e-02
-1.34155965e+00 -7.22278535e-01 -3.89292836e-01 1.95638287e+00
9.86045361e-01 4.43782985e-01 -8.71966928e-02 3.69420826e-01
4.13997382e-01 -1.69635341e-01 -7.88517714e-01 -6.65392637e-01
-3.20146620e-01 7.29949594e-01 7.65805900e-01 3.53556484e-01
-4.70651746e-01 4.84529823e-01 7.22423840e+00 7.93331563e-01
-1.01509976e+00 -2.90217429e-01 5.12434840e-01 5.04491106e-02
-7.20660985e-01 1.75522372e-01 -1.50280014e-01 2.05161065e-01
1.34001780e+00 -2.38354489e-01 5.81039488e-01 5.79266310e-01
4.46929723e-01 -3.27288389e-01 -1.61485410e+00 6.18631124e-01
-5.20292044e-01 -1.65650535e+00 -4.35021631e-02 -1.14260511e-02
7.43682325e-01 -3.54604185e-01 2.76873738e-01 -7.92816356e-02
1.72829211e-01 -8.05241048e-01 1.00317299e+00 8.49721789e-01
7.60922551e-01 -4.52039033e-01 2.82623976e-01 -3.85552675e-01
-1.11915624e+00 -5.89742362e-01 2.29539461e-02 1.20117038e-01
1.90212995e-01 7.70299852e-01 -9.87115800e-01 1.07890952e+00
4.88965452e-01 5.80424309e-01 -5.79089284e-01 5.96159160e-01
4.44924206e-01 5.51318645e-01 -3.94084632e-01 9.71045569e-02
4.79132272e-02 -3.70906264e-01 2.59818614e-01 5.44826925e-01
3.64133090e-01 -9.19216648e-02 -1.36389598e-01 1.22740149e+00
4.56486017e-01 2.01541614e-02 -3.73614818e-01 -2.94531971e-01
3.86270702e-01 5.93091011e-01 -4.83004779e-01 1.69923514e-01
-1.30857199e-01 2.74072677e-01 -4.10387754e-01 4.10158873e-01
-4.69960511e-01 3.32544714e-01 6.43373728e-01 7.02671170e-01
-1.69928819e-02 -5.25077462e-01 -6.20785892e-01 -2.31601179e-01
-2.71636248e-01 -6.87318623e-01 -4.19771224e-02 -9.90086675e-01
-1.55987728e+00 6.10543042e-03 4.19993758e-01 -8.13333213e-01
1.39063029e-02 -9.35079277e-01 -5.32805800e-01 8.24200094e-01
-8.16721082e-01 -1.16290355e+00 4.04273480e-01 1.63027987e-01
3.88160706e-01 -6.11690581e-01 7.09900439e-01 2.24535353e-02
-3.89714271e-01 2.73746908e-01 8.56772184e-01 -6.86835229e-01
6.12866282e-01 -1.06962252e+00 -4.07571681e-02 -6.35779798e-02
-5.47318876e-01 3.66355568e-01 1.25092685e+00 -1.08713865e+00
-2.23803592e+00 -7.54319310e-01 -6.59855455e-02 -3.82567436e-01
9.85701561e-01 -1.71888903e-01 -8.44103098e-01 2.51030892e-01
-1.32435769e-01 -7.70276606e-01 6.72321796e-01 -1.07886735e-03
-9.64734256e-02 -1.86751023e-01 -1.09057474e+00 4.46429968e-01
4.84675914e-01 -6.71629250e-01 -6.89593136e-01 3.20431322e-01
5.29969156e-01 -1.02102742e-01 -1.70654929e+00 6.36885524e-01
7.11215734e-01 -1.39088333e-01 1.08909047e+00 -8.37028205e-01
8.83295596e-01 6.73325956e-02 -4.64739382e-01 -1.17374003e+00
-6.97304249e-01 -4.85523164e-01 1.56661887e-02 9.92292404e-01
7.84433842e-01 -5.19427955e-01 7.05935657e-01 1.50533664e+00
-4.89700377e-01 -9.80570674e-01 -9.20751631e-01 -8.50005388e-01
4.35512483e-01 -4.12660390e-01 8.66546214e-01 5.79961300e-01
-1.75028220e-02 -1.69340745e-01 1.01769820e-01 -1.40019003e-02
5.05355120e-01 1.55026913e-01 1.10191613e-01 -1.55230200e+00
-2.57496089e-01 -5.69920480e-01 -4.53653574e-01 9.59086642e-02
-1.90188605e-02 -9.10573125e-01 1.23645015e-01 -1.00025821e+00
-1.52743444e-01 -8.25427771e-01 -5.46654940e-01 3.00397109e-02
2.71560699e-01 -5.18072069e-01 3.55862118e-02 6.08123243e-01
3.75354737e-01 8.24748397e-01 1.34315801e+00 -6.43607557e-01
-3.84995341e-01 -1.96683154e-01 -3.86579275e-01 1.22442171e-01
7.47712255e-01 -4.76613581e-01 -3.14444691e-01 3.24148327e-01
3.22903931e-01 2.22754553e-01 7.13956431e-02 -1.35484517e+00
2.75816191e-02 -4.96913135e-01 5.37048280e-01 -5.86670220e-01
4.25754011e-01 -1.21930337e+00 1.06872594e+00 5.44160664e-01
-3.75790149e-01 2.38618553e-01 3.01324964e-01 5.81826448e-01
3.16220783e-02 -1.13363057e-01 5.55107415e-01 -2.32038781e-01
-6.96349263e-01 1.95354089e-01 -5.29454052e-01 -6.36275709e-01
1.26871216e+00 -6.43346727e-01 -1.03465714e-01 2.50402987e-01
-1.01964188e+00 -1.03031427e-01 7.50506401e-01 4.31028515e-01
5.30996919e-01 -1.33238900e+00 -5.21846116e-01 2.65257299e-01
1.30886450e-01 -1.80217236e-01 4.93736416e-01 5.62106967e-01
-7.79477119e-01 2.13117585e-01 -2.79848397e-01 -6.73391938e-01
-6.89671993e-01 8.46229255e-01 4.35173184e-01 -4.24050242e-02
-4.20877904e-01 4.57871675e-01 -7.71814883e-02 -3.54946852e-01
-7.81028330e-01 -3.09776366e-01 3.68508399e-01 -7.75821954e-02
-2.99827844e-01 3.91937256e-01 4.98492122e-01 -4.80293810e-01
-1.88088298e-01 7.19416440e-01 -2.10003272e-01 9.85542536e-02
1.86995196e+00 1.69886529e-01 -1.83855787e-01 1.00642312e+00
1.38949776e+00 -4.98556167e-01 -1.04915047e+00 -1.28685251e-01
4.93600219e-01 2.27066781e-02 -6.70106933e-02 -8.44071865e-01
-1.03056455e+00 2.73622215e-01 8.53988707e-01 1.05387503e-02
7.49391794e-01 3.65382850e-01 6.27954543e-01 4.07858461e-01
1.61703259e-01 -1.80072796e+00 -6.48698658e-02 2.38758460e-01
1.07359517e+00 -8.71268094e-01 5.86953223e-01 -2.02904895e-01
-7.44658783e-02 1.52819800e+00 4.71984267e-01 1.71529159e-01
1.11801648e+00 4.04944092e-01 -2.92353272e-01 -8.39576364e-01
-4.65224057e-01 7.88742900e-01 3.29473436e-01 2.29650930e-01
2.00858265e-01 2.53049344e-01 -7.59384185e-02 4.11796898e-01
-4.89456862e-01 -3.23693514e-01 1.67856261e-01 1.50355792e+00
-4.01639700e-01 -1.27999544e+00 -7.83073902e-01 9.31458950e-01
2.81567782e-01 5.97682521e-02 -2.07631946e-01 9.69056308e-01
-1.73337162e-01 8.33895385e-01 -1.05117463e-01 -8.78773868e-01
6.15003943e-01 5.52870572e-01 6.09912694e-01 -3.74130249e-01
-5.56443334e-01 -2.49069557e-01 2.78128684e-02 -4.82109487e-01
-2.00219765e-01 -8.42315555e-01 -1.29362559e+00 -4.16657239e-01
-2.84864038e-01 4.96194884e-02 9.85770166e-01 1.12911856e+00
5.45153916e-02 7.49188721e-01 9.87715185e-01 -1.19144976e+00
-4.48550075e-01 -7.95254409e-01 -1.00914681e+00 6.46734416e-01
1.61710098e-01 -1.18871796e+00 -2.68557429e-01 3.47107798e-02] | [5.282450199127197, 5.266504287719727] |
056ffa2e-6eba-45df-a5c1-2009aafbbd59 | centered-self-attention-layers | 2306.01610 | null | https://arxiv.org/abs/2306.01610v1 | https://arxiv.org/pdf/2306.01610v1.pdf | Centered Self-Attention Layers | The self-attention mechanism in transformers and the message-passing mechanism in graph neural networks are repeatedly applied within deep learning architectures. We show that this application inevitably leads to oversmoothing, i.e., to similar representations at the deeper layers for different tokens in transformers and different nodes in graph neural networks. Based on our analysis, we present a correction term to the aggregating operator of these mechanisms. Empirically, this simple term eliminates much of the oversmoothing problem in visual transformers, obtaining performance in weakly supervised segmentation that surpasses elaborate baseline methods that introduce multiple auxiliary networks and training phrases. In graph neural networks, the correction term enables the training of very deep architectures more effectively than many recent solutions to the same problem. | ['Lior Wolf', 'Tomer Galanti', 'Ameen Ali'] | 2023-06-02 | null | null | null | null | ['weakly-supervised-segmentation'] | ['computer-vision'] | [ 1.48390442e-01 7.97535658e-01 1.47415310e-01 -4.05024618e-01
-4.95346695e-01 -7.62638688e-01 7.26615071e-01 2.94765174e-01
-4.43260610e-01 4.10383731e-01 2.41482705e-01 -7.55648851e-01
2.75759041e-01 -7.45372951e-01 -9.63295162e-01 -5.38063645e-01
-3.62668559e-02 5.11262715e-01 5.92940450e-01 -1.08985849e-01
-8.51358548e-02 4.40661132e-01 -9.88654971e-01 2.11805448e-01
7.97860026e-01 7.57818699e-01 1.50177822e-01 8.23648930e-01
-4.65447724e-01 8.89356852e-01 -6.15152001e-01 -8.43758464e-01
1.75584152e-01 -2.95259625e-01 -1.05236483e+00 1.63018480e-01
1.15224576e+00 -3.62577409e-01 -4.35936391e-01 1.18329060e+00
1.09647773e-01 1.54786780e-01 5.59824944e-01 -1.08432221e+00
-1.04128075e+00 1.28013241e+00 -6.31962955e-01 3.84559482e-01
-2.22757593e-01 2.20156431e-01 1.35844791e+00 -7.32808769e-01
4.40858781e-01 1.31858313e+00 1.15050101e+00 3.48505080e-01
-1.41493106e+00 -1.17611833e-01 5.10500968e-01 -1.85859472e-01
-1.12304115e+00 -1.74202740e-01 6.79215908e-01 -2.71817297e-01
1.24493206e+00 4.17860188e-02 7.99694359e-01 8.14451694e-01
1.14011541e-02 9.18312907e-01 4.91482705e-01 -1.14403002e-01
-9.59472656e-02 6.66541010e-02 4.06771183e-01 1.16389394e+00
3.52673948e-01 -2.72225201e-01 -3.43146212e-02 3.37140821e-02
9.70484376e-01 -1.08432949e-01 -1.99983165e-01 -4.60751772e-01
-8.67076278e-01 8.26733887e-01 8.89191806e-01 3.52217495e-01
-1.80426836e-01 6.64452910e-01 3.96038145e-01 2.80177534e-01
5.61403215e-01 5.49681187e-01 -5.92305779e-01 2.20052302e-01
-8.99128675e-01 -1.90980248e-02 7.53472924e-01 9.82120633e-01
8.12903941e-01 4.78780806e-01 -3.31406206e-01 5.93910992e-01
2.08981901e-01 4.27937470e-02 2.99832672e-01 -8.97328138e-01
5.93946159e-01 6.72605217e-01 -3.67370516e-01 -6.63511872e-01
-5.37110567e-01 -8.80984545e-01 -7.38010824e-01 8.24984759e-02
8.59498501e-01 -2.82338232e-01 -1.48637664e+00 1.85466111e+00
-1.62356570e-02 6.69023767e-02 -1.26394659e-01 4.83869404e-01
8.12514663e-01 5.81019104e-01 3.71742725e-01 1.55051410e-01
1.14412963e+00 -1.31482530e+00 -4.63756979e-01 -5.83530307e-01
7.43442893e-01 -4.71946776e-01 1.34635556e+00 1.72305271e-01
-1.35194814e+00 -4.25766230e-01 -8.77196312e-01 -5.26981354e-01
-4.43048120e-01 -2.21432447e-01 9.96817946e-01 4.28921938e-01
-1.65689492e+00 1.05827379e+00 -9.00651455e-01 -5.02315700e-01
9.17488098e-01 5.06016612e-01 -1.54883936e-01 1.75448552e-01
-8.71924818e-01 9.90025699e-01 2.36704618e-01 1.70652911e-01
-8.54787707e-01 -8.31110477e-01 -1.06618786e+00 5.49563587e-01
2.96520352e-01 -7.83162832e-01 1.47887683e+00 -1.45106113e+00
-1.29339659e+00 9.45909262e-01 -3.52648576e-03 -7.46246457e-01
5.06020010e-01 -2.31924936e-01 2.46840507e-01 8.91013891e-02
-1.58031315e-01 7.68968225e-01 8.94249141e-01 -1.20104277e+00
-3.06994319e-01 -7.57462606e-02 2.91399866e-01 2.25268230e-01
-2.42136821e-01 -2.41383120e-01 -7.04489887e-01 -8.00244510e-01
1.99009199e-02 -7.63800085e-01 -5.02894819e-01 -5.83335273e-02
-6.48527026e-01 -2.71219760e-01 5.32172799e-01 -7.61354685e-01
8.10262024e-01 -1.96638298e+00 3.44810843e-01 3.35524857e-01
6.31013215e-01 2.67108768e-01 -4.92707640e-01 1.45546451e-01
-3.12454939e-01 3.04585904e-01 -5.52194297e-01 -5.09574831e-01
-2.93063145e-04 4.49071586e-01 -7.64609054e-02 3.76224756e-01
3.62510026e-01 1.47066784e+00 -9.24848914e-01 -3.37574840e-01
7.18176812e-02 4.04511362e-01 -5.65828800e-01 -1.17116414e-01
-4.69733983e-01 -9.74228457e-02 -4.81017195e-02 4.72285867e-01
5.26292801e-01 -6.84210241e-01 2.31588617e-01 -1.95538089e-01
1.52471021e-01 7.29342401e-01 -5.88808179e-01 1.75924563e+00
-3.44302148e-01 7.23395705e-01 3.75153840e-01 -1.11944211e+00
3.04622084e-01 1.87590733e-01 1.06158294e-01 -5.57527721e-01
1.67406335e-01 3.92514467e-03 1.85126841e-01 -3.00931662e-01
4.54473853e-01 -2.42974877e-01 2.16057926e-01 3.97468120e-01
5.67918897e-01 -2.47084081e-01 1.54256016e-01 6.44731581e-01
1.09422946e+00 1.65388539e-01 -2.46887922e-01 -3.97815913e-01
4.25019860e-02 -4.45694141e-02 3.50389779e-01 1.12403083e+00
-1.49560973e-01 5.89909434e-01 9.97779012e-01 -3.75106722e-01
-1.18941379e+00 -9.67595696e-01 3.61842811e-01 1.29029453e+00
-1.99047983e-01 -4.98853683e-01 -9.32592869e-01 -1.01385367e+00
1.32664546e-01 4.89290386e-01 -7.03675151e-01 -2.74365246e-01
-6.54196441e-01 -6.87400997e-01 6.16632938e-01 9.77711260e-01
2.45143965e-01 -1.00097013e+00 -8.72922614e-02 1.74639016e-01
2.43689716e-01 -1.15770638e+00 -3.46402824e-01 6.20333374e-01
-1.18850899e+00 -7.63636410e-01 -8.79500508e-01 -1.03883862e+00
8.05001557e-01 6.96888641e-02 1.63560915e+00 6.07630372e-01
8.99204165e-02 2.68705726e-01 5.75311109e-02 -2.44474262e-01
-3.90272409e-01 5.81709564e-01 -6.60803020e-01 -3.84463221e-01
1.85614335e-03 -7.62624741e-01 -3.93333018e-01 -3.02220047e-01
-7.50173211e-01 1.09147523e-02 6.16683424e-01 7.33583748e-01
1.15839794e-01 -1.27285272e-01 5.16129062e-02 -1.29683721e+00
7.50642002e-01 -3.30186307e-01 -6.61454082e-01 1.96587995e-01
-4.66788203e-01 5.06263077e-01 5.96089303e-01 -4.18893605e-01
-6.75691128e-01 3.43170688e-02 -4.00517434e-01 -2.58421123e-01
2.45559542e-03 4.90704924e-01 -3.18553559e-02 -2.31150493e-01
4.55251962e-01 -8.33941251e-02 -4.18069251e-02 -4.23280776e-01
7.40239501e-01 -1.55342907e-01 6.02694333e-01 -3.59816462e-01
1.12688804e+00 4.01634008e-01 -1.70741588e-01 -6.21181607e-01
-1.05569291e+00 2.44363192e-02 -6.13487244e-01 2.31772557e-01
9.30006385e-01 -7.96239495e-01 -3.93638909e-01 5.61745405e-01
-1.29950428e+00 -8.83350670e-01 -6.68600857e-01 2.03742489e-01
-1.47562653e-01 3.93539399e-01 -1.23855329e+00 -4.04142737e-01
-9.66926888e-02 -1.09065509e+00 8.27451646e-01 2.40205616e-01
-1.32997692e-01 -1.34956396e+00 -2.90070236e-01 -6.68816268e-02
5.38377583e-01 -5.75801358e-02 1.17516792e+00 -8.89718533e-01
-7.48004556e-01 6.20608702e-02 -5.05843580e-01 4.05724108e-01
1.17352843e-01 1.90303519e-01 -1.07939792e+00 -8.98873135e-02
-2.16260597e-01 -1.63136572e-01 1.47210801e+00 5.52335918e-01
1.09093678e+00 -4.88686323e-01 -1.21285848e-01 7.79102385e-01
1.26340854e+00 -1.46274373e-01 4.81027037e-01 1.51421860e-01
1.34869611e+00 5.36210835e-01 -3.96443188e-01 -2.91451484e-01
4.92580622e-01 1.51044741e-01 6.78511262e-01 -7.67835021e-01
-2.42443055e-01 -2.25111991e-01 2.45684385e-01 7.10657775e-01
4.57706079e-02 -2.67290533e-01 -1.00280809e+00 7.91732311e-01
-1.77352595e+00 -7.80461848e-01 -2.57156730e-01 1.78958416e+00
6.49954915e-01 6.21575654e-01 1.40693098e-01 -1.36643201e-01
6.09555840e-01 4.58282381e-01 -5.08316457e-01 -7.89711535e-01
-2.38652021e-01 5.44107378e-01 7.93950438e-01 9.52342451e-01
-1.04578066e+00 1.39043856e+00 7.50733423e+00 6.79484785e-01
-9.77102518e-01 1.67750388e-01 7.64196992e-01 1.91531792e-01
-8.01893055e-01 6.62878156e-02 -3.25125605e-01 -3.26529741e-02
8.71872962e-01 1.92094848e-01 5.68477631e-01 6.84227943e-01
-1.24749117e-01 8.39331374e-02 -1.16180670e+00 4.37281013e-01
-8.33821669e-02 -1.49384677e+00 3.20150524e-01 -4.00112346e-02
5.02589762e-01 5.60920179e-01 1.73012465e-01 3.31469804e-01
9.38546956e-01 -1.12097716e+00 7.72171617e-01 -8.03978890e-02
2.98248619e-01 -5.50262153e-01 4.86374617e-01 -2.81530786e-02
-1.14427924e+00 -2.53991559e-02 -3.37110817e-01 -1.84378341e-01
1.76379532e-01 6.41113997e-01 -8.27874184e-01 3.09805572e-01
3.66365969e-01 6.31682515e-01 -6.78182781e-01 8.44527423e-01
-5.85829258e-01 7.39780426e-01 -4.13891971e-01 9.63386074e-02
7.81098604e-01 -3.08468938e-01 4.20106351e-01 1.30887222e+00
-8.16331357e-02 -2.76181817e-01 -1.11126922e-01 9.73024726e-01
-5.15328467e-01 -1.92015693e-01 -5.82799256e-01 -3.84287387e-01
1.04054119e-02 1.26725268e+00 -1.02534366e+00 -6.65104866e-01
-6.17268682e-01 1.05958235e+00 8.48045409e-01 8.36516500e-01
-8.61156166e-01 -3.00029337e-01 4.33026642e-01 2.78285891e-01
7.37850845e-01 -3.22880924e-01 -4.36334521e-01 -9.50392962e-01
-3.20511870e-02 -5.91178596e-01 4.32157636e-01 -6.85376525e-01
-1.06447959e+00 7.17589736e-01 -3.93635929e-01 -3.24092150e-01
-4.99451235e-02 -6.74456120e-01 -9.89360809e-01 8.92246544e-01
-1.56195068e+00 -1.31138110e+00 4.33246717e-02 5.38945377e-01
3.59745324e-01 2.09177643e-01 3.69187355e-01 2.98776567e-01
-5.74275553e-01 4.62502182e-01 -3.08548957e-01 4.41122532e-01
1.74543440e-01 -1.81172121e+00 1.14425099e+00 1.15538013e+00
5.32497585e-01 7.29796231e-01 7.20780253e-01 -4.99109536e-01
-9.58570719e-01 -1.05483496e+00 8.90005887e-01 -4.88185465e-01
1.00013280e+00 -6.22821450e-01 -1.38087511e+00 1.27086246e+00
6.71452880e-01 -7.12545216e-02 1.49873465e-01 4.42273259e-01
-5.84581017e-01 1.19164363e-01 -7.52027452e-01 6.35767639e-01
9.96522248e-01 -7.38142669e-01 -7.04602420e-01 3.82482499e-01
1.07395864e+00 -4.43962812e-01 -4.65951204e-01 7.93541893e-02
2.02320307e-01 -7.98019946e-01 9.25392985e-01 -9.05138195e-01
2.63179451e-01 -2.13837624e-02 3.42504114e-01 -1.28276598e+00
-3.55299175e-01 -7.52377033e-01 -2.14604899e-01 1.18464243e+00
8.40641081e-01 -6.40133500e-01 9.98420119e-01 7.62954533e-01
-4.44355220e-01 -6.17992520e-01 -6.52776361e-01 -5.17814040e-01
4.46691036e-01 -1.82025477e-01 3.73820156e-01 1.01713598e+00
-2.04270691e-01 6.03041947e-01 -9.40863136e-03 1.16095662e-01
5.33134699e-01 -2.12553740e-01 4.00518119e-01 -1.18182170e+00
-4.00119483e-01 -7.70317078e-01 -9.04496685e-02 -1.37423921e+00
3.13001484e-01 -1.12316155e+00 2.83722654e-02 -1.99420047e+00
5.68041466e-02 -2.82032639e-01 -3.10606003e-01 7.62114823e-01
-3.30196470e-01 2.90223777e-01 2.05719635e-01 -1.36344746e-01
-3.94743294e-01 3.40741813e-01 1.21110618e+00 -4.05175984e-01
-2.81805605e-01 -1.33298844e-01 -1.03543401e+00 1.08474827e+00
7.95193672e-01 -6.22084320e-01 -4.85660404e-01 -1.03998685e+00
4.57567573e-01 -4.35053289e-01 5.78645587e-01 -6.71382308e-01
1.25963777e-01 1.83534339e-01 3.56277466e-01 -2.69051433e-01
8.93711001e-02 -6.77854717e-01 -2.69109815e-01 5.45044959e-01
-3.53778839e-01 2.68982291e-01 4.99426752e-01 3.60088050e-01
4.76602092e-02 -1.77102610e-01 6.81777656e-01 -5.49623668e-01
-5.34089684e-01 3.92607599e-01 -6.34248674e-01 2.33539969e-01
4.19217408e-01 -3.21942359e-01 -3.87286156e-01 -4.28440511e-01
-8.65399122e-01 3.64410669e-01 3.81773829e-01 9.15747061e-02
2.95847923e-01 -1.04829502e+00 -5.15039861e-01 -3.67909372e-02
-3.97834301e-01 3.19912612e-01 -1.46787599e-01 9.73800480e-01
-6.10625923e-01 -1.43281028e-01 4.78940979e-02 -5.63908756e-01
-9.47283149e-01 4.01464313e-01 7.03231514e-01 -4.78599459e-01
-8.25036108e-01 1.34646070e+00 4.01421070e-01 -3.48260343e-01
4.35120404e-01 -8.31708550e-01 -1.60228158e-03 3.79115455e-02
2.92340736e-03 -8.59797224e-02 1.50137201e-01 -2.99062103e-01
-2.81769454e-01 2.52771735e-01 -3.49423170e-01 -1.04779884e-01
1.44746518e+00 -6.83872700e-02 -1.05826691e-01 3.43555242e-01
9.79556501e-01 -2.79733557e-02 -1.45379603e+00 -1.68505028e-01
1.99676752e-01 2.09673136e-01 8.15116912e-02 -5.93958437e-01
-1.49500144e+00 1.11413944e+00 1.33102596e-01 4.76195663e-01
7.80818045e-01 2.26777568e-01 8.15807641e-01 4.18311447e-01
-2.82145172e-01 -9.94187832e-01 -9.48163494e-02 6.46885633e-01
5.45278907e-01 -1.08325958e+00 -2.25294139e-02 -3.76343191e-01
-4.75841105e-01 9.07637537e-01 5.91918945e-01 -3.82154793e-01
4.73235995e-01 5.53029895e-01 1.07387923e-01 -4.63787705e-01
-7.03453064e-01 -3.81411582e-01 1.52353689e-01 6.89267397e-01
5.04203200e-01 -2.63136923e-01 -4.61354181e-02 2.57044971e-01
-7.33618140e-02 -4.25323069e-01 6.12666011e-01 7.90633857e-01
-4.64375257e-01 -7.86355913e-01 1.24046996e-01 4.15319026e-01
-6.06959283e-01 -7.66223490e-01 -6.65540040e-01 1.09138060e+00
-2.20070168e-01 5.36519408e-01 4.61332411e-01 -2.12886697e-03
2.90380388e-01 6.13897294e-02 5.44738352e-01 -6.48533106e-01
-1.08539391e+00 1.65269494e-01 2.04599127e-01 -6.09160841e-01
-2.46387735e-01 -2.08262235e-01 -1.38138044e+00 -2.54029572e-01
-2.82868028e-01 2.84704775e-01 3.37141812e-01 9.79592383e-01
2.31388196e-01 8.62932384e-01 1.25599196e-02 -8.77700210e-01
-4.49798644e-01 -9.06637430e-01 -5.62503040e-01 4.84912783e-01
6.12038612e-01 -2.12872401e-01 -4.06758308e-01 -2.67916247e-02] | [6.887908458709717, 6.120258331298828] |
16b8afbf-9f4f-4b5f-be2f-70d3c97cb77e | mapping-research-topics-in-software-testing-a | 2109.04086 | null | https://arxiv.org/abs/2109.04086v4 | https://arxiv.org/pdf/2109.04086v4.pdf | Mapping the Structure and Evolution of Software Testing Research Over the Past Three Decades | Background: The field of software testing is growing and rapidly-evolving. Aims: Based on keywords assigned to publications, we seek to identify predominant research topics and understand how they are connected and have evolved. Method: We apply co-word analysis to map the topology of testing research as a network where author-assigned keywords are connected by edges indicating co-occurrence in publications. Keywords are clustered based on edge density and frequency of connection. We examine the most popular keywords, summarize clusters into high-level research topics, examine how topics connect, and examine how the field is changing. Results: Testing research can be divided into 16 high-level topics and 18 subtopics. Creation guidance, automated test generation, evolution and maintenance, and test oracles have particularly strong connections to other topics, highlighting their multidisciplinary nature. Emerging keywords relate to web and mobile apps, machine learning, energy consumption, automated program repair and test generation, while emerging connections have formed between web apps, test oracles, and machine learning with many topics. Random and requirements-based testing show potential decline. Conclusions: Our observations, advice, and map data offer a deeper understanding of the field and inspiration regarding challenges and connections to explore. | ['Ehsan Mohammadi', 'Gregory Gay', 'Alireza Salahirad'] | 2021-09-09 | null | null | null | null | ['program-repair', 'program-repair'] | ['computer-code', 'reasoning'] | [-1.08754799e-01 -3.82063836e-02 -7.15243876e-01 1.50203943e-01
-4.13037419e-01 -9.47969139e-01 2.42204934e-01 3.62059236e-01
1.92672566e-01 3.60505879e-01 3.23249549e-01 -9.66883421e-01
-7.46335745e-01 -5.63694179e-01 -8.81541431e-01 9.11390111e-02
-1.66154623e-01 2.35329643e-01 1.21562935e-01 -4.41072211e-02
9.52897191e-01 -3.29780765e-03 -1.57978773e+00 2.23266575e-02
1.20893764e+00 6.08151436e-01 8.06500316e-02 3.22635680e-01
-4.36704844e-01 4.90181506e-01 -1.02067840e+00 -2.69735098e-01
-3.56584519e-01 -5.64029336e-01 -7.83245683e-01 -1.58070493e-02
2.54225507e-02 5.70666552e-01 6.62601292e-02 1.03795731e+00
3.03146988e-01 -4.75042641e-01 1.68725789e-01 -1.87672532e+00
-7.60198116e-01 7.77198255e-01 -6.95256174e-01 4.36486214e-01
7.40006387e-01 -9.64699965e-03 8.55523825e-01 -5.42518973e-01
1.04251409e+00 5.67350149e-01 9.40132976e-01 1.20456710e-01
-1.19359660e+00 -7.44107604e-01 2.09193602e-01 4.15450633e-01
-1.38589871e+00 1.44474506e-01 6.48097515e-01 -8.84737611e-01
1.30200827e+00 3.51036161e-01 1.18071377e+00 1.11462069e+00
5.72137892e-01 7.37269893e-02 7.86649764e-01 -4.79282886e-01
2.78057784e-01 3.91999274e-01 1.83960766e-01 7.56806433e-01
8.14811528e-01 -7.22007453e-01 -8.62478316e-01 -5.21647453e-01
4.50493842e-02 -2.65178755e-02 -4.82876301e-01 2.34902017e-02
-1.34011471e+00 4.50282633e-01 -5.25084913e-01 3.89593810e-01
-2.41016150e-01 -2.13074591e-02 2.07538709e-01 5.67710459e-01
5.65489471e-01 9.04889405e-01 -7.22813547e-01 -9.72444892e-01
-1.23572433e+00 -1.53967023e-01 1.25586820e+00 1.35236990e+00
6.52343094e-01 1.18165195e-01 1.19500466e-01 8.75868917e-01
4.61785436e-01 -2.41348483e-02 6.04543686e-01 -8.42497826e-01
5.59430778e-01 1.10732138e+00 -3.49577576e-01 -1.09083915e+00
-2.95127034e-01 -7.26382434e-01 -9.21179727e-03 -1.43646181e-01
-4.44012225e-01 -4.49704021e-01 -5.34640849e-01 1.30778372e+00
-3.65207493e-01 3.47079813e-01 -5.05625069e-01 6.11430593e-02
5.81110239e-01 3.70660484e-01 -1.70095608e-01 -2.23038629e-01
1.24344611e+00 -8.18712294e-01 -8.07773530e-01 -4.97121736e-02
6.51961446e-01 -7.39694834e-01 1.02519119e+00 6.13717377e-01
-1.18232381e+00 -3.68710995e-01 -1.30058718e+00 5.87233245e-01
-8.77144277e-01 -3.69870514e-01 3.09871137e-01 9.84730899e-01
-1.68242979e+00 4.99350488e-01 -6.34416103e-01 -6.58401489e-01
4.06945467e-01 1.12576053e-01 8.92461613e-02 -1.85675785e-01
-6.84162378e-01 8.54906380e-01 -2.95669194e-02 -4.45298553e-01
-4.53006268e-01 -1.02932525e+00 -5.62980235e-01 -5.74085452e-02
6.06570542e-01 -6.71370745e-01 8.33522975e-01 8.40474758e-03
-9.18698430e-01 4.17838633e-01 -8.72283522e-03 2.28242218e-01
-9.00662169e-02 2.78630376e-01 -1.23662126e+00 -7.38810673e-02
3.97293210e-01 8.95783678e-02 7.56166428e-02 -8.98054838e-01
-9.69543040e-01 -1.98270828e-01 -2.99039602e-01 -2.08626404e-01
-9.16374326e-01 2.55812079e-01 -7.04542756e-01 -4.32381600e-01
1.06647231e-01 -6.66985631e-01 5.43768071e-02 -7.07070172e-01
-3.69838625e-01 -3.99370909e-01 1.11718595e+00 -5.77938914e-01
1.95486736e+00 -1.95005012e+00 1.04027562e-01 3.19934994e-01
4.15602446e-01 -6.68392062e-01 -6.13336777e-03 8.36062312e-01
-1.06714675e-02 1.04447877e+00 3.52138251e-01 5.55867136e-01
2.05343336e-01 -3.10679764e-01 1.15049161e-01 3.54645401e-01
2.23117724e-01 1.05038214e+00 -8.72470677e-01 -4.41803068e-01
-1.75021827e-01 3.43826175e-01 -5.12241304e-01 -3.90609503e-01
-2.84340322e-01 -5.36661297e-02 -3.07072788e-01 1.27391672e+00
2.66600877e-01 -6.02473199e-01 1.95559278e-01 8.73051137e-02
-5.52057743e-01 5.47589540e-01 -8.05523753e-01 1.35612333e+00
-4.63526845e-01 1.01973867e+00 2.81366631e-02 -8.88414800e-01
1.06277561e+00 2.37233847e-01 7.29920924e-01 -5.25963902e-01
-3.54802608e-01 3.46381664e-01 3.88219476e-01 -8.15092802e-01
2.33098656e-01 7.36352980e-01 -1.28662333e-01 7.99926996e-01
1.15434993e-02 -3.07780564e-01 4.28080738e-01 2.23215535e-01
1.87501085e+00 7.98313916e-02 -2.29456946e-01 -6.59832597e-01
-1.23570912e-01 3.72532964e-01 3.91415149e-01 3.83083999e-01
2.27182582e-01 1.98691219e-01 9.96072352e-01 1.88949659e-01
-6.63572192e-01 -7.16671705e-01 -3.10622841e-01 9.19601858e-01
-3.25183347e-02 -1.08541477e+00 -6.83243036e-01 -6.18644297e-01
2.38788649e-01 7.36332178e-01 -5.36590874e-01 -1.14387400e-01
-1.15640327e-01 -3.43665600e-01 2.60483921e-01 2.85135895e-01
2.44085267e-01 -9.74252760e-01 -6.21756613e-01 1.71033889e-01
-1.23417810e-01 -7.55036533e-01 -4.02041733e-01 4.06531483e-01
-1.02772439e+00 -1.14160562e+00 -4.68414873e-01 -9.11709189e-01
6.08388960e-01 1.28503278e-01 1.62740278e+00 1.48510769e-01
-6.34681523e-01 1.07937229e+00 -4.18495893e-01 -4.22843337e-01
-1.23253889e-01 6.08148873e-02 -2.76428491e-01 -8.69629204e-01
6.66914463e-01 -5.46126783e-01 -3.55617762e-01 7.35610664e-01
-6.42466962e-01 -6.93280697e-01 7.27073967e-01 4.19089615e-01
2.96651870e-01 5.63572168e-01 5.65863967e-01 -9.18149173e-01
1.10222638e+00 -1.28336370e+00 -2.90430188e-01 4.41056162e-01
-1.54502642e+00 -2.87419081e-01 -1.19112164e-01 -3.69157195e-01
-2.03947335e-01 -9.08080518e-01 2.91402459e-01 -8.88213962e-02
-5.03250770e-02 1.27577114e+00 -1.44171536e-01 -1.23334855e-01
6.41473413e-01 5.20898998e-02 -1.85161725e-01 -2.28097051e-01
-1.50545314e-01 7.35202789e-01 -3.61127257e-02 -7.40419209e-01
5.52218378e-01 -2.04900637e-01 -3.82046282e-01 -6.86062515e-01
1.44196346e-01 -3.30243647e-01 1.29688606e-01 -6.84338808e-01
7.27120459e-01 -3.22610199e-01 -5.82174361e-01 -2.75521874e-01
-7.82788396e-01 -1.74989134e-01 -3.25689286e-01 4.83116478e-01
-2.87939478e-02 6.70492947e-02 -4.98323292e-01 -4.78348523e-01
1.14560105e-01 -1.37872732e+00 7.30145454e-01 2.15491369e-01
-8.92995059e-01 -1.16119230e+00 1.35457352e-01 2.93704450e-01
7.34730303e-01 2.46486202e-01 1.37241542e+00 -6.32365465e-01
-3.57362181e-01 -3.06560174e-02 1.11562811e-01 -4.27276380e-02
4.36133742e-01 4.80870008e-01 -1.50488928e-01 -1.54558852e-01
-3.63497585e-02 2.55207300e-01 2.99800903e-01 2.72333920e-01
1.42649293e+00 -2.34009355e-01 -8.03129971e-01 2.09383905e-01
1.33006394e+00 7.61908591e-01 6.18130267e-01 7.31520057e-01
3.36654186e-01 8.74015391e-01 5.03564775e-01 2.66962647e-01
3.71195525e-01 2.53570914e-01 3.12106043e-01 2.51002282e-01
2.33888552e-01 -1.60450682e-01 4.47934896e-01 1.51644456e+00
-1.44341081e-01 -3.89359593e-01 -1.62198246e+00 7.46906161e-01
-1.47300005e+00 -6.71196222e-01 -5.63964903e-01 1.81823468e+00
5.57177782e-01 5.98855019e-01 1.25717953e-01 -2.93160472e-02
9.00548100e-01 -3.26724082e-01 -5.51796854e-01 -4.14868146e-01
1.59118995e-01 4.28652078e-01 3.63950789e-01 -1.45986229e-01
-9.45699140e-02 2.99401432e-01 7.53556824e+00 4.22943294e-01
-7.23458230e-01 2.30138525e-01 6.61949396e-01 -2.80804969e-02
-8.89795780e-01 2.77918912e-02 -4.97760147e-01 8.16764057e-01
1.27875376e+00 -7.33572781e-01 2.21629515e-01 9.45637703e-01
-6.32661581e-02 -6.91080168e-02 -1.01661611e+00 8.50040615e-01
2.97989398e-01 -1.57687032e+00 -3.73391479e-01 3.57784569e-01
1.15060484e+00 1.62373111e-01 2.64546156e-01 3.61138105e-01
3.07193935e-01 -8.34446132e-01 5.00028372e-01 5.17855763e-01
8.29904079e-01 -7.67212033e-01 6.57543719e-01 -2.24855781e-01
-1.06151688e+00 -1.47855893e-01 3.44931670e-02 -6.47011725e-03
-2.71110862e-01 7.51549721e-01 -5.57972193e-01 5.62223077e-01
1.22943509e+00 8.32565367e-01 -6.65426612e-01 1.21149480e+00
4.65243235e-02 1.03294694e+00 2.38965929e-01 -5.61454594e-01
-1.58488870e-01 -3.99761736e-01 3.35455865e-01 1.45123696e+00
8.95131826e-01 -4.21369255e-01 -5.57979718e-02 1.03332043e+00
-1.13863736e-01 -1.44190341e-02 -6.86220407e-01 -9.16411936e-01
7.94305742e-01 1.23413837e+00 -1.31909657e+00 -2.49087978e-02
-7.84703493e-01 5.14763415e-01 -6.18842959e-01 5.75987399e-01
-8.75230849e-01 -9.10246789e-01 7.55706787e-01 5.83859444e-01
2.55892366e-01 -2.35885203e-01 -5.71635306e-01 -4.92443502e-01
2.78948426e-01 -1.07400477e+00 1.42162591e-01 -5.73987782e-01
-1.03699398e+00 4.23702031e-01 -2.32223392e-01 -1.23010767e+00
4.65727895e-02 -3.68746072e-01 -7.95592546e-01 6.95207894e-01
-9.27346945e-01 -1.89136475e-01 -4.79656845e-01 1.62596866e-01
5.10266483e-01 -5.15352786e-01 4.71293479e-01 4.04442459e-01
-5.44951141e-01 6.32919192e-01 -1.13318756e-01 -5.65656841e-01
6.18709743e-01 -1.13148427e+00 6.11837029e-01 3.79134178e-01
-2.80850995e-02 8.60689163e-01 4.28914338e-01 -1.28285170e+00
-1.86666787e+00 -7.19152510e-01 7.30595589e-01 -7.38920987e-01
1.29545736e+00 -3.93928260e-01 -6.94805741e-01 3.93819869e-01
6.64216816e-01 -5.65404415e-01 1.03717875e+00 5.51152229e-01
-1.54953852e-01 5.42426221e-02 -9.61602688e-01 6.11801267e-01
1.29075444e+00 -6.53612316e-01 -1.28530383e-01 5.28229296e-01
9.21225488e-01 1.41428411e-01 -1.01782715e+00 2.58689940e-01
2.62934834e-01 -6.30856216e-01 3.39184046e-01 -3.71916622e-01
5.50108671e-01 -1.14602707e-01 1.75731570e-01 -1.34805810e+00
-3.51001889e-01 -1.03871071e+00 9.93214489e-05 1.30389273e+00
1.15008247e+00 -9.22148705e-01 8.26419055e-01 4.72919673e-01
-6.78224385e-01 -1.05038130e+00 -6.00275934e-01 -8.41349423e-01
-1.49063647e-01 -6.22985482e-01 5.17524302e-01 1.56980383e+00
1.04410672e+00 2.63080209e-01 5.45564234e-01 -3.74178231e-01
1.40423521e-01 -2.56426185e-01 1.88024104e-01 -1.65159500e+00
-4.39751267e-01 -1.00291777e+00 -7.67138004e-01 -5.02192080e-01
-1.85426190e-01 -1.02295732e+00 -2.61051178e-01 -1.97525311e+00
3.03299725e-01 -1.67089701e-01 -2.44434059e-01 4.58746761e-01
1.70218706e-01 4.33382541e-02 -5.18534422e-01 -7.48763829e-02
-6.71879888e-01 -3.39124560e-01 7.58109510e-01 -1.93915069e-01
-2.02472478e-01 -1.64785936e-01 -1.28762400e+00 3.14844340e-01
8.66810977e-01 -3.11978996e-01 -4.50399190e-01 -2.66830981e-01
1.19908774e+00 -1.96160838e-01 -8.07696655e-02 -1.07268500e+00
6.98127270e-01 -1.31446775e-02 5.09079918e-02 -6.14912391e-01
-4.32652801e-01 -8.66861641e-01 5.88550627e-01 4.98528808e-01
2.66851578e-02 9.85782683e-01 6.54819608e-01 3.36664766e-01
-2.17487693e-01 -3.49939615e-01 -1.40861824e-01 2.20782995e-01
-3.64495039e-01 -1.80268869e-01 -7.42023528e-01 2.28317022e-01
1.18041015e+00 -5.40696800e-01 -7.66366899e-01 -4.26713884e-01
-6.32840633e-01 3.21369797e-01 4.78986442e-01 6.89496458e-01
6.73397720e-01 -1.21405554e+00 -4.47468519e-01 -1.03206679e-01
9.05348063e-02 -7.55387485e-01 -1.87325943e-02 1.06787634e+00
-3.07947189e-01 4.21049714e-01 -1.02508247e-01 -6.51262641e-01
-1.01221609e+00 4.33151156e-01 -3.08632344e-01 2.64063776e-01
-2.36107275e-01 7.42148042e-01 -6.87683761e-01 -1.34472698e-01
4.58592862e-01 -5.08437634e-01 -2.76770871e-02 4.02561128e-01
6.53540939e-02 1.00600946e+00 3.32608581e-01 2.24651754e-01
-4.57123339e-01 6.46803141e-01 9.54727083e-02 -2.39125520e-01
1.68490791e+00 2.12375537e-01 -7.50549316e-01 1.02539051e+00
1.10741937e+00 2.42522955e-01 -3.92732114e-01 4.95189399e-01
6.19419754e-01 -1.48396939e-01 1.65785488e-03 -1.17290533e+00
-9.49673772e-01 3.15233558e-01 3.71617198e-01 9.21075225e-01
1.05473924e+00 4.31143492e-01 7.14000314e-02 6.37056231e-02
4.63136613e-01 -1.49369657e+00 4.34357315e-01 2.51074404e-01
6.70507491e-01 -5.04305661e-01 -7.38988305e-03 -5.35313249e-01
-1.09769061e-01 1.13595688e+00 8.17340016e-01 5.15971422e-01
1.08468533e+00 9.24311996e-01 -3.76138330e-01 -7.32396781e-01
-8.01788688e-01 3.52899134e-01 1.93840116e-01 4.00063068e-01
6.29274726e-01 -2.14926690e-01 -4.88545656e-01 5.97353220e-01
-3.46074492e-01 -1.14858218e-01 6.34663820e-01 1.18914723e+00
-3.70844394e-01 -1.16413796e+00 -3.40373963e-01 1.19385338e+00
-4.12036777e-01 -3.16576660e-01 -7.96662927e-01 8.61766398e-01
2.16034591e-01 1.24393547e+00 2.46439412e-01 -1.15711880e+00
5.09947956e-01 3.48005354e-01 6.98566586e-02 -8.52177918e-01
-6.45472229e-01 2.24981811e-02 1.49677962e-01 -5.19707859e-01
1.55049384e-01 -9.41436648e-01 -8.83550346e-01 -2.65312970e-01
-7.96327174e-01 7.19830036e-01 1.24468672e+00 4.94200885e-01
1.01784551e+00 1.36652040e+00 3.65678132e-01 5.86211085e-02
1.74523309e-01 -1.01747608e+00 -3.58574599e-01 -4.30697888e-01
-2.76469469e-01 -5.67082405e-01 -5.25688767e-01 -8.62212926e-02] | [7.690410137176514, 7.848615646362305] |
f69de2d8-b6df-4e27-8dc9-725d244b38d2 | cerec-a-corpus-for-entity-resolution-in-email | 2105.10606 | null | https://arxiv.org/abs/2105.10606v2 | https://arxiv.org/pdf/2105.10606v2.pdf | CEREC: A Corpus for Entity Resolution in Email Conversations | We present the first large scale corpus for entity resolution in email conversations (CEREC). The corpus consists of 6001 email threads from the Enron Email Corpus containing 36,448 email messages and 60,383 entity coreference chains. The annotation is carried out as a two-step process with minimal manual effort. Experiments are carried out for evaluating different features and performance of four baselines on the created corpus. For the task of mention identification and coreference resolution, a best performance of 59.2 F1 is reported, highlighting the room for improvement. An in-depth qualitative and quantitative error analysis is presented to understand the limitations of the baselines considered. | ['Dan I. Moldovan', 'Parag Pravin Dakle'] | 2021-05-21 | null | https://aclanthology.org/2020.coling-main.30 | https://aclanthology.org/2020.coling-main.30.pdf | coling-2020-8 | ['entity-resolution'] | ['natural-language-processing'] | [ 3.17056865e-01 7.56456792e-01 -2.41941363e-01 -4.45149869e-01
-1.49822342e+00 -7.44397581e-01 9.66995001e-01 3.64027739e-01
-8.47285390e-01 9.71074045e-01 8.96989465e-01 -3.92191172e-01
-1.97242424e-02 -1.41685933e-01 -3.09120238e-01 -2.83610225e-01
1.57344621e-02 9.99679923e-01 2.35545039e-01 -2.58072138e-01
3.31223845e-01 4.63012308e-01 -9.57928181e-01 7.31962323e-01
4.95610505e-01 3.53096128e-01 1.51172262e-02 1.02974284e+00
-5.08516788e-01 7.36577749e-01 -1.03197968e+00 -1.03937912e+00
-4.30586696e-01 -3.52332711e-01 -1.66968071e+00 -2.79493392e-01
3.20311874e-01 3.23486477e-01 -1.47170588e-01 7.84937143e-01
7.31275558e-01 1.11551933e-01 5.69799781e-01 -9.52412307e-01
-2.92880982e-02 1.12889385e+00 -3.68627787e-01 5.56829035e-01
7.66129136e-01 -3.95816922e-01 1.34923828e+00 -7.33206093e-01
1.23420179e+00 1.51927531e+00 7.19220877e-01 6.91017091e-01
-1.27348328e+00 -7.63374329e-01 -8.13446864e-02 7.05429167e-02
-1.06069946e+00 -7.65855432e-01 1.97023019e-01 -2.62509793e-01
1.46963441e+00 5.33199847e-01 5.81258498e-02 1.37003458e+00
-2.33674839e-01 5.15330911e-01 9.11332726e-01 -6.91641152e-01
-2.37184241e-01 4.05311584e-01 6.18847668e-01 2.37785116e-01
2.61869580e-01 -1.57501966e-01 -6.39434755e-01 -5.76707304e-01
2.60917544e-01 -7.44683027e-01 -3.72229308e-01 2.03686431e-01
-1.07975924e+00 7.92344034e-01 1.24449536e-01 7.09904373e-01
-3.58839720e-01 -2.34912068e-01 6.94817901e-01 4.11791086e-01
3.69538248e-01 7.60190427e-01 -6.26185358e-01 -7.05838799e-01
-5.22853494e-01 3.09395373e-01 1.40863466e+00 1.29879200e+00
3.51002425e-01 -7.48267591e-01 -1.39498770e-01 9.78764951e-01
2.27738827e-01 3.55529279e-01 2.20474065e-03 -1.33197904e+00
9.35871959e-01 5.60576022e-01 2.54832864e-01 -9.78149533e-01
-6.69980884e-01 2.42602620e-02 -4.17136759e-01 -5.77514529e-01
4.40226436e-01 -4.69267219e-01 -2.16212839e-01 1.48591328e+00
2.68286735e-01 -1.36236668e-01 1.77996561e-01 3.49079967e-01
1.30050969e+00 3.49158496e-01 4.47991818e-01 -5.39298475e-01
1.82585061e+00 -8.47034395e-01 -1.20536947e+00 -9.45415050e-02
7.77384162e-01 -1.31281126e+00 2.86441386e-01 -3.48204792e-01
-1.09878027e+00 -2.21275985e-01 -4.65075135e-01 -9.73840877e-02
-2.57542342e-01 -3.73427004e-01 4.60436553e-01 6.23086691e-01
-7.46374965e-01 3.35975647e-01 -5.54172397e-01 -8.26008618e-01
-1.76497191e-01 2.72595406e-01 -4.96875882e-01 3.94963175e-01
-1.36904776e+00 1.07577240e+00 4.58808422e-01 -1.31962746e-01
1.12497471e-01 -6.80981338e-01 -6.46630764e-01 -1.12945437e-02
2.62050718e-01 -6.93388134e-02 1.61668956e+00 -3.40169668e-01
-1.13159108e+00 1.19482327e+00 -4.55299139e-01 -5.52693188e-01
5.53787470e-01 -4.16350245e-01 -6.87585175e-01 3.05514574e-01
2.65044481e-01 5.48287392e-01 -5.27058057e-02 -1.29106069e+00
-1.04530251e+00 6.02606945e-02 -1.36669695e-01 1.56426921e-01
3.61191750e-01 9.27359998e-01 -7.85874248e-01 -5.98441422e-01
-2.09103048e-01 -1.19987977e+00 6.59991726e-02 -1.13678455e+00
-3.27409416e-01 -5.92833996e-01 5.37227213e-01 -8.47784162e-01
1.60916805e+00 -1.85936999e+00 -5.64940050e-02 1.34492680e-01
1.06229063e-03 1.65732265e-01 -3.75244841e-02 8.60511243e-01
-1.74506471e-01 3.18584561e-01 -9.67206880e-02 -6.76207423e-01
2.80602518e-02 -6.78378344e-03 -2.06802249e-01 9.52036977e-02
5.11077307e-02 9.44394588e-01 -8.18817377e-01 -8.32299352e-01
-2.04098329e-01 4.99825478e-01 -3.74291182e-01 2.79771626e-01
1.16868638e-01 4.32641387e-01 -3.95454347e-01 3.48994523e-01
4.67389047e-01 -2.54259109e-01 7.42533207e-01 -9.70906988e-02
-8.94782096e-02 9.21993256e-01 -1.08728969e+00 1.44651318e+00
-4.74208146e-01 9.06645477e-01 5.76530039e-01 -3.50377947e-01
7.70093918e-01 7.06171930e-01 2.97890842e-01 -3.20366025e-01
2.58480847e-01 2.66312420e-01 7.98973441e-02 -3.44941825e-01
9.27727640e-01 5.48748635e-02 -5.92581272e-01 7.22718120e-01
1.43668532e-01 2.38491148e-01 1.73364982e-01 7.99269199e-01
1.17660832e+00 -2.66469449e-01 3.46317530e-01 -3.46018553e-01
6.87033832e-01 2.18596503e-01 6.68087363e-01 7.69170523e-01
-3.58583361e-01 4.48646307e-01 4.81043190e-01 -1.44473941e-03
-6.31426990e-01 -4.16589051e-01 -3.13178986e-01 1.15322113e+00
-1.28511310e-01 -7.11438477e-01 -9.60254192e-01 -9.59050775e-01
-1.65366828e-01 8.04670095e-01 -5.73135257e-01 3.34610343e-01
-1.11729717e+00 -6.96312487e-01 9.07144964e-01 5.54325640e-01
3.03121448e-01 -1.38853407e+00 -3.72464418e-01 3.81406605e-01
-1.04628682e+00 -1.58702302e+00 -6.99394941e-01 1.01980373e-01
-4.78317708e-01 -1.47145557e+00 -2.56060362e-01 -8.84367526e-01
1.61617562e-01 -2.98303855e-03 1.53724372e+00 3.13354611e-01
-8.17177445e-02 5.56211948e-01 -3.82356316e-01 -5.96793965e-02
-8.24886382e-01 6.35257959e-01 -1.88084468e-01 -5.17380536e-01
1.08671081e+00 -1.70269296e-01 -2.58624732e-01 2.72095382e-01
-6.63034618e-02 -2.45609060e-01 6.19206429e-01 9.35943484e-01
-1.06645934e-01 -5.73397160e-01 7.97105372e-01 -1.56631446e+00
7.76769698e-01 -3.42064232e-01 -7.65480846e-02 2.74216264e-01
-6.21566355e-01 -9.97366533e-02 -4.95954752e-02 -1.56229123e-01
-1.74903333e+00 -1.80212051e-01 -4.24485326e-01 5.48371255e-01
-3.62302631e-01 9.51812342e-02 -1.03724658e-01 3.28024805e-01
6.00019336e-01 -4.50755835e-01 -2.70345956e-01 -9.12843287e-01
3.65271807e-01 1.33594930e+00 9.26533163e-01 -8.78017664e-01
3.59102011e-01 -1.84341714e-01 -5.82208157e-01 -8.86828125e-01
-9.78912830e-01 -1.01937127e+00 -9.37769115e-01 -8.66251513e-02
9.53829169e-01 -7.54445553e-01 -1.08191431e+00 1.74881220e-01
-1.52560556e+00 -1.61806613e-01 8.36514831e-02 4.40133214e-01
-2.14266002e-01 2.56508946e-01 -1.11481893e+00 -7.81619370e-01
-5.28911054e-01 -8.63249779e-01 9.18031037e-01 2.34519109e-01
-9.74579096e-01 -1.02606487e+00 2.22392112e-01 8.13547015e-01
2.02525809e-01 2.35388987e-02 6.35852218e-01 -1.38493347e+00
-1.17086850e-01 -2.36581638e-01 -3.40802699e-01 -3.65097880e-01
4.65640053e-03 -6.27357587e-02 -8.26030910e-01 -2.98284262e-01
-4.98196036e-01 -2.53684670e-01 6.19743943e-01 -2.47122541e-01
3.29133481e-01 -2.23013520e-01 -8.28549445e-01 6.69796672e-03
9.35629368e-01 4.11896646e-01 6.20952904e-01 7.45268404e-01
4.50634122e-01 9.82609570e-01 5.48505306e-01 7.76003022e-03
5.55420816e-01 7.11295605e-01 -1.80708498e-01 2.78948396e-01
-2.20098451e-01 -1.19547859e-01 -1.28584996e-01 1.06179190e+00
-2.27096677e-01 -1.72795609e-01 -1.00838459e+00 6.77925885e-01
-1.80248618e+00 -1.04541326e+00 -4.86942738e-01 1.77198064e+00
1.27946651e+00 2.90237904e-01 1.07412085e-01 -2.40977511e-01
1.28496659e+00 9.09824669e-02 1.90833718e-01 -4.43471491e-01
-1.15385443e-01 1.97838277e-01 3.78644168e-01 1.03194809e+00
-1.14767933e+00 9.62600470e-01 7.31403875e+00 2.81983554e-01
-2.41571993e-01 -2.10651476e-02 3.48539293e-01 1.32856935e-01
-6.62724748e-02 1.51911154e-01 -1.31850839e+00 3.93867284e-01
1.58749747e+00 -1.70072228e-01 6.22203015e-02 3.72721404e-01
-1.95572376e-01 2.05249283e-02 -9.67793882e-01 6.49801314e-01
-1.29551753e-01 -1.39343965e+00 -3.85561824e-01 5.16756140e-02
5.49776256e-01 6.25102669e-02 -7.70139694e-01 7.15214431e-01
7.71776378e-01 -6.79686487e-01 2.78252453e-01 3.33373904e-01
6.21406376e-01 -9.35284734e-01 1.17477262e+00 1.57205075e-01
-9.97971058e-01 2.72014290e-01 6.71237409e-02 3.27894390e-01
5.46795547e-01 1.12883084e-01 -8.05434823e-01 6.40461028e-01
9.20392215e-01 2.87077010e-01 -4.06546682e-01 4.59685385e-01
-1.60930589e-01 6.92596555e-01 -2.36401916e-01 -1.54059976e-01
1.08038150e-01 2.68646032e-02 7.00818002e-01 2.18029118e+00
-3.58909249e-01 5.06947339e-01 -9.65194926e-02 1.87355042e-01
-6.95805430e-01 8.47833753e-02 -1.29507795e-01 -5.88495508e-02
1.24697447e+00 1.48757827e+00 -6.86387360e-01 -4.54992622e-01
-4.58314627e-01 7.91071713e-01 5.59453487e-01 2.09171653e-01
-6.01297259e-01 -7.24827290e-01 6.12590730e-01 -2.26047203e-01
2.34143317e-01 1.75976187e-01 -1.64654050e-02 -8.33837390e-01
-4.13610101e-01 -1.14236307e+00 7.07707882e-01 -8.50876644e-02
-1.14984286e+00 7.47973323e-01 8.26366618e-02 -3.88640672e-01
-5.26725888e-01 -2.46599317e-01 -6.02972090e-01 9.26749945e-01
-1.08768809e+00 -7.92905807e-01 -1.39920428e-01 1.30215630e-01
5.25987506e-01 -1.24358147e-01 1.14823961e+00 5.57473123e-01
-6.56003296e-01 8.13088596e-01 -9.75963380e-03 5.48460841e-01
1.21052146e+00 -1.43059134e+00 7.71202266e-01 4.44455862e-01
-1.08380854e-01 1.02906549e+00 9.58377182e-01 -7.11898029e-01
-9.69210207e-01 -7.50107646e-01 1.94444454e+00 -8.41621518e-01
8.29261124e-01 -4.59378004e-01 -1.20358729e+00 9.97153580e-01
8.82534027e-01 -6.48711562e-01 8.71841729e-01 6.33914769e-01
-2.26994351e-01 5.08398533e-01 -1.21028364e+00 3.09498489e-01
1.07892716e+00 -6.31649256e-01 -1.15898764e+00 9.82654020e-02
8.26893687e-01 -6.06337965e-01 -1.37951815e+00 3.05696428e-01
3.82235855e-01 -6.58402324e-01 7.96066582e-01 -7.78011620e-01
-8.85658637e-02 2.60007232e-01 -9.15030316e-02 -8.58595908e-01
-4.00884062e-01 -1.16615295e+00 -1.97184637e-01 2.05377173e+00
6.69944048e-01 -5.81211567e-01 4.78756011e-01 8.36499453e-01
9.37555805e-02 -2.27266923e-01 -8.02378595e-01 -2.44945362e-01
1.31225556e-01 -1.83464065e-01 5.50143957e-01 1.09613514e+00
3.03205431e-01 1.17537558e+00 -4.19139639e-02 -1.39525518e-01
5.17285705e-01 4.44004387e-02 6.41885221e-01 -1.53169584e+00
-1.00270491e-02 -3.24217677e-01 3.16367447e-01 -9.78739798e-01
5.10939181e-01 -6.74162686e-01 1.70223042e-01 -1.36500323e+00
4.41203177e-01 -3.97503972e-01 7.16263652e-02 1.92189127e-01
-4.81492281e-01 7.76487077e-03 1.71750844e-01 4.65225220e-01
-7.84831524e-01 1.15487717e-01 5.33885956e-01 1.38515726e-01
-1.37871593e-01 -1.80466343e-02 -9.29021060e-01 5.71716547e-01
6.31380439e-01 -6.20628119e-01 3.01552981e-01 4.00380082e-02
-8.81651118e-02 1.36875987e-01 -3.78279090e-01 -2.11330384e-01
4.12505001e-01 6.59983754e-02 -4.31531668e-02 -8.89665663e-01
1.45516366e-01 -5.24882436e-01 6.05466068e-02 2.68655539e-01
-7.51773179e-01 2.79669642e-01 2.62657851e-01 6.38489187e-01
-1.50066212e-01 -4.34706748e-01 6.58506811e-01 -9.31012630e-02
-4.89008635e-01 -4.70942497e-01 -5.12109101e-01 7.59602964e-01
7.02427983e-01 3.64810914e-01 -6.16047859e-01 -2.76174605e-01
-7.25302994e-01 4.73125845e-01 2.68902034e-01 6.60369396e-01
-7.42868111e-02 -1.21695054e+00 -9.14358616e-01 -3.47345978e-01
3.34808752e-02 -3.33796829e-01 -2.04679184e-02 6.21681750e-01
-1.62143633e-01 7.99338758e-01 2.74779707e-01 -2.78541893e-01
-1.91950512e+00 2.46656135e-01 1.76823869e-01 -5.63785791e-01
-8.15384805e-01 7.18184650e-01 -4.27895278e-01 -7.46630251e-01
5.52518845e-01 3.37415993e-01 -5.84200442e-01 4.54264253e-01
7.24956155e-01 6.42153800e-01 1.83739379e-01 -1.04804122e+00
-7.56462991e-01 1.70904025e-01 -4.39165413e-01 -2.99180418e-01
1.21664405e+00 -5.82407892e-01 -3.00828040e-01 2.73583025e-01
1.30287504e+00 3.65802497e-01 -5.97983301e-01 -4.05259788e-01
8.81711364e-01 -1.25776798e-01 -3.73700768e-01 -9.71533358e-01
-4.39890534e-01 3.32741350e-01 2.34296039e-01 2.75301427e-01
5.09479284e-01 3.81921738e-01 7.80074716e-01 5.71241498e-01
7.68841431e-02 -1.26501489e+00 -3.49842757e-01 9.17145073e-01
8.05665076e-01 -1.35487926e+00 -1.30563065e-01 -5.27600110e-01
-7.80263186e-01 9.68947589e-01 5.72744370e-01 3.77565324e-01
4.78214711e-01 5.34841537e-01 3.34932655e-01 -4.53509092e-01
-1.00990832e+00 -7.27613568e-02 2.46563941e-01 3.07094991e-01
1.09001815e+00 -1.10516489e-01 -8.24544787e-01 7.36285567e-01
-3.34991276e-01 -5.98640382e-01 3.91934395e-01 9.34823811e-01
-3.00564408e-01 -1.32016742e+00 -2.15766519e-01 1.48112833e-01
-1.17301059e+00 -3.94898243e-02 -9.20822322e-01 1.05893147e+00
-4.83315736e-01 1.35861433e+00 1.89635634e-01 -5.38544357e-02
6.14649057e-01 4.74077463e-01 1.63329080e-01 -6.36365950e-01
-1.12114608e+00 3.17460410e-02 1.05856633e+00 -3.42869759e-01
-8.70624244e-01 -1.00716043e+00 -1.56288314e+00 -5.59543371e-01
-4.98132318e-01 9.39453661e-01 4.08375472e-01 7.60156572e-01
4.55800653e-01 5.23445368e-01 4.19568241e-01 -4.76281047e-01
-2.25520611e-01 -1.41453624e+00 -3.05115152e-02 7.79889286e-01
6.50684088e-02 -5.09180069e-01 -4.99149144e-01 -5.15173562e-02] | [9.390941619873047, 9.371644020080566] |
0e011554-9f2d-4419-a8ee-ab52e857cb92 | facial-action-unit-detection-using-3d-facial | 2005.08343 | null | https://arxiv.org/abs/2005.08343v1 | https://arxiv.org/pdf/2005.08343v1.pdf | Facial Action Unit Detection using 3D Facial Landmarks | In this paper, we propose to detect facial action units (AU) using 3D facial landmarks. Specifically, we train a 2D convolutional neural network (CNN) on 3D facial landmarks, tracked using a shape index-based statistical shape model, for binary and multi-class AU detection. We show that the proposed approach is able to accurately model AU occurrences, as the movement of the facial landmarks corresponds directly to the movement of the AUs. By training a CNN on 3D landmarks, we can achieve accurate AU detection on two state-of-the-art emotion datasets, namely BP4D and BP4D+. Using the proposed method, we detect multiple AUs on over 330,000 frames, reporting improved results over state-of-the-art methods. | ['Shaun Canavan', 'Saurabh Hinduja'] | 2020-05-17 | null | null | null | null | ['action-unit-detection', 'facial-action-unit-detection'] | ['computer-vision', 'computer-vision'] | [-2.20776960e-01 1.01397663e-01 7.15228692e-02 -2.63131201e-01
-6.43519044e-01 -3.15094590e-01 5.76562822e-01 -7.55040199e-02
-5.37704468e-01 -2.09851451e-02 -8.49128589e-02 5.03621757e-01
4.61416155e-01 -5.37260413e-01 -7.70543516e-01 -4.93394315e-01
-3.21748734e-01 1.66928262e-01 -1.33093134e-01 -2.51625568e-01
2.11727154e-03 1.07930338e+00 -1.78211284e+00 1.50730968e-01
-6.19623475e-02 1.82665086e+00 -6.32243931e-01 5.32949090e-01
6.81268424e-02 3.28408420e-01 -5.99986851e-01 -4.52337652e-01
3.48832190e-01 -3.71057153e-01 -5.19620299e-01 1.14822447e-01
7.35617876e-01 -4.34203774e-01 -2.80760735e-01 1.03934085e+00
4.55256999e-01 -1.73477586e-02 8.46682966e-01 -1.32699299e+00
-5.51874518e-01 -2.09010273e-01 -8.63503993e-01 -7.67347366e-02
5.06979406e-01 -2.87482888e-01 1.07401526e+00 -1.24933600e+00
6.28536642e-01 1.51213717e+00 9.03041363e-01 9.18265402e-01
-8.50685656e-01 -7.01042175e-01 -1.25116378e-01 4.51784953e-02
-1.68202233e+00 -6.93125188e-01 9.39955771e-01 -2.39263818e-01
1.10673237e+00 -1.33804128e-01 7.68137932e-01 1.07068288e+00
1.03770614e-01 9.67037499e-01 5.92276812e-01 -4.05133247e-01
1.43248886e-01 -3.71014237e-01 -2.19534144e-01 1.35134363e+00
-3.46284598e-01 7.90267885e-02 -6.06989264e-01 -2.19270617e-01
8.03327918e-01 -1.29255094e-02 1.62367985e-01 -2.42413040e-02
-4.50209409e-01 6.88245714e-01 4.09586221e-01 5.44239044e-01
-5.64327478e-01 4.70019966e-01 5.24320543e-01 9.20563936e-02
9.17738795e-01 -6.02462552e-02 -2.78458357e-01 -4.28430796e-01
-8.99686754e-01 4.57814857e-02 3.32108259e-01 7.20910430e-01
8.12148988e-01 1.14844218e-01 -7.23562986e-02 7.87357509e-01
3.92272949e-01 5.34152985e-01 3.49394321e-01 -1.01110435e+00
-2.03354403e-01 8.74513984e-01 -3.40854637e-02 -1.28448403e+00
-8.79802167e-01 3.00877482e-01 -7.11215079e-01 4.36495543e-01
3.25270981e-01 -1.86507672e-01 -9.60908711e-01 1.78634667e+00
3.25186372e-01 5.51683426e-01 -6.17160685e-02 8.32679510e-01
1.13555920e+00 3.90942812e-01 -7.57042915e-02 -1.13824226e-01
1.25507987e+00 -6.44032061e-01 -5.84748566e-01 2.01969802e-01
7.39586651e-01 -4.99751091e-01 5.87052345e-01 2.68993348e-01
-1.15975821e+00 -7.09966481e-01 -6.92336857e-01 7.61708245e-02
-3.35744858e-01 6.31059468e-01 3.14484894e-01 6.06443346e-01
-1.37592101e+00 5.56153417e-01 -9.80639040e-01 -4.90695894e-01
8.25886309e-01 4.08201188e-01 -7.60814905e-01 2.29773790e-01
-9.17449296e-01 7.37712324e-01 -1.42296836e-01 4.09876972e-01
-1.02798593e+00 -3.79859984e-01 -1.18524981e+00 7.68879652e-02
-2.13530183e-01 -7.56331310e-02 9.91642177e-01 -1.09516394e+00
-1.70476592e+00 1.38355172e+00 -3.10932636e-01 -1.38421178e-01
3.91070426e-01 -3.08962047e-01 -4.65544194e-01 5.21634161e-01
-2.15930477e-01 9.49019432e-01 1.15139270e+00 -1.10690153e+00
-4.66457576e-01 -5.28464556e-01 -4.64360565e-02 -2.02797890e-01
-5.54054797e-01 5.27870238e-01 -5.65361261e-01 -2.46102199e-01
8.91152248e-02 -9.50593233e-01 1.78992718e-01 5.85744679e-01
-1.10664479e-01 -8.19436431e-01 9.15834069e-01 -4.15705830e-01
8.29123437e-01 -2.53625584e+00 4.81027141e-02 1.98363483e-01
3.44063371e-01 2.82565832e-01 -4.57789004e-01 -1.47105873e-01
-7.92316571e-02 -2.44965125e-02 8.61561149e-02 -1.11676979e+00
1.76881164e-01 1.76741794e-01 1.49302751e-01 9.06551480e-01
5.72325349e-01 1.00992596e+00 -7.44588196e-01 -3.36754173e-01
2.62082487e-01 7.67808914e-01 -3.46280128e-01 1.78924575e-01
1.35326028e-01 1.21788913e-02 -1.38731152e-01 1.21044743e+00
6.75340414e-01 1.58158958e-01 -1.95869714e-01 -3.25491816e-01
5.91296591e-02 -3.34527135e-01 -7.83246398e-01 1.63047934e+00
-5.23605347e-01 8.13807607e-01 7.69434422e-02 -8.93931568e-01
1.44063580e+00 4.17302996e-01 9.13883269e-01 -6.34468794e-01
6.11638427e-01 1.94285825e-01 -5.52092612e-01 -2.70321280e-01
1.73268273e-01 4.73227911e-03 -9.22302995e-03 3.81125212e-01
5.27208149e-01 1.08842388e-01 -1.03492469e-01 -3.56428981e-01
1.02027786e+00 7.00476617e-02 2.54953951e-01 -8.46212804e-02
5.56061924e-01 -5.41917264e-01 4.39564705e-01 4.11308438e-01
-7.32583284e-01 5.64176738e-01 7.46771336e-01 -8.83383393e-01
-7.53202379e-01 -8.08276892e-01 -1.45234764e-01 1.15871942e+00
-4.39838693e-02 -4.17800367e-01 -7.98650861e-01 -9.15175319e-01
9.97737795e-02 1.59562394e-01 -1.26136494e+00 -2.92577714e-01
-4.79970545e-01 -5.07603943e-01 1.02867448e+00 5.86136580e-01
5.37725627e-01 -1.15957224e+00 -8.30907166e-01 8.31957087e-02
2.43335158e-01 -1.24128103e+00 -2.63874918e-01 -1.44146189e-01
-4.13902760e-01 -1.23000979e+00 -7.14498222e-01 -7.39644885e-01
7.09035218e-01 -4.09721509e-02 8.96706641e-01 1.63961247e-01
-5.65446019e-01 7.21595824e-01 -4.99099791e-01 -5.18800378e-01
-3.09541434e-01 -4.39174592e-01 3.95673752e-01 5.29806554e-01
7.70934165e-01 -3.87294054e-01 -3.96936357e-01 1.96880773e-01
-5.61085582e-01 -4.91465807e-01 2.04301998e-01 6.82644129e-01
5.40159225e-01 -3.59482527e-01 1.86434746e-01 -6.52499124e-02
3.16601396e-01 -1.75599948e-01 -4.98967648e-01 1.37576491e-01
-2.41480786e-02 -3.50818753e-01 2.80264735e-01 -3.99017006e-01
-5.13563514e-01 4.61197883e-01 -5.36264598e-01 -1.15081561e+00
-5.48600078e-01 6.54365448e-03 1.70868963e-01 -6.45117998e-01
5.17384410e-01 1.66587949e-01 2.49797657e-01 -2.84197807e-01
3.37162048e-01 5.35455704e-01 3.03742468e-01 -3.97920281e-01
3.35360467e-01 7.90567994e-01 2.86100119e-01 -9.53383327e-01
-7.87935793e-01 -4.78925586e-01 -1.00840771e+00 -6.90430284e-01
1.08751678e+00 -8.76823843e-01 -1.12873673e+00 7.77468503e-01
-1.40022373e+00 -4.06802684e-01 -1.90979186e-02 1.53657630e-01
-7.24307716e-01 1.53698534e-01 -7.42544830e-01 -1.00741899e+00
-5.34753084e-01 -1.11375701e+00 1.90084314e+00 2.10036337e-01
-3.19646299e-01 -9.00354087e-01 1.42363936e-01 -2.84203380e-01
2.62089260e-02 6.71767116e-01 4.67442840e-01 -4.04065698e-01
1.68763682e-01 -4.50387269e-01 -3.20178270e-01 4.58064675e-01
3.63803148e-01 4.99607265e-01 -1.09525383e+00 -2.74041537e-02
-2.66306847e-01 -6.82155550e-01 7.62540102e-01 3.77167672e-01
1.08828366e+00 -6.95747957e-02 -2.19859004e-01 6.54774785e-01
1.07019734e+00 2.62733381e-02 5.22701204e-01 6.36796653e-02
5.65430284e-01 3.60544771e-01 5.70154786e-01 8.83424222e-01
1.73226193e-01 7.67813683e-01 9.08005953e-01 -2.43411094e-01
-1.26158334e-02 1.04672417e-01 3.96560639e-01 2.06783682e-01
-4.83878702e-01 1.39486298e-01 -8.18356633e-01 5.53075314e-01
-1.67553806e+00 -8.68763149e-01 3.40218574e-01 1.61131203e+00
4.58820701e-01 -1.29170865e-01 2.51630008e-01 -1.69875529e-02
5.16520739e-01 3.49898368e-01 -3.58530283e-01 -5.15048563e-01
-1.22418277e-01 5.16145825e-01 8.66234303e-02 2.75808722e-01
-1.39809513e+00 1.27739215e+00 7.00623035e+00 5.54863214e-01
-1.37383604e+00 5.21185994e-02 7.18547285e-01 -1.54674366e-01
4.36262280e-01 -8.55468929e-01 -9.60944891e-01 1.33181542e-01
7.08654523e-01 1.39359608e-01 1.21105425e-01 1.07381094e+00
1.90493107e-01 1.00862198e-01 -9.89993930e-01 1.44236660e+00
6.36670351e-01 -1.22593987e+00 -9.52984765e-02 -1.05325274e-01
6.92810237e-01 2.95167398e-02 1.20889798e-01 1.33115202e-01
3.53250373e-03 -1.06300461e+00 7.29482830e-01 5.52475095e-01
1.00367236e+00 -8.98861229e-01 7.04345107e-01 -1.51634574e-01
-1.39272285e+00 4.22208086e-02 -4.84659195e-01 5.67427576e-02
-1.17153510e-01 7.74414986e-02 -5.88367224e-01 1.78643033e-01
1.12533605e+00 1.09639955e+00 -5.05925238e-01 5.51859021e-01
-1.69548914e-01 4.00155872e-01 -5.49592137e-01 -8.97441506e-02
4.85178322e-01 -1.68999210e-01 2.00401455e-01 1.21473217e+00
5.88497460e-01 2.96210855e-01 -1.52986035e-01 6.69026494e-01
-3.43266398e-01 8.77998471e-02 -6.87688649e-01 8.11749995e-02
-1.43887633e-02 1.53063047e+00 -6.05149686e-01 -1.33528203e-01
-4.53995734e-01 1.22819602e+00 5.80977201e-01 4.71249260e-02
-8.72190416e-01 -5.19052632e-02 1.28234494e+00 -4.23677325e-01
2.21801803e-01 -3.93204272e-01 1.75921261e-01 -7.24385321e-01
-1.98334664e-01 -5.69713831e-01 3.35681438e-01 -7.92571604e-01
-1.15916514e+00 7.88005710e-01 -5.05511701e-01 -1.01037681e+00
-1.03241041e-01 -1.07467318e+00 -6.66461051e-01 3.67999852e-01
-1.31905782e+00 -1.30433047e+00 -4.94414598e-01 8.29771221e-01
2.23875731e-01 -1.63117245e-01 1.17084241e+00 2.49088943e-01
-4.92315024e-01 1.00653756e+00 -1.78120658e-01 6.94071114e-01
5.81281245e-01 -9.19566333e-01 3.51317048e-01 4.41402465e-01
5.54120660e-01 8.34983662e-02 3.47016424e-01 -1.55102864e-01
-1.23493791e+00 -1.14801264e+00 6.87909842e-01 -4.34537470e-01
6.54668450e-01 -4.42322165e-01 -5.68675697e-01 7.06116498e-01
-1.64660275e-01 6.19770706e-01 6.71229839e-01 -1.01529829e-01
-4.47037458e-01 -1.10809438e-01 -1.30951297e+00 3.88820142e-01
1.27682412e+00 -5.85244238e-01 -2.99233198e-01 3.06358129e-01
2.04797253e-01 -5.60624540e-01 -9.05625045e-01 5.11367023e-01
7.79782236e-01 -1.03212106e+00 8.14045310e-01 -7.23139882e-01
2.92365223e-01 1.41667441e-01 -1.91414326e-01 -1.25668156e+00
-2.19409205e-02 -4.76461530e-01 -4.44001228e-01 8.69616151e-01
2.34114006e-02 -2.80345887e-01 9.45387900e-01 2.69576639e-01
-1.27783090e-01 -9.09058034e-01 -1.60985303e+00 -5.90746105e-01
-1.85054526e-01 -5.20228565e-01 4.37668711e-01 6.70432031e-01
-5.01528345e-02 -1.71461999e-01 -3.90746176e-01 2.03406528e-01
4.00277704e-01 -6.03685454e-02 6.36149645e-01 -1.30825520e+00
4.27957922e-01 -7.67276227e-01 -1.13816166e+00 -8.65124762e-01
8.62472713e-01 -5.22367656e-01 4.99360897e-02 -1.08427811e+00
-1.46862611e-01 -1.64693311e-01 -3.86837602e-01 1.05255139e+00
1.33268788e-01 9.08813477e-01 8.87750238e-02 -2.29011729e-01
-7.08173633e-01 9.64924634e-01 9.46380496e-01 -1.60731673e-01
-1.15288191e-01 -1.66253939e-01 -2.98873428e-02 9.38053370e-01
5.81110775e-01 -2.57652670e-01 3.53324294e-01 -1.49763018e-01
1.71238445e-02 -2.06002936e-01 5.49239039e-01 -1.10185015e+00
-1.66248493e-02 2.39557058e-01 6.54233217e-01 -4.65865314e-01
8.24401855e-01 -9.70555127e-01 -4.96749282e-01 3.32302690e-01
-1.34584382e-01 4.30150926e-02 5.60529768e-01 2.71776319e-01
-3.56027901e-01 -3.31620686e-02 9.10876989e-01 8.31214339e-02
-8.24441135e-01 7.90727556e-01 -3.44422519e-01 -3.23837727e-01
1.26699197e+00 4.77269944e-03 2.78143603e-02 -4.41834271e-01
-8.61800849e-01 -1.88199282e-01 2.55730122e-01 5.76772511e-01
9.02795851e-01 -1.62705266e+00 -6.26015246e-01 4.03907984e-01
2.48053581e-01 -3.17728996e-01 9.28402841e-02 8.13286841e-01
-5.03486335e-01 1.51219100e-01 -4.62852240e-01 -7.93301821e-01
-1.62601984e+00 2.47837484e-01 7.77244151e-01 2.76100248e-01
-2.98438162e-01 1.16427159e+00 -8.79864767e-02 -3.60374302e-01
2.39525661e-01 -1.11863546e-01 -3.45929205e-01 4.23201352e-01
7.46851802e-01 1.33288935e-01 4.70141806e-02 -1.18021679e+00
-5.21362424e-01 1.03453672e+00 2.84138978e-01 1.97451040e-01
1.42836106e+00 1.25835285e-01 -2.21105218e-01 2.13797212e-01
1.63987958e+00 -3.81044745e-01 -1.45150554e+00 -1.37330025e-01
-2.80075312e-01 -3.18041265e-01 1.11624613e-01 -1.27663970e-01
-1.32854664e+00 9.49944377e-01 9.32107747e-01 5.90765383e-03
1.25229001e+00 2.86582470e-01 6.91510737e-01 2.99441844e-01
4.31126535e-01 -1.03514969e+00 3.97623956e-01 6.35804772e-01
8.74532819e-01 -1.41388535e+00 -4.50087100e-01 -2.05864295e-01
-3.53540480e-01 1.52444184e+00 6.82241082e-01 -2.94521153e-01
9.65188682e-01 4.72032204e-02 3.17668945e-01 -4.91870612e-01
-3.79087001e-01 -5.29136002e-01 4.50091213e-01 3.91105741e-01
1.57018423e-01 -1.24205999e-01 1.90502197e-01 2.73088545e-01
8.84112343e-02 -8.40410665e-02 7.06204250e-02 7.41826057e-01
-3.29533041e-01 -5.73252678e-01 -1.25553638e-01 1.34986535e-01
-4.86419350e-01 5.74886166e-02 -5.73405147e-01 9.44297075e-01
2.59139478e-01 7.44962037e-01 5.64357579e-01 -4.94670540e-01
4.30999368e-01 2.83174753e-01 7.68780649e-01 -3.84385437e-01
-2.34810725e-01 -7.03638792e-02 -1.59109429e-01 -1.05574095e+00
-7.32404590e-01 -6.98601902e-01 -1.30436099e+00 -1.90807536e-01
1.39374426e-02 -3.34306419e-01 6.40738189e-01 9.55930173e-01
5.68443298e-01 5.96387573e-02 8.40402007e-01 -1.31583560e+00
-1.43905923e-01 -9.94519055e-01 -7.57557094e-01 4.89556670e-01
3.91675681e-01 -1.08987737e+00 -4.57174450e-01 -1.27382308e-01] | [13.552587509155273, 1.5417859554290771] |
e64340f2-06c2-40b1-821d-1031041bd547 | dense-recurrent-neural-networks-for-scene | 1801.06831 | null | http://arxiv.org/abs/1801.06831v1 | http://arxiv.org/pdf/1801.06831v1.pdf | Dense Recurrent Neural Networks for Scene Labeling | Recently recurrent neural networks (RNNs) have demonstrated the ability to
improve scene labeling through capturing long-range dependencies among image
units. In this paper, we propose dense RNNs for scene labeling by exploring
various long-range semantic dependencies among image units. In comparison with
existing RNN based approaches, our dense RNNs are able to capture richer
contextual dependencies for each image unit via dense connections between each
pair of image units, which significantly enhances their discriminative power.
Besides, to select relevant and meanwhile restrain irrelevant dependencies for
each unit from dense connections, we introduce an attention model into dense
RNNs. The attention model enables automatically assigning more importance to
helpful dependencies while less weight to unconcerned dependencies. Integrating
with convolutional neural networks (CNNs), our method achieves state-of-the-art
performances on the PASCAL Context, MIT ADE20K and SiftFlow benchmarks. | ['Heng Fan', 'Haibin Ling'] | 2018-01-21 | null | null | null | null | ['scene-labeling'] | ['computer-vision'] | [ 3.00135791e-01 1.33538857e-01 -3.14157397e-01 -6.67170584e-01
-3.73145677e-02 -2.53434420e-01 5.12711406e-01 -1.65504292e-01
-6.16302848e-01 4.54410911e-01 8.33905339e-01 -7.54161403e-02
-1.10908616e-02 -9.39087152e-01 -8.43211949e-01 -4.57985252e-01
1.53008789e-01 2.50356287e-01 2.31758222e-01 -2.33392254e-01
8.66205096e-02 6.23206973e-01 -1.51128030e+00 4.90904659e-01
6.81461990e-01 7.38952696e-01 5.96766651e-01 4.24087942e-01
-1.70163289e-01 1.60965431e+00 -5.65344632e-01 -2.14898750e-01
-1.86791599e-01 -1.35787219e-01 -1.07073176e+00 4.74760607e-02
5.73295414e-01 -4.13557142e-01 -6.90667272e-01 1.08741665e+00
1.77401960e-01 3.23483735e-01 4.44727391e-01 -6.72029316e-01
-1.09325552e+00 8.15302312e-01 -5.72334230e-01 2.90816784e-01
5.62761836e-02 4.50072624e-02 1.41917300e+00 -7.96297193e-01
5.37650824e-01 1.37813437e+00 2.67874151e-01 4.78879690e-01
-1.02476287e+00 -4.96845394e-01 8.05276871e-01 5.30302525e-01
-1.04735625e+00 -2.30796918e-01 9.52884376e-01 -1.38650954e-01
1.32402289e+00 4.60037626e-02 5.36978245e-01 1.29038775e+00
-5.40251657e-02 1.26813030e+00 4.95807499e-01 -2.09913582e-01
-2.27575526e-01 -1.57697305e-01 4.26923186e-01 6.26373649e-01
6.89627305e-02 -1.70418158e-01 -1.49824917e-01 5.12836993e-01
9.95639563e-01 4.34062809e-01 -1.54401168e-01 -1.81148782e-01
-1.07406783e+00 8.23803246e-01 1.33784175e+00 3.88948321e-01
-4.14958507e-01 5.32329082e-01 5.31048298e-01 -1.09244227e-01
2.52672106e-01 5.32189071e-01 -4.60386038e-01 2.36401886e-01
-3.40614051e-01 -2.36384213e-01 3.09461504e-01 1.07041919e+00
1.03093839e+00 1.43508762e-01 -6.13716424e-01 1.02123511e+00
1.57437757e-01 1.62716717e-01 5.31116486e-01 -7.10503280e-01
4.12838012e-01 9.78749812e-01 -3.27548593e-01 -1.05662775e+00
-4.13616598e-01 -6.62616253e-01 -1.24165082e+00 -1.32113948e-01
-1.85910642e-01 9.15465653e-02 -1.29722178e+00 1.66000593e+00
-1.42045811e-01 2.91331530e-01 1.18888862e-01 9.09691334e-01
1.16370070e+00 6.75246358e-01 4.31077838e-01 2.76665449e-01
1.13139236e+00 -1.57162154e+00 -5.62575638e-01 -7.27302432e-01
6.20603323e-01 -6.02022231e-01 1.33679783e+00 5.35428934e-02
-7.17128754e-01 -1.03142822e+00 -7.87645161e-01 -5.60393453e-01
-5.09459615e-01 2.02160344e-01 9.60006356e-01 -4.56822515e-02
-1.13275158e+00 3.40529710e-01 -7.19929457e-01 -3.55917156e-01
7.05374479e-01 5.31287193e-01 -2.33114615e-01 -3.52217555e-01
-1.18604743e+00 8.01569402e-01 4.20522720e-01 3.71769696e-01
-1.10761225e+00 -3.66458416e-01 -1.37794566e+00 2.49922574e-01
3.50189537e-01 -5.76802373e-01 1.02469742e+00 -1.31947732e+00
-1.29761267e+00 6.55776322e-01 -1.54066205e-01 -6.09891593e-01
-5.26369549e-02 -6.78808391e-01 -2.27698728e-01 1.81801990e-01
1.21063679e-01 1.29359531e+00 6.16997838e-01 -1.04984128e+00
-4.70928609e-01 -7.72292167e-02 4.20346916e-01 2.35658020e-01
-4.76547629e-01 -1.32213950e-01 -7.79857814e-01 -7.16687679e-01
-1.12774692e-01 -7.48007059e-01 -7.00031281e-01 -4.78170812e-01
-7.07920790e-01 -4.30097103e-01 1.23637033e+00 -1.99448064e-01
1.00053632e+00 -2.06267500e+00 3.80923748e-01 -1.82907268e-01
1.80796057e-01 5.00983059e-01 -5.83859146e-01 -4.27245302e-03
-2.40329340e-01 -9.94486436e-02 -3.98423932e-02 -4.31916565e-01
-7.16852844e-02 8.81214261e-01 -3.56171876e-01 3.05909187e-01
5.39267123e-01 1.43896091e+00 -9.33073401e-01 -3.27710122e-01
8.01501930e-01 7.27259815e-01 -7.43859828e-01 3.22163701e-01
-3.26782912e-01 2.99901843e-01 -4.70946699e-01 5.17975807e-01
4.16055351e-01 -5.18809319e-01 -1.11389337e-02 -4.49933022e-01
9.44012478e-02 3.93732637e-01 -5.95125198e-01 1.86025238e+00
-7.97667444e-01 7.88241684e-01 -2.16025606e-01 -1.22531283e+00
9.89875138e-01 -1.54961392e-01 1.76069766e-01 -1.02363217e+00
2.36301363e-01 -3.40836197e-01 -3.11256219e-02 -4.16769356e-01
7.22571194e-01 2.54324555e-01 -1.86705559e-01 1.31480351e-01
4.57906812e-01 2.64594257e-01 1.16604447e-01 3.09163570e-01
8.86458933e-01 1.10938944e-01 2.49508638e-02 -3.11811715e-01
7.18918443e-01 -3.95128131e-01 5.13650954e-01 6.82569206e-01
-2.06890866e-01 6.21147096e-01 4.18894261e-01 -7.95368910e-01
-7.97693789e-01 -6.93841159e-01 8.74800086e-02 1.47752237e+00
4.19598699e-01 -4.92802978e-01 -4.46532279e-01 -9.20196533e-01
-1.31558746e-01 6.93968892e-01 -9.99447286e-01 -2.71929562e-01
-6.69314921e-01 -4.59751576e-01 4.88266259e-01 1.02947748e+00
8.64894390e-01 -1.61462474e+00 -5.18942952e-01 1.44818038e-01
-7.48647377e-02 -1.38598406e+00 -3.74714136e-01 5.69518685e-01
-6.47486269e-01 -1.04240847e+00 -5.29719532e-01 -1.05430424e+00
8.97933900e-01 3.34823191e-01 1.40475857e+00 3.72271170e-03
-4.06819314e-01 2.38544017e-01 -4.44914311e-01 -1.21193893e-01
1.62246674e-01 4.68592644e-01 -3.42409670e-01 -2.44372860e-02
5.63841999e-01 -5.37991583e-01 -6.76290095e-01 3.44004244e-01
-9.32087481e-01 1.97739348e-01 7.92553425e-01 9.33697045e-01
7.66681671e-01 -2.36359879e-01 3.73698920e-01 -1.21619630e+00
1.88109413e-01 -2.50964671e-01 -4.45155084e-01 1.99036255e-01
-1.06969513e-01 4.12466735e-01 7.82972455e-01 -2.10043326e-01
-1.25030327e+00 3.92055362e-01 -2.95204639e-01 -6.43404603e-01
-4.19242054e-01 2.66739249e-01 -2.48603091e-01 1.34865642e-02
3.64628673e-01 9.77518484e-02 -5.34784317e-01 -3.96875352e-01
7.53521740e-01 2.18902841e-01 7.66073644e-01 -5.99245727e-01
3.51946175e-01 3.56476605e-01 -2.12236375e-01 -6.66553855e-01
-1.45605719e+00 -5.32152951e-01 -1.00059521e+00 1.26349464e-01
1.30181122e+00 -1.00572014e+00 -6.03204608e-01 3.01840454e-01
-1.20482314e+00 -4.51458454e-01 -5.15703738e-01 4.24290061e-01
-2.42003456e-01 8.42412189e-02 -9.07893956e-01 -3.55895311e-01
-2.26802215e-01 -1.40842307e+00 1.21584773e+00 4.35555398e-01
-1.40926570e-01 -1.13906932e+00 -8.09715390e-02 1.93639323e-01
3.46558332e-01 1.17888190e-01 6.68426633e-01 -4.41458762e-01
-5.95347881e-01 -3.36991739e-03 -7.12346077e-01 4.79325056e-01
3.05169046e-01 -7.97797441e-02 -1.11950195e+00 -5.90125769e-02
-2.88341284e-01 -5.91737449e-01 1.63939953e+00 5.47710717e-01
1.55371404e+00 -3.28232586e-01 -2.59846002e-01 9.52312529e-01
1.31142664e+00 -1.25436455e-01 1.00351632e+00 2.49226272e-01
1.33040905e+00 4.98816639e-01 2.52649069e-01 7.25722387e-02
2.94822872e-01 2.58371025e-01 8.00234079e-01 -6.48446262e-01
-2.30050623e-01 -4.28589195e-01 2.23650336e-01 5.94098628e-01
4.68062013e-02 -1.85640514e-01 -6.00532413e-01 6.34190321e-01
-2.00812602e+00 -7.55005479e-01 2.60502156e-02 1.58199751e+00
4.32332158e-01 3.16874444e-01 -2.67833084e-01 -3.99626225e-01
7.36599147e-01 6.13651872e-01 -6.89588904e-01 -4.28108245e-01
-3.55921239e-01 3.18651199e-01 5.94860971e-01 4.92941648e-01
-1.33147395e+00 1.56739855e+00 6.16295624e+00 5.71477711e-01
-1.07278609e+00 -1.37187421e-01 1.00041687e+00 1.90474018e-02
-3.87716293e-01 -2.37301528e-01 -8.69654298e-01 -1.09393969e-02
5.67948163e-01 3.92378181e-01 7.72129446e-02 1.13628924e+00
-1.81306556e-01 7.37703294e-02 -1.01597619e+00 9.05866325e-01
8.79684538e-02 -1.41701162e+00 4.16947782e-01 -5.81949428e-02
9.04780447e-01 4.93265688e-01 8.53335932e-02 5.24349332e-01
8.46017659e-01 -1.44390321e+00 2.51994938e-01 3.23442310e-01
6.48284376e-01 -1.05558610e+00 9.56180155e-01 -2.20431294e-02
-1.36880362e+00 -1.66050717e-01 -8.33458841e-01 -3.90161693e-01
-7.30465576e-02 5.30940533e-01 -5.00696123e-01 3.73850226e-01
7.30734110e-01 1.45544863e+00 -7.99233317e-01 5.07227957e-01
-8.34197581e-01 3.91039789e-01 8.94718338e-04 1.59313940e-02
8.87128413e-01 -1.15114167e-01 3.00825220e-02 1.61723471e+00
-2.61871248e-01 3.98491547e-02 1.89291507e-01 9.50785041e-01
-4.11801457e-01 -1.50806218e-01 -7.02820718e-01 1.18291929e-01
9.06851664e-02 1.49998569e+00 -8.28568935e-01 -5.35788536e-01
-4.77440894e-01 1.28055298e+00 8.04730535e-01 6.22281730e-01
-6.63204014e-01 -4.91248697e-01 9.09179091e-01 -3.11085403e-01
5.36342800e-01 -1.79357946e-01 -2.28273019e-01 -1.19184792e+00
-2.78271228e-01 -4.51951444e-01 6.27261460e-01 -8.41581225e-01
-1.22441685e+00 9.27981913e-01 -2.73118138e-01 -7.30707407e-01
-1.13671347e-02 -9.00163293e-01 -6.57023251e-01 7.86382735e-01
-1.74974394e+00 -1.43572211e+00 -3.84965271e-01 8.66684020e-01
8.02386582e-01 3.37414034e-02 7.82800198e-01 3.30723137e-01
-8.23945284e-01 3.10908765e-01 -2.82729626e-01 5.95483840e-01
4.98305261e-01 -1.25041866e+00 5.13373435e-01 8.81854832e-01
4.98838723e-01 6.83555305e-01 1.98744640e-01 -3.37228686e-01
-1.08370817e+00 -1.68182099e+00 5.89668810e-01 -3.12861651e-01
5.52103579e-01 -4.85757530e-01 -8.16556215e-01 1.06146884e+00
4.85091150e-01 3.59375179e-01 4.42952245e-01 5.54603934e-01
-7.04833150e-01 -6.48339316e-02 -5.64926624e-01 5.84689736e-01
1.23067689e+00 -7.59217143e-01 -6.68866396e-01 2.90308565e-01
1.07072020e+00 -2.44618401e-01 -3.79916489e-01 6.55482054e-01
2.55315393e-01 -9.77349579e-01 1.16100407e+00 -6.72012269e-01
7.19497919e-01 -2.23310709e-01 -1.80255786e-01 -1.18865776e+00
-7.22175956e-01 -1.72852397e-01 1.33400530e-01 1.14121222e+00
4.05438095e-01 -2.31962487e-01 9.02462602e-01 4.10820037e-01
-4.99082416e-01 -7.32092977e-01 -3.06055069e-01 -2.99131006e-01
-1.12132415e-01 -4.86145049e-01 4.65778112e-01 9.72549498e-01
-3.50209177e-01 9.06871617e-01 -5.49154043e-01 2.19536290e-01
3.79119545e-01 2.25738883e-01 6.29623175e-01 -9.07492518e-01
-3.07641745e-01 -5.73739350e-01 -5.25137126e-01 -1.57993162e+00
6.78126931e-01 -7.43476808e-01 5.08424118e-02 -1.78388715e+00
3.64190131e-01 -1.25759900e-01 -7.66385376e-01 8.49051356e-01
-3.49721611e-01 4.94488120e-01 2.01349348e-01 -2.21933033e-02
-1.09939063e+00 6.98594272e-01 1.41836166e+00 -4.73801553e-01
-2.24521026e-01 -3.28316927e-01 -7.94081509e-01 8.91956687e-01
5.19522488e-01 -1.02347046e-01 -5.94358385e-01 -9.77719188e-01
1.36774078e-01 -4.70431656e-01 3.82309139e-01 -9.00446355e-01
3.82972538e-01 3.45839038e-02 7.62357056e-01 -8.38749468e-01
9.55396667e-02 -6.96292758e-01 -2.79721767e-01 2.67869145e-01
-5.70341647e-01 -2.01255113e-01 3.73229712e-01 4.99360234e-01
-6.41357958e-01 -7.64770992e-03 7.13123083e-01 -3.49085271e-01
-1.30963063e+00 4.08092916e-01 -2.12266296e-01 -1.44585565e-01
7.11685419e-01 5.58134280e-02 -2.93146580e-01 -2.59944618e-01
-8.56739044e-01 4.84490961e-01 1.34551331e-01 6.16552651e-01
6.89781189e-01 -1.26475906e+00 -4.30751652e-01 2.39425868e-01
1.28596038e-01 5.45570731e-01 4.25976872e-01 2.88931578e-01
-3.12944740e-01 6.19351625e-01 -3.78850073e-01 -7.33865261e-01
-1.12569940e+00 7.51922071e-01 2.80238479e-01 -4.20904249e-01
-6.47985816e-01 1.41559136e+00 9.90160525e-01 -4.31832254e-01
4.15434867e-01 -7.69426167e-01 -5.50630927e-01 -3.08108460e-02
5.42021751e-01 -3.92432846e-02 -1.33830950e-01 -7.97986567e-01
-3.25362474e-01 5.82491815e-01 -4.22288597e-01 5.40518165e-01
1.30681741e+00 -8.94690603e-02 -1.65408835e-01 1.36647940e-01
1.48739743e+00 -3.79709095e-01 -1.57424057e+00 -3.34528774e-01
-1.05769746e-01 -1.03364445e-01 1.00345165e-01 -6.60131872e-01
-1.52477038e+00 1.23180056e+00 1.16793692e-01 -2.08696246e-01
1.43991804e+00 2.84773111e-01 6.85349643e-01 6.53119922e-01
-1.02644466e-01 -8.61369789e-01 4.69796300e-01 9.74658310e-01
7.85307109e-01 -1.30377376e+00 -1.64076120e-01 -2.77861506e-01
-8.41515243e-01 1.13464820e+00 1.06839478e+00 -4.85079110e-01
4.24415380e-01 2.59050369e-01 -7.29169920e-02 -3.70723188e-01
-6.60875559e-01 -6.68163538e-01 4.36188817e-01 4.79088634e-01
6.22965515e-01 1.93799995e-02 3.42833668e-01 4.30578768e-01
3.66910324e-02 -2.96929955e-01 2.33494550e-01 4.93423790e-01
-5.08795977e-01 -1.01253545e+00 1.45647705e-01 2.72720814e-01
-3.17038327e-01 -4.27423626e-01 -5.52558064e-01 7.18362629e-01
1.16361924e-01 6.23638749e-01 3.99841130e-01 -3.80726218e-01
4.07850385e-01 -3.22986186e-01 2.32122108e-01 -8.01761568e-01
-6.62369728e-01 1.95919991e-01 2.30797636e-03 -8.22769940e-01
-6.17237449e-01 -9.35695395e-02 -1.36836731e+00 6.31297082e-02
-2.04613581e-01 6.55088723e-02 6.89630210e-02 9.99594927e-01
3.97668332e-01 1.03446674e+00 4.54106152e-01 -8.08298469e-01
2.19911546e-01 -9.72349644e-01 -3.95825267e-01 5.85169256e-01
3.58920217e-01 -4.93951380e-01 1.92496199e-02 1.07768271e-02] | [9.570086479187012, 0.4655357003211975] |
67800219-3cd1-4382-bcc9-df028fd043e9 | backdoor-learning-on-sequence-to-sequence | 2305.02424 | null | https://arxiv.org/abs/2305.02424v1 | https://arxiv.org/pdf/2305.02424v1.pdf | Backdoor Learning on Sequence to Sequence Models | Backdoor learning has become an emerging research area towards building a trustworthy machine learning system. While a lot of works have studied the hidden danger of backdoor attacks in image or text classification, there is a limited understanding of the model's robustness on backdoor attacks when the output space is infinite and discrete. In this paper, we study a much more challenging problem of testing whether sequence-to-sequence (seq2seq) models are vulnerable to backdoor attacks. Specifically, we find by only injecting 0.2\% samples of the dataset, we can cause the seq2seq model to generate the designated keyword and even the whole sentence. Furthermore, we utilize Byte Pair Encoding (BPE) to create multiple new triggers, which brings new challenges to backdoor detection since these backdoors are not static. Extensive experiments on machine translation and text summarization have been conducted to show our proposed methods could achieve over 90\% attack success rate on multiple datasets and models. | ['Heng Huang', 'Minhao Cheng', 'Lichang Chen'] | 2023-05-03 | null | null | null | null | ['text-summarization'] | ['natural-language-processing'] | [ 5.98338187e-01 -1.34480044e-01 -5.41225851e-01 2.21688628e-01
-1.04146445e+00 -1.37735248e+00 7.03159392e-01 6.85994029e-02
8.25982988e-02 5.56122839e-01 -1.06665902e-01 -1.16586006e+00
6.48765802e-01 -5.54185748e-01 -1.13995886e+00 -7.30713904e-01
-1.71691477e-01 -2.21119702e-01 2.72962004e-01 -1.07184060e-01
5.40434420e-01 3.22535366e-01 -7.14695573e-01 5.74773371e-01
4.50252503e-01 6.02462590e-01 -4.25758988e-01 1.03106427e+00
1.76222488e-01 6.07179046e-01 -1.10240793e+00 -8.08948874e-01
7.12280273e-01 -5.25813103e-01 -7.68208385e-01 -4.38832611e-01
6.66535258e-01 -6.27141535e-01 -6.26431346e-01 1.50714433e+00
5.64896464e-01 -6.61098301e-01 2.80989021e-01 -1.31553447e+00
-5.61694622e-01 9.90382016e-01 -7.90050864e-01 4.82434154e-01
5.23556888e-01 4.95445788e-01 8.09599996e-01 -4.65559691e-01
5.02282977e-01 1.27001595e+00 3.16105455e-01 5.97325981e-01
-9.26143527e-01 -1.27968347e+00 -3.13178986e-01 1.19256474e-01
-1.22225463e+00 -3.67359668e-01 5.98380685e-01 -1.79548353e-01
8.99256706e-01 7.65544772e-01 2.22494714e-02 1.63104296e+00
9.69734490e-01 7.03864098e-01 1.29737425e+00 -5.29556215e-01
-3.25316601e-02 6.68222904e-01 1.79464132e-01 7.13566959e-01
4.10577267e-01 4.70791221e-01 -7.44618118e-01 -9.15067196e-01
2.31618240e-01 -2.96616882e-01 -4.13188130e-01 3.39484125e-01
-1.05047226e+00 1.10390699e+00 -1.02987170e-01 -2.56430358e-02
3.09003770e-01 4.18441534e-01 5.43437898e-01 5.72439611e-01
1.50225714e-01 4.94872779e-01 -1.60727620e-01 -6.95014894e-02
-7.93424964e-01 2.44307011e-01 9.05262649e-01 8.37664723e-01
2.57818520e-01 1.08680688e-01 -1.24090075e-01 -9.85025615e-03
4.95438352e-02 8.59180391e-01 4.93211448e-01 -3.20620805e-01
1.08209169e+00 -1.98847458e-01 -1.28991038e-01 -1.49797356e+00
2.63365358e-01 -3.91923964e-01 -4.74870503e-01 -2.46974215e-01
1.38246492e-01 -4.15834397e-01 -6.04426742e-01 1.49870527e+00
2.78859675e-01 3.14086407e-01 8.10760185e-02 6.32375121e-01
2.71567136e-01 1.00545228e+00 -4.65489089e-01 -2.64974952e-01
1.44434237e+00 -7.22031593e-01 -7.54530549e-01 -4.45940197e-02
8.35316598e-01 -1.01995993e+00 9.34285223e-01 3.13369900e-01
-6.86572313e-01 -6.27159849e-02 -1.41256523e+00 2.84063101e-01
-2.42849112e-01 -4.59223419e-01 6.87398255e-01 1.54721177e+00
-4.12745357e-01 1.87332392e-01 -5.67757607e-01 2.07241043e-01
3.72380495e-01 3.10736716e-01 -2.99912423e-01 1.08090773e-01
-1.63107824e+00 5.42036712e-01 2.81606346e-01 -1.27516970e-01
-1.28364182e+00 -3.99114370e-01 -4.54873294e-01 -2.18702972e-01
3.43650818e-01 -3.14589798e-01 1.01242697e+00 -4.27696466e-01
-1.37648606e+00 6.45702422e-01 5.36299758e-02 -9.24252510e-01
7.22111404e-01 -2.89830089e-01 -6.34039462e-01 3.74057084e-01
-1.55412495e-01 1.01817332e-01 1.54456151e+00 -9.51466560e-01
-2.17136890e-01 -3.77264589e-01 -6.54419288e-02 -3.15851629e-01
-7.85803378e-01 4.62637275e-01 9.76186991e-02 -8.82372320e-01
-3.27084988e-01 -1.24012852e+00 -3.36049125e-02 -6.16323471e-01
-1.28041911e+00 2.38609582e-01 1.17955804e+00 -4.98310447e-01
1.42431700e+00 -2.10544896e+00 -4.57818091e-01 3.56148958e-01
5.81252761e-02 6.85179353e-01 -8.48114118e-02 7.87798464e-01
-8.72717872e-02 6.93370640e-01 -8.49538669e-02 1.38086691e-01
-2.52569228e-01 -1.39435619e-01 -1.50897145e+00 7.82441676e-01
-1.50309175e-01 9.18478251e-01 -5.37403166e-01 -2.39394471e-01
-2.59212464e-01 1.10132813e-01 -4.38687950e-01 -4.48932536e-02
-2.01835826e-01 1.52243316e-01 -5.94533026e-01 5.87458849e-01
8.40395153e-01 -5.51662110e-02 7.31417239e-02 2.69620210e-01
2.33351305e-01 3.52655113e-01 -8.20729792e-01 8.83199751e-01
-3.20667699e-02 9.04353917e-01 -3.59474272e-01 -4.89598662e-01
6.85565531e-01 3.72032851e-01 -1.43236250e-01 -3.70937228e-01
2.98827887e-01 1.67196095e-01 8.80999118e-02 -5.26744545e-01
6.74730062e-01 2.28573214e-02 -2.80772686e-01 8.60537648e-01
-2.46083274e-01 1.06634639e-01 -4.16712880e-01 3.24764729e-01
1.23962188e+00 -5.89680314e-01 -6.61120489e-02 2.21198779e-02
4.27635044e-01 3.72339599e-02 2.71494061e-01 1.27104259e+00
-6.91575259e-02 3.30747157e-01 7.39624798e-01 -1.95239350e-01
-8.11887503e-01 -9.43128943e-01 -6.50421456e-02 5.99707305e-01
2.56668627e-01 -6.74911261e-01 -9.97236609e-01 -9.00365949e-01
3.71709764e-02 6.56471848e-01 -2.53467411e-01 -5.20955503e-01
-5.86057603e-01 -8.33081305e-01 1.65341377e+00 1.11953542e-02
5.12864947e-01 -3.65977764e-01 -4.98513579e-01 -6.49989992e-02
-1.62964463e-01 -1.52621758e+00 -8.75153601e-01 -1.83196142e-01
-9.32889521e-01 -1.06453323e+00 -3.53972375e-01 -4.60106373e-01
5.91258347e-01 3.66253078e-01 4.61049944e-01 -4.29004990e-02
-4.61239100e-01 -5.18356748e-02 -1.52654335e-01 -5.36758006e-01
-1.04159951e+00 2.91918479e-02 2.30660275e-01 2.96166502e-02
1.47715285e-01 -1.71999231e-01 -3.53963375e-01 2.88523674e-01
-1.04357374e+00 -4.37990129e-01 3.78406793e-01 6.52951181e-01
1.18663140e-01 4.11596358e-01 3.58339787e-01 -1.15392232e+00
9.91007328e-01 -4.78967905e-01 -7.42829382e-01 2.54126877e-01
-3.39019477e-01 2.82914285e-02 9.21549916e-01 -8.65881860e-01
-4.11489457e-01 -2.07677558e-01 -1.53624937e-01 -6.74927711e-01
2.73170978e-01 1.75266728e-01 -7.37240463e-02 -3.68116051e-01
7.68731475e-01 5.62127531e-01 3.11618317e-02 -1.20967165e-01
5.49798071e-01 8.73522043e-01 4.81031388e-01 -4.29061860e-01
1.31764543e+00 5.20864010e-01 -1.96794365e-02 -9.79888439e-01
-2.32396156e-01 -4.38588187e-02 2.58447468e-01 6.97272494e-02
4.79538709e-01 -6.42095804e-01 -8.31035316e-01 5.30948341e-01
-1.44146025e+00 1.69929117e-01 4.59043264e-01 2.99727052e-01
-1.02694675e-01 9.91892040e-01 -6.92738533e-01 -8.53845179e-01
-6.78281665e-01 -1.38821197e+00 7.79410481e-01 -2.24634767e-01
-1.38643190e-01 -7.55791247e-01 -1.12967193e-01 5.65952420e-01
8.35293718e-03 3.72070819e-01 1.19346952e+00 -1.01425838e+00
-7.48286724e-01 -6.52075589e-01 4.05807495e-01 3.67216557e-01
-1.36941493e-01 2.68829358e-03 -1.03199804e+00 -5.54314673e-01
6.08837843e-01 -2.43854016e-01 7.56423116e-01 -1.56980917e-01
1.20289016e+00 -1.12063920e+00 -4.48139995e-01 5.96739233e-01
1.19997835e+00 1.28950328e-01 6.90768182e-01 6.90967739e-02
5.46518207e-01 3.93768311e-01 4.71375734e-01 3.40939194e-01
-2.72019833e-01 5.64832032e-01 5.10191500e-01 2.82062352e-01
4.49745268e-01 -6.87386334e-01 1.05350256e+00 5.44114649e-01
6.37712479e-01 -5.14072418e-01 -7.00924575e-01 6.88057244e-02
-1.24919105e+00 -1.04080153e+00 -3.04652244e-01 2.31624198e+00
9.25381184e-01 5.12178779e-01 -1.08894847e-01 2.44846761e-01
8.01339686e-01 3.77025515e-01 -2.83221036e-01 -7.90730774e-01
-6.85942695e-02 2.26616040e-01 1.17703128e+00 3.99906129e-01
-1.08083844e+00 1.20729446e+00 6.97254705e+00 1.16084564e+00
-1.58990037e+00 1.09064527e-01 6.97250843e-01 -1.13903604e-01
-5.34701824e-01 2.29841709e-01 -1.16692579e+00 6.81570947e-01
1.00492716e+00 -2.55625278e-01 2.23489434e-01 6.15729630e-01
-1.07121699e-01 1.77067533e-01 -1.01019168e+00 8.75363529e-01
2.80528516e-01 -1.57534206e+00 2.74186939e-01 4.51046884e-01
7.67384231e-01 -2.33040839e-01 7.12923467e-01 -5.76528683e-02
1.83580086e-01 -1.14989662e+00 5.62101305e-01 -1.72706023e-01
7.96707153e-01 -1.04990256e+00 5.21489143e-01 7.47066438e-01
-4.76846546e-01 -1.82963207e-01 -4.25594062e-01 2.98099875e-01
-1.52703509e-01 5.96017420e-01 -1.25116444e+00 3.92375588e-01
3.75836611e-01 -1.50164574e-01 -4.78778839e-01 3.54179412e-01
-2.02976808e-01 1.18547988e+00 -4.06260401e-01 -2.78689265e-01
4.66895401e-01 2.23837450e-01 1.03551459e+00 1.09012794e+00
2.24860400e-01 -1.13826916e-01 5.31566271e-04 7.26996839e-01
-3.85957807e-01 4.93450696e-03 -1.21136284e+00 -5.42259932e-01
6.16234004e-01 7.51477122e-01 -5.37461579e-01 4.03907634e-02
5.40843373e-03 1.07600343e+00 -3.78223568e-01 9.00673643e-02
-1.06587279e+00 -5.00790775e-01 6.32366955e-01 -2.86635142e-02
9.65225548e-02 -4.32581127e-01 -1.32059380e-01 -1.13432014e+00
1.49024412e-01 -1.43390882e+00 1.84035912e-01 -1.78929210e-01
-9.00773466e-01 4.22821134e-01 -2.38566682e-01 -1.13909769e+00
-2.82848001e-01 -3.26026678e-01 -7.92590618e-01 6.55344427e-01
-1.18623662e+00 -9.33362961e-01 6.55437648e-01 5.93455315e-01
3.47641021e-01 -5.44316590e-01 8.16787899e-01 -3.33504677e-02
-6.34620786e-01 1.31043756e+00 2.22089633e-01 3.76550704e-01
5.93608677e-01 -5.81802011e-01 8.19408834e-01 1.28261495e+00
5.60351610e-01 1.17093766e+00 9.41517472e-01 -8.98033500e-01
-2.14841771e+00 -9.51440513e-01 7.38677800e-01 -6.79700553e-01
6.47115409e-01 -7.98757374e-01 -5.87167859e-01 7.05584764e-01
2.46381298e-01 -6.06425256e-02 8.95684004e-01 -4.74060684e-01
-7.95141041e-01 2.36461639e-01 -1.16157579e+00 7.57517934e-01
5.23106813e-01 -9.52972591e-01 -3.29353631e-01 6.23811662e-01
1.03047132e+00 -5.37769973e-01 -3.86002094e-01 1.17960788e-01
4.86035347e-01 -6.80641532e-01 9.02666807e-01 -8.32401395e-01
4.68434870e-01 -1.57575801e-01 -2.43230417e-01 -8.08228493e-01
5.01061320e-01 -1.47284794e+00 -4.86569881e-01 1.19855475e+00
4.43891674e-01 -9.34163928e-01 7.66506910e-01 -1.17943101e-01
4.52319175e-01 -5.76056659e-01 -1.09274125e+00 -1.07441568e+00
2.97564059e-01 -5.31213105e-01 5.79649389e-01 1.03652287e+00
1.24743141e-01 1.92861691e-01 -9.62534487e-01 6.89816117e-01
7.56163180e-01 -5.60303777e-02 1.00433195e+00 -2.62975216e-01
-5.35509586e-01 4.72097062e-02 -3.20937485e-01 -1.10730803e+00
1.65745914e-01 -1.14723492e+00 -3.88608843e-01 -3.63116264e-01
1.79614872e-01 -4.44578938e-02 -2.60735657e-02 1.75655216e-01
-1.14827320e-01 3.38633031e-01 2.50182688e-01 2.96291471e-01
1.10874481e-01 2.94433147e-01 9.28526163e-01 -3.67576063e-01
1.19119011e-01 2.00000495e-01 -7.51147211e-01 4.21352118e-01
9.46044981e-01 -9.91289556e-01 -5.51514208e-01 -1.65654853e-01
3.74888599e-01 3.00784409e-01 3.56579363e-01 -7.04805493e-01
3.00465256e-01 -2.76475176e-02 -1.02249742e-01 -6.08932018e-01
3.69632896e-03 -7.53550053e-01 5.01479618e-02 8.42209935e-01
-4.93097633e-01 2.49322101e-01 1.91502213e-01 8.40683758e-01
7.07340911e-02 -2.57369518e-01 5.13647914e-01 -4.07348312e-02
-3.35139669e-02 2.91581988e-01 -5.40783107e-01 1.01900272e-01
1.14946365e+00 -1.94852680e-01 -6.21216118e-01 -3.14257383e-01
-3.38415802e-02 -2.87616020e-03 2.64173985e-01 6.85651660e-01
7.64155865e-01 -8.93261969e-01 -6.37576878e-01 4.09771353e-01
-2.48053223e-02 -7.93307245e-01 2.06127558e-02 4.67219740e-01
-5.04471421e-01 7.22272336e-01 1.14980437e-01 -6.80549979e-01
-1.88920903e+00 6.95849895e-01 1.47946373e-01 -2.13417262e-01
-6.55735612e-01 8.55628133e-01 -1.07880928e-01 -2.01140940e-01
1.50231317e-01 1.16596103e-01 4.21768159e-01 -2.80344605e-01
6.53171122e-01 2.66857326e-01 1.01380181e-02 -4.42544580e-01
-3.17752063e-01 2.51875103e-01 -5.16937554e-01 -3.96227628e-01
5.36823392e-01 2.67718524e-01 -1.98562503e-01 5.12246415e-02
1.65208399e+00 5.78520894e-01 -4.98184890e-01 -5.49136549e-02
-8.64530206e-02 -8.17591786e-01 -1.47003084e-01 -3.73541355e-01
-6.76075876e-01 1.05878568e+00 5.25171697e-01 3.55184615e-01
5.00975788e-01 -2.63600349e-01 1.41999030e+00 5.56811392e-01
4.96480316e-01 -8.92995894e-01 1.88447714e-01 4.84888881e-01
5.26464403e-01 -7.29546607e-01 4.59832232e-03 -5.64108431e-01
-5.45522749e-01 9.66051340e-01 1.94411814e-01 6.40400499e-02
4.14818823e-01 4.39033628e-01 2.46669855e-02 1.49296939e-01
-8.58464420e-01 8.18182290e-01 -3.07007909e-01 3.68532419e-01
-1.48946449e-01 1.27455235e-01 -1.85823724e-01 3.33095908e-01
-7.22007573e-01 -2.65867472e-01 9.45436239e-01 9.05240297e-01
-3.21107626e-01 -1.34391928e+00 -8.37860703e-01 2.50262648e-01
-1.28447306e+00 -3.06966722e-01 -6.42855644e-01 4.50074136e-01
-3.26733917e-01 1.05873811e+00 -3.30842376e-01 -9.12063599e-01
-2.65769273e-01 7.34028295e-02 3.47290933e-01 -5.43179989e-01
-8.82691681e-01 -1.65018529e-01 4.68312651e-02 -4.61192429e-01
4.11591619e-01 -4.08429354e-01 -1.09913123e+00 -7.37651169e-01
-7.75366247e-01 2.65773147e-01 8.32446158e-01 7.10707545e-01
3.99228722e-01 2.60926113e-02 1.41252124e+00 4.76561040e-02
-1.30577612e+00 -4.75750476e-01 -4.44042027e-01 -5.70330769e-02
5.69479108e-01 3.54429684e-03 -6.50798917e-01 -4.32618894e-02] | [5.886063575744629, 7.807079315185547] |
59bb943b-82e8-45bb-b245-8ae40db323e8 | automated-grain-boundary-gb-segmentation-and | 2305.07790 | null | https://arxiv.org/abs/2305.07790v1 | https://arxiv.org/pdf/2305.07790v1.pdf | Automated Grain Boundary (GB) Segmentation and Microstructural Analysis in 347H Stainless Steel Using Deep Learning and Multimodal Microscopy | Austenitic 347H stainless steel offers superior mechanical properties and corrosion resistance required for extreme operating conditions such as high temperature. The change in microstructure due to composition and process variations is expected to impact material properties. Identifying microstructural features such as grain boundaries thus becomes an important task in the process-microstructure-properties loop. Applying convolutional neural network (CNN) based deep-learning models is a powerful technique to detect features from material micrographs in an automated manner. Manual labeling of the images for the segmentation task poses a major bottleneck for generating training data and labels in a reliable and reproducible way within a reasonable timeframe. In this study, we attempt to overcome such limitations by utilizing multi-modal microscopy to generate labels directly instead of manual labeling. We combine scanning electron microscopy (SEM) images of 347H stainless steel as training data and electron backscatter diffraction (EBSD) micrographs as pixel-wise labels for grain boundary detection as a semantic segmentation task. We demonstrate that despite producing instrumentation drift during data collection between two modes of microscopy, this method performs comparably to similar segmentation tasks that used manual labeling. Additionally, we find that na\"ive pixel-wise segmentation results in small gaps and missing boundaries in the predicted grain boundary map. By incorporating topological information during model training, the connectivity of the grain boundary network and segmentation performance is improved. Finally, our approach is validated by accurate computation on downstream tasks of predicting the underlying grain morphology distributions which are the ultimate quantities of interest for microstructural characterization. | ['Keerti S Kappagantula', 'Alan L Schemer-Kohrn', 'Ram Devanathan', 'Madison Wenzlick', 'Marissa Masden', 'Jing Wang', 'M. F. N. Taufique', 'Shoieb Ahmed Chowdhury'] | 2023-05-12 | null | null | null | null | ['boundary-detection'] | ['computer-vision'] | [ 4.23477590e-01 1.20948406e-03 4.23867851e-01 -5.40296972e-01
-7.82310426e-01 -4.07193631e-01 2.79402286e-01 3.41919422e-01
-6.30911648e-01 4.76309270e-01 -5.14939308e-01 -4.60748732e-01
-6.89671785e-02 -1.04964316e+00 -7.53392220e-01 -8.65917265e-01
2.62428969e-01 9.13313925e-01 3.04282069e-01 -7.36722425e-02
6.75729096e-01 1.02005303e+00 -1.46843910e+00 3.45255554e-01
2.92936325e-01 1.13184547e+00 6.91713631e-01 5.75849891e-01
-5.56704774e-02 4.48836565e-01 -1.37466788e-01 3.19388330e-01
1.91568866e-01 -1.22333780e-01 -1.26266718e+00 1.76162228e-01
2.68788785e-01 -9.39248204e-02 3.06583792e-01 8.53771210e-01
1.47558838e-01 8.44779983e-02 8.98089707e-01 -6.40404582e-01
-5.14624059e-01 4.66729462e-01 -5.43906212e-01 2.28230357e-01
-5.19578308e-02 3.30175102e-01 9.10323620e-01 -8.69078159e-01
7.99981415e-01 8.42916548e-01 7.93260872e-01 4.36799973e-01
-1.40671837e+00 -2.78419286e-01 -3.53658408e-01 2.31826991e-01
-1.13843298e+00 -4.56565619e-01 9.67348993e-01 -7.28884637e-01
1.08571506e+00 1.38804331e-01 2.96363205e-01 4.10182446e-01
5.97473145e-01 2.72208571e-01 1.21801972e+00 -5.41102052e-01
5.92541218e-01 -3.22890997e-01 2.17481017e-01 6.24750853e-01
9.96594951e-02 -9.25195068e-02 1.48870930e-01 4.08970773e-01
9.49363053e-01 4.56561781e-02 3.57918330e-02 -4.36242893e-02
-1.06448984e+00 6.15516901e-01 6.17640674e-01 4.42942023e-01
-3.75221103e-01 4.89351273e-01 3.17602605e-01 1.81734174e-01
5.44064164e-01 9.82604504e-01 -4.57298994e-01 8.30201060e-02
-1.23263919e+00 2.94832438e-01 4.24436659e-01 4.79380816e-01
1.30544508e+00 -8.24651972e-04 3.82164657e-01 7.12399542e-01
3.51201117e-01 4.61144358e-01 2.67126232e-01 -1.17478967e+00
-3.56099345e-02 6.19042516e-01 -3.12312040e-02 -1.03155541e+00
-8.39480817e-01 1.86142802e-01 -5.77351272e-01 6.43818855e-01
5.63337088e-01 1.07095353e-01 -1.11954367e+00 1.11686921e+00
2.89530605e-01 -5.88394344e-01 -4.08699900e-01 9.42894936e-01
6.44238949e-01 2.94430524e-01 5.34994826e-02 2.10346758e-01
1.28906596e+00 -5.51204860e-01 -3.86806041e-01 -2.72979647e-01
7.54401326e-01 -6.80786252e-01 1.00131559e+00 3.15589854e-03
-9.94448423e-01 -3.89850289e-01 -1.21450758e+00 -1.24251470e-01
-5.33368170e-01 -3.85021372e-03 5.09758353e-01 1.77172139e-01
-1.10270846e+00 1.17196369e+00 -1.26954079e+00 -1.86307117e-01
5.63124299e-01 8.25476050e-01 -4.18255329e-01 4.73182164e-02
-7.14305043e-01 8.07474256e-01 3.48207384e-01 1.41012475e-01
-6.37841403e-01 -7.40647674e-01 -8.70576680e-01 -1.36015192e-01
8.22051689e-02 -3.44493270e-01 1.26360071e+00 -4.64196116e-01
-1.40029752e+00 1.18713033e+00 -1.41474321e-01 -1.95083216e-01
3.90289187e-01 1.65426567e-01 2.26113591e-02 4.67881262e-01
3.54146242e-01 7.12275028e-01 5.83642185e-01 -1.48225069e+00
-5.46295047e-01 -3.37181002e-01 -2.38310859e-01 -1.99046806e-01
2.42705390e-01 -9.94105563e-02 -1.10507309e-02 -1.25055686e-01
4.73924190e-01 -7.28701591e-01 -4.70103800e-01 -3.82234007e-01
-4.93365049e-01 -2.29806781e-01 1.21783042e+00 -6.68605685e-01
4.96033728e-01 -1.76492488e+00 -1.82217330e-01 5.09388983e-01
5.43177545e-01 -1.93899408e-01 3.26284915e-01 2.42727473e-01
-3.84135582e-02 1.75918415e-01 -4.88476634e-01 -4.83150959e-01
-2.39977509e-01 -1.77815817e-02 3.22247744e-01 7.11014092e-01
3.08974594e-01 1.05528629e+00 -6.07245684e-01 -3.97822559e-01
5.06667495e-01 1.21167846e-01 -4.20714289e-01 9.43809152e-02
-4.07155454e-01 7.26861954e-01 -2.61239439e-01 7.98539162e-01
5.55520892e-01 -4.41654444e-01 -9.97344181e-02 -4.34958756e-01
-4.09586489e-01 2.70991586e-02 -6.79330051e-01 1.49088299e+00
-6.79447830e-01 5.53741753e-01 3.80089968e-01 -1.19607687e+00
1.07235467e+00 1.91232726e-01 5.16651213e-01 -9.11259711e-01
4.27102715e-01 5.06893039e-01 -6.94227889e-02 -5.23498535e-01
4.69457060e-01 -6.68628454e-01 4.15679440e-02 6.49128675e-01
-2.54585773e-01 -6.71807647e-01 1.56496510e-01 -3.07020277e-01
1.02866292e+00 6.51307106e-02 -3.50759715e-01 -7.83022106e-01
2.58997619e-01 4.33904171e-01 2.41726622e-01 6.33766115e-01
1.22295581e-01 1.00120902e+00 5.37568405e-02 -6.86059892e-01
-1.88055909e+00 -8.08379114e-01 -3.01048517e-01 5.51349640e-01
2.48488620e-01 3.38212043e-01 -8.78892124e-01 -8.52357596e-02
-8.41148421e-02 4.36233371e-01 -6.88251913e-01 -5.66803440e-02
-8.51462066e-01 -7.84131646e-01 2.12144569e-01 7.35448658e-01
4.78410721e-01 -1.41589785e+00 -8.61296177e-01 4.85748202e-01
2.28757814e-01 -1.14996505e+00 4.98426817e-02 5.81217706e-01
-8.65953743e-01 -1.19789159e+00 -3.09938222e-01 -9.70922589e-01
8.53407919e-01 -3.47333215e-02 1.11055923e+00 2.26901382e-01
-7.12059379e-01 1.77015349e-01 -8.87627155e-02 -1.09233987e-02
-6.94874823e-01 2.34245375e-01 -1.59959570e-01 -2.95093209e-01
2.67639637e-01 -5.89633763e-01 -9.05453742e-01 3.16052318e-01
-9.12130415e-01 3.59292150e-01 4.25259888e-01 7.12508261e-01
9.13998842e-01 2.52487183e-01 6.12927139e-01 -1.29475045e+00
4.17606503e-01 -3.49333704e-01 -4.57844406e-01 -1.91942811e-01
-9.12673652e-01 -1.52004331e-01 1.01824856e+00 2.46473983e-01
-1.12177682e+00 -3.73409921e-03 -5.13462901e-01 -1.46962628e-01
-8.46565008e-01 5.58242977e-01 1.21123996e-02 -2.04806298e-01
4.89944518e-01 -7.53479749e-02 2.81070232e-01 -4.40210700e-01
-3.12660486e-02 6.05003476e-01 7.22958088e-01 -8.19022715e-01
5.22743404e-01 8.96156549e-01 3.21817964e-01 -9.82071221e-01
-5.08724988e-01 -5.67561567e-01 -9.86581743e-01 -3.37917268e-01
1.20246553e+00 -4.11773533e-01 -8.13230217e-01 6.26109838e-01
-9.57402527e-01 -5.82927883e-01 -1.93536088e-01 2.12735966e-01
-7.62638450e-01 2.62053490e-01 -1.06385708e+00 -3.56248200e-01
-4.91543144e-01 -1.48874652e+00 1.13479936e+00 1.74474120e-01
-3.96181077e-01 -1.33558607e+00 -1.85494006e-01 5.69018006e-01
5.45759618e-01 4.59242046e-01 1.37481523e+00 -4.61401016e-01
-4.86885220e-01 -3.78559679e-01 -3.28486115e-01 2.97405988e-01
5.65231740e-01 3.09105497e-02 -9.44858551e-01 -1.59648627e-01
2.89884448e-01 -1.72779545e-01 7.98417985e-01 6.78436875e-01
1.20097637e+00 3.62779170e-01 -3.08683902e-01 6.39590204e-01
1.78711402e+00 4.65832144e-01 5.13768375e-01 8.32233965e-01
1.22892511e+00 6.95311844e-01 2.34123915e-01 -1.27855111e-02
2.10371718e-01 3.90309453e-01 3.23220164e-01 -4.11962181e-01
-1.47611335e-01 3.40528369e-01 -2.40622595e-01 6.99413359e-01
-2.06722766e-01 1.24238171e-01 -1.47079551e+00 6.95602000e-01
-1.27689791e+00 -6.72912776e-01 -4.70372230e-01 1.74717760e+00
8.52898777e-01 2.26516351e-01 -1.31699562e-01 3.79253507e-01
8.23896527e-01 -3.30378741e-01 -6.45800948e-01 -4.54543561e-01
-2.97379401e-02 4.99138236e-01 7.61857688e-01 5.81548750e-01
-1.02431190e+00 8.26453805e-01 6.24930334e+00 5.22128761e-01
-1.47047925e+00 -1.47519559e-01 9.48874056e-01 2.81414330e-01
-2.59815365e-01 -5.91402762e-02 -6.22976422e-01 4.24564749e-01
9.25400496e-01 3.89072895e-01 3.07324678e-01 5.53603947e-01
5.03208160e-01 -3.87333959e-01 -1.17396832e+00 6.67472184e-01
-3.49833131e-01 -1.88771272e+00 -3.15982282e-01 2.08059065e-02
6.29295528e-01 2.16833353e-01 -9.88893583e-02 -4.12554502e-01
2.59369135e-01 -1.00921190e+00 8.14077795e-01 3.40045482e-01
9.56834078e-01 -7.02353179e-01 7.77723551e-01 -6.09345958e-02
-1.03257787e+00 1.20178171e-01 -3.70120376e-01 -1.27178937e-01
3.12260211e-01 8.67560983e-01 -1.09346771e+00 3.86559993e-01
8.66866767e-01 4.80641037e-01 -4.66452032e-01 6.55723333e-01
2.01081082e-01 6.12032771e-01 -4.31777239e-01 3.06184769e-01
4.34546828e-01 -3.58447254e-01 5.77733517e-02 1.09431934e+00
6.87626656e-03 -2.84342974e-01 1.10146880e-01 1.31695032e+00
2.84885317e-02 -1.96245536e-01 -4.69524950e-01 4.16903831e-02
3.52771521e-01 1.27564609e+00 -1.69547617e+00 -6.84257150e-02
-1.88214809e-01 5.11978805e-01 2.77120978e-01 1.69038445e-01
-4.28578258e-01 -3.91217560e-01 2.63068885e-01 5.34185529e-01
2.56695300e-01 -4.65926707e-01 -9.97253418e-01 -4.60300535e-01
-1.90391004e-01 -2.94553638e-01 -1.71962500e-01 -5.19621313e-01
-1.26213944e+00 2.73284048e-01 -1.06193833e-01 -4.73907173e-01
3.49881686e-02 -8.69460940e-01 -7.97190011e-01 8.35987508e-01
-1.37653220e+00 -1.15832901e+00 -1.78467423e-01 7.13510737e-02
4.82419699e-01 2.58391112e-01 5.50672293e-01 1.23171523e-01
-5.67854404e-01 5.26311398e-02 4.65041965e-01 6.84041083e-02
1.61856651e-01 -1.55518126e+00 5.53077996e-01 5.80519736e-01
-5.17216444e-01 4.29608822e-01 9.27221060e-01 -8.00952792e-01
-1.21054602e+00 -1.21669793e+00 4.67933893e-01 -2.74946868e-01
6.45604670e-01 -2.70448804e-01 -1.05581641e+00 5.35576224e-01
-1.31209329e-01 -2.19759881e-01 5.39670110e-01 -2.07020566e-01
3.24354827e-01 2.66875505e-01 -1.28892028e+00 4.06824976e-01
5.39475620e-01 -5.56012332e-01 -4.05018330e-01 2.69105911e-01
2.73403943e-01 -2.13086963e-01 -1.15219641e+00 4.31361467e-01
3.99615586e-01 -8.96978736e-01 7.73465037e-01 -9.67055932e-02
9.12814617e-01 -2.27646351e-01 -9.79795605e-02 -1.11850476e+00
-3.43503773e-01 -9.12523493e-02 5.86411655e-01 1.15125597e+00
5.29831827e-01 -4.77399617e-01 1.26119995e+00 8.66785944e-01
-6.67597473e-01 -8.76357317e-01 -8.28677058e-01 -5.67290127e-01
4.91456836e-01 -3.10665220e-01 5.86687982e-01 7.61036217e-01
-3.72950763e-01 6.76314384e-02 4.19852018e-01 3.33651491e-02
7.00228989e-01 3.17145079e-01 1.75179228e-01 -1.30765080e+00
1.08914025e-01 -4.22026336e-01 -4.17478800e-01 -5.78311324e-01
2.88040221e-01 -9.33433294e-01 4.15366828e-01 -1.70944941e+00
-4.79973555e-02 -1.01571226e+00 -8.88579413e-02 4.61557060e-01
2.42271021e-01 5.50587118e-01 -2.91281611e-01 6.11667514e-01
-2.20157251e-01 1.38961911e-01 1.38401103e+00 -2.60538757e-01
3.91132012e-02 -3.53220224e-01 -2.85265386e-01 7.26784229e-01
1.00545156e+00 -3.34851891e-01 -1.34178787e-01 -4.97573137e-01
1.38786167e-01 -8.71252939e-02 3.95709872e-01 -1.09174442e+00
2.32300714e-01 -2.20564823e-03 5.19351900e-01 -5.23015738e-01
7.74090085e-03 -8.25755715e-01 2.11486027e-01 3.68734002e-01
-2.80973941e-01 4.18987274e-02 7.36368150e-02 1.67669162e-01
-7.71734864e-02 -3.73844981e-01 1.00651419e+00 -6.22160196e-01
-8.44924748e-01 2.48611927e-01 -7.53957570e-01 -3.74703467e-01
9.65512633e-01 -7.28994727e-01 -2.26170897e-01 2.75301337e-01
-9.05532539e-01 -3.96324769e-02 1.03804743e+00 -1.14032380e-01
6.76290512e-01 -9.26063180e-01 -2.90036678e-01 2.96160907e-01
-1.33467972e-01 6.31692827e-01 5.19448817e-01 6.16287112e-01
-1.32026458e+00 2.02278793e-01 -3.03316772e-01 -9.25925672e-01
-7.59989619e-01 2.16581300e-01 4.56451923e-01 -1.25633091e-01
-8.37297916e-01 9.45536911e-01 2.01232448e-01 -4.60693926e-01
-5.94094276e-01 -4.83527690e-01 -1.28786549e-01 -2.80639291e-01
3.91370319e-02 3.94233912e-01 5.73186278e-01 -5.90423763e-01
-1.91803247e-01 6.02439106e-01 -3.08273375e-01 2.46745333e-01
1.67558336e+00 -3.49636227e-01 -5.14428020e-01 6.59230947e-01
1.39710402e+00 -5.59300303e-01 -1.37457764e+00 1.90608472e-01
2.91510731e-01 -2.69092657e-02 1.94200531e-01 -4.33932573e-01
-1.24808180e+00 8.54471982e-01 5.31850398e-01 2.95649499e-01
6.68734193e-01 2.25430369e-01 9.68468785e-01 3.20567131e-01
3.17359000e-01 -1.34286666e+00 -2.00914666e-01 3.51005256e-01
3.76007318e-01 -1.32300568e+00 -9.00649875e-02 -3.74764860e-01
4.62957025e-02 1.38389742e+00 6.58409238e-01 -2.84225613e-01
7.42653728e-01 6.25567377e-01 3.37623835e-01 -1.02898240e+00
-3.32204312e-01 3.57749790e-01 -2.20392793e-01 5.52241564e-01
4.29221153e-01 5.80140091e-02 8.98586586e-02 3.04012801e-02
-4.35676068e-01 -2.46432960e-01 5.07070601e-01 1.35420358e+00
-6.08071744e-01 -8.86561632e-01 -4.55201000e-01 8.18433821e-01
-5.98277092e-01 1.06771864e-01 8.52556378e-02 5.63947320e-01
5.43693034e-03 8.37341905e-01 4.77396488e-01 -3.48963708e-01
8.42779800e-02 -7.20626786e-02 4.12322015e-01 -7.99038947e-01
-4.49687719e-01 -1.59403265e-01 -1.13079667e-01 -3.50427091e-01
-4.76714522e-01 -6.40891194e-01 -1.82313728e+00 -3.08020473e-01
-3.53325844e-01 5.24686649e-02 8.60932887e-01 1.44891500e+00
4.20358218e-02 8.17970455e-01 5.94285190e-01 -1.49117887e+00
-5.01968004e-02 -7.06281364e-01 -9.06714439e-01 8.27824056e-01
1.60519019e-01 -6.32112443e-01 -4.80732590e-01 4.74376559e-01] | [14.173625946044922, -2.87034273147583] |
632a298a-31a2-4c44-bf2c-a1f3ff4d9994 | speech-gesture-generation-from-the-trimodal | 2009.02119 | null | https://arxiv.org/abs/2009.02119v1 | https://arxiv.org/pdf/2009.02119v1.pdf | Speech Gesture Generation from the Trimodal Context of Text, Audio, and Speaker Identity | For human-like agents, including virtual avatars and social robots, making proper gestures while speaking is crucial in human--agent interaction. Co-speech gestures enhance interaction experiences and make the agents look alive. However, it is difficult to generate human-like gestures due to the lack of understanding of how people gesture. Data-driven approaches attempt to learn gesticulation skills from human demonstrations, but the ambiguous and individual nature of gestures hinders learning. In this paper, we present an automatic gesture generation model that uses the multimodal context of speech text, audio, and speaker identity to reliably generate gestures. By incorporating a multimodal context and an adversarial training scheme, the proposed model outputs gestures that are human-like and that match with speech content and rhythm. We also introduce a new quantitative evaluation metric for gesture generation models. Experiments with the introduced metric and subjective human evaluation showed that the proposed gesture generation model is better than existing end-to-end generation models. We further confirm that our model is able to work with synthesized audio in a scenario where contexts are constrained, and show that different gesture styles can be generated for the same speech by specifying different speaker identities in the style embedding space that is learned from videos of various speakers. All the code and data is available at https://github.com/ai4r/Gesture-Generation-from-Trimodal-Context. | ['Joo-Haeng Lee', 'Youngwoo Yoon', 'Jaehong Kim', 'Minsu Jang', 'Jaeyeon Lee', 'Geehyuk Lee', 'Bok Cha'] | 2020-09-04 | null | null | null | null | ['gesture-generation'] | ['robots'] | [ 1.94300681e-01 1.74728304e-01 6.88567460e-02 -3.81438643e-01
-6.70113742e-01 -7.71391273e-01 1.09635615e+00 -7.81123698e-01
-2.51572371e-01 4.89113420e-01 5.22042215e-01 -2.40622014e-02
3.59279662e-01 -4.04980272e-01 -5.20846486e-01 -8.45650196e-01
3.73392133e-04 5.25801361e-01 -1.60655200e-01 -4.39357072e-01
-4.87273559e-02 3.87843698e-01 -1.59438109e+00 3.48537028e-01
3.96570385e-01 6.13487124e-01 1.89147875e-01 1.31207573e+00
-1.66500673e-01 6.82556689e-01 -1.06448388e+00 -2.75718272e-01
2.83060193e-01 -8.29025745e-01 -6.54354095e-01 1.56257004e-02
1.98402047e-01 -7.63732374e-01 -5.29021978e-01 7.57257223e-01
7.73954749e-01 3.23295772e-01 8.80904078e-01 -1.76099372e+00
-5.64173400e-01 7.72525728e-01 1.50255486e-01 -5.63306332e-01
8.84500325e-01 5.78681529e-01 9.20337677e-01 -7.53746092e-01
7.70262480e-01 1.59653521e+00 3.07846665e-01 1.38704467e+00
-9.81836081e-01 -9.51322198e-01 1.01856984e-01 1.05600238e-01
-1.43448806e+00 -5.57150185e-01 7.93288827e-01 -3.74075562e-01
7.85693944e-01 5.92994094e-01 6.44307375e-01 2.06126714e+00
-4.09281105e-01 9.97787774e-01 7.80441105e-01 -5.20723999e-01
1.56028315e-01 -3.53722484e-03 -4.01390314e-01 4.32939768e-01
-5.79427063e-01 6.25980973e-01 -7.67125428e-01 -8.45894814e-02
8.76321018e-01 -1.39067680e-01 -3.75197887e-01 -2.11491808e-01
-1.69857156e+00 7.48370409e-01 2.54717708e-01 3.61454189e-01
-3.03079724e-01 3.80071759e-01 2.74257749e-01 4.70877707e-01
-2.12110385e-01 2.69546419e-01 -5.67664579e-03 -5.81088781e-01
-6.44755423e-01 4.46603715e-01 9.80396926e-01 1.18768060e+00
7.34281391e-02 3.76133561e-01 -1.71186164e-01 6.57557547e-01
4.46027070e-01 1.07366884e+00 7.74811566e-01 -1.15674388e+00
3.94460469e-01 1.75799638e-01 2.42977560e-01 -7.38843262e-01
-2.73471743e-01 3.40528458e-01 -7.78691888e-01 6.75331354e-01
6.30629063e-01 -4.26241368e-01 -9.39664721e-01 2.09883976e+00
3.53511661e-01 3.26079577e-01 4.86456841e-01 1.17516840e+00
1.22317338e+00 7.47255027e-01 6.22916743e-02 -1.02837272e-01
1.11473620e+00 -1.02082765e+00 -1.05496740e+00 1.13140740e-01
2.40718231e-01 -8.41283262e-01 1.45430231e+00 2.59923846e-01
-7.99565077e-01 -5.95524430e-01 -1.03637862e+00 1.60751536e-01
-3.71252596e-02 1.25341952e-01 5.05722880e-01 9.21604872e-01
-9.00793076e-01 1.87702000e-01 -9.55586493e-01 -5.53671479e-01
-9.39992890e-02 3.70593041e-01 -3.70315611e-01 4.19420749e-01
-9.98093128e-01 6.25993669e-01 7.52351806e-02 -8.56865272e-02
-1.13882864e+00 -4.88206334e-02 -9.16189909e-01 -2.96083272e-01
1.47887126e-01 -6.21249557e-01 1.71833014e+00 -1.16087008e+00
-2.31669211e+00 3.35811883e-01 -1.37861401e-01 -2.28328854e-01
7.79563785e-01 -1.12902492e-01 -3.85384500e-01 3.04917693e-01
-3.35725546e-01 1.03343630e+00 1.07535982e+00 -1.65293598e+00
-5.95763385e-01 -6.48922548e-02 9.90060568e-02 5.58209538e-01
-2.08449155e-01 1.12476638e-02 -2.49459371e-01 -9.19124961e-01
-9.97219533e-02 -1.25996423e+00 -1.00527816e-02 1.16802407e-02
-4.25419867e-01 -1.76409129e-02 9.79256690e-01 -6.29712760e-01
8.27480018e-01 -2.16966128e+00 4.70624298e-01 2.75960922e-01
-1.47983015e-01 1.54290736e-01 -4.50188577e-01 5.67447186e-01
3.82589757e-01 1.84736419e-02 -2.15744525e-02 -4.74909902e-01
3.97138447e-01 3.08108807e-01 -4.54356641e-01 8.24042261e-02
-5.06198928e-02 8.04989576e-01 -9.83612597e-01 -3.74445468e-01
4.10111815e-01 8.43658388e-01 -5.30450046e-01 6.20631695e-01
-2.13623405e-01 1.01822305e+00 -4.21006292e-01 5.95003188e-01
6.89940229e-02 3.44583780e-01 3.15116048e-01 2.73556937e-03
2.57501721e-01 1.54304996e-01 -1.37294114e+00 1.75838768e+00
-6.19898736e-01 8.06120396e-01 2.20674813e-01 -4.96974021e-01
8.51928294e-01 9.46196973e-01 1.94127843e-01 -4.29509193e-01
2.87515938e-01 2.29531527e-01 1.91380888e-01 -8.02741170e-01
3.05838585e-01 -1.81372896e-01 -2.29068443e-01 7.47628808e-01
6.58667162e-02 -6.41826272e-01 -1.33561403e-01 -1.08009158e-02
8.69631171e-01 3.19478035e-01 1.29435539e-01 6.27324104e-01
3.53026599e-01 -3.12121153e-01 1.22002319e-01 6.54253066e-01
-2.16025800e-01 6.99480236e-01 2.14712657e-02 -2.06288114e-01
-8.95722151e-01 -1.23160100e+00 4.63731647e-01 1.37459195e+00
1.72207370e-01 -8.15808699e-02 -8.76153767e-01 -5.79969585e-01
-4.06774879e-01 9.35135365e-01 -3.83898973e-01 1.47282258e-02
-7.52697945e-01 -1.34882137e-01 1.09648573e+00 4.61708963e-01
2.37857461e-01 -1.62720251e+00 -7.56450117e-01 8.30212235e-02
-4.91215199e-01 -1.13838112e+00 -7.75523722e-01 -3.24011624e-01
-4.32766527e-01 -8.39333236e-01 -9.76061046e-01 -9.82113004e-01
4.21589375e-01 7.88209308e-03 6.62490964e-01 4.88842949e-02
-5.67405149e-02 6.93059564e-01 -7.53271103e-01 -4.26434577e-01
-1.10122454e+00 -1.92201421e-01 2.90921688e-01 -1.12968953e-02
2.24936694e-01 -7.09527016e-01 -3.02116483e-01 5.56454420e-01
-8.87895763e-01 2.35086992e-01 4.24078643e-01 1.07194042e+00
-6.66380450e-02 -5.05102813e-01 5.68779588e-01 -1.24371424e-01
7.61600018e-01 -6.68543903e-03 -5.99999987e-02 8.45831558e-02
2.13418994e-02 6.13330379e-02 3.78357649e-01 -1.15233207e+00
-1.08201396e+00 3.81008267e-01 -2.56189048e-01 -4.46772605e-01
-6.33842468e-01 -9.25591961e-02 -4.83248383e-01 1.72963291e-01
7.35933542e-01 2.90920138e-01 3.67455810e-01 -6.98897019e-02
7.58042753e-01 1.14747417e+00 6.21122301e-01 -5.66478193e-01
8.72772098e-01 3.73345912e-01 -4.27631438e-01 -1.06212437e+00
2.30310425e-01 -1.37391776e-01 -4.79436010e-01 -4.60299224e-01
8.61036837e-01 -7.60022044e-01 -1.06838369e+00 6.20834291e-01
-1.39220631e+00 -8.65827560e-01 -1.83581382e-01 8.27049315e-01
-9.41057384e-01 1.35846481e-01 -5.56366920e-01 -1.11157358e+00
-3.09875369e-01 -1.44395280e+00 1.21398878e+00 1.28567383e-01
-7.22255290e-01 -6.78252339e-01 -7.87413940e-02 2.89911330e-01
2.68705875e-01 3.72120827e-01 4.92812842e-01 -7.30199695e-01
-4.57576782e-01 1.80776510e-02 3.86974245e-01 1.96726128e-01
4.60157514e-01 6.62115440e-02 -9.97598469e-01 -4.55505103e-02
-2.48816177e-01 -5.38708091e-01 2.85586238e-01 1.48898199e-01
3.97554487e-01 -7.20971644e-01 -6.59331307e-02 4.83626723e-01
6.39424860e-01 5.87508678e-01 5.47834516e-01 -2.43691355e-02
8.07497084e-01 7.60998011e-01 5.96725285e-01 4.00584668e-01
3.24401706e-01 1.09105730e+00 3.68625373e-01 1.47936687e-01
-3.88243109e-01 -4.23278660e-01 8.51453900e-01 7.37977028e-01
-3.78005654e-01 -4.20713782e-01 -6.47020638e-01 4.57635403e-01
-1.77252269e+00 -1.42811763e+00 1.62347823e-01 1.89051890e+00
1.06353390e+00 -2.80163080e-01 5.41650891e-01 1.69505656e-01
6.87942684e-01 8.21080804e-03 -4.39287364e-01 -4.80319947e-01
-6.51685894e-02 2.06285834e-01 8.08205281e-04 8.55887413e-01
-1.01219499e+00 1.08115816e+00 5.83869648e+00 6.63373291e-01
-1.34039259e+00 3.15026604e-02 4.70165759e-02 -3.55807841e-01
-2.55711466e-01 -5.24251938e-01 -3.70015591e-01 2.90061444e-01
6.41603589e-01 7.90692642e-02 8.51675391e-01 6.45560205e-01
4.50594664e-01 6.40360832e-01 -1.49001479e+00 1.08949506e+00
2.93067902e-01 -9.88271117e-01 2.04014838e-01 -1.16249152e-01
5.62871218e-01 -3.70905876e-01 2.04066038e-01 3.46435338e-01
4.80466068e-01 -1.30616248e+00 1.10025311e+00 2.12593287e-01
9.67744112e-01 -4.64739919e-01 4.80547041e-01 3.53509754e-01
-1.18847370e+00 -7.18033016e-02 3.48599881e-01 -3.99924442e-02
4.40126389e-01 -6.90956652e-01 -1.35142410e+00 1.39182076e-01
3.30086648e-01 1.05137169e-01 -6.12059003e-03 5.35210311e-01
-5.08802354e-01 4.76265639e-01 -3.11766982e-01 -5.10777652e-01
9.56180990e-02 6.58351630e-02 9.13551331e-01 1.24838519e+00
6.21903062e-01 2.49876097e-01 9.03031603e-02 6.62965715e-01
1.48294732e-01 1.08821057e-01 -8.04777980e-01 -2.02973753e-01
7.16705203e-01 8.71939063e-01 -4.35346693e-01 -2.09310010e-01
-1.96813121e-01 1.28170145e+00 -2.99720526e-01 6.58382535e-01
-8.13855052e-01 -2.76389390e-01 8.78465474e-01 -3.23154628e-01
1.50083199e-01 -3.51991266e-01 -1.09846726e-01 -9.01300013e-01
-1.04921803e-01 -1.49211764e+00 -4.57527302e-02 -7.74182200e-01
-1.00832689e+00 8.97569180e-01 5.17891683e-02 -1.52975404e+00
-1.26468217e+00 -5.13087213e-01 -6.85069680e-01 4.70387340e-01
-6.47532642e-01 -1.59564722e+00 -3.39830875e-01 6.82513118e-01
9.12683189e-01 -6.39416635e-01 1.21346998e+00 4.45870049e-02
3.14377859e-04 9.69542384e-01 -2.07259670e-01 3.69215906e-01
6.98223531e-01 -1.06644225e+00 4.64842230e-01 4.97466028e-01
5.94553888e-01 5.76115668e-01 1.01624155e+00 -3.93732905e-01
-1.30531585e+00 -6.96348488e-01 5.52131593e-01 -4.16511834e-01
4.66575980e-01 -3.89211446e-01 -4.83804643e-01 6.63208663e-01
4.47707802e-01 -4.96793985e-01 7.50109911e-01 -3.16914886e-01
-3.65849942e-01 3.59108478e-01 -1.18859076e+00 1.06823111e+00
1.28223550e+00 -4.80951756e-01 -7.15523243e-01 4.68601808e-02
8.27388048e-01 -5.15681624e-01 -3.85210127e-01 1.59594283e-01
1.05906606e+00 -7.12167919e-01 1.05369878e+00 -5.19971430e-01
1.94814786e-01 -3.37606519e-01 -4.73021209e-01 -1.45411813e+00
3.34877700e-01 -1.04429626e+00 -6.12227917e-02 1.17556727e+00
4.28695112e-01 -4.28867608e-01 5.98308146e-01 4.93006140e-01
-2.94815819e-03 -8.02041143e-02 -8.95496726e-01 -9.28157628e-01
-7.05641881e-02 -4.89527196e-01 9.61410344e-01 8.51088583e-01
2.64480203e-01 3.44447970e-01 -5.79542994e-01 3.77632529e-01
3.44946831e-01 -3.27370055e-02 1.39861345e+00 -8.32462549e-01
-5.53749740e-01 -6.78119481e-01 -4.89990771e-01 -1.05553126e+00
4.25371110e-01 -7.05537260e-01 3.72882068e-01 -1.25333643e+00
-2.15244904e-01 -2.86710083e-01 3.65278512e-01 4.33532000e-01
-2.09992453e-02 3.41689110e-01 6.49726033e-01 1.78370655e-01
-1.47943959e-01 8.27733219e-01 1.33363760e+00 -3.61345828e-01
-6.12592638e-01 1.42903611e-01 -1.61545068e-01 7.81609952e-01
8.58947515e-01 -1.73451290e-01 -5.38651228e-01 -3.56709301e-01
-3.29245359e-01 2.27060944e-01 5.94239891e-01 -8.89813662e-01
1.09872393e-01 -3.62806201e-01 3.09382454e-02 -1.24944761e-01
9.15841162e-01 -8.41745973e-01 2.36630499e-01 4.74154621e-01
-5.90644836e-01 -1.39687851e-01 -6.10985467e-03 3.71446967e-01
-1.29239306e-01 -1.17066942e-01 2.81427920e-01 9.92584750e-02
-5.83363712e-01 -7.27940202e-02 -5.53163767e-01 -1.83421642e-01
9.08593297e-01 -1.63702950e-01 -2.64498945e-02 -1.20912719e+00
-9.07734275e-01 3.88809964e-02 4.46037143e-01 7.90465891e-01
8.15507889e-01 -1.68156433e+00 -8.82682860e-01 2.09311247e-01
8.78810957e-02 -2.27797836e-01 -1.36220409e-02 2.61585146e-01
-5.33471227e-01 1.09888114e-01 -3.44380468e-01 -6.56722546e-01
-1.63834715e+00 2.33766422e-01 1.59545824e-01 2.99156964e-01
-4.89415914e-01 7.45018423e-01 2.12555856e-01 -7.15980470e-01
7.81838655e-01 -3.92787158e-01 -1.30309567e-01 -2.34699160e-01
7.77805865e-01 2.07380176e-01 -5.51172435e-01 -1.07594454e+00
-2.00836405e-01 3.89436632e-01 4.78173852e-01 -1.10785007e+00
1.05024052e+00 -4.94991951e-02 5.75243175e-01 5.24328947e-01
8.64441574e-01 1.25682756e-01 -1.34234917e+00 3.91656309e-02
-4.91259575e-01 -3.66242021e-01 -5.52854598e-01 -7.87927985e-01
-8.16140831e-01 8.86687279e-01 7.27068126e-01 9.58031937e-02
9.20804858e-01 -3.86771839e-03 8.49311531e-01 5.93316853e-01
5.41433275e-01 -8.90045464e-01 5.60805559e-01 4.94717389e-01
1.46284103e+00 -1.18991685e+00 -5.84067464e-01 -2.39399880e-01
-1.15695262e+00 8.95379484e-01 4.58303988e-01 2.84075022e-01
3.04625154e-01 3.66448760e-01 7.48049676e-01 2.57256478e-01
-4.70787942e-01 -2.22065449e-01 1.33682624e-01 1.28075206e+00
2.70138919e-01 4.00153548e-01 -6.31387830e-02 7.12831438e-01
-8.72455299e-01 -2.15474799e-01 4.76400048e-01 7.50558496e-01
-9.79830921e-02 -1.33089447e+00 -7.02937663e-01 -2.89935321e-01
-1.25845522e-01 1.28387183e-01 -6.86689198e-01 8.49213719e-01
-1.30715087e-01 1.41148591e+00 -1.26351461e-01 -7.14607716e-01
2.83005297e-01 4.06133592e-01 6.07483625e-01 -4.55254972e-01
-5.21038115e-01 1.19393826e-01 2.17972830e-01 -5.01178622e-01
-5.54222465e-01 -7.89596975e-01 -1.62493873e+00 -2.52917171e-01
-2.78777275e-02 -1.31582335e-01 7.88435400e-01 7.79937208e-01
2.99594291e-02 3.64645332e-01 6.61902964e-01 -1.57900155e+00
-6.26258850e-01 -1.25572896e+00 -1.72864005e-01 8.02065909e-01
4.60716546e-01 -5.75018525e-01 -4.49444175e-01 5.09430647e-01] | [5.6110944747924805, -0.10605797171592712] |
cc6672f4-93c7-4ff8-b4c9-73d4cb4066ac | deep-tiling-texture-tile-synthesis-using-a | 2103.07992 | null | https://arxiv.org/abs/2103.07992v1 | https://arxiv.org/pdf/2103.07992v1.pdf | Deep Tiling: Texture Tile Synthesis Using a Deep Learning Approach | Texturing is a fundamental process in computer graphics. Texture is leveraged to enhance the visualization outcome for a 3D scene. In many cases a texture image cannot cover a large 3D model surface because of its small resolution. Conventional techniques like repeating, mirror repeating or clamp to edge do not yield visually acceptable results. Deep learning based texture synthesis has proven to be very effective in such cases. All deep texture synthesis methods trying to create larger resolution textures are limited in terms of GPU memory resources. In this paper, we propose a novel approach to example-based texture synthesis by using a robust deep learning process for creating tiles of arbitrary resolutions that resemble the structural components of an input texture. In this manner, our method is firstly much less memory limited owing to the fact that a new texture tile of small size is synthesized and merged with the original texture and secondly can easily produce missing parts of a large texture. | ['Ioannis Fudos', 'Vasilis Toulatzis'] | 2021-03-14 | null | null | null | null | ['texture-synthesis'] | ['computer-vision'] | [ 4.95020717e-01 1.28670827e-01 4.14017230e-01 2.15206861e-01
-2.71617323e-01 -1.80457816e-01 7.04321742e-01 -3.74586657e-02
1.53885916e-01 7.53417134e-01 4.45889235e-02 -3.86495918e-01
1.43303707e-01 -1.20589638e+00 -6.98352277e-01 -7.51282215e-01
3.10593367e-01 4.21863973e-01 4.41905022e-01 -2.96648145e-01
3.70725453e-01 7.74218857e-01 -2.01675248e+00 6.80930793e-01
5.00715613e-01 1.02724814e+00 2.26280823e-01 5.43981194e-01
-5.54337025e-01 4.08542186e-01 -4.91082191e-01 -5.54191247e-02
4.59669501e-01 -4.40394908e-01 -4.60847110e-01 4.43350643e-01
6.26868844e-01 -2.25571364e-01 1.17753670e-01 5.83053350e-01
3.82521689e-01 -6.78979829e-02 5.87659478e-01 -7.80162752e-01
-7.17030391e-02 -4.85203080e-02 -9.75341320e-01 -4.02811408e-01
3.17841172e-01 -1.17893934e-01 3.60411495e-01 -7.85295606e-01
7.64028192e-01 1.29951739e+00 6.93708718e-01 3.84008229e-01
-1.45501876e+00 -4.80945796e-01 -3.16540569e-01 -3.35855871e-01
-9.92534578e-01 -2.68650800e-01 1.03334284e+00 -2.33130991e-01
9.53092158e-01 6.23013020e-01 8.77082765e-01 8.66116285e-01
7.01117635e-01 4.14685190e-01 1.60803127e+00 -7.47251987e-01
2.81075478e-01 4.51329090e-02 -5.90526640e-01 7.13433206e-01
1.79064617e-01 2.53287345e-01 -5.15595317e-01 3.61117907e-02
1.50954902e+00 -7.23009855e-02 -1.08795641e-02 -5.61279655e-01
-1.06901228e+00 5.87659538e-01 3.22999179e-01 1.29356712e-01
-2.22728372e-01 2.55840063e-01 3.85969311e-01 4.65442270e-01
8.12662482e-01 4.99524385e-01 -2.09488496e-01 -1.43837392e-01
-9.25787807e-01 3.08933020e-01 7.55747259e-01 7.30046034e-01
8.86872709e-01 3.86156678e-01 2.04231486e-01 7.04140902e-01
-1.93130642e-01 2.58674949e-01 6.88607618e-02 -9.10391152e-01
-3.44762467e-02 7.64029801e-01 7.20307976e-02 -1.06844127e+00
-3.74048263e-01 -3.99869345e-02 -1.09283292e+00 1.34531140e+00
4.08301175e-01 1.67377129e-01 -1.12732589e+00 8.83461654e-01
7.05237746e-01 -1.34177938e-01 -3.57263505e-01 6.00632787e-01
7.85792112e-01 6.14427269e-01 -4.13631558e-01 2.76454717e-01
1.19024777e+00 -7.70408571e-01 -5.55389225e-01 7.52612129e-02
4.55291241e-01 -1.26666272e+00 1.27336240e+00 5.65477788e-01
-1.09047186e+00 -4.82443810e-01 -1.15479040e+00 -2.21626863e-01
-3.00442159e-01 -2.51601428e-01 6.70703948e-01 6.70512438e-01
-9.94455457e-01 6.31074309e-01 -5.58153570e-01 -1.54770672e-01
6.50676131e-01 3.00876498e-01 -4.66404259e-01 8.52089003e-02
-3.43310803e-01 7.81319559e-01 1.45745113e-01 -1.55517355e-01
-3.51423502e-01 -7.05626845e-01 -6.30062163e-01 9.86360312e-02
2.66023815e-01 -7.58781791e-01 9.30063784e-01 -9.55284178e-01
-1.85559440e+00 1.04696560e+00 -1.03680164e-01 -1.11082338e-01
8.87230694e-01 -4.23153266e-02 4.09305766e-02 -2.38574520e-01
-2.42848411e-01 5.84118962e-01 1.27860391e+00 -1.42524672e+00
-6.02863133e-01 -2.16976508e-01 -1.70945346e-01 2.04540297e-01
1.35843620e-01 -4.80916291e-01 -9.30646658e-02 -8.71841609e-01
2.88055569e-01 -6.01646066e-01 -3.72468263e-01 3.70619208e-01
-2.82571554e-01 1.67117342e-01 1.24920833e+00 -2.68723220e-01
7.21744478e-01 -2.09398031e+00 -9.60311443e-02 4.13194418e-01
3.90519023e-01 6.35610446e-02 7.41644278e-02 4.74311233e-01
1.65272787e-01 8.07684511e-02 -8.44728053e-02 -3.22112918e-01
-1.65044785e-01 7.09831044e-02 -3.01089883e-01 5.79985939e-02
1.94694042e-01 5.81665397e-01 -4.11139339e-01 -2.91250557e-01
4.74069417e-01 6.45072162e-01 -6.54792428e-01 -4.00431491e-02
-4.63277638e-01 6.71770096e-01 -4.30203617e-01 5.99044144e-01
8.88410807e-01 -1.49908299e-02 1.41647846e-01 -1.39754355e-01
-4.76039916e-01 -6.30620029e-03 -1.22114146e+00 1.53301871e+00
-7.51817167e-01 7.36863971e-01 1.97283048e-02 -6.81724310e-01
1.38175786e+00 2.72740871e-01 1.93691105e-01 -1.10677814e+00
2.06433371e-01 4.31431651e-01 -2.36251950e-01 -1.83120266e-01
6.79583549e-01 -3.08892459e-01 1.24014542e-01 7.91438103e-01
-6.67741716e-01 -7.28439689e-01 -2.28080302e-01 -3.62029701e-01
8.63201082e-01 4.87443686e-01 1.94047034e-01 -3.87803882e-01
2.32245520e-01 1.25910446e-01 5.28332479e-02 7.06806064e-01
4.71919686e-01 8.69132936e-01 6.23714805e-01 -1.14200664e+00
-1.90883863e+00 -8.57005179e-01 -2.13520095e-01 7.04200149e-01
7.43550137e-02 -1.73009470e-01 -7.32650280e-01 -5.83500452e-02
3.20269517e-03 2.49900252e-01 -8.62180531e-01 2.23061502e-01
-8.22344780e-01 -3.49726290e-01 1.73781410e-01 3.10448587e-01
6.94690347e-01 -1.27411091e+00 -1.24214160e+00 3.07913303e-01
3.82679582e-01 -6.74748302e-01 -8.68873578e-03 2.22297370e-01
-1.23052561e+00 -8.50026071e-01 -7.67760634e-01 -9.15956676e-01
6.96799219e-01 3.31936121e-01 1.15841186e+00 3.46315295e-01
-5.09981871e-01 -2.10535318e-01 -1.72761694e-01 -4.45471555e-01
-5.85419297e-01 -1.24589376e-01 -2.83669233e-01 -1.24613151e-01
1.49139287e-02 -6.22238159e-01 -4.98254955e-01 2.60356396e-01
-1.12397397e+00 9.01852787e-01 4.64049399e-01 9.08527017e-01
8.08629990e-01 2.94243783e-01 1.67480558e-01 -1.00382221e+00
6.39149606e-01 1.01077467e-01 -5.67438424e-01 -2.48766644e-03
-1.57534942e-01 1.22370668e-01 5.35062432e-01 -5.33833742e-01
-1.15959370e+00 4.88055609e-02 9.00542736e-02 -2.96760172e-01
-2.92878211e-01 2.52771229e-01 1.59838021e-01 -1.92800537e-01
7.09118664e-01 3.21813859e-03 2.64965445e-01 -6.47238314e-01
-1.59459584e-03 2.28444472e-01 1.33375645e-01 -6.29240155e-01
6.58953726e-01 8.31210852e-01 4.20304716e-01 -9.78388786e-01
-3.02888960e-01 3.00058246e-01 -7.28532910e-01 -3.06338280e-01
5.82673728e-01 -4.22754884e-01 -5.55990756e-01 6.52670205e-01
-1.01240456e+00 -7.17836559e-01 -4.11085606e-01 3.45937684e-02
-7.19921410e-01 1.38740968e-02 -3.87398869e-01 -6.18383110e-01
-2.26148382e-01 -1.15649855e+00 1.24243784e+00 1.93803102e-01
-4.16066527e-01 -9.19545352e-01 2.50138547e-02 -1.37097563e-03
7.61236072e-01 7.63526678e-01 1.27164412e+00 3.88475686e-01
-6.36767685e-01 -1.47804201e-01 -4.00457352e-01 -2.10250124e-01
3.41534257e-01 1.49053201e-01 -1.02918100e+00 -1.03708439e-01
3.57617475e-02 -1.23135410e-01 6.49068296e-01 4.21932846e-01
1.06372607e+00 -1.31165655e-02 -2.73405224e-01 5.88021517e-01
1.49946892e+00 3.12026173e-01 9.22886431e-01 8.15562010e-01
7.52405524e-01 6.93595707e-01 3.30156296e-01 4.18057650e-01
-2.80636281e-01 8.68294179e-01 1.47338644e-01 -6.50638223e-01
-4.57229644e-01 -1.46959424e-01 -1.76857457e-01 3.57182145e-01
-3.95447195e-01 -1.26820371e-01 -9.59403396e-01 1.02088973e-01
-1.50610721e+00 -7.51085162e-01 -3.30857009e-01 2.28512573e+00
6.68303549e-01 3.23003918e-01 -1.49705127e-01 5.01684666e-01
5.01261115e-01 -2.28937492e-02 -2.26807863e-01 -1.05336702e+00
-2.76519358e-01 7.13542879e-01 2.35385284e-01 4.69616354e-01
-8.57581675e-01 9.92097139e-01 6.33709383e+00 1.00878298e+00
-1.45538485e+00 -2.70229548e-01 7.33592868e-01 -1.12642951e-01
-4.77344722e-01 -1.98297307e-01 -3.35317731e-01 1.94797978e-01
2.61477202e-01 4.59565334e-02 2.14952826e-01 5.35640299e-01
3.12781602e-01 -6.32220149e-01 -9.00757313e-01 1.03800380e+00
-2.99978644e-01 -1.65899742e+00 3.93833458e-01 3.39220911e-01
9.46672380e-01 -5.83841324e-01 3.13003421e-01 -2.71501660e-01
-2.41246484e-02 -1.39179718e+00 7.42598712e-01 4.11736131e-01
1.14574838e+00 -9.67534840e-01 5.75851798e-01 2.10450187e-01
-9.85252738e-01 3.74107450e-01 -5.12441099e-01 -3.47134829e-01
8.08042437e-02 7.53133774e-01 -7.18542755e-01 2.08525434e-01
8.91783655e-01 2.03498796e-01 -3.00768286e-01 9.81639624e-01
2.21158117e-01 1.36628300e-01 -5.71808755e-01 7.58038983e-02
3.46782416e-01 -2.76057452e-01 3.82638872e-01 9.71221030e-01
5.79643726e-01 -1.19539931e-01 -8.20753425e-02 8.51123512e-01
1.64158687e-01 1.18090451e-01 -1.12213492e+00 4.05013651e-01
2.54176825e-01 9.87558246e-01 -1.21242177e+00 -3.50797176e-01
-3.50656956e-01 1.02760136e+00 1.52963087e-01 1.28870770e-01
-4.31313813e-01 -2.22393647e-01 2.79610217e-01 5.18042684e-01
4.80289221e-01 -3.22695643e-01 -9.63140368e-01 -7.06699133e-01
-2.72482280e-02 -9.85529304e-01 -3.24852765e-01 -8.50017309e-01
-8.38276386e-01 7.74284303e-01 -3.68529320e-01 -1.50796211e+00
1.60243675e-01 -5.72337329e-01 -5.74126124e-01 9.84567821e-01
-1.10026228e+00 -1.25229490e+00 -4.76608157e-01 4.50154722e-01
6.66602135e-01 -1.20854713e-01 1.01603687e+00 4.77966368e-02
-3.92746367e-02 2.34875694e-01 2.11314574e-01 -4.58866000e-01
4.30681497e-01 -1.02142644e+00 5.53339362e-01 3.58691335e-01
-7.51932487e-02 1.36993900e-01 7.55158305e-01 -8.12550545e-01
-1.36928713e+00 -6.34996116e-01 6.27981722e-01 -2.71060824e-01
2.19622955e-01 -5.11751235e-01 -1.08239198e+00 1.56535998e-01
1.96484908e-01 -3.56252074e-01 2.91529864e-01 -9.77182910e-02
-1.80589482e-01 6.89101890e-02 -1.17750704e+00 1.01274741e+00
7.64723659e-01 -3.08443516e-01 -1.38979331e-01 -1.84526294e-01
6.34679720e-02 -8.20912480e-01 -6.10379279e-01 4.03280407e-01
9.77050781e-01 -1.36166787e+00 7.50140786e-01 -2.86196113e-01
8.14414918e-01 -2.98020124e-01 1.87618420e-01 -1.14996994e+00
-2.79337168e-01 -5.97916543e-01 1.98131859e-01 7.29831219e-01
2.59084314e-01 -5.34082115e-01 1.10111165e+00 3.43914658e-01
-1.66084141e-01 -8.36535096e-01 -1.00153339e+00 -6.05354667e-01
1.26946092e-01 -1.61259651e-01 8.18754613e-01 9.50268745e-01
-3.41401488e-01 -6.07141666e-02 -4.72527295e-01 -2.66896039e-01
6.19754672e-01 4.97000128e-01 1.15703607e+00 -1.34612322e+00
-7.18534365e-02 -7.20413685e-01 -3.60597879e-01 -6.73027515e-01
-5.08147657e-01 -4.66849059e-01 -1.25005424e-01 -1.33885968e+00
6.04446754e-02 -9.62265193e-01 3.21538746e-01 2.47308165e-01
2.04953164e-01 7.37562180e-01 -6.57690689e-02 1.51909038e-01
2.47600138e-01 3.23768497e-01 1.83644474e+00 3.42850797e-02
-3.60317677e-01 -2.15847805e-01 -2.89530873e-01 8.19797218e-01
9.29396510e-01 -3.28529954e-01 -3.02684605e-01 -5.52237809e-01
2.45822787e-01 2.71171499e-02 2.67651707e-01 -9.37372327e-01
-1.34073883e-01 1.84038207e-02 7.97948003e-01 -5.87092161e-01
3.49868178e-01 -8.02682757e-01 6.55272961e-01 3.84187967e-01
-5.46164811e-02 1.12418771e-01 5.26522219e-01 2.87884444e-01
-1.02622516e-01 1.53015209e-02 8.17174196e-01 -3.59004766e-01
-4.70558763e-01 2.18243040e-02 -6.23028457e-01 -4.84688848e-01
9.27270472e-01 -9.62601066e-01 -9.50572640e-02 -1.60204709e-01
-6.85900450e-01 -4.52544034e-01 1.05149794e+00 2.91852206e-01
6.47093117e-01 -1.38700128e+00 -5.30588746e-01 5.70323229e-01
-2.14580491e-01 3.13061297e-01 2.14861691e-01 4.04206693e-01
-1.32705307e+00 8.08583871e-02 -7.06844449e-01 -6.61020219e-01
-1.34305239e+00 3.88883054e-01 2.46740118e-01 -1.78268835e-01
-1.27796257e+00 6.71961188e-01 6.49342656e-01 -1.79364964e-01
7.71589652e-02 -2.43993610e-01 3.46016400e-02 -2.85616606e-01
5.15033185e-01 2.52671450e-01 3.36048186e-01 -1.79483354e-01
2.32056916e-01 9.79047477e-01 -1.06570624e-01 -1.57614991e-01
1.38748944e+00 7.91684240e-02 -7.72329047e-02 2.64340371e-01
8.55667770e-01 1.94838509e-01 -1.48338246e+00 -4.59982231e-02
-1.65120676e-01 -8.07498693e-01 1.62631676e-01 -4.26472962e-01
-1.06807220e+00 9.64317739e-01 5.17744780e-01 3.80724400e-01
1.13952005e+00 -2.48598039e-01 7.77734935e-01 1.28235430e-01
6.12464666e-01 -1.05098283e+00 1.81877896e-01 4.64903772e-01
1.04690850e+00 -9.59558070e-01 2.00538158e-01 -5.34028590e-01
-3.69823337e-01 1.58637309e+00 4.73025560e-01 -3.88792574e-01
5.46194613e-01 8.66329312e-01 1.79565489e-01 -4.58605170e-01
-6.76267147e-01 1.42526507e-01 1.89180642e-01 6.43725693e-01
5.64483821e-01 -1.34910112e-02 -1.74850196e-01 -3.72398049e-01
-4.49465036e-01 -2.68080309e-02 5.95203102e-01 1.14214551e+00
-5.04660547e-01 -1.24780512e+00 -5.83765209e-01 4.34638053e-01
-2.88952053e-01 1.01752795e-01 -2.21380442e-01 8.25417638e-01
-2.28989660e-03 3.63709420e-01 4.43907350e-01 -2.35016242e-01
3.04976255e-01 -2.32373834e-01 8.81178737e-01 -4.49542046e-01
-4.50732946e-01 4.81176257e-01 -1.48273632e-02 -5.06519794e-01
-2.85798550e-01 -2.77374208e-01 -8.48709464e-01 -7.30197668e-01
1.29656447e-02 -3.95085394e-01 5.93800843e-01 6.55754983e-01
2.89541155e-01 4.47581708e-01 4.88952398e-01 -1.44574952e+00
3.70990843e-01 -5.58267474e-01 -7.54923820e-01 3.53416681e-01
2.13316754e-01 -8.09783041e-01 -9.94204506e-02 1.69074982e-01] | [9.601944923400879, -3.032437324523926] |
524b9d2c-b46e-48b1-96e2-8dec9b6d2d86 | icsde-an-indicator-for-constrained-multi | 2305.18734 | null | https://arxiv.org/abs/2305.18734v1 | https://arxiv.org/pdf/2305.18734v1.pdf | IcSDE+ -- An Indicator for Constrained Multi-Objective Optimization | The effectiveness of Constrained Multi-Objective Evolutionary Algorithms (CMOEAs) depends on their ability to reach the different feasible regions during evolution, by exploiting the information present in infeasible solutions, in addition to optimizing the several conflicting objectives. Over the years, researchers have proposed several CMOEAs to handle CMOPs. However, among the different CMOEAs proposed most of them are either decomposition-based or Pareto-based, with little focus on indicator-based CMOEAs. In literature, most indicator-based CMOEAs employ - a) traditional indicators used to solve unconstrained multi-objective problems to find the indicator values using objectives values and combine them with overall constraint violation to solve Constrained Multi-objective Optimization Problem (CMOP) as a single objective constraint problem, or b) consider each constraint or the overall constraint violation as objective(s) in addition to the actual objectives. In this paper, we propose an effective single-population indicator-based CMOEA referred to as IcSDE+ that can explore the different feasible regions in the search space. IcSDE+ is an (I)ndicator, that is an efficient fusion of constraint violation (c), shift-based density estimation (SDE) and sum of objectives (+). The performance of CMOEA with IcSDE+ is favorably compared against 9 state-of-the-art CMOEAs on 6 different benchmark suites with diverse characteristics | ['Sri Srinivasa Raju M', 'Rammohan Mallipeddi', 'Oladayo S. Ajani'] | 2023-05-30 | null | null | null | null | ['density-estimation'] | ['methodology'] | [ 8.75674114e-02 -4.99264449e-01 -1.15823954e-01 1.95461467e-01
-2.12443426e-01 -4.29102480e-01 -2.63051331e-01 8.20712969e-02
-3.78944039e-01 1.24335361e+00 -8.92439783e-02 -5.64251877e-02
-1.04181385e+00 -7.24891305e-01 -2.94222683e-01 -8.95394325e-01
-2.59546816e-01 6.69153392e-01 1.58726513e-01 -2.84943223e-01
6.63886189e-01 3.37405890e-01 -1.84702408e+00 -1.17566146e-01
1.57250321e+00 8.97032201e-01 2.00282842e-01 4.47201401e-01
-2.30638325e-01 -8.73388052e-02 -1.19759893e+00 -5.02223596e-02
1.03574939e-01 -4.00376111e-01 -3.43841255e-01 -8.07089284e-02
-7.47183979e-01 3.14978063e-01 5.28094947e-01 1.00612974e+00
7.05607295e-01 1.55123413e-01 6.99397326e-01 -1.66344452e+00
-5.91336548e-01 3.96847725e-01 -8.06756437e-01 3.31665188e-01
4.05460387e-01 -2.54114736e-02 6.14892840e-01 -7.63452947e-01
6.12726808e-01 1.15994430e+00 3.59472632e-01 2.55252749e-01
-9.68058705e-01 -5.26400685e-01 2.75021672e-01 3.37204456e-01
-1.73559296e+00 1.48865625e-01 6.56284034e-01 -7.90569559e-02
1.28281069e+00 8.08290124e-01 8.68704021e-01 3.90488893e-01
5.13062119e-01 5.27754784e-01 1.04619277e+00 -5.62376559e-01
6.60461724e-01 1.42523617e-01 4.02781693e-03 3.50707620e-01
9.02491689e-01 1.52494878e-01 -3.78144622e-01 -3.08441341e-01
-1.29675627e-01 -4.70321774e-01 -4.74992961e-01 -1.48783341e-01
-8.15304637e-01 8.00216675e-01 3.19870710e-02 3.53374749e-01
-4.86520231e-01 -2.54826486e-01 1.56948596e-01 2.51720045e-02
2.07875431e-01 8.95877004e-01 -4.85763222e-01 -2.78931350e-01
-8.81048918e-01 4.44243222e-01 8.28424335e-01 6.35766268e-01
3.17501694e-01 4.25842226e-01 -3.50527704e-01 5.39679229e-01
4.47614461e-01 4.13275063e-01 3.80283237e-01 -3.81628782e-01
5.45378149e-01 9.89488363e-01 1.78686544e-01 -1.09034908e+00
-3.40502828e-01 -8.70439708e-01 -4.77715164e-01 3.88874441e-01
-2.37662002e-01 -5.20997524e-01 -8.03926885e-01 1.52751327e+00
3.06265891e-01 -1.58445403e-01 9.78897735e-02 7.29408503e-01
3.13639522e-01 8.65670919e-01 -2.71614075e-01 -7.48510122e-01
7.66894460e-01 -8.07810128e-01 -9.18044090e-01 1.51662360e-04
3.20051938e-01 -6.79756463e-01 2.57866591e-01 8.15589905e-01
-1.22465789e+00 -2.10761994e-01 -1.39461279e+00 9.51002359e-01
-6.33317113e-01 -5.73360920e-03 4.96462137e-01 1.19087434e+00
-9.50759470e-01 1.62511721e-01 -4.05215889e-01 4.51874221e-03
5.22461981e-02 7.92951703e-01 1.68541685e-01 -7.00669736e-02
-9.60434258e-01 1.05113745e+00 8.96291971e-01 2.26209000e-01
-5.14128923e-01 -6.25790596e-01 -5.50821662e-01 5.33433139e-01
8.25582385e-01 -5.98215580e-01 1.93473130e-01 -8.51118922e-01
-1.44540310e+00 -1.71978474e-02 -1.80932388e-01 1.36773691e-01
3.17274243e-01 2.85754204e-01 -6.27005696e-01 -2.87902355e-01
-2.92467117e-01 2.66634732e-01 2.17458427e-01 -1.50514650e+00
-6.65087581e-01 -5.21920472e-02 -3.11024692e-02 2.94830918e-01
-4.72643167e-01 4.40144747e-01 -2.45054409e-01 -4.85977381e-01
-1.40401259e-01 -7.21786499e-01 -2.49062955e-01 -4.15862709e-01
-4.66803491e-01 3.33298743e-02 9.68742788e-01 -5.69255292e-01
2.22165561e+00 -1.77310014e+00 1.02638030e+00 6.40270293e-01
-3.40116620e-01 5.95229983e-01 -6.29152730e-02 5.60176969e-01
1.00369617e-01 5.02698779e-01 -7.66431808e-01 -1.59393623e-01
2.46227905e-01 2.93053746e-01 5.45990944e-01 1.77749231e-01
2.33276680e-01 5.77850521e-01 -6.76469564e-01 -4.47773606e-01
2.22751617e-01 2.47328967e-01 -6.88523889e-01 -1.11782365e-01
-2.85608768e-01 1.77428707e-01 -4.87173915e-01 1.12152612e+00
1.06858969e+00 3.13176483e-01 3.07369024e-01 1.20616384e-01
-5.17274141e-01 -6.72763109e-01 -1.72475445e+00 1.19472229e+00
-3.68341804e-01 2.49085262e-01 2.43556157e-01 -9.79262471e-01
9.21614945e-01 2.65811026e-01 8.32622290e-01 -4.04334843e-01
5.89404464e-01 6.36366248e-01 4.94881988e-01 -3.88564736e-01
4.25238997e-01 2.40003049e-01 -4.73748296e-02 3.37975860e-01
-5.66520952e-02 1.30436525e-01 8.99076343e-01 -2.75536060e-01
7.04408526e-01 1.55914441e-01 5.00530899e-01 -5.49850047e-01
9.51420665e-01 4.13873605e-02 1.26473665e+00 2.30963171e-01
-3.15815173e-02 4.99691844e-01 5.16363204e-01 6.32936209e-02
-6.62582695e-01 -6.58044159e-01 -2.38273934e-01 3.17660302e-01
4.29653317e-01 -1.79201439e-01 -4.39198375e-01 -3.75731736e-01
1.09617665e-01 1.03850698e+00 -3.10078532e-01 -2.59199291e-01
-5.38470805e-01 -1.52144814e+00 1.57067075e-01 1.13197856e-01
6.07800364e-01 -6.81558728e-01 -9.87105906e-01 5.65417886e-01
-1.27215823e-02 -4.55758095e-01 -2.01055974e-01 2.19254732e-01
-5.07593215e-01 -8.81493866e-01 -7.41826594e-01 -3.78742874e-01
7.88649142e-01 -3.11159283e-01 1.08201957e+00 3.08783591e-01
-6.52891278e-01 3.27047333e-02 -5.88221610e-01 -8.01031291e-01
-4.27242815e-02 -1.76426470e-01 -8.13882239e-03 6.52388781e-02
1.23195089e-01 -5.15949309e-01 -1.99611038e-01 6.03496671e-01
-8.41369629e-01 -5.38155019e-01 5.95507145e-01 8.95500183e-01
6.49254143e-01 6.57875776e-01 6.71884537e-01 -2.28179276e-01
9.96157944e-01 -8.05510104e-01 -9.25226748e-01 8.81045401e-01
-8.42773497e-01 6.79362863e-02 6.24975801e-01 -5.00250936e-01
-1.15071166e+00 -4.22313333e-01 2.81715821e-02 -5.26664197e-01
4.22621965e-01 9.43949938e-01 -4.74125177e-01 -2.25972593e-01
1.37794912e-01 2.14826405e-01 -4.79902267e-01 -2.63138354e-01
-3.92500639e-01 6.46379530e-01 6.16995618e-02 -9.95792806e-01
5.58862388e-01 -1.32101536e-01 3.68792325e-01 -1.56077206e-01
-1.21365704e-01 -3.46929938e-01 -1.59904450e-01 -4.71464843e-01
9.11929429e-01 -4.95331883e-02 -9.06728208e-01 2.95180887e-01
-1.03502333e+00 5.07328689e-01 4.97846007e-02 4.90001053e-01
7.81612247e-02 1.41147837e-01 3.04716140e-01 -1.46672857e+00
-3.39590997e-01 -1.45216858e+00 4.34063822e-01 7.66524613e-01
-9.11381245e-02 -8.64193022e-01 1.17159829e-01 -1.12932855e-02
5.87620497e-01 9.01970744e-01 1.01019776e+00 -1.76490232e-01
-3.75641674e-01 -1.34202108e-01 2.36290798e-01 6.45546913e-02
4.31045294e-02 5.31064034e-01 -2.04704151e-01 -5.10129571e-01
-5.27425036e-02 2.42089164e-02 2.44455919e-01 7.01734543e-01
9.03382957e-01 -1.95443019e-01 -5.85454702e-01 7.88086474e-01
1.97026682e+00 9.56061661e-01 6.11331463e-01 6.05705142e-01
-1.51787549e-01 3.94138634e-01 1.08440804e+00 9.00905848e-01
6.70464411e-02 7.83961892e-01 6.70317233e-01 1.84199199e-01
2.52526641e-01 5.37633121e-01 1.45489767e-01 7.70900488e-01
-4.09917712e-01 -1.21427190e+00 -8.49987745e-01 6.00064814e-01
-2.01124144e+00 -7.25535095e-01 -5.27214855e-02 2.13374281e+00
2.50254482e-01 1.94614679e-01 1.65758610e-01 3.63086909e-01
1.12827110e+00 -3.05723548e-01 -7.02729464e-01 -1.19645858e+00
-3.56139481e-01 3.28768134e-01 4.86853749e-01 3.28918815e-01
-5.49896121e-01 4.67545912e-02 5.65143633e+00 8.32311869e-01
-1.02176642e+00 -1.01968765e-01 3.97277147e-01 -4.88507807e-01
-5.57045519e-01 6.59945905e-02 -7.35415757e-01 1.00393772e+00
6.77967012e-01 -7.28327572e-01 5.20660281e-01 5.06550670e-01
8.82514566e-02 -5.24965644e-01 -5.57929873e-01 7.20804095e-01
1.91936776e-01 -1.10159218e+00 -2.48528004e-01 3.49956691e-01
1.34460473e+00 -6.36565208e-01 7.28078708e-02 1.35887966e-01
2.08500564e-01 -1.06708670e+00 7.81097472e-01 5.99056125e-01
6.01829290e-01 -1.35842764e+00 1.25720930e+00 2.44703427e-01
-1.30872047e+00 -6.24257505e-01 -2.48494521e-01 1.34283751e-01
6.36804402e-01 5.27929127e-01 -1.52841926e-01 1.55684912e+00
7.90548205e-01 3.14356647e-02 -2.19376668e-01 1.74232292e+00
1.79770926e-04 2.03062162e-01 -4.33517814e-01 -4.74431157e-01
3.42775762e-01 -4.85160768e-01 1.15322280e+00 6.91545486e-01
9.22610819e-01 -7.20334426e-02 5.09520620e-02 1.13595366e+00
5.32414258e-01 1.63908768e-02 1.71391234e-01 -1.16395235e-01
8.54350269e-01 9.26790118e-01 -7.59611964e-01 2.02229694e-01
-7.23805055e-02 4.95622218e-01 -3.10225904e-01 2.68273503e-01
-1.23665285e+00 -8.11245918e-01 6.02539480e-01 -5.55674970e-01
3.08640033e-01 -1.05656520e-01 -3.86429638e-01 -8.05907130e-01
1.71312630e-01 -8.62831950e-01 5.17861605e-01 -5.16199350e-01
-8.18715572e-01 6.23960078e-01 3.81648690e-01 -1.26813161e+00
-6.09718412e-02 -6.07302785e-01 -8.75154912e-01 1.22344244e+00
-1.48176718e+00 -5.62288523e-01 -1.66424155e-01 1.53925210e-01
4.90304619e-01 -3.67337137e-01 5.57140350e-01 4.74327415e-01
-1.09571373e+00 5.08513570e-01 3.74280721e-01 -9.38132584e-01
2.08508968e-01 -9.06183660e-01 -5.75884581e-01 1.11132574e+00
-6.19248807e-01 5.29704630e-01 8.82159233e-01 -7.48059392e-01
-1.49436653e+00 -3.91920894e-01 6.73738480e-01 1.05026484e-01
1.72189608e-01 -3.27572525e-02 -5.76439083e-01 -3.21341008e-02
2.66050130e-01 -2.51512378e-01 6.91551745e-01 -2.15876341e-01
6.20598197e-01 -6.78235218e-02 -1.62996244e+00 3.84161144e-01
9.49298739e-01 6.04174435e-01 -1.38487995e-01 -9.13836136e-02
4.30426180e-01 -4.23796773e-01 -8.91765594e-01 7.14550376e-01
4.69779223e-01 -8.81522298e-01 9.51390982e-01 -1.58584148e-01
2.92116046e-01 -6.70809984e-01 -2.37074584e-01 -1.59184396e+00
-4.73529756e-01 -4.16616797e-01 -2.99794555e-01 1.30363905e+00
4.57734942e-01 -9.83139992e-01 4.18704629e-01 5.74007273e-01
-5.04559398e-01 -1.31846189e+00 -1.20116401e+00 -1.14733338e+00
-2.10484087e-01 2.63303697e-01 1.18039393e+00 8.13019872e-01
-7.98125565e-02 -4.43816006e-01 -2.51556963e-01 1.98771000e-01
4.40640599e-01 2.14915629e-02 1.79659843e-01 -1.20224679e+00
-6.43568277e-01 -7.53267765e-01 -2.50060230e-01 1.43031284e-01
-1.81491345e-01 -6.00486994e-01 -2.66739726e-01 -1.62128115e+00
6.29811957e-02 -5.58777809e-01 -4.64870572e-01 9.67626274e-02
-3.11312735e-01 -2.22989723e-01 3.80926937e-01 -2.80705869e-01
-3.02755147e-01 6.99901342e-01 9.85422790e-01 -1.30479544e-01
-4.28796560e-01 -6.84071258e-02 -6.00502729e-01 2.87974715e-01
6.23317242e-01 -8.08308721e-01 -5.70605695e-01 -3.78403932e-01
5.06301820e-01 1.85267061e-01 -3.52685779e-01 -1.22770798e+00
2.26927385e-01 -7.01011598e-01 1.77349642e-01 -8.03428531e-01
2.82560289e-01 -1.01773977e+00 7.92877853e-01 8.29364657e-01
4.78677958e-01 5.44266462e-01 5.80209792e-01 3.54297996e-01
-4.85910624e-01 -7.56227255e-01 5.41708767e-01 -1.64961386e-02
-7.73806214e-01 -1.12062693e-01 -4.44838405e-01 -2.13704914e-01
1.64678228e+00 -8.06794584e-01 -4.00759816e-01 5.95222190e-02
-4.28053617e-01 6.41831636e-01 3.46598655e-01 1.03823200e-01
5.25824010e-01 -9.59833145e-01 -8.45229447e-01 -2.58686930e-01
-2.51622617e-01 -1.94818854e-01 2.86451131e-01 7.03904688e-01
-7.26001620e-01 5.93281329e-01 -5.70456803e-01 -3.69360656e-01
-1.35969746e+00 4.99749660e-01 2.98431277e-01 -6.75596833e-01
3.64424825e-01 1.03923500e+00 -5.56122065e-01 -2.32532561e-01
-4.58096005e-02 1.86788395e-01 -4.28855687e-01 1.62644625e-01
4.12106477e-02 9.99548793e-01 -7.10492313e-04 -3.85827214e-01
-8.37022960e-01 7.76472926e-01 6.23871207e-01 -9.72000733e-02
1.50563836e+00 -8.09088051e-02 -3.77984047e-01 -1.14288837e-01
8.82020831e-01 1.96568251e-01 -7.40919411e-01 5.77251673e-01
-1.74976364e-02 -8.23504567e-01 1.57447547e-01 -1.37784004e+00
-1.11397767e+00 1.89115420e-01 5.29173136e-01 -7.54411593e-02
1.91234565e+00 -8.41776609e-01 4.77209210e-01 -2.38636378e-02
5.88066876e-01 -1.32733727e+00 -3.21396142e-02 3.01987022e-01
9.11250830e-01 -7.61202931e-01 4.57465976e-01 -6.01896584e-01
-4.75038052e-01 1.23926151e+00 8.35517645e-01 1.39032826e-01
4.01353896e-01 4.71907943e-01 -5.75425804e-01 9.92969349e-02
-7.86823034e-01 -5.63899539e-02 4.37998414e-01 5.14366567e-01
4.08088900e-02 -1.92402333e-01 -1.22017479e+00 7.39962578e-01
1.20033450e-01 9.53297596e-03 4.12824929e-01 1.26921511e+00
-3.26195866e-01 -1.44606078e+00 -9.86067474e-01 3.50311041e-01
-1.98471338e-01 7.76488557e-02 -3.12120080e-01 8.21061850e-01
7.57097423e-01 1.32726836e+00 -1.43986210e-01 -2.88498819e-01
1.89093143e-01 -2.31555805e-01 4.59721297e-01 -2.70687699e-01
-9.27755594e-01 -2.02003211e-01 3.87323439e-01 -2.21250311e-01
-2.92523116e-01 -7.07300901e-01 -1.09596443e+00 -3.31827670e-01
-8.80936325e-01 3.23978305e-01 9.61169779e-01 8.44658136e-01
4.16549474e-01 1.08749938e+00 7.00842321e-01 -6.10230982e-01
-6.12248518e-02 -5.32680631e-01 -3.57144505e-01 -3.15437496e-01
-2.47009277e-01 -1.25066209e+00 -4.18507963e-01 -8.46294701e-01] | [5.709583759307861, 3.5072433948516846] |
549ef9d1-c5df-433c-82ab-647e3561eb82 | taploss-a-temporal-acoustic-parameter-loss | 2302.08088 | null | https://arxiv.org/abs/2302.08088v1 | https://arxiv.org/pdf/2302.08088v1.pdf | TAPLoss: A Temporal Acoustic Parameter Loss for Speech Enhancement | Speech enhancement models have greatly progressed in recent years, but still show limits in perceptual quality of their speech outputs. We propose an objective for perceptual quality based on temporal acoustic parameters. These are fundamental speech features that play an essential role in various applications, including speaker recognition and paralinguistic analysis. We provide a differentiable estimator for four categories of low-level acoustic descriptors involving: frequency-related parameters, energy or amplitude-related parameters, spectral balance parameters, and temporal features. Unlike prior work that looks at aggregated acoustic parameters or a few categories of acoustic parameters, our temporal acoustic parameter (TAP) loss enables auxiliary optimization and improvement of many fine-grain speech characteristics in enhancement workflows. We show that adding TAPLoss as an auxiliary objective in speech enhancement produces speech with improved perceptual quality and intelligibility. We use data from the Deep Noise Suppression 2020 Challenge to demonstrate that both time-domain models and time-frequency domain models can benefit from our method. | ['Bhiksha Raj', 'Shinji Watanabe', 'Anurag Kumar', 'Muqiao Yang', 'David Bick', 'Shuo Han', 'Joseph Konan', 'Yunyang Zeng'] | 2023-02-16 | null | null | null | null | ['speaker-recognition', 'speech-enhancement'] | ['speech', 'speech'] | [ 2.31856272e-01 -2.33660400e-01 3.19286793e-01 -4.66917336e-01
-1.30049253e+00 -4.70944256e-01 6.74103737e-01 2.92486280e-01
-4.80818272e-01 2.78553605e-01 8.24063659e-01 -2.98678786e-01
-4.17620361e-01 -1.66613489e-01 -1.69221789e-01 -7.45311260e-01
-4.31852341e-01 -4.82753813e-01 1.47044033e-01 -4.84066516e-01
2.79814396e-02 3.89547646e-01 -1.80999506e+00 5.94127225e-03
8.10057938e-01 1.14579546e+00 3.15503210e-01 1.06369603e+00
2.48516455e-01 1.67820811e-01 -9.35162187e-01 -8.35127980e-02
2.11473301e-01 -3.86171341e-01 -3.47583711e-01 -1.75443977e-01
3.35236311e-01 -2.39607394e-01 -4.03884441e-01 1.15618062e+00
1.10690200e+00 7.24850655e-01 3.41383457e-01 -1.03985119e+00
-2.41425663e-01 4.86508310e-01 -1.35624722e-01 4.90949810e-01
1.89754948e-01 4.43442672e-01 1.02650344e+00 -7.24311709e-01
6.14986122e-02 1.45528078e+00 8.22000206e-01 4.14689213e-01
-1.15256274e+00 -6.25180244e-01 1.34218195e-02 3.61716241e-01
-9.16163146e-01 -1.20931661e+00 5.62827587e-01 -1.37283668e-01
1.25010431e+00 3.69758397e-01 2.36248031e-01 8.61647189e-01
9.55174193e-02 4.51216847e-01 1.06993413e+00 -5.46246946e-01
1.61227077e-01 -1.02380015e-01 2.82138139e-01 1.13894589e-01
-6.61245108e-01 9.74386573e-01 -9.87973452e-01 1.73017792e-02
3.39805454e-01 -7.50035763e-01 -3.88849050e-01 3.12754035e-01
-9.59508896e-01 4.43652064e-01 -3.86902946e-03 2.34848633e-01
-1.72658652e-01 3.21938545e-01 5.01566887e-01 5.63356638e-01
6.39829695e-01 5.23995221e-01 -5.67996621e-01 -9.14189100e-01
-1.03486228e+00 7.02939481e-02 5.48293173e-01 6.26326859e-01
3.73750687e-01 4.96411860e-01 -6.33284748e-01 1.45906675e+00
2.75535107e-01 6.96275294e-01 3.54367256e-01 -1.32899785e+00
2.27689400e-01 -5.04823029e-01 1.46984905e-02 -6.39863968e-01
-5.97194195e-01 -7.36522138e-01 -5.96621811e-01 3.56154770e-01
2.17409402e-01 -1.05697833e-01 -1.00345123e+00 2.04348898e+00
2.06863478e-01 1.37066454e-01 -1.76795915e-01 9.00519788e-01
5.63694835e-01 6.27465546e-01 2.39831954e-01 -4.26499397e-01
1.46437347e+00 -8.94066989e-01 -1.29209602e+00 -2.13978946e-01
8.86997432e-02 -1.14968908e+00 1.15085900e+00 5.47032893e-01
-1.46514904e+00 -8.24646056e-01 -1.06163597e+00 -5.06612435e-02
-1.73896685e-01 -1.25478610e-01 3.15968066e-01 1.25016892e+00
-1.48041594e+00 8.75102460e-01 -8.14572692e-01 -2.43887752e-02
-1.45330831e-01 4.13264543e-01 -3.20647061e-02 4.53550905e-01
-1.27584696e+00 8.88021529e-01 -2.35997140e-01 -9.25322399e-02
-8.39764357e-01 -1.11002982e+00 -8.73300016e-01 3.69505137e-01
2.43069232e-02 -3.40439230e-01 1.74084663e+00 -2.26405293e-01
-2.00822306e+00 1.34855777e-01 -4.49008048e-01 -4.79222029e-01
-1.18818777e-02 -3.40351760e-01 -1.02833760e+00 1.81380615e-01
-3.06917220e-01 6.50985599e-01 1.26486385e+00 -8.58270109e-01
-6.57140613e-01 6.54241517e-02 -2.57077992e-01 3.59368682e-01
-8.06233466e-01 3.83345425e-01 -3.16997170e-01 -9.45658147e-01
6.39766306e-02 -5.96877754e-01 -3.68191823e-02 7.13367388e-02
-1.00519918e-01 1.01218730e-01 9.45198715e-01 -1.25600576e+00
1.30606461e+00 -2.45205665e+00 -1.16124839e-01 -1.70553438e-02
5.36314445e-03 4.76777911e-01 -4.74673778e-01 8.77868384e-02
-7.55664557e-02 2.77025372e-01 -1.86643645e-01 -7.66745865e-01
4.32696730e-01 -3.57188106e-01 -2.42258057e-01 1.61165804e-01
3.58199507e-01 4.75432783e-01 -8.29809010e-01 -1.16275311e-01
5.43791294e-01 9.13246930e-01 -8.40839386e-01 1.93166330e-01
1.91973627e-01 2.71605194e-01 2.00094089e-01 3.08989853e-01
6.90235376e-01 4.95338470e-01 -3.14158916e-01 -5.58332205e-01
-1.63532272e-01 8.72665882e-01 -9.90218341e-01 1.51689649e+00
-7.41102397e-01 9.63806331e-01 7.31294930e-01 -4.74601477e-01
6.99159324e-01 6.24898195e-01 4.07591045e-01 -9.23127949e-01
-1.21741347e-01 1.10295624e-01 3.67812276e-01 -1.62142828e-01
7.81928003e-01 -2.28955299e-01 3.83550853e-01 2.71936834e-01
3.15424860e-01 -7.56700277e-01 5.37677445e-02 1.56514775e-02
1.23328221e+00 -2.17845604e-01 -1.14406824e-01 -2.69069880e-01
3.67669374e-01 -7.24272072e-01 3.68688762e-01 6.23997808e-01
-7.33161867e-01 6.93766952e-01 1.15019806e-01 5.80757260e-01
-1.21823668e+00 -1.41533601e+00 -4.31846827e-01 1.44283485e+00
-2.60691464e-01 -5.89249909e-01 -7.62983263e-01 3.71087082e-02
-2.14299455e-01 7.74498403e-01 -1.32000566e-01 -4.23983663e-01
-4.67665821e-01 -5.02133548e-01 8.04050744e-01 4.35847431e-01
2.00268760e-01 -8.35970223e-01 -1.69208143e-02 3.57264161e-01
-1.69171482e-01 -1.19762933e+00 -1.03459954e+00 4.74082381e-01
-6.56035125e-01 -1.72820300e-01 -6.00245953e-01 -4.10221130e-01
-8.86218995e-02 1.82061389e-01 9.80560005e-01 -2.41211429e-01
-1.80841655e-01 6.34031594e-01 -3.25451910e-01 -4.76200432e-01
-8.37790787e-01 -1.56182602e-01 5.16966283e-01 -2.04858065e-01
-1.08115554e-01 -9.09221113e-01 -7.28038728e-01 4.39767540e-01
-7.43456960e-01 -3.82642835e-01 2.62212127e-01 8.18546891e-01
2.94941217e-01 1.69331074e-01 1.06911492e+00 1.21317506e-01
1.13728654e+00 7.71264825e-03 -3.91049385e-01 -2.05621123e-01
-6.34137273e-01 2.98099238e-02 4.25102741e-01 -5.00718892e-01
-1.35146332e+00 -4.92355019e-01 -6.79536402e-01 -4.06472981e-01
-6.63506687e-02 4.26513821e-01 -2.17070818e-01 5.98186627e-02
7.76297867e-01 5.89443557e-02 2.33548597e-01 -6.21187210e-01
3.32256824e-01 9.07856345e-01 7.96693563e-01 -5.54999828e-01
8.25527191e-01 1.17283478e-01 -1.41643062e-01 -1.29534531e+00
-4.73243952e-01 -8.57783794e-01 -1.17658034e-01 -6.49556890e-02
4.42555159e-01 -8.16871107e-01 -7.31793642e-01 4.97969121e-01
-9.77248013e-01 -4.27020997e-01 -4.90563989e-01 6.89946473e-01
-6.31073236e-01 4.66000736e-01 -7.96258271e-01 -1.33529294e+00
-4.31552738e-01 -1.27829599e+00 1.27347946e+00 1.04835302e-01
-2.11995453e-01 -9.50448394e-01 -1.26139984e-01 1.83003366e-01
1.08870065e+00 -5.57020724e-01 7.29297400e-01 -2.18480811e-01
-9.42906216e-02 2.04249322e-01 6.58664107e-02 9.16844964e-01
3.83731663e-01 -4.95065786e-02 -1.54536557e+00 -4.13855046e-01
3.42585593e-01 5.76232150e-02 8.57506931e-01 8.76476705e-01
1.05789590e+00 -1.89615844e-03 1.41745061e-01 6.90041125e-01
6.75825536e-01 4.47415888e-01 6.92813039e-01 -1.44472584e-01
1.00260049e-01 6.46380723e-01 5.63609123e-01 4.74829644e-01
-1.95024282e-01 9.36501622e-01 6.01919293e-02 -2.22821370e-01
-7.06580937e-01 8.00096169e-02 7.97199547e-01 1.15490103e+00
1.07208036e-01 -2.61245996e-01 -6.48027241e-01 6.96672380e-01
-1.24885571e+00 -1.06444299e+00 3.33212942e-01 2.22449183e+00
1.21163297e+00 1.09561890e-01 2.02938929e-01 5.03134549e-01
7.64716446e-01 3.37084711e-01 -5.16616821e-01 -6.29441500e-01
-2.13561356e-01 6.28199279e-01 3.13673943e-01 8.53404641e-01
-1.00446200e+00 6.54322803e-01 7.25230312e+00 1.15181589e+00
-9.77806509e-01 1.87607586e-01 4.38644558e-01 -3.63114476e-01
-3.43396008e-01 -3.19773048e-01 -6.44616961e-01 1.41146839e-01
1.55174863e+00 -3.15290928e-01 9.12001550e-01 3.54335278e-01
9.81439590e-01 1.15276359e-01 -9.29070532e-01 8.43297780e-01
-3.01988840e-01 -9.69415843e-01 -4.67429668e-01 7.11921304e-02
4.48883921e-01 1.55517263e-02 7.47472763e-01 3.71489644e-01
3.35064717e-02 -1.08075190e+00 8.95126581e-01 2.82328874e-01
9.53209281e-01 -7.07129598e-01 2.41939798e-01 -5.03259189e-02
-1.28686821e+00 -2.00807884e-01 -8.43020454e-02 2.25528106e-01
5.09492159e-01 8.39605212e-01 -8.66026163e-01 2.90626973e-01
8.03624451e-01 4.10272866e-01 -1.69911802e-01 1.23768413e+00
-8.70310515e-02 1.03919899e+00 -4.53432381e-01 2.10213721e-01
-1.61097690e-01 -2.06116643e-02 1.01599324e+00 1.42946553e+00
4.90367353e-01 1.40492633e-01 -4.49031293e-01 5.90664923e-01
1.38087362e-01 -1.60308376e-01 4.77535017e-02 -3.03066790e-01
7.16664553e-01 1.19920182e+00 -2.01615915e-01 1.90514848e-02
-7.75865912e-02 5.80902755e-01 -3.61971915e-01 6.31216824e-01
-7.99259663e-01 -5.85804880e-01 1.49711299e+00 -8.79143458e-03
2.45806366e-01 -5.06940842e-01 -1.55655891e-01 -5.37164390e-01
-5.02119996e-02 -1.33280087e+00 -1.50361627e-01 -7.30634451e-01
-1.02976298e+00 5.73260844e-01 -2.04389498e-01 -1.22267401e+00
-3.97002250e-01 -7.31942117e-01 -4.69356686e-01 1.23153734e+00
-1.44890904e+00 -6.83400035e-01 -3.84037849e-03 4.36060011e-01
6.51236176e-01 -2.89821401e-02 8.37203741e-01 5.82543314e-01
-2.12090105e-01 8.99455667e-01 1.96275264e-01 -2.68734306e-01
9.14732635e-01 -1.27746534e+00 6.77328169e-01 9.63994324e-01
-1.15892440e-01 6.55626714e-01 1.20233381e+00 -2.65227288e-01
-1.06697857e+00 -8.12173784e-01 7.13569999e-01 -1.20460667e-01
6.80115938e-01 -4.00329202e-01 -7.32092202e-01 2.08459403e-02
2.90822595e-01 -1.20965071e-01 4.79361743e-01 4.40806568e-01
-1.56145856e-01 -2.20345736e-01 -1.03495419e+00 6.43653929e-01
1.26720786e+00 -1.06481314e+00 -2.26158097e-01 2.12482393e-01
1.29118967e+00 -3.61776829e-01 -1.02941477e+00 4.28315014e-01
4.78388339e-01 -9.38088775e-01 1.20766985e+00 -3.29040319e-01
2.55719740e-02 -3.77013594e-01 -3.79053295e-01 -1.88657689e+00
-1.39281273e-01 -1.41796958e+00 -2.14312375e-01 1.41291165e+00
4.61745203e-01 -4.83038247e-01 2.43911967e-01 4.67905313e-01
-8.13677907e-01 -2.08976462e-01 -1.11237645e+00 -1.18548524e+00
8.76773894e-02 -1.06972122e+00 5.93103588e-01 2.83210605e-01
1.58726275e-01 1.83074757e-01 -2.57478625e-01 3.41808826e-01
5.49114645e-01 -6.26467109e-01 2.74855733e-01 -8.13815236e-01
-5.01355171e-01 -8.16610098e-01 -2.83860624e-01 -1.22520101e+00
-8.99880975e-02 -4.17046815e-01 4.74939972e-01 -1.15782881e+00
-4.23013091e-01 -4.67292219e-01 -5.74945033e-01 1.32623687e-01
-2.97067463e-01 -1.26756996e-01 1.24963857e-01 -4.21655595e-01
-1.57845348e-01 9.44310308e-01 1.15109599e+00 -3.15731078e-01
-4.16945398e-01 1.35156542e-01 -4.74589288e-01 3.83886665e-01
6.60610974e-01 -2.39855513e-01 -5.67667484e-01 -2.89412111e-01
-3.24222982e-01 2.25727156e-01 3.60070646e-01 -1.19724154e+00
3.10412914e-01 -6.42835721e-02 -9.35299695e-02 -3.82971615e-01
9.75252867e-01 -4.18902934e-01 -1.07996710e-01 1.12417608e-01
-5.71672082e-01 -9.84181240e-02 6.25049114e-01 4.62615848e-01
-4.67230767e-01 1.29046202e-01 9.14322436e-01 6.10668361e-01
-5.54030061e-01 1.16864271e-01 -4.66238469e-01 -4.01547477e-02
9.03941765e-02 -5.91897108e-02 -4.51786965e-01 -8.71941924e-01
-9.79056716e-01 -2.71499008e-01 -1.24260224e-01 5.73049843e-01
5.46948195e-01 -1.16127443e+00 -8.41256201e-01 2.00952798e-01
-2.36607805e-01 -7.61731744e-01 5.42722225e-01 9.22746360e-01
1.27460375e-01 4.36109155e-01 1.01485536e-01 -7.72141874e-01
-1.30941391e+00 1.55233741e-01 7.19551682e-01 -3.57405245e-02
-2.70262182e-01 9.93787408e-01 1.89993098e-01 -3.57475013e-01
6.23095810e-01 -4.78703856e-01 1.41277583e-02 -1.08493350e-01
6.32290781e-01 5.35395741e-01 6.67839885e-01 -5.29072881e-01
-2.94922382e-01 1.61523998e-01 2.85961509e-01 -9.40582037e-01
1.20345688e+00 -5.63649476e-01 3.29916775e-01 1.91859856e-01
1.28505301e+00 2.71314204e-01 -1.40186727e+00 -2.71709144e-01
-2.09389836e-01 -3.87600034e-01 6.88100398e-01 -1.09973383e+00
-8.01082015e-01 1.15809476e+00 1.11189425e+00 3.88473898e-01
1.54894829e+00 -3.18521738e-01 8.97420287e-01 -6.96018059e-03
6.08015507e-02 -1.38283432e+00 2.38424197e-01 8.40889394e-01
1.05104315e+00 -9.92735565e-01 -3.55433881e-01 -4.04807925e-01
-2.55849659e-01 8.93055201e-01 2.53896892e-01 7.08794296e-01
9.57036316e-01 8.25089931e-01 3.28559428e-01 2.64471889e-01
-8.42608213e-01 -5.10603011e-01 5.45620203e-01 9.95504439e-01
6.34640753e-01 6.49927929e-02 6.58767596e-02 4.92433369e-01
-7.36923277e-01 -6.53516769e-01 9.71807167e-02 4.30305809e-01
-6.42783701e-01 -1.24785221e+00 -4.77222055e-01 3.40039372e-01
-4.78370786e-01 -6.64992094e-01 2.25942716e-01 2.04064786e-01
-3.05496424e-01 1.71292889e+00 9.36958641e-02 -6.40253544e-01
5.46691298e-01 2.42574606e-03 2.50101507e-01 -4.63332802e-01
-7.20514476e-01 4.43495393e-01 4.72138643e-01 -4.33429539e-01
-1.47935286e-01 -7.70985365e-01 -1.07458150e+00 -3.76443624e-01
-4.57677692e-01 8.87139738e-02 1.14736545e+00 8.60513031e-01
3.04463625e-01 1.03667009e+00 8.27504516e-01 -1.03967202e+00
-8.47097516e-01 -1.19133282e+00 -6.02935135e-01 2.73486435e-01
8.34294617e-01 -5.00763416e-01 -8.22027206e-01 9.28941816e-02] | [15.014670372009277, 5.951858997344971] |
f43b4f50-fb02-4987-885e-897c464dddbc | synthesizer-rethinking-self-attention-in | 2005.00743 | null | https://arxiv.org/abs/2005.00743v3 | https://arxiv.org/pdf/2005.00743v3.pdf | Synthesizer: Rethinking Self-Attention in Transformer Models | The dot product self-attention is known to be central and indispensable to state-of-the-art Transformer models. But is it really required? This paper investigates the true importance and contribution of the dot product-based self-attention mechanism on the performance of Transformer models. Via extensive experiments, we find that (1) random alignment matrices surprisingly perform quite competitively and (2) learning attention weights from token-token (query-key) interactions is useful but not that important after all. To this end, we propose \textsc{Synthesizer}, a model that learns synthetic attention weights without token-token interactions. In our experiments, we first show that simple Synthesizers achieve highly competitive performance when compared against vanilla Transformer models across a range of tasks, including machine translation, language modeling, text generation and GLUE/SuperGLUE benchmarks. When composed with dot product attention, we find that Synthesizers consistently outperform Transformers. Moreover, we conduct additional comparisons of Synthesizers against Dynamic Convolutions, showing that simple Random Synthesizer is not only $60\%$ faster but also improves perplexity by a relative $3.5\%$. Finally, we show that simple factorized Synthesizers can outperform Linformers on encoding only tasks. | ['Che Zheng', 'Da-Cheng Juan', 'Zhe Zhao', 'Yi Tay', 'Donald Metzler', 'Dara Bahri'] | 2020-05-02 | null | null | null | null | ['linguistic-acceptability'] | ['natural-language-processing'] | [ 2.37282455e-01 6.83181807e-02 -6.44302294e-02 -2.49690846e-01
-1.06134462e+00 -6.88234448e-01 8.33973229e-01 -2.50200391e-01
-3.63023102e-01 6.25071287e-01 3.73224407e-01 -8.27736914e-01
2.34895751e-01 -7.18116999e-01 -1.14260507e+00 -4.43212450e-01
1.82842836e-01 7.18844056e-01 -1.58023059e-01 -5.33714771e-01
7.52656385e-02 2.74455816e-01 -1.13912237e+00 4.79459584e-01
1.03037989e+00 9.03773904e-01 3.40760797e-01 6.91738248e-01
-3.61141860e-01 9.25852001e-01 -6.74627304e-01 -8.80401492e-01
3.72638494e-01 -3.30397576e-01 -9.06871855e-01 -2.44668543e-01
6.05133832e-01 -4.57726091e-01 -3.97504807e-01 8.65830600e-01
5.06797314e-01 -9.27734822e-02 5.40530205e-01 -9.84548867e-01
-1.17172110e+00 1.40623355e+00 -3.62278670e-01 3.35718542e-01
-1.62026491e-02 5.51149070e-01 1.66354680e+00 -1.11128438e+00
5.57894826e-01 1.23173487e+00 6.48519099e-01 3.38228703e-01
-1.29725194e+00 -8.76414716e-01 2.75400817e-01 -9.08518881e-02
-1.21034992e+00 -7.88362503e-01 2.59195387e-01 -2.57381976e-01
1.81659770e+00 3.28517467e-01 3.87974471e-01 1.19574130e+00
2.82778502e-01 9.35272932e-01 8.28153014e-01 -3.28986287e-01
-3.56937110e-01 -7.64832795e-02 9.89997536e-02 8.92545104e-01
1.83222126e-02 -6.71476349e-02 -6.39464736e-01 1.89358667e-01
6.81242943e-01 -3.30535829e-01 -2.22911909e-01 2.22475275e-01
-1.57293105e+00 7.34540582e-01 3.96675408e-01 4.87015545e-01
-2.59439319e-01 8.08071375e-01 3.00707847e-01 5.83235741e-01
5.15823722e-01 8.70577931e-01 -6.01311445e-01 -3.50673825e-01
-9.92689967e-01 2.00364947e-01 5.40009260e-01 1.29090250e+00
5.78093410e-01 5.28396785e-01 -4.99403596e-01 7.37272680e-01
-2.20208168e-02 7.45814204e-01 6.23572767e-01 -7.68798590e-01
8.66156816e-01 2.53972769e-01 -8.33657533e-02 -5.14472246e-01
-1.85474679e-01 -8.41869533e-01 -7.73932874e-01 -3.90088499e-01
4.91043270e-01 -2.66444951e-01 -1.01043820e+00 1.90970373e+00
-2.50973761e-01 -1.23425379e-01 -4.94095795e-02 5.62184989e-01
7.35408604e-01 8.77493203e-01 -1.30366609e-01 -1.87566888e-03
1.22524130e+00 -1.20558870e+00 -7.15418696e-01 -4.35909986e-01
7.56981075e-01 -1.21458888e+00 1.26667380e+00 2.31714711e-01
-1.55591846e+00 -5.81412852e-01 -9.93490696e-01 -4.78869081e-01
-2.17031434e-01 2.26886734e-01 8.61360848e-01 6.32159531e-01
-1.26319456e+00 8.67844701e-01 -8.25355053e-01 -1.88422143e-01
1.73281491e-01 4.75644231e-01 -2.43915096e-01 3.44111659e-02
-1.21021223e+00 1.04252505e+00 -6.25271499e-02 4.45689335e-02
-8.47228408e-01 -1.04111028e+00 -7.84243345e-01 3.92434835e-01
7.15848580e-02 -8.55549634e-01 1.67576182e+00 -9.73477960e-01
-1.61083579e+00 4.96006101e-01 -4.27706510e-01 -8.36228073e-01
3.49812597e-01 -4.56957042e-01 -1.61223307e-01 -2.27278799e-01
3.47484350e-02 5.82866967e-01 6.76531911e-01 -7.36420155e-01
-5.24699748e-01 7.51551539e-02 3.60033363e-02 1.80691585e-01
-1.45047635e-01 2.06065178e-01 -4.13768500e-01 -7.84290373e-01
-1.83612749e-01 -7.24837959e-01 -3.59766968e-02 -5.23228765e-01
-6.26538217e-01 -2.82990813e-01 2.55323708e-01 -6.80841625e-01
1.30024469e+00 -1.92065597e+00 1.76391155e-01 2.49009598e-02
2.72543162e-01 1.89916372e-01 -3.93296301e-01 5.49900591e-01
-2.88426667e-01 3.77554536e-01 -3.45142223e-02 -5.59908688e-01
3.82905841e-01 7.58205028e-03 -6.67881727e-01 1.40738204e-01
5.52869856e-01 1.46940899e+00 -6.72375798e-01 -1.67236626e-01
-1.95674673e-02 2.65422910e-01 -7.53738463e-01 9.86106228e-03
-5.28528214e-01 -1.47913992e-01 -1.63849294e-01 5.47670364e-01
2.85104603e-01 -5.20848453e-01 2.10873529e-01 -3.13759059e-01
-2.19352797e-01 9.19926286e-01 -5.99580169e-01 1.62956429e+00
-8.31325889e-01 8.32054555e-01 -1.37954831e-01 -7.88011909e-01
6.61556840e-01 3.49153578e-01 2.23522231e-01 -9.54149723e-01
2.39076823e-01 4.46326226e-01 4.56807911e-01 -6.54667541e-02
8.93119752e-01 -1.33159399e-01 -4.45294008e-02 8.95731747e-01
4.44657564e-01 -2.62482941e-01 3.93201053e-01 4.69574451e-01
1.24313390e+00 1.38331249e-01 -9.45673361e-02 -3.46984893e-01
-2.57051103e-02 -1.02537960e-01 3.50640506e-01 8.23886871e-01
3.35470438e-01 6.63927078e-01 5.98358512e-01 -3.06170344e-01
-1.29296100e+00 -1.08536017e+00 3.57103467e-01 1.33439410e+00
-3.74185026e-01 -6.77234113e-01 -7.44256973e-01 -4.54920262e-01
-9.51675475e-02 1.00549674e+00 -5.17344117e-01 -7.88504928e-02
-8.16143453e-01 -9.27275240e-01 8.85300398e-01 7.27225184e-01
3.82499605e-01 -9.21746433e-01 -1.92644909e-01 4.44751829e-01
-4.15967047e-01 -1.18472481e+00 -8.78004193e-01 5.25742829e-01
-7.26261497e-01 -5.79984665e-01 -8.25834632e-01 -6.04248762e-01
3.67966443e-01 1.40305564e-01 1.57022583e+00 -7.48048676e-03
9.91398692e-02 -8.84899218e-03 -2.92401105e-01 -2.20123276e-01
-4.46499228e-01 6.23852491e-01 -1.28310680e-01 -1.29740953e-01
2.05274239e-01 -6.56625390e-01 -3.80060196e-01 1.15043826e-01
-6.32903159e-01 2.13516966e-01 9.58850920e-01 9.61480260e-01
2.60807931e-01 -4.06144381e-01 2.98370212e-01 -1.04209828e+00
8.45483303e-01 -2.08991155e-01 -4.48021144e-01 3.48469913e-01
-5.78621268e-01 4.19188440e-01 7.90598631e-01 -3.77796054e-01
-7.90546596e-01 -3.20609301e-01 -3.34743708e-01 -3.67878348e-01
4.84020680e-01 4.30623949e-01 -2.52182782e-02 2.89319515e-01
6.54596925e-01 4.41800773e-01 7.06961425e-03 -4.49627638e-01
5.61231792e-01 3.76977682e-01 2.76991814e-01 -9.34093058e-01
9.15860295e-01 1.49989739e-01 -3.94976318e-01 -5.24978518e-01
-6.63324475e-01 -3.25121842e-02 -4.61091399e-02 3.68646413e-01
7.11648464e-01 -1.06860280e+00 -5.05576611e-01 4.46020365e-01
-1.44579887e+00 -9.29956675e-01 -3.06462914e-01 3.24188173e-01
-6.12318337e-01 4.95778397e-02 -1.04036915e+00 -3.98724318e-01
-6.84047103e-01 -1.52828705e+00 1.01371956e+00 -1.51349589e-01
-3.45049500e-01 -7.53239274e-01 -3.16249311e-01 3.19334447e-01
8.42745662e-01 -3.82039279e-01 1.17310286e+00 -6.92647874e-01
-1.03578174e+00 2.69397020e-01 -4.29504037e-01 3.85360688e-01
4.49442044e-02 1.40203699e-01 -8.33079815e-01 7.38645671e-03
-4.09355342e-01 -1.11272313e-01 9.84880626e-01 2.32641295e-01
9.11412597e-01 -4.11514461e-01 -1.03956439e-01 8.40757489e-01
1.14017427e+00 1.33637309e-01 7.98931539e-01 -1.74288582e-02
7.15794265e-01 1.81597754e-01 7.56587088e-02 7.12084174e-02
4.25878793e-01 5.54620743e-01 1.84500158e-01 -1.78085834e-01
-4.90662456e-01 -3.58898014e-01 7.02288806e-01 1.25405598e+00
-4.67936285e-02 -4.38687921e-01 -8.28356206e-01 6.48284197e-01
-1.57201767e+00 -8.92656863e-01 -1.03349648e-01 1.88308775e+00
1.02792954e+00 3.34052414e-01 -1.21832043e-01 -8.09960440e-02
3.32378745e-01 1.31412223e-01 -3.28483373e-01 -8.09809327e-01
-2.99420476e-01 1.03562951e+00 7.68224418e-01 7.00442433e-01
-7.29264319e-01 1.45393348e+00 6.67747307e+00 1.02095377e+00
-1.26686943e+00 2.39594653e-01 6.66727364e-01 -2.89229184e-01
-7.39613116e-01 -2.17347994e-01 -7.47598827e-01 3.34907293e-01
1.09603941e+00 -2.45614395e-01 7.83710837e-01 6.31683946e-01
-5.26212491e-02 3.13676953e-01 -1.36549020e+00 7.36013889e-01
-7.28786364e-02 -1.52435899e+00 2.27068886e-01 2.79924963e-02
7.35321999e-01 3.25729609e-01 4.80224073e-01 5.56242466e-01
1.00921750e+00 -1.27319360e+00 9.77590203e-01 1.13709576e-01
9.76179898e-01 -6.40686512e-01 5.34204721e-01 -4.09148932e-02
-1.07466495e+00 9.09360871e-02 -2.48082995e-01 1.26303791e-03
2.13080600e-01 5.49402475e-01 -8.55949163e-01 5.45638144e-01
4.96902466e-01 4.94517773e-01 -4.23194438e-01 6.29410923e-01
-2.87570298e-01 7.16343701e-01 -2.70276904e-01 -1.07656844e-01
5.55751562e-01 -1.28260637e-02 2.27567032e-01 1.40207779e+00
6.92725599e-01 -1.08698815e-01 -4.67201799e-01 1.02861166e+00
-4.30011928e-01 9.32487175e-02 -5.19465983e-01 -5.37789643e-01
3.12108308e-01 9.84669626e-01 -4.95533973e-01 -4.77118611e-01
-3.45595777e-01 1.14256275e+00 3.72617006e-01 5.40186405e-01
-9.30935204e-01 -5.39272904e-01 9.75559592e-01 1.07773446e-01
6.07019126e-01 -4.49180633e-01 -3.09621930e-01 -1.26218998e+00
-7.64527619e-02 -1.34762633e+00 -1.71112552e-01 -9.37440038e-01
-1.04468417e+00 8.16427290e-01 -4.26442236e-01 -6.88378513e-01
-3.82675290e-01 -6.73091054e-01 -5.12420118e-01 1.20176375e+00
-1.38373566e+00 -1.16809368e+00 1.97792500e-01 4.12414074e-01
7.09834337e-01 -2.30286211e-01 7.28984773e-01 4.27843630e-01
-5.59994340e-01 1.03673720e+00 1.72002375e-01 4.04831707e-01
6.88095212e-01 -1.36098635e+00 1.36865842e+00 9.53976870e-01
4.65111941e-01 1.16068757e+00 6.27350926e-01 -4.19254482e-01
-1.70389950e+00 -9.92916465e-01 1.35988986e+00 -5.11951089e-01
9.77304935e-01 -6.55826330e-01 -5.11342227e-01 9.92734075e-01
6.61669552e-01 -2.79873043e-01 4.16471392e-01 2.39828885e-01
-7.63073742e-01 -1.62443116e-01 -6.55118883e-01 9.12179708e-01
1.21210587e+00 -6.01056755e-01 -4.14236844e-01 4.63977516e-01
1.37046897e+00 -4.97130811e-01 -7.30183065e-01 1.96238443e-01
5.82700074e-01 -8.78304362e-01 8.54843915e-01 -6.48394644e-01
7.93365836e-01 8.51820037e-02 -3.96358192e-01 -1.53142190e+00
-4.44762826e-01 -9.27178144e-01 8.17526877e-02 1.22533703e+00
9.73075449e-01 -7.15661764e-01 6.42021835e-01 3.62876266e-01
-5.18838227e-01 -7.10360944e-01 -7.24080324e-01 -8.50034177e-01
6.20880365e-01 -5.96448958e-01 9.11309183e-01 7.37524271e-01
-6.28867745e-02 6.22094393e-01 -2.76887000e-01 -2.30684966e-01
2.64305204e-01 8.52691680e-02 6.65857852e-01 -6.45298362e-01
-7.22082138e-01 -8.42580676e-01 1.81065276e-01 -1.44761848e+00
1.13454007e-01 -1.15960038e+00 -7.34181702e-02 -1.61456680e+00
-4.03133221e-02 -5.78945518e-01 -8.30915719e-02 4.34445411e-01
-1.63639262e-01 2.54724443e-01 4.10301715e-01 -8.00885484e-02
-2.77564079e-01 5.25160134e-01 1.35055816e+00 -3.19903255e-01
2.07091033e-01 -1.13959543e-01 -1.06282020e+00 1.01034544e-01
7.34790027e-01 -1.12359054e-01 -2.89150000e-01 -1.16997278e+00
5.25948644e-01 -1.42050534e-01 -4.80728876e-03 -6.79146588e-01
5.03843911e-02 4.37800698e-02 -2.03906139e-03 -2.99562335e-01
4.25242364e-01 -3.42607737e-01 -1.08747995e-02 2.79602408e-01
-3.93805772e-01 6.17288411e-01 2.55993366e-01 1.33767733e-02
3.24216858e-02 2.98890043e-02 3.60802233e-01 -4.31219548e-01
-1.75000057e-01 3.11442941e-01 -4.43309188e-01 2.60559410e-01
2.69327432e-01 1.56317353e-01 -4.69040573e-01 -5.03184140e-01
-3.93107057e-01 4.05555069e-02 2.58242726e-01 3.42988431e-01
2.76371866e-01 -1.25658131e+00 -7.53303945e-01 2.09194347e-01
-1.52414367e-01 -1.71388030e-01 -5.03564253e-02 8.17934573e-01
-5.68898976e-01 7.87429035e-01 1.26935452e-01 -2.55568683e-01
-8.72345090e-01 4.46165025e-01 3.05717975e-01 -5.18519044e-01
-2.66657501e-01 1.26394236e+00 2.57445306e-01 -5.51857471e-01
1.44124880e-01 -9.57605124e-01 6.31887138e-01 -1.04310282e-01
3.49219561e-01 6.86552450e-02 3.14940065e-01 -4.01892543e-01
-1.34469047e-01 3.79546016e-01 -3.78288150e-01 -3.66323262e-01
1.27615058e+00 2.72283465e-01 -1.40917018e-01 2.56547391e-01
1.13335299e+00 7.50959618e-04 -9.26929533e-01 -4.55804318e-01
-2.23614499e-01 -6.14224859e-02 -1.72241986e-01 -9.07278836e-01
-1.19732654e+00 1.22666776e+00 -1.31939724e-03 1.14345700e-01
8.39062870e-01 -2.45763808e-01 1.12115073e+00 3.85214955e-01
2.78906554e-01 -8.85384798e-01 7.06109628e-02 8.25287402e-01
8.74085367e-01 -9.05613661e-01 -3.15448850e-01 -2.58614361e-01
-5.68768680e-01 8.34236920e-01 4.51664180e-01 -6.56813011e-02
3.87649149e-01 6.84694171e-01 -9.66935828e-02 5.13689592e-02
-1.27889907e+00 -3.29584777e-01 9.69692692e-02 3.59074891e-01
8.44305158e-01 1.21430475e-02 -1.16009779e-01 5.57530165e-01
-6.91519916e-01 -2.32440785e-01 3.08555186e-01 5.02458036e-01
-2.63713777e-01 -1.37996769e+00 -1.53630242e-01 5.10617256e-01
-7.25765824e-01 -8.76022756e-01 -3.81807894e-01 8.37341785e-01
-9.21948552e-02 7.83641517e-01 2.52716929e-01 -5.12160838e-01
1.68198809e-01 3.00716370e-01 7.55021870e-01 -5.78872383e-01
-1.28155684e+00 3.00570149e-02 2.81772256e-01 -4.72345322e-01
1.77510917e-01 -4.71063465e-01 -9.46048141e-01 -7.42280483e-01
-2.62944877e-01 1.08879276e-01 6.78210080e-01 8.53254557e-01
4.60203946e-01 7.48679698e-01 2.83488244e-01 -5.02911031e-01
-7.87916183e-01 -1.19887614e+00 -7.12295920e-02 -2.99959797e-02
2.84842014e-01 -2.08042666e-01 -2.10664093e-01 6.37648348e-03] | [10.875831604003906, 7.643625736236572] |
489d5bdc-e886-424e-bbef-3b20330b7eff | aligning-bag-of-regions-for-open-vocabulary | 2302.13996 | null | https://arxiv.org/abs/2302.13996v1 | https://arxiv.org/pdf/2302.13996v1.pdf | Aligning Bag of Regions for Open-Vocabulary Object Detection | Pre-trained vision-language models (VLMs) learn to align vision and language representations on large-scale datasets, where each image-text pair usually contains a bag of semantic concepts. However, existing open-vocabulary object detectors only align region embeddings individually with the corresponding features extracted from the VLMs. Such a design leaves the compositional structure of semantic concepts in a scene under-exploited, although the structure may be implicitly learned by the VLMs. In this work, we propose to align the embedding of bag of regions beyond individual regions. The proposed method groups contextually interrelated regions as a bag. The embeddings of regions in a bag are treated as embeddings of words in a sentence, and they are sent to the text encoder of a VLM to obtain the bag-of-regions embedding, which is learned to be aligned to the corresponding features extracted by a frozen VLM. Applied to the commonly used Faster R-CNN, our approach surpasses the previous best results by 4.6 box AP50 and 2.8 mask AP on novel categories of open-vocabulary COCO and LVIS benchmarks, respectively. Code and models are available at https://github.com/wusize/ovdet. | ['Chen Change Loy', 'Wentao Liu', 'Sheng Jin', 'Wenwei Zhang', 'Size Wu'] | 2023-02-27 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Wu_Aligning_Bag_of_Regions_for_Open-Vocabulary_Object_Detection_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Wu_Aligning_Bag_of_Regions_for_Open-Vocabulary_Object_Detection_CVPR_2023_paper.pdf | cvpr-2023-1 | ['open-vocabulary-object-detection'] | ['computer-vision'] | [-8.59761983e-02 -4.03070226e-02 -3.56502324e-01 -4.66129154e-01
-6.37824237e-01 -5.59054613e-01 7.82293499e-01 1.95915222e-01
-6.24264896e-01 2.01483607e-01 3.86272311e-01 2.38659792e-02
5.45242012e-01 -6.28250897e-01 -9.86944258e-01 -7.25240469e-01
3.52812558e-01 8.65960121e-02 3.90046686e-01 1.77423973e-02
2.88454026e-01 3.43172938e-01 -1.36676538e+00 4.66587186e-01
3.19784880e-01 1.10586047e+00 5.62468231e-01 4.20778185e-01
-5.01320064e-01 6.43346310e-01 -2.80880928e-01 -1.14506885e-01
2.57607639e-01 -1.75970480e-01 -4.78977799e-01 2.00578898e-01
1.05475438e+00 -3.08932126e-01 -7.07650542e-01 1.01916957e+00
1.86372876e-01 8.52199569e-02 8.02452028e-01 -1.03543377e+00
-1.29945755e+00 3.52337688e-01 -6.83118761e-01 2.77947813e-01
-9.13610756e-02 6.49578124e-02 1.26085591e+00 -1.36186743e+00
6.21142626e-01 1.15144122e+00 3.02086949e-01 5.41583955e-01
-1.19014978e+00 -5.76661289e-01 5.80626726e-01 2.60385573e-01
-1.78398108e+00 -5.02029717e-01 6.43793285e-01 -6.22491658e-01
1.20671010e+00 -1.99666128e-01 8.74410510e-01 1.00215912e+00
3.69313419e-01 7.58546412e-01 7.54867494e-01 -3.74726146e-01
1.32568985e-01 4.44351971e-01 3.28810304e-01 7.91450858e-01
2.73397863e-01 -1.34403363e-01 -5.64567029e-01 1.92217559e-01
7.16076672e-01 4.80572462e-01 -1.43346116e-01 -8.43214095e-01
-1.27253640e+00 1.09384906e+00 1.02491236e+00 3.58267546e-01
-3.21169555e-01 3.60841990e-01 3.41598213e-01 -8.33004564e-02
6.06421411e-01 -3.85855921e-02 -2.82138288e-01 6.01887882e-01
-8.17563474e-01 1.73369750e-01 2.57478833e-01 1.23485541e+00
1.08450389e+00 -1.38473362e-01 -3.76713604e-01 8.46841693e-01
7.48247862e-01 6.14433706e-01 5.62369943e-01 -4.20425177e-01
5.23241520e-01 7.29129553e-01 -7.23864362e-02 -9.01372552e-01
-1.19035400e-01 -3.33034605e-01 -7.05033779e-01 -2.47777194e-01
2.45817214e-01 4.39532727e-01 -1.06488895e+00 1.55647981e+00
2.47788519e-01 1.82395339e-01 1.33635521e-01 8.51460755e-01
1.01968896e+00 8.37499440e-01 1.21524610e-01 1.87827513e-01
1.66765404e+00 -1.53377104e+00 -5.63332200e-01 -8.23392630e-01
6.50107503e-01 -6.47035003e-01 1.00763309e+00 -2.31155306e-01
-7.76072323e-01 -8.25878441e-01 -9.51101899e-01 -5.65225244e-01
-7.45800614e-01 2.38748103e-01 1.10901035e-01 1.53368324e-01
-1.18585610e+00 -3.23486552e-02 -4.28485304e-01 -4.49316680e-01
8.06379318e-01 6.85463697e-02 -5.55334330e-01 -3.26042086e-01
-8.77405226e-01 8.73606622e-01 4.51789290e-01 -3.43302414e-02
-1.00341284e+00 -5.63658774e-01 -1.21495330e+00 -9.81291533e-02
2.29713187e-01 -5.17852008e-01 8.76636446e-01 -1.01927078e+00
-7.96234727e-01 1.28288913e+00 -5.28243840e-01 -4.50901687e-01
4.50524931e-05 -2.98449844e-01 -1.58869684e-01 3.21240962e-01
3.68545115e-01 1.27525866e+00 1.12616992e+00 -1.34417999e+00
-5.62614501e-01 -5.91791511e-01 1.14887767e-01 1.65678546e-01
-4.12218302e-01 1.26997337e-01 -7.11810350e-01 -6.03886247e-01
1.22678563e-01 -8.37284684e-01 -1.89054042e-01 4.51141983e-01
-1.95376173e-01 -3.07942450e-01 7.10723579e-01 -6.59655750e-01
9.03409898e-01 -2.40410972e+00 2.33945206e-01 -3.00062180e-01
3.74679565e-01 9.04792994e-02 -4.30963218e-01 2.07747445e-01
-8.87570083e-02 -1.21494569e-01 -9.43609774e-02 -5.70893705e-01
-2.12109327e-01 2.59623051e-01 -6.81068003e-01 8.51842046e-01
3.85328412e-01 1.27509522e+00 -8.01194131e-01 -7.33813763e-01
4.20794636e-01 4.83989209e-01 -5.58800101e-01 3.37995440e-01
-2.66910672e-01 2.14967161e-01 -3.95845354e-01 4.38316554e-01
5.74839771e-01 -2.97316939e-01 -2.26380333e-01 -3.02426755e-01
-1.57431915e-01 1.79934368e-01 -7.39475489e-01 1.95220280e+00
-5.47776759e-01 8.33600521e-01 -1.64519250e-01 -1.18467915e+00
1.02744710e+00 1.08486481e-01 2.95128584e-01 -5.83764613e-01
4.46483456e-02 -1.11854693e-03 -1.83633968e-01 -4.39436972e-01
5.55430830e-01 5.59642613e-02 -1.11277767e-01 1.83115721e-01
4.30749267e-01 -1.14901237e-01 -1.30677193e-01 3.12644809e-01
6.52498364e-01 2.66745202e-02 5.76045394e-01 -2.27624267e-01
5.02836406e-01 -1.82882085e-01 1.96169555e-01 6.48507357e-01
-3.37763280e-01 7.69350648e-01 1.52097523e-01 -4.38542992e-01
-1.23657334e+00 -1.30819225e+00 -3.19336087e-01 1.11724877e+00
2.79947221e-01 -4.97570097e-01 -4.67872947e-01 -7.35582411e-01
1.42393827e-01 6.78494930e-01 -7.86623418e-01 -2.34462410e-01
-3.17091495e-01 -1.59430504e-01 2.67749816e-01 7.41618812e-01
1.96850061e-01 -1.08490086e+00 -4.75640208e-01 4.08592150e-02
-1.49203977e-02 -1.43417645e+00 -7.50911832e-01 2.14552909e-01
-6.84257925e-01 -7.90137053e-01 -6.47390485e-01 -1.23566628e+00
9.35442209e-01 7.67046571e-01 9.43526030e-01 -2.24644113e-02
-3.53992194e-01 4.77216840e-01 -4.36396241e-01 -4.28872406e-01
-5.36056124e-02 -1.62546605e-01 -4.72104028e-02 2.93166518e-01
7.68012822e-01 -1.14660136e-01 -7.26342976e-01 9.91061851e-02
-8.09452891e-01 1.45088091e-01 5.03965676e-01 9.31563199e-01
9.29454386e-01 -7.34161019e-01 4.42183286e-01 -4.02924508e-01
1.09764459e-02 -5.20808995e-01 -5.24954796e-01 2.47614846e-01
-2.57128298e-01 1.04960287e-03 5.84179223e-01 -5.17333865e-01
-6.00794315e-01 1.84287623e-01 1.41889393e-01 -7.84375787e-01
-4.14455205e-01 5.44790477e-02 -1.41813323e-01 7.83259869e-02
3.18262130e-01 5.63886106e-01 -2.60862336e-02 -3.79265726e-01
9.54166532e-01 8.82408321e-01 1.92474738e-01 -2.44997159e-01
7.27770567e-01 7.78660357e-01 -4.89290684e-01 -9.43538547e-01
-1.12466311e+00 -9.33019042e-01 -9.57808256e-01 -1.83658674e-02
1.31934702e+00 -1.39167392e+00 2.97762062e-02 1.84669957e-01
-1.47877717e+00 -8.89112130e-02 -4.08013403e-01 2.89157182e-01
-4.63605225e-01 2.96582192e-01 -3.40557665e-01 -4.69417274e-01
-1.92504704e-01 -1.19042301e+00 1.42728472e+00 2.73154706e-01
-3.55587937e-02 -9.30453539e-01 -2.64890622e-02 4.55335498e-01
1.61079556e-01 -3.84424061e-01 6.02345288e-01 -7.84397066e-01
-6.12640798e-01 -2.45907426e-01 -5.93805492e-01 5.38954258e-01
1.49493128e-01 -2.23007575e-01 -1.24924898e+00 -3.66484076e-01
-9.65532288e-03 -4.58686978e-01 1.29404986e+00 4.22559559e-01
1.39481270e+00 -2.12605447e-01 -4.26350057e-01 6.55938208e-01
1.58243990e+00 -1.02537699e-01 4.53621268e-01 2.08255604e-01
9.94657695e-01 6.81806087e-01 4.41933066e-01 2.09129810e-01
4.11315233e-01 7.05298781e-01 4.93765324e-01 -1.42173067e-01
-2.04503000e-01 -3.79285693e-01 6.30684078e-01 8.58368158e-01
4.52816039e-01 -1.26056686e-01 -6.85164213e-01 9.23841238e-01
-1.73011971e+00 -7.78815508e-01 1.28991619e-01 1.93004930e+00
8.07270586e-01 -1.23945801e-02 -2.36352637e-01 -6.18366063e-01
6.27590299e-01 6.89344049e-01 -4.86648202e-01 -3.35738391e-01
-1.84361383e-01 4.47629802e-02 5.29822171e-01 3.74197900e-01
-1.30856192e+00 1.30432737e+00 5.13032389e+00 7.73749650e-01
-1.06068003e+00 4.34489131e-01 5.34629643e-01 -3.80010396e-01
-2.37420142e-01 7.41340518e-02 -1.18453217e+00 1.86676636e-01
8.00528288e-01 4.47467938e-02 2.71105915e-01 9.46939528e-01
1.25233725e-01 1.27590075e-02 -1.37907839e+00 9.99043703e-01
5.98698318e-01 -1.39749479e+00 3.34354758e-01 1.68412268e-01
7.50358164e-01 4.42267329e-01 9.89969522e-02 3.70210886e-01
-1.39052674e-01 -1.12461245e+00 9.33586776e-01 4.43426967e-01
6.94391429e-01 -1.99625850e-01 5.54095984e-01 1.66397408e-01
-1.38289189e+00 -7.16354325e-02 -8.40078413e-01 5.72373904e-02
5.32543799e-03 2.18169808e-01 -6.09463573e-01 3.22057098e-01
7.58979261e-01 1.18485117e+00 -7.73167372e-01 6.42771006e-01
-1.74230069e-01 3.35917473e-01 3.39321382e-02 -2.82817148e-02
5.06331027e-01 -2.43389741e-01 2.76116580e-01 1.17879844e+00
1.67686362e-02 -3.19715023e-01 3.22996140e-01 1.19121408e+00
-1.82249695e-01 3.47728103e-01 -1.01873362e+00 -4.09140624e-02
3.12246382e-01 1.34702277e+00 -5.20949543e-01 -5.88254511e-01
-1.05022383e+00 9.46946323e-01 6.82882071e-01 3.79917502e-01
-8.75348508e-01 7.44994059e-02 9.14757192e-01 1.12203635e-01
7.80567527e-01 -2.56783158e-01 -8.39357600e-02 -1.28274941e+00
8.47545043e-02 -4.60274696e-01 1.59380317e-01 -9.18572724e-01
-1.45159972e+00 5.13615370e-01 -9.33416421e-04 -1.24945760e+00
1.30678028e-01 -9.01455224e-01 -5.07041693e-01 9.18017507e-01
-1.67339838e+00 -1.41410792e+00 -2.98692018e-01 5.69950163e-01
1.00773740e+00 -3.07912260e-01 7.75441468e-01 2.87104095e-03
-4.66402829e-01 4.98977542e-01 2.65910029e-01 4.40950304e-01
7.80812800e-01 -9.70071316e-01 4.20201421e-01 7.28499591e-01
6.71788633e-01 7.54926443e-01 3.35575938e-01 -4.38426942e-01
-1.19042361e+00 -1.54168141e+00 9.46072876e-01 -5.87211370e-01
9.49391663e-01 -8.05783689e-01 -1.09877765e+00 8.84111762e-01
4.73536402e-01 6.24307096e-01 5.15372634e-01 -1.97794423e-01
-8.15432668e-01 -1.99048117e-01 -6.44728184e-01 7.05654919e-01
9.86438692e-01 -1.00386691e+00 -9.00550783e-01 4.51861858e-01
9.97605085e-01 -1.16071723e-01 -6.45832956e-01 -8.97089466e-02
3.95138025e-01 -5.79268217e-01 1.25231743e+00 -6.80185437e-01
5.20369887e-01 -3.86903226e-01 -5.21423340e-01 -1.17156351e+00
-3.57277840e-01 3.32520694e-01 -9.37204435e-02 1.04439414e+00
1.21834002e-01 -6.35441661e-01 2.53348172e-01 3.08858305e-02
-7.71241859e-02 -8.64347100e-01 -9.98999774e-01 -6.81097806e-01
2.62149394e-01 -3.27898085e-01 1.98648155e-01 7.88363576e-01
-3.45007777e-01 5.23024142e-01 -5.24373278e-02 2.61831939e-01
6.41305387e-01 1.99760318e-01 6.00316823e-01 -7.38673389e-01
-4.83914427e-02 -4.82799619e-01 -5.81652641e-01 -1.29844677e+00
5.38832068e-01 -1.26490974e+00 1.29439890e-01 -1.55009758e+00
4.95048642e-01 -1.96172878e-01 -4.89983410e-01 5.73462069e-01
-6.84392303e-02 4.04874384e-01 3.66963893e-01 2.11780623e-01
-7.98639357e-01 8.97050858e-01 1.10615885e+00 -5.75912058e-01
1.26928270e-01 -5.09335637e-01 -5.54202139e-01 7.63514280e-01
8.35239410e-01 -4.53538299e-01 -3.52072537e-01 -6.02902651e-01
-1.37040734e-01 -5.00911951e-01 8.47892344e-01 -6.81824088e-01
1.28642902e-01 -1.03983805e-01 4.92681861e-01 -6.80533290e-01
4.53083456e-01 -8.56092215e-01 -4.46714908e-01 2.75134414e-01
-5.15489221e-01 -1.10650629e-01 2.75456727e-01 7.27624655e-01
-3.02378207e-01 -3.02514941e-01 8.20187867e-01 -1.19300701e-01
-1.00415742e+00 4.32270855e-01 -1.38891205e-01 8.91190208e-03
1.21350789e+00 -2.29084864e-01 -4.14044172e-01 -8.34915563e-02
-4.87960130e-01 1.96001723e-01 5.94002604e-01 9.24765706e-01
1.03562713e+00 -1.35974813e+00 -5.72959542e-01 4.45290059e-01
7.60714352e-01 2.67677546e-01 3.62201035e-01 6.93045020e-01
-2.62086272e-01 6.53346360e-01 -1.79078683e-01 -9.15173411e-01
-1.25021100e+00 9.77082551e-01 4.13064361e-01 6.35754243e-02
-9.60775554e-01 9.67369735e-01 1.04214561e+00 -2.74508625e-01
2.96320409e-01 -5.26314020e-01 -3.59841883e-01 1.71772018e-01
6.41189456e-01 -4.74540383e-01 -3.49328756e-01 -1.13494134e+00
-4.84746635e-01 8.51538599e-01 -3.49185735e-01 1.89199090e-01
1.31818473e+00 -2.83937365e-01 -3.09056640e-01 7.39255130e-01
1.56620824e+00 -1.62490323e-01 -1.37436187e+00 -7.64723897e-01
-1.99395016e-01 -3.62366527e-01 1.04099318e-01 -1.33009151e-01
-1.15659690e+00 1.11771011e+00 5.94489753e-01 -3.58949870e-01
7.86535442e-01 7.41985500e-01 5.91222525e-01 1.49427384e-01
2.51511067e-01 -8.90899837e-01 3.28059524e-01 4.50268716e-01
1.12537479e+00 -1.44267356e+00 -4.69896644e-02 -1.86782703e-01
-7.22534060e-01 9.22805369e-01 7.66277254e-01 -5.42902172e-01
7.45374978e-01 -1.49314389e-01 -5.11709377e-02 -1.17500462e-01
-7.66740441e-01 -2.27717131e-01 5.20241141e-01 4.91542041e-01
2.93700844e-01 1.86164945e-01 9.17772651e-02 5.33488095e-01
1.59288526e-01 -3.81399691e-01 3.34889293e-01 6.52131021e-01
-6.08731270e-01 -6.87752008e-01 -3.50875437e-01 4.60210353e-01
-5.38388640e-02 -3.15155268e-01 -2.03439027e-01 5.84759414e-01
2.46755227e-01 6.43822253e-01 6.03414536e-01 -1.95319772e-01
1.20653898e-01 1.11734703e-01 4.02656585e-01 -9.94736016e-01
-2.89265454e-01 2.96070781e-02 -2.80628771e-01 -6.45407200e-01
-3.83244008e-01 -6.09511018e-01 -1.23794937e+00 3.45230907e-01
-1.27379313e-01 -2.37744287e-01 5.65720916e-01 9.86622989e-01
2.60019451e-01 4.86228108e-01 4.50658947e-01 -7.66418338e-01
-5.26112616e-01 -1.01783395e+00 -5.70027053e-01 5.31566083e-01
5.36720514e-01 -7.57546902e-01 -2.72573799e-01 2.40122363e-01] | [9.982748031616211, 1.6044068336486816] |
44036254-2e97-494b-95ad-bc1fe5f24ce6 | patternrank-leveraging-pretrained-language | 2210.05245 | null | https://arxiv.org/abs/2210.05245v2 | https://arxiv.org/pdf/2210.05245v2.pdf | PatternRank: Leveraging Pretrained Language Models and Part of Speech for Unsupervised Keyphrase Extraction | Keyphrase extraction is the process of automatically selecting a small set of most relevant phrases from a given text. Supervised keyphrase extraction approaches need large amounts of labeled training data and perform poorly outside the domain of the training data. In this paper, we present PatternRank, which leverages pretrained language models and part-of-speech for unsupervised keyphrase extraction from single documents. Our experiments show PatternRank achieves higher precision, recall and F1-scores than previous state-of-the-art approaches. In addition, we present the KeyphraseVectorizers package, which allows easy modification of part-of-speech patterns for candidate keyphrase selection, and hence adaptation of our approach to any domain. | ['Florian Matthes', 'Simon Klimek', 'Tim Schopf'] | 2022-10-11 | null | null | null | null | ['keyphrase-extraction'] | ['natural-language-processing'] | [ 1.65592298e-01 1.46760372e-02 -6.75725937e-01 7.22026676e-02
-1.27536297e+00 -1.18699634e+00 1.05570173e+00 8.70242953e-01
-7.92836666e-01 7.08751440e-01 5.45845747e-01 -3.58953476e-01
-2.52181720e-02 -8.18832099e-01 -5.11430383e-01 -2.42689893e-01
-1.35964364e-01 5.86370826e-01 5.91902256e-01 -1.37173429e-01
5.63505173e-01 5.53111732e-01 -1.39446926e+00 5.61596811e-01
2.83968925e-01 7.61130869e-01 6.11262321e-02 9.48963583e-01
-6.07266963e-01 9.14731503e-01 -5.90035379e-01 -2.20154583e-01
1.71324298e-01 -1.60651609e-01 -9.30381060e-01 -9.15133581e-02
6.21445060e-01 -3.29376429e-01 -4.40954298e-01 9.14384365e-01
1.67049929e-01 -4.80458513e-02 6.06532395e-01 -1.08592963e+00
-2.53749996e-01 1.05943072e+00 -4.15481031e-01 5.37487805e-01
6.19005620e-01 -4.22481447e-01 1.76946819e+00 -1.41291845e+00
7.66318083e-01 9.22903955e-01 2.46452987e-01 6.41913190e-02
-1.16251957e+00 -7.57791698e-01 1.24994002e-01 -2.13375643e-01
-1.40116966e+00 -2.82748133e-01 6.76512599e-01 -2.23929226e-01
1.26389349e+00 4.11221981e-01 5.79112172e-01 8.19314182e-01
6.82992190e-02 1.44238138e+00 1.00662684e+00 -8.09938908e-01
9.90015455e-03 2.87488818e-01 6.06050491e-01 5.15977561e-01
4.67674971e-01 -1.36726737e-01 -8.57910573e-01 -8.26805353e-01
3.66182268e-01 -1.44759834e-01 2.59969123e-02 -1.31456971e-01
-1.47048831e+00 7.33682454e-01 -4.85569209e-01 3.86590332e-01
-4.64096069e-01 1.08906941e-03 3.67164075e-01 3.57330561e-01
5.19560575e-01 1.01814663e+00 -1.00687253e+00 -1.57373130e-01
-1.37643635e+00 7.38608837e-01 1.19005084e+00 1.17169666e+00
8.39213073e-01 -5.93922853e-01 -4.78076935e-01 6.40466571e-01
6.81975260e-02 6.34029508e-01 3.29820663e-01 -4.80677754e-01
6.43620253e-01 7.34118521e-01 2.59377688e-01 -8.96714032e-01
-2.97364473e-01 -1.48144722e-01 -1.17373489e-01 -7.07708716e-01
7.38773942e-02 -2.96431035e-01 -1.00098455e+00 1.09366953e+00
2.41408572e-01 -3.58632088e-01 -2.03446597e-02 1.65691331e-01
8.64553928e-01 8.50828350e-01 5.54600693e-02 -4.23201412e-01
1.54539013e+00 -8.07939172e-01 -7.39857435e-01 -4.14857328e-01
3.73872876e-01 -1.18884122e+00 8.37543666e-01 5.13518810e-01
-8.75013530e-01 -4.63060886e-02 -7.14444101e-01 -1.14324316e-01
-7.72515357e-01 3.43968034e-01 6.50409222e-01 2.73717970e-01
-6.33187950e-01 4.23965007e-01 -5.54814816e-01 -3.72588515e-01
1.20905250e-01 2.90609807e-01 -2.80665606e-01 2.71063358e-01
-1.22642958e+00 5.47191262e-01 9.08711076e-01 -8.03385198e-01
-6.41758084e-01 -8.77071261e-01 -6.39534891e-01 -5.59885018e-02
1.17969656e+00 -2.86190748e-01 1.72555161e+00 -2.98247725e-01
-1.03736424e+00 8.46746147e-01 -2.63063520e-01 -4.89356101e-01
-4.32506315e-02 -8.03633451e-01 -5.18218517e-01 5.68337142e-01
3.82284760e-01 4.16384310e-01 1.12256265e+00 -9.23745990e-01
-1.26193631e+00 1.41254649e-01 -1.01727709e-01 5.79026453e-02
-6.17861450e-01 5.30696392e-01 -6.51458263e-01 -1.08002794e+00
-3.59660499e-02 -8.36106837e-01 -1.65726542e-01 -5.65963984e-01
-7.35554874e-01 -3.37856740e-01 8.12998176e-01 -5.95713437e-01
1.87687910e+00 -1.84611654e+00 -2.18807667e-01 6.47934854e-01
2.34122232e-01 8.11290219e-02 -9.00069699e-02 1.04746246e+00
-4.26808670e-02 4.62062240e-01 2.41664886e-01 1.59796283e-01
6.59736097e-02 4.94754687e-02 -9.34277892e-01 -1.15940580e-02
1.36925817e-01 7.32564509e-01 -1.20174348e+00 -9.54975545e-01
-1.70213506e-01 -2.38992795e-01 -2.91903824e-01 2.44937703e-01
-6.19380772e-01 -4.55876052e-01 -8.12837422e-01 6.76118374e-01
1.46860033e-01 -2.23068118e-01 8.80104527e-02 -7.26321116e-02
-2.07798600e-01 9.97875631e-01 -1.39532530e+00 1.13636529e+00
-1.64433405e-01 4.28801984e-01 -1.23399831e-01 -4.28533554e-01
6.15433633e-01 3.96326363e-01 7.91846395e-01 -7.30090216e-02
-1.12585099e-02 4.38698202e-01 -4.75814521e-01 -1.05723329e-01
1.15271282e+00 2.80578226e-01 -4.16888177e-01 9.20995831e-01
3.81136239e-01 -4.87050503e-01 9.23499286e-01 8.69842231e-01
1.31718850e+00 -2.80635238e-01 9.58750188e-01 -3.76144499e-01
4.06736851e-01 3.28848034e-01 3.78679149e-02 1.14945030e+00
3.49704951e-01 2.93852806e-01 3.83050829e-01 -3.27023894e-01
-1.03575146e+00 -7.18237996e-01 1.05280839e-01 1.34586036e+00
-3.82387727e-01 -1.30420530e+00 -4.16661114e-01 -1.18828118e+00
2.81396717e-01 6.43549860e-01 -2.83543944e-01 3.42140943e-01
-6.14416301e-01 -6.35548651e-01 5.80712974e-01 2.20479637e-01
-8.20363835e-02 -1.00074065e+00 -2.52109259e-01 3.80127996e-01
-9.69175696e-02 -1.40486872e+00 -6.97907150e-01 5.46979010e-01
-5.50591111e-01 -1.26942205e+00 -5.75136006e-01 -8.31378281e-01
6.74590647e-01 4.47921842e-01 1.40662384e+00 -7.72192776e-02
7.05716833e-02 5.53659916e-01 -6.97175205e-01 -6.10819221e-01
-5.64133883e-01 7.17934430e-01 7.73032978e-02 -4.11767691e-01
8.88548017e-01 -2.11631417e-01 -1.04656436e-01 -1.35286391e-01
-8.94899547e-01 -1.12155236e-01 7.46421337e-01 7.19730616e-01
7.36319602e-01 4.94819999e-01 6.94833100e-02 -1.15145755e+00
1.10868454e+00 -2.54203737e-01 -5.92291772e-01 4.38973635e-01
-9.09162104e-01 2.60657907e-01 6.51456118e-01 -6.06575310e-01
-5.15513062e-01 6.50724098e-02 2.74947256e-01 6.46469072e-02
-7.49346092e-02 7.50259101e-01 2.41841570e-01 2.27604911e-01
6.11005664e-01 3.78837109e-01 -7.19097435e-01 -5.90732157e-01
6.09449983e-01 8.15067172e-01 2.89793283e-01 -7.75767624e-01
1.27564466e+00 2.48386234e-01 -2.69836426e-01 -1.07306087e+00
-1.20953429e+00 -1.17582011e+00 -7.71162033e-01 1.54004350e-01
2.06689134e-01 -1.03125131e+00 -1.24873802e-01 1.82316467e-01
-9.53358591e-01 1.30148411e-01 -5.11021197e-01 3.45964640e-01
-2.27183625e-02 5.05017519e-01 -5.51585734e-01 -7.16348827e-01
-8.60194981e-01 -4.62097526e-01 1.24043036e+00 2.32965320e-01
-7.35071957e-01 -5.21581948e-01 1.75548330e-01 -8.33387151e-02
-1.71338201e-01 -3.33764553e-01 8.00750017e-01 -1.48233259e+00
-4.59828824e-01 -7.98495471e-01 -9.17429700e-02 4.38248925e-02
4.36736405e-01 1.72082379e-01 -5.89992225e-01 -9.97486040e-02
-4.23368603e-01 -3.49066496e-01 1.09190595e+00 2.56029367e-02
8.42043579e-01 -7.95345902e-01 -6.32363796e-01 -9.39319748e-03
9.94615853e-01 -9.47289448e-03 3.72767746e-02 5.92939794e-01
5.29094934e-01 5.82416058e-01 8.92810822e-01 6.31287396e-01
2.95143604e-01 2.64158756e-01 -4.48282182e-01 6.54043332e-02
1.78389058e-01 -8.91009092e-01 3.31474155e-01 8.35316718e-01
5.00068843e-01 -1.86915383e-01 -8.70486617e-01 9.35615063e-01
-1.55730414e+00 -8.99309456e-01 1.00186132e-01 1.79332852e+00
1.61451757e+00 6.95950508e-01 4.11943972e-01 2.52566904e-01
3.57528180e-01 3.87308836e-01 9.46224332e-02 -1.05118148e-01
-3.26315276e-02 5.99408805e-01 1.02352893e+00 4.48788553e-01
-1.46707356e+00 1.36335039e+00 6.96695757e+00 1.11927640e+00
-6.18833840e-01 -3.36164176e-01 1.00831814e-01 1.93836968e-02
-4.34215933e-01 4.03248727e-01 -1.46464598e+00 4.85214293e-02
8.76172006e-01 -4.25339103e-01 2.92607754e-01 9.97697532e-01
-4.87602875e-02 -3.18079203e-01 -1.13597536e+00 6.50936544e-01
-8.27419758e-03 -1.26614606e+00 2.29627714e-01 -5.72553687e-02
6.44391537e-01 8.05976242e-02 -3.54709715e-01 8.82584080e-02
8.59205544e-01 -4.37445998e-01 6.38445556e-01 2.00541049e-01
4.56224531e-01 -7.21495807e-01 4.62506443e-01 3.67860854e-01
-1.41034353e+00 1.99201722e-02 -6.32819757e-02 3.12322259e-01
-4.08983342e-02 9.23320591e-01 -1.28490531e+00 2.66231328e-01
6.56916916e-01 5.27391255e-01 -7.33067751e-01 6.61783695e-01
-7.29449630e-01 8.43044996e-01 -6.23957276e-01 -5.80614269e-01
3.12431544e-01 3.40913385e-01 7.31206000e-01 1.85065866e+00
-8.33969377e-03 2.01178104e-01 7.99772084e-01 4.80797112e-01
-1.86149150e-01 4.57448095e-01 -3.87014180e-01 -7.96421885e-01
8.21218073e-01 1.44123363e+00 -1.08326006e+00 -8.04474115e-01
-4.30135638e-01 5.30092061e-01 -1.32490441e-01 2.70799279e-01
1.60877984e-02 -7.74902463e-01 3.90010685e-01 2.24231124e-01
6.21699452e-01 -4.40360665e-01 -4.95784245e-02 -1.35383606e+00
4.14869189e-03 -1.31387794e+00 6.71152413e-01 -3.35519284e-01
-1.17754364e+00 2.29271531e-01 5.07984638e-01 -9.67175543e-01
-4.74630803e-01 -6.43642247e-01 -2.39230514e-01 5.27758777e-01
-1.36256921e+00 -1.11565697e+00 3.58968347e-01 2.40329266e-01
6.41668618e-01 -3.03799450e-01 9.18834805e-01 -1.77148074e-01
-1.69726923e-01 2.15911344e-01 -1.77707076e-02 7.04698384e-01
8.77238631e-01 -1.51916850e+00 1.04054487e+00 1.09250867e+00
8.29324543e-01 1.07830048e+00 9.48722243e-01 -9.68773007e-01
-1.57950568e+00 -7.48734832e-01 1.64190638e+00 -5.84409356e-01
1.30436611e+00 -6.01736486e-01 -5.67666948e-01 5.24431527e-01
2.49761656e-01 -2.32430816e-01 8.01948071e-01 3.43882769e-01
-5.95821977e-01 2.89190263e-02 -6.00088120e-01 8.73467267e-01
5.83918929e-01 -1.01083004e+00 -1.04732990e+00 4.33820486e-01
8.74345303e-01 -2.25579336e-01 -7.11752713e-01 2.04614982e-01
5.96225441e-01 -9.82095972e-02 9.89497840e-01 -5.80244005e-01
1.31942868e-01 -2.45532632e-01 -1.65966406e-01 -1.00068331e+00
-1.98090196e-01 -1.13313687e+00 -6.98632538e-01 1.26095784e+00
8.67066145e-01 -8.90427604e-02 6.83748007e-01 3.19764525e-01
4.68979686e-01 -4.86451209e-01 -3.63671273e-01 -6.67320073e-01
-1.13410652e-01 -6.01357281e-01 4.44795102e-01 8.29245627e-01
4.93188262e-01 7.71010935e-01 -2.68900245e-01 -9.50561743e-03
5.23961127e-01 3.08590084e-01 9.24356461e-01 -1.13772058e+00
-2.88548112e-01 -2.63594687e-01 2.92544216e-01 -1.22848845e+00
4.28525209e-02 -8.27995598e-01 -6.88054785e-02 -1.52603495e+00
2.20081806e-01 -6.12609647e-02 -2.91343123e-01 9.15150464e-01
-2.92934984e-01 -1.33433729e-01 4.61972840e-02 4.20757741e-01
-7.32949674e-01 1.27050951e-01 7.97102094e-01 -4.63099867e-01
-5.62546670e-01 1.11782223e-01 -8.42114270e-01 6.69536412e-01
6.18465602e-01 -1.01172256e+00 -3.05220336e-01 2.43241504e-01
6.20418429e-01 -3.58619779e-01 -1.36301890e-01 -4.43765938e-01
3.15388978e-01 -4.72082287e-01 2.40909457e-01 -1.03543055e+00
-1.20444536e-01 -6.36302233e-01 -5.24952114e-01 1.46614820e-01
-5.33174634e-01 2.54197419e-01 3.35556746e-01 3.60022753e-01
-1.79924160e-01 -4.58402246e-01 1.66891307e-01 -5.06362557e-01
-7.16863155e-01 2.67158747e-01 -8.20028484e-01 4.74944025e-01
4.55872595e-01 2.38674760e-01 -1.34916216e-01 -2.52567202e-01
-2.53222704e-01 8.07861388e-02 2.12656900e-01 5.26583433e-01
5.80943942e-01 -1.04297781e+00 -7.84949720e-01 6.29736930e-02
5.25925457e-01 -2.56664842e-01 -7.66759455e-01 3.68726224e-01
-3.17060590e-01 7.26998270e-01 5.12950480e-01 5.03450457e-04
-1.49432933e+00 6.05901003e-01 -2.96366781e-01 -8.11633289e-01
-6.07641637e-01 6.65374935e-01 -3.20793808e-01 -2.76735157e-01
2.28769228e-01 -7.73663163e-01 -3.50282729e-01 6.05112076e-01
9.06754732e-01 -3.64230908e-02 1.95676297e-01 -3.23456466e-01
-4.79553312e-01 2.19158322e-01 -8.22184741e-01 -5.76188147e-01
1.24643171e+00 1.51807547e-01 -3.72569799e-01 3.39934796e-01
9.59793985e-01 6.69951618e-01 -5.16205311e-01 -7.64195561e-01
7.23514497e-01 -2.16541439e-01 1.12774797e-01 -7.52657175e-01
-3.69751841e-01 1.15862496e-01 -2.01313242e-01 4.10012186e-01
9.03584123e-01 2.16154948e-01 9.22214687e-01 9.68043566e-01
4.63296235e-01 -1.53468657e+00 1.52694672e-01 7.64060020e-01
4.65395570e-01 -1.08726501e+00 8.01287472e-01 -3.70910704e-01
-3.40445846e-01 1.13124430e+00 7.11073130e-02 -4.92291562e-02
1.10034454e+00 6.69817746e-01 -3.52828801e-02 -3.87837678e-01
-9.41573083e-01 -4.11860228e-01 7.88150370e-01 1.32204797e-02
5.56846380e-01 -1.04129933e-01 -4.57704037e-01 7.07670569e-01
-5.11595666e-01 -1.86016455e-01 4.11114991e-01 1.50846004e+00
-8.05497110e-01 -1.43186247e+00 -3.93969506e-01 6.66028619e-01
-1.13299489e+00 -6.47882640e-01 -1.17480171e+00 7.26276875e-01
-2.19575599e-01 9.84883249e-01 -4.48186576e-01 -3.65487009e-01
2.74518102e-01 2.21216932e-01 2.36647248e-01 -1.09738481e+00
-8.27354252e-01 5.75028956e-01 4.50164586e-01 -2.32771471e-01
-3.79501492e-01 -6.78554237e-01 -1.10203195e+00 -7.45346323e-02
-5.97401440e-01 6.34574115e-01 3.18371385e-01 1.05345237e+00
2.93513745e-01 -2.68041855e-03 5.44744074e-01 -3.29776675e-01
-5.87497771e-01 -1.07681513e+00 -4.67592835e-01 -1.89310890e-02
4.79600191e-01 -8.99292305e-02 -1.50442719e-01 2.82476336e-01] | [12.244593620300293, 8.911954879760742] |
e78f2609-34a6-4e95-acf1-f3f483a0c77c | inconsistency-aware-uncertainty-estimation | 2110.08762 | null | https://arxiv.org/abs/2110.08762v1 | https://arxiv.org/pdf/2110.08762v1.pdf | Inconsistency-aware Uncertainty Estimation for Semi-supervised Medical Image Segmentation | In semi-supervised medical image segmentation, most previous works draw on the common assumption that higher entropy means higher uncertainty. In this paper, we investigate a novel method of estimating uncertainty. We observe that, when assigned different misclassification costs in a certain degree, if the segmentation result of a pixel becomes inconsistent, this pixel shows a relative uncertainty in its segmentation. Therefore, we present a new semi-supervised segmentation model, namely, conservative-radical network (CoraNet in short) based on our uncertainty estimation and separate self-training strategy. In particular, our CoraNet model consists of three major components: a conservative-radical module (CRM), a certain region segmentation network (C-SN), and an uncertain region segmentation network (UC-SN) that could be alternatively trained in an end-to-end manner. We have extensively evaluated our method on various segmentation tasks with publicly available benchmark datasets, including CT pancreas, MR endocardium, and MR multi-structures segmentation on the ACDC dataset. Compared with the current state of the art, our CoraNet has demonstrated superior performance. In addition, we have also analyzed its connection with and difference from conventional methods of uncertainty estimation in semi-supervised medical image segmentation. | ['Yang Gao', 'Lei Qi', 'Qian Yu', 'Yefeng Zheng', 'Jiwen Lu', 'Tong Ling', 'Jian Zhang', 'Yinghuan Shi'] | 2021-10-17 | null | null | null | null | ['semi-supervised-medical-image-segmentation'] | ['computer-vision'] | [ 3.96113694e-01 7.57381380e-01 -1.29191488e-01 -6.66868210e-01
-9.58373964e-01 -4.46360856e-01 2.85366386e-01 4.46289629e-01
-4.44965303e-01 9.24832463e-01 -1.86125651e-01 -4.18969125e-01
-7.73942769e-02 -5.76172352e-01 -8.20188582e-01 -7.71008670e-01
-7.12284297e-02 8.35428238e-01 4.79876906e-01 4.37607199e-01
1.22725531e-01 7.98934549e-02 -8.21431875e-01 1.75584480e-01
1.42582750e+00 1.29334176e+00 -1.55770585e-01 2.66128242e-01
5.59741482e-02 6.54837251e-01 -4.52971488e-01 -2.88189948e-01
1.91098005e-01 -5.60907543e-01 -1.18582344e+00 1.80906907e-01
1.00114740e-01 -1.94527566e-01 1.79725707e-01 1.32597697e+00
4.20094013e-01 -9.10735037e-03 8.81574869e-01 -1.00597906e+00
-3.06078494e-02 1.56853557e+00 -5.79913676e-01 1.34687886e-01
-6.82303011e-02 -2.93142051e-02 4.85576332e-01 -3.79281223e-01
6.28818512e-01 8.90423954e-01 8.48436832e-01 4.61718678e-01
-1.34796441e+00 -2.68863738e-01 1.12796076e-01 -2.72837728e-01
-1.29755807e+00 1.17932804e-01 6.23985052e-01 -4.62401360e-01
4.77717936e-01 1.63113505e-01 5.74211776e-01 7.64019728e-01
4.74307209e-01 9.66790497e-01 1.47955573e+00 -2.64198601e-01
6.14223719e-01 2.51962185e-01 3.41858566e-01 6.50495291e-01
2.52573878e-01 2.80683845e-01 2.06822664e-01 -2.97826886e-01
5.53394139e-01 -3.52413863e-01 -3.87576491e-01 -6.80470169e-01
-1.24556792e+00 6.07524335e-01 5.08875847e-01 3.94530952e-01
-2.13511392e-01 5.27868718e-02 5.90612471e-01 -3.82463373e-02
6.31305337e-01 2.94000089e-01 -5.28838754e-01 1.47156358e-01
-1.36614716e+00 -1.72005832e-01 1.04660296e+00 7.51390338e-01
3.75305355e-01 -3.46135139e-01 -3.57974201e-01 5.26324272e-01
5.47129035e-01 9.68033168e-03 7.41897047e-01 -9.18208539e-01
1.74408674e-01 3.56621176e-01 -2.54160762e-01 -4.76020634e-01
-7.14236379e-01 -6.23556018e-01 -9.68962908e-01 3.13954353e-01
4.67418998e-01 -4.27799821e-01 -1.22479403e+00 1.59615207e+00
4.41702574e-01 3.81764024e-01 1.40090734e-01 8.48766744e-01
8.28251898e-01 2.67471254e-01 4.05672900e-02 -3.20943594e-01
1.06452322e+00 -1.07613337e+00 -7.25756168e-01 1.36671588e-01
5.57291687e-01 -4.80634570e-01 4.80440348e-01 5.44519901e-01
-1.11682618e+00 -4.27285999e-01 -1.23555946e+00 3.23046774e-01
-2.53469795e-01 -2.45431792e-02 4.50801879e-01 8.06421459e-01
-9.24226701e-01 1.13744724e+00 -1.27475154e+00 -1.15278549e-02
5.92394531e-01 2.37446994e-01 4.07785736e-02 3.21018189e-01
-1.23237753e+00 9.11662877e-01 9.25136566e-01 2.34161928e-01
-8.82053971e-01 -6.61973953e-01 -1.01241291e+00 -2.69611210e-01
5.98376036e-01 -7.14263201e-01 1.22277415e+00 -9.80839014e-01
-1.64397311e+00 8.98832262e-01 4.15189683e-01 -9.27797079e-01
1.15641272e+00 2.25196332e-02 -1.05243221e-01 2.50244975e-01
1.37325749e-01 8.03062260e-01 8.96569431e-01 -1.49406791e+00
-4.51824874e-01 -3.90084445e-01 -1.97500169e-01 1.10424407e-01
4.26204771e-01 -2.75094151e-01 -3.89769405e-01 -6.90938056e-01
5.76577425e-01 -1.19675148e+00 -5.73760390e-01 -6.58170581e-02
-9.14878547e-01 -4.82807010e-02 4.42707449e-01 -4.99446958e-01
1.19636917e+00 -2.07375813e+00 2.27984503e-01 6.62513316e-01
2.56469488e-01 7.25503564e-02 4.62034166e-01 -4.97709930e-01
-2.00217530e-01 2.65249729e-01 -1.26971209e+00 -4.24884498e-01
-1.72524258e-01 3.08550179e-01 1.73029140e-01 5.30318320e-01
1.03391975e-01 7.69316435e-01 -1.12504685e+00 -1.15425599e+00
3.51355314e-01 2.69624561e-01 -3.31474543e-01 -2.68338826e-02
-4.47535574e-01 7.51045048e-01 -4.42683011e-01 7.01671004e-01
9.21517611e-01 -3.20653498e-01 1.74852088e-01 -2.22426698e-01
3.79352383e-02 -4.07804511e-02 -1.30546188e+00 1.94255149e+00
-1.45306841e-01 1.63239673e-01 -1.41766250e-01 -9.25034881e-01
4.86310005e-01 2.85790592e-01 6.80731595e-01 -5.33477589e-02
3.82131100e-01 4.56648976e-01 9.69668776e-02 -3.25049460e-01
2.58223623e-01 -1.01405792e-01 4.33521485e-03 3.07985902e-01
1.72296926e-01 -4.47286516e-01 2.94502467e-01 2.56116003e-01
7.69292712e-01 4.75795448e-01 4.04556751e-01 -6.71050191e-01
6.62737966e-01 1.37505263e-01 5.69688082e-01 7.07317591e-01
-6.78013146e-01 9.99952316e-01 8.13703656e-01 1.82472691e-02
-4.56002414e-01 -1.26334810e+00 -9.59959269e-01 2.88531601e-01
3.56192410e-01 -2.52936147e-02 -1.16721964e+00 -1.20590496e+00
1.08049726e-02 6.91537440e-01 -7.13737369e-01 1.61690433e-02
-4.58319962e-01 -9.83055770e-01 6.23614192e-01 6.00518763e-01
7.04834342e-01 -9.41370130e-01 -7.67047584e-01 1.95751965e-01
-2.01890379e-01 -1.09295058e+00 -6.15227640e-01 5.71993232e-01
-1.32051623e+00 -1.10275388e+00 -8.52363765e-01 -5.82369208e-01
9.72186089e-01 -5.11077106e-01 1.16341138e+00 -1.25322267e-01
-1.07565977e-01 2.04291165e-01 -2.09374055e-01 -2.77550578e-01
-6.39761925e-01 4.78714406e-02 -3.07763755e-01 -2.66768187e-01
-2.01223388e-01 -2.83499092e-01 -6.46984518e-01 2.72697061e-01
-9.96739268e-01 -1.60909057e-01 5.23081720e-01 9.06136870e-01
1.10775471e+00 1.34454980e-01 4.08889204e-01 -1.47767031e+00
3.10607821e-01 -5.87769687e-01 -5.69678962e-01 3.27773422e-01
-1.05820882e+00 2.49570668e-01 2.97052532e-01 -1.44216001e-01
-1.30539298e+00 3.10409158e-01 -2.21545249e-01 -3.47430736e-01
-1.49586752e-01 6.07936263e-01 1.57338262e-01 2.67630024e-03
6.81772351e-01 -5.05507290e-02 2.22909465e-01 -1.38507277e-01
3.40135872e-01 6.00627124e-01 6.44328058e-01 -5.41762114e-01
2.91697055e-01 5.65196455e-01 -4.24653031e-02 -1.48133695e-01
-8.43837500e-01 -3.43051523e-01 -8.39107633e-01 -1.18631788e-01
1.02294636e+00 -6.68570399e-01 -1.61066219e-01 5.13639033e-01
-8.48785520e-01 -3.57999623e-01 -6.32390141e-01 6.57525599e-01
-6.35587275e-01 8.07810962e-01 -7.29470015e-01 -5.99478304e-01
-5.65860450e-01 -1.52710128e+00 1.00873363e+00 3.42745095e-01
-6.18031174e-02 -1.22767091e+00 1.03521645e-01 2.40077212e-01
2.70546854e-01 7.57665694e-01 6.65557086e-01 -8.61740470e-01
-3.16755950e-01 1.59478158e-01 -1.38340577e-01 7.12581396e-01
4.73263189e-02 9.40737035e-03 -7.81590760e-01 -1.24212876e-01
3.19431692e-01 -2.46078968e-01 1.17501652e+00 8.84584367e-01
1.42256272e+00 1.50868222e-01 -4.75368291e-01 6.89338088e-01
1.50475180e+00 1.67107821e-01 5.34375548e-01 1.64016694e-01
4.25634414e-01 3.77483189e-01 7.07142413e-01 2.59227633e-01
2.96931535e-01 1.90880686e-01 5.42287409e-01 -2.17246354e-01
4.31099646e-02 1.22672930e-01 3.27151008e-02 5.87200344e-01
6.72983052e-03 -4.21444103e-02 -8.74757171e-01 4.20941144e-01
-2.00384903e+00 -3.22971314e-01 -1.56654358e-01 2.03299117e+00
1.14879024e+00 5.28717875e-01 -1.05106838e-01 6.29991293e-02
7.37187743e-01 4.53792810e-02 -8.08272958e-01 -3.15350384e-01
1.35345920e-03 2.79428419e-02 5.60754120e-01 4.49097693e-01
-1.46431017e+00 6.03345096e-01 6.11152649e+00 7.03537524e-01
-9.42032516e-01 1.30302712e-01 1.16640973e+00 3.63779306e-01
-2.45765150e-01 -6.85053542e-02 -3.72314155e-01 6.00012541e-01
6.72139764e-01 2.23987058e-01 -1.00244686e-01 9.62455273e-01
-2.73916155e-01 -5.75074613e-01 -1.32789409e+00 4.92745996e-01
1.55226111e-01 -1.05762219e+00 -3.16971272e-01 -3.67530048e-01
9.54855978e-01 1.41589195e-01 -5.78387231e-02 1.49727479e-01
1.88952133e-01 -8.27099919e-01 6.14531636e-01 4.97425944e-01
6.41994476e-01 -5.64671099e-01 1.16606283e+00 3.03723931e-01
-7.79010773e-01 3.80724490e-01 -1.24580311e-02 6.68951690e-01
2.68815130e-01 1.05832088e+00 -6.99182868e-01 8.85940075e-01
6.70066297e-01 6.31180465e-01 -4.17522192e-01 1.10752678e+00
-3.74633640e-01 5.34343898e-01 -4.15474206e-01 3.21920007e-01
3.21859211e-01 -3.85867476e-01 5.90817511e-01 1.27352357e+00
-1.04435235e-02 3.33220251e-02 1.24888346e-01 1.13542509e+00
-3.64066241e-03 -3.66789177e-02 -3.61042060e-02 3.05218071e-01
1.53686702e-01 1.24695873e+00 -1.30855381e+00 -6.22210026e-01
-2.67502386e-02 9.13859606e-01 1.40139658e-03 -1.12934925e-01
-1.04761207e+00 -1.71793982e-01 -1.13345861e-01 -2.59634435e-01
9.96751189e-02 3.12828213e-01 -6.71491623e-01 -1.09342325e+00
-8.34678784e-02 -5.38936317e-01 7.28953600e-01 -4.89822686e-01
-1.33675468e+00 9.02858377e-01 2.33079568e-01 -1.18650901e+00
-2.30494410e-01 -4.25687194e-01 -4.58946079e-01 6.42514408e-01
-1.68337893e+00 -7.62746036e-01 -9.04036760e-02 4.22869503e-01
3.47352266e-01 3.72334152e-01 5.72874367e-01 6.97709396e-02
-6.05350494e-01 4.04202282e-01 1.16566375e-01 1.82005301e-01
6.23872757e-01 -1.84814680e+00 -2.49022022e-02 7.90484548e-01
-1.53035387e-01 3.52948904e-01 5.15290022e-01 -8.65004063e-01
-5.79998851e-01 -1.18488526e+00 4.44869697e-01 -3.86656016e-01
5.27627945e-01 1.23684652e-01 -8.98797095e-01 8.48824739e-01
2.02606887e-01 3.97728324e-01 5.42728543e-01 -2.83788770e-01
1.44629046e-01 2.06229329e-01 -1.77246153e+00 3.05400789e-01
7.73467422e-01 -5.05319834e-02 -8.26886892e-01 4.28848952e-01
9.61498082e-01 -1.11798966e+00 -1.33039606e+00 8.35660398e-01
2.82752544e-01 -1.02394402e+00 7.71446764e-01 -5.14832251e-02
4.90464598e-01 -3.45966816e-01 2.26787403e-01 -1.37890303e+00
1.97722182e-01 -4.65019017e-01 -1.03410453e-01 9.98060942e-01
6.59447014e-01 -6.89775586e-01 6.95526183e-01 7.41729736e-01
-4.88296300e-01 -1.00417066e+00 -1.10359752e+00 -7.04099774e-01
3.39206368e-01 -4.65003669e-01 4.24152225e-01 8.92024755e-01
2.42553711e-01 -1.78929031e-01 5.05521372e-02 1.88657552e-01
1.02412474e+00 2.24631131e-01 -1.60475582e-01 -1.13460827e+00
-2.62491167e-01 -3.72349918e-01 -2.29123056e-01 -6.76666498e-01
1.38718084e-01 -1.03738797e+00 5.33878684e-01 -1.41633344e+00
3.13948363e-01 -7.63242543e-01 -5.84164798e-01 3.01156282e-01
-2.49231353e-01 1.42243002e-02 -1.08860962e-01 2.05667645e-01
-6.39422655e-01 1.41269028e-01 1.43940020e+00 -1.47504106e-01
-3.44695419e-01 3.55923831e-01 -4.38160360e-01 9.68623042e-01
8.19001853e-01 -6.75070167e-01 -3.18991184e-01 -9.20547247e-02
-1.91511959e-01 3.32491666e-01 2.00854227e-01 -1.03822649e+00
2.48329103e-01 3.80558223e-01 2.80758590e-01 -7.62333691e-01
-3.00386488e-01 -9.84114945e-01 -8.07247385e-02 7.98963487e-01
-4.53946024e-01 -4.47538704e-01 1.78910866e-01 5.51646948e-01
-3.01709473e-01 -6.57744884e-01 1.10124135e+00 -3.55467647e-01
-4.03880000e-01 2.04911187e-01 -1.72316298e-01 1.55133039e-01
1.21228087e+00 -2.39546858e-02 -4.66529801e-02 2.95320094e-01
-1.17783785e+00 5.88930666e-01 1.89122647e-01 -9.93124843e-02
5.34533024e-01 -8.97481501e-01 -6.20555699e-01 -7.19769020e-03
-1.63304247e-02 6.52872980e-01 2.66368181e-01 1.19936097e+00
-5.92445314e-01 3.17863226e-01 5.94354235e-03 -1.24223781e+00
-8.00499260e-01 5.23143888e-01 6.76409245e-01 -5.62977314e-01
-5.37798941e-01 7.63275206e-01 -2.06735671e-01 -6.30747736e-01
4.63835090e-01 -1.10824919e+00 -2.60783255e-01 -1.98778883e-02
-3.36010233e-02 4.21459258e-01 7.60999471e-02 -4.08121794e-01
-3.86951268e-01 4.50130612e-01 5.75991161e-02 -9.74100381e-02
9.09161568e-01 -8.31134990e-02 -3.23651910e-01 5.46982586e-01
8.45169961e-01 -3.95093501e-01 -1.33179426e+00 -3.43303323e-01
1.96341470e-01 -2.48837918e-02 2.47846752e-01 -1.15408587e+00
-1.29432034e+00 4.99975592e-01 9.18746650e-01 1.29297361e-01
1.12035656e+00 1.84105799e-01 5.98611772e-01 7.69081190e-02
4.89624649e-01 -1.30426300e+00 -3.37654382e-01 2.07180813e-01
6.32522762e-01 -1.69489551e+00 1.16144381e-01 -6.56205237e-01
-9.10635591e-01 9.72849548e-01 4.53000665e-01 -6.92910925e-02
1.17219949e+00 5.19015193e-01 9.46401954e-02 -2.41778102e-02
-2.45001003e-01 -1.13227844e-01 3.43053848e-01 3.17816794e-01
3.52116287e-01 3.03531468e-01 -6.15651488e-01 7.05941319e-01
1.19718745e-01 2.17306510e-01 3.87649596e-01 1.00245547e+00
-3.06206018e-01 -8.87124240e-01 -3.11490029e-01 4.96380359e-01
-6.37374759e-01 -1.00050613e-01 4.32838732e-03 8.31094384e-01
3.72853279e-01 7.30827987e-01 -2.14072838e-02 -2.79865693e-02
1.16302602e-01 -6.27777576e-02 4.78897840e-01 -6.58820331e-01
-9.19903100e-01 2.19862089e-01 -1.34013128e-02 -6.33178353e-01
-7.34950066e-01 -9.21779990e-01 -1.94728863e+00 4.47306246e-01
-5.07070422e-01 1.98255852e-01 7.39851534e-01 1.01767254e+00
-1.18904501e-01 8.83876562e-01 4.37207252e-01 -5.63187599e-01
-7.86145926e-01 -1.06476712e+00 -6.34018362e-01 2.53947467e-01
9.57563296e-02 -4.55749273e-01 -5.16845942e-01 -3.34767834e-03] | [14.621150016784668, -2.134460687637329] |
887d75e2-4c4a-4e32-a630-af2260e3ea7a | iterative-knowledge-exchange-between-deep | 2012.07123 | null | https://arxiv.org/abs/2012.07123v1 | https://arxiv.org/pdf/2012.07123v1.pdf | Iterative Knowledge Exchange Between Deep Learning and Space-Time Spectral Clustering for Unsupervised Segmentation in Videos | We propose a dual system for unsupervised object segmentation in video, which brings together two modules with complementary properties: a space-time graph that discovers objects in videos and a deep network that learns powerful object features. The system uses an iterative knowledge exchange policy. A novel spectral space-time clustering process on the graph produces unsupervised segmentation masks passed to the network as pseudo-labels. The net learns to segment in single frames what the graph discovers in video and passes back to the graph strong image-level features that improve its node-level features in the next iteration. Knowledge is exchanged for several cycles until convergence. The graph has one node per each video pixel, but the object discovery is fast. It uses a novel power iteration algorithm computing the main space-time cluster as the principal eigenvector of a special Feature-Motion matrix without actually computing the matrix. The thorough experimental analysis validates our theoretical claims and proves the effectiveness of the cyclical knowledge exchange. We also perform experiments on the supervised scenario, incorporating features pretrained with human supervision. We achieve state-of-the-art level on unsupervised and supervised scenarios on four challenging datasets: DAVIS, SegTrack, YouTube-Objects, and DAVSOD. | ['Marius Leordeanu', 'Adina Magda Florea', 'Emanuela Haller'] | 2020-12-13 | null | null | null | null | ['unsupervised-object-segmentation'] | ['computer-vision'] | [ 1.32988736e-01 3.06742281e-01 -2.71987736e-01 -2.11462304e-01
-2.61996895e-01 -5.47394216e-01 2.22165748e-01 -1.60380051e-01
-5.33236504e-01 3.79274666e-01 -2.00892985e-01 1.02584779e-01
-3.95799696e-01 -5.73554754e-01 -9.86111343e-01 -9.24366832e-01
-5.53274810e-01 5.28760612e-01 5.52494526e-01 3.11827779e-01
1.28437281e-01 3.88020068e-01 -1.37511230e+00 1.34965762e-01
5.69572747e-01 1.00132608e+00 2.26870567e-01 8.32128048e-01
8.88542533e-02 1.11960804e+00 -9.05282274e-02 -2.84669369e-01
7.84415960e-01 -3.05091083e-01 -1.15525186e+00 6.56369746e-01
3.58348131e-01 -1.37800112e-01 -7.02573419e-01 1.28684795e+00
-7.81500563e-02 2.09158763e-01 4.33412284e-01 -1.48409832e+00
-5.43384433e-01 8.62269282e-01 -7.57750094e-01 5.51933229e-01
1.53876662e-01 2.58259654e-01 1.01718760e+00 -7.90424228e-01
9.80426073e-01 9.35438931e-01 5.78825057e-01 2.29261294e-01
-9.46156442e-01 -3.71157974e-01 1.71675637e-01 4.26417083e-01
-1.36501741e+00 -6.26169741e-02 7.90344417e-01 -6.20981216e-01
6.97086155e-01 1.45002425e-01 1.08465278e+00 6.32308185e-01
-1.66259781e-01 1.05164528e+00 7.94505656e-01 -1.20423995e-01
2.81933129e-01 -4.64603258e-03 2.55692154e-01 1.20044863e+00
6.07436001e-02 -7.76926652e-02 -6.71582043e-01 1.77331373e-01
1.01388133e+00 2.46067315e-01 -3.64190370e-01 -5.64840317e-01
-1.27240896e+00 6.81006789e-01 6.27344847e-01 4.46235061e-01
-4.62620884e-01 4.01292086e-01 1.68239638e-01 3.97996873e-01
3.01578701e-01 1.75320655e-01 -5.01233876e-01 -2.78246850e-02
-1.07047164e+00 -2.26180494e-01 1.05210674e+00 1.00565588e+00
1.18210506e+00 -2.01439306e-01 5.67796826e-02 2.14004502e-01
3.00104916e-01 2.73419917e-01 3.88697445e-01 -1.28163695e+00
-1.94405187e-02 7.75033116e-01 -2.99532741e-01 -1.18808949e+00
-5.29918551e-01 -6.84039235e-01 -6.91906214e-01 -5.98945543e-02
3.81299257e-01 -1.93222895e-01 -9.55357432e-01 1.37746668e+00
4.89903152e-01 7.22862303e-01 -9.28007215e-02 1.23818827e+00
7.11977303e-01 5.56584954e-01 -2.09037244e-01 -4.82453257e-01
1.08130264e+00 -1.32258046e+00 -4.61119443e-01 1.25349805e-01
5.51608264e-01 -3.35871518e-01 6.56628132e-01 2.79805064e-01
-1.04875290e+00 -6.82443023e-01 -7.63217926e-01 2.73077637e-01
-2.03586444e-01 1.04969844e-01 7.92729139e-01 3.67620885e-01
-1.22251964e+00 8.56783986e-01 -1.03637767e+00 -3.91801894e-01
6.47472620e-01 5.73333561e-01 -3.99210423e-01 2.29484469e-01
-7.50509024e-01 1.90629646e-01 7.61131108e-01 9.53242034e-02
-1.17006934e+00 -6.28550828e-01 -6.15470707e-01 -2.96891760e-02
7.46441126e-01 -6.74953640e-01 8.06176484e-01 -1.63440013e+00
-1.43963909e+00 9.77028191e-01 1.32297441e-01 -6.25474155e-01
6.21487916e-01 -2.29958341e-01 -6.04909323e-02 8.57846916e-01
3.25332969e-01 8.66684318e-01 1.27518499e+00 -1.20954907e+00
-8.39325607e-01 -4.03510660e-01 1.95944980e-01 2.62667120e-01
-4.26473230e-01 -2.90923446e-01 -1.09022939e+00 -5.92763126e-01
3.65569115e-01 -1.13041532e+00 -3.64089698e-01 -1.77874103e-01
-4.83430237e-01 -2.49240577e-01 1.01038778e+00 -4.62136567e-01
1.16520882e+00 -2.27133465e+00 5.28312504e-01 5.43513417e-01
6.09226644e-01 -6.18693456e-02 -4.37438972e-02 4.99459244e-02
-6.88393116e-02 -6.23484589e-02 -3.10660094e-01 -2.34625228e-02
-3.63734484e-01 2.64135033e-01 1.01295725e-01 8.87283683e-01
-3.12459935e-02 1.19724023e+00 -1.04188323e+00 -7.16711223e-01
4.42128479e-02 1.51091903e-01 -6.37169480e-01 9.03056487e-02
-2.49239504e-01 4.76557553e-01 -5.02092838e-01 4.56696510e-01
4.62814957e-01 -8.77270460e-01 1.51249290e-01 -4.48214948e-01
-1.09326385e-01 -4.71992284e-01 -1.27098131e+00 1.88489068e+00
3.94719750e-01 6.20770752e-01 3.47397150e-03 -1.46302247e+00
4.12622273e-01 4.58785780e-02 1.03600597e+00 -2.86501646e-01
2.77515560e-01 -3.60813253e-02 -2.20844850e-01 -8.52343440e-01
1.46466762e-01 3.94450873e-01 3.29773605e-01 3.69184285e-01
6.32689238e-01 2.70493150e-01 5.57311296e-01 7.31024086e-01
1.24610162e+00 1.33747786e-01 -2.55323201e-01 -5.00884056e-01
4.86639351e-01 1.17550820e-01 4.88601506e-01 8.11521411e-01
-1.49907276e-01 3.36544067e-01 4.55274016e-01 -5.55347443e-01
-6.97678685e-01 -1.03535676e+00 2.40392745e-01 1.11382985e+00
5.35406888e-01 -5.29467046e-01 -1.02961135e+00 -1.05942523e+00
-1.34772956e-01 -4.80197854e-02 -7.93987155e-01 -3.20335664e-02
-4.78449106e-01 -4.80583072e-01 2.18556970e-01 4.32565480e-01
7.35968649e-01 -1.03583848e+00 -7.47860372e-01 -1.56335328e-02
-1.09836444e-01 -1.35549366e+00 -6.15142822e-01 2.77249426e-01
-9.27137256e-01 -1.37950909e+00 -5.83025634e-01 -1.21297932e+00
1.00137186e+00 3.27107370e-01 8.20755720e-01 2.16979846e-01
-4.36715275e-01 9.74052131e-01 -4.00121868e-01 1.39526978e-01
2.48181857e-02 7.25140199e-02 1.86181907e-02 5.77419102e-01
1.68994486e-01 -6.72423184e-01 -7.29084432e-01 2.22816661e-01
-8.26410592e-01 1.59174338e-01 5.24516344e-01 5.56015193e-01
7.89523304e-01 2.27692425e-01 2.70345002e-01 -8.10775399e-01
-1.09550446e-01 -5.25944710e-01 -6.47707224e-01 1.30257875e-01
-2.38643289e-01 -4.07598205e-02 3.64486277e-01 -4.25133765e-01
-6.31802380e-01 7.25065291e-01 4.88985300e-01 -8.79450560e-01
-7.92243034e-02 5.58216453e-01 2.14930192e-01 -2.69577593e-01
4.75837827e-01 3.04494023e-01 -5.29350825e-02 -3.19394857e-01
5.83127201e-01 1.90133840e-01 9.05729473e-01 -2.87958205e-01
1.09294522e+00 8.45604658e-01 -4.81248237e-02 -1.00629222e+00
-9.64517951e-01 -8.43005896e-01 -1.03635144e+00 -6.25180960e-01
1.24275088e+00 -9.76016045e-01 -9.20062542e-01 3.45309466e-01
-8.02532077e-01 -4.47822303e-01 -7.58614600e-01 5.97711146e-01
-7.47037709e-01 4.76329237e-01 -7.44148910e-01 -4.09233302e-01
-3.21158841e-02 -8.19444835e-01 8.04544747e-01 4.60534632e-01
2.08477035e-01 -9.71007466e-01 -6.75704107e-02 3.76313567e-01
-2.28326380e-01 1.58158422e-01 4.36838210e-01 -6.60739660e-01
-1.08820307e+00 1.66089252e-01 -3.17591876e-01 3.45123738e-01
-1.31847262e-01 7.84166604e-02 -7.46070147e-01 -2.93096811e-01
1.10457994e-01 -1.84724092e-01 1.08057320e+00 5.60003459e-01
1.25764680e+00 -3.64808321e-01 -4.10927176e-01 9.24017131e-01
1.37961781e+00 4.56185676e-02 2.24579275e-01 1.51434496e-01
1.13368142e+00 3.09947580e-01 3.75203907e-01 3.20381075e-01
2.95509189e-01 2.32985929e-01 4.77231801e-01 -2.59424895e-01
-1.49952948e-01 7.30206743e-02 4.64304924e-01 1.18310308e+00
-4.44587797e-01 9.04429555e-02 -8.39175642e-01 7.28995979e-01
-2.18744969e+00 -9.45188344e-01 -1.99101165e-01 1.72308409e+00
6.02208495e-01 1.87391207e-01 2.04443902e-01 4.77904491e-02
6.62606716e-01 -4.43801563e-03 -6.63901210e-01 2.98979640e-01
-1.66067064e-01 -5.70301004e-02 7.25587785e-01 3.46235603e-01
-1.43010545e+00 1.14746130e+00 5.96893406e+00 6.06880426e-01
-1.02148640e+00 2.09605560e-01 6.40866518e-01 -8.86083469e-02
5.86304814e-02 3.15252952e-02 -3.28912556e-01 2.54126370e-01
5.38064063e-01 -2.09942162e-01 5.47188222e-01 8.52961898e-01
-1.04100138e-01 -2.22156242e-01 -1.19413853e+00 1.13587046e+00
1.42319709e-01 -1.60630059e+00 -2.86306255e-02 -6.72921985e-02
1.02967858e+00 4.24976975e-01 -2.57336963e-02 -8.81838053e-02
1.12160370e-01 -6.39016211e-01 7.31314123e-01 5.92175663e-01
4.04887855e-01 -5.87462664e-01 3.55316490e-01 3.16561133e-01
-1.32888007e+00 -1.95889771e-01 -9.95356888e-02 1.00618377e-01
2.64360271e-02 5.47834694e-01 -6.40445530e-01 6.00315809e-01
1.03543174e+00 1.27517080e+00 -6.13436699e-01 9.48195100e-01
-1.27540911e-02 7.80242264e-01 -5.10084927e-01 2.54252404e-01
4.31923747e-01 -5.12593091e-01 7.89436996e-01 1.27216613e+00
1.47026004e-02 3.75463724e-01 6.17023706e-01 6.84242129e-01
-2.15042189e-01 -6.37925602e-03 -3.51595193e-01 -2.22873524e-01
-8.68464187e-02 1.44109380e+00 -1.50374722e+00 -6.56166375e-01
-4.47923958e-01 1.36714673e+00 2.34567717e-01 6.43032074e-01
-8.41041386e-01 -9.92380902e-02 2.05339104e-01 1.28233647e-02
7.59466231e-01 -3.77982557e-01 5.93498908e-02 -1.28053463e+00
-3.97388972e-02 -4.43333566e-01 5.75493455e-01 -5.63946784e-01
-1.18327081e+00 3.25496525e-01 -1.10299490e-01 -1.02937508e+00
1.17136255e-01 -3.68345469e-01 -3.87360007e-01 2.88908221e-02
-1.28230166e+00 -9.16076183e-01 -4.51882541e-01 1.13718510e+00
5.79208076e-01 -1.85022011e-01 2.50025004e-01 3.27425003e-01
-5.24899840e-01 2.49087542e-01 -1.41350348e-02 4.99321014e-01
1.46095753e-01 -1.22518706e+00 3.18181515e-02 8.28007102e-01
6.50091708e-01 2.82457709e-01 3.85386437e-01 -8.04012716e-01
-1.79449737e+00 -1.10619962e+00 2.97126800e-01 -3.26258689e-01
9.28547800e-01 -2.08749056e-01 -7.11627722e-01 8.99428129e-01
1.62451163e-01 4.46033925e-01 4.06697989e-01 -2.91240662e-01
-7.98595250e-02 -9.61279962e-03 -7.84034848e-01 2.06450000e-01
1.39271986e+00 -5.34009337e-01 -4.06040490e-01 6.62645638e-01
9.36793447e-01 -3.12663317e-01 -8.98625135e-01 2.47204572e-01
3.68064165e-01 -8.87129545e-01 7.42467046e-01 -7.38480031e-01
2.18755975e-01 -6.01749480e-01 9.38213319e-02 -9.12050068e-01
-4.52049166e-01 -1.07124209e+00 -4.04656678e-01 7.71485507e-01
3.39774311e-01 -2.82101005e-01 1.04587805e+00 1.32377058e-01
-9.69041046e-03 -7.48421669e-01 -7.58683741e-01 -7.55880058e-01
-5.98032713e-01 -4.38409775e-01 -5.12352493e-03 1.17062891e+00
-1.04234383e-01 3.04722756e-01 -1.16810441e-01 4.00124431e-01
9.37167466e-01 2.75097191e-01 6.44284248e-01 -1.22105885e+00
-3.69383603e-01 -2.32365325e-01 -7.95985460e-01 -1.22165918e+00
9.84835997e-02 -1.14226401e+00 -5.34315147e-02 -1.22734869e+00
5.47398448e-01 -1.27038896e-01 -4.96365011e-01 4.69902664e-01
1.70064736e-02 4.09690708e-01 3.38425428e-01 2.87353992e-01
-1.31400084e+00 2.67022699e-01 1.41734087e+00 -2.48346940e-01
-3.72774482e-01 -2.19873652e-01 -3.27605098e-01 9.93108451e-01
5.50845325e-01 -5.91678202e-01 -4.69457448e-01 -3.22856247e-01
2.76453346e-01 -8.62290412e-02 4.49107170e-01 -1.16977298e+00
6.27552688e-01 -4.20790687e-02 3.77173930e-01 -5.22162735e-01
3.36149428e-03 -1.00934327e+00 1.35958552e-01 5.84757328e-01
-2.31993496e-01 -3.27021033e-01 -1.53319448e-01 8.50812018e-01
-1.38378575e-01 2.96222605e-02 7.01324105e-01 -1.63283885e-01
-9.55636740e-01 6.98813558e-01 -1.86292574e-01 1.20148331e-01
1.32383311e+00 -6.16573274e-01 1.34324595e-01 -2.47116223e-01
-1.27147686e+00 4.64804381e-01 3.51430207e-01 3.67248088e-01
5.66403508e-01 -1.16673827e+00 -5.38819492e-01 2.37408847e-01
-2.30699822e-01 6.41529560e-02 2.19954908e-01 1.25225353e+00
-5.18854499e-01 9.99747030e-03 -1.89502612e-01 -1.22671366e+00
-1.01151168e+00 6.56990349e-01 2.32360810e-01 5.26911840e-02
-7.96774507e-01 1.02538693e+00 3.86314601e-01 1.21559761e-01
4.00059074e-01 -2.27272719e-01 -1.97864398e-01 2.28139773e-01
2.84415781e-01 4.71316814e-01 -3.75035107e-01 -8.12409222e-01
-2.66645461e-01 7.13209510e-01 7.28294579e-03 7.26147592e-02
1.45156300e+00 -3.36759567e-01 -3.98623317e-01 5.06622314e-01
1.45390952e+00 -3.81361634e-01 -1.65394175e+00 -3.13981414e-01
1.33327484e-01 -3.78440380e-01 3.44455540e-02 -2.49169961e-01
-1.74851882e+00 3.65013480e-01 6.67323351e-01 4.72794563e-01
1.20720887e+00 3.89609307e-01 6.66445255e-01 6.79586589e-01
2.70598620e-01 -1.32403064e+00 5.19804001e-01 5.80100596e-01
5.26086152e-01 -1.10620975e+00 -1.14493882e-02 -5.60043633e-01
-5.22566199e-01 1.09093392e+00 4.80490565e-01 -3.99526149e-01
1.03076029e+00 1.52177617e-01 -1.00356728e-01 -6.28729284e-01
-4.09116387e-01 -3.88688833e-01 3.00990403e-01 3.68507236e-01
-1.44422635e-01 -3.54329795e-02 1.36123216e-02 3.93682599e-01
-3.50626791e-03 1.32071197e-01 3.40799958e-01 8.15030873e-01
-5.82842052e-01 -4.74469960e-01 2.35046893e-02 5.55375218e-01
-3.76643956e-01 2.44307324e-01 -3.43083948e-01 6.76441729e-01
2.26495564e-01 6.79862678e-01 2.23055854e-01 -3.42293710e-01
-1.16307355e-01 -1.50275767e-01 5.30170560e-01 -4.80466634e-01
-4.89051968e-01 2.08511695e-01 -3.38984430e-01 -1.06858790e+00
-9.97128963e-01 -7.22793162e-01 -1.68438625e+00 7.02440068e-02
-3.87958914e-01 3.15307081e-01 4.09506261e-01 9.55905139e-01
2.63443828e-01 4.85242277e-01 7.60594249e-01 -1.04163730e+00
1.23346284e-01 -4.68122005e-01 -6.62303507e-01 6.79408848e-01
3.69073451e-01 -4.16277021e-01 -3.22872579e-01 7.97363341e-01] | [9.108854293823242, -0.24289478361606598] |
79f6491d-3d40-4177-ba21-7cf1ab01c0a6 | building-goal-oriented-dialogue-systems-with | 2111.11576 | null | https://arxiv.org/abs/2111.11576v1 | https://arxiv.org/pdf/2111.11576v1.pdf | Building Goal-Oriented Dialogue Systems with Situated Visual Context | Most popular goal-oriented dialogue agents are capable of understanding the conversational context. However, with the surge of virtual assistants with screen, the next generation of agents are required to also understand screen context in order to provide a proper interactive experience, and better understand users' goals. In this paper, we propose a novel multimodal conversational framework, where the dialogue agent's next action and their arguments are derived jointly conditioned both on the conversational and the visual context. Specifically, we propose a new model, that can reason over the visual context within a conversation and populate API arguments with visual entities given the user query. Our model can recognize visual features such as color and shape as well as the metadata based features such as price or star rating associated with a visual entity. In order to train our model, due to a lack of suitable multimodal conversational datasets, we also propose a novel multimodal dialog simulator to generate synthetic data and also collect realistic user data from MTurk to improve model robustness. The proposed model achieves a reasonable 85% model accuracy, without high inference latency. We also demonstrate the proposed approach in a prototypical furniture shopping experience for a multimodal virtual assistant. | ['Tagyoung Chung', 'Shuyang Gao', 'Emre Barut', 'Arijit Biswas', 'Jan Jezabek', 'Sanchit Agarwal'] | 2021-11-22 | null | null | null | null | ['goal-oriented-dialogue-systems'] | ['natural-language-processing'] | [-6.62105456e-02 9.57057104e-02 2.50393689e-01 -6.18796527e-01
-3.99478495e-01 -9.81429279e-01 8.40713322e-01 2.23166212e-01
-3.20573866e-01 5.33761561e-01 3.32927436e-01 -3.46608192e-01
1.69068590e-01 -6.60975635e-01 -3.92856508e-01 -2.13811010e-01
2.47258008e-01 6.32342458e-01 1.34172723e-01 -5.60548306e-01
2.31283084e-01 2.29432359e-01 -1.80457747e+00 7.14653313e-01
9.70537305e-01 1.00633800e+00 6.69944644e-01 1.04537094e+00
-5.46271741e-01 8.05703282e-01 -8.55950713e-01 -4.12037820e-01
-6.99380785e-02 -2.91674078e-01 -7.38397717e-01 3.09249341e-01
8.32378194e-02 -7.12623298e-01 -2.30136976e-01 6.78630710e-01
4.70370293e-01 4.26205546e-01 5.95941126e-01 -1.61408401e+00
-4.14027482e-01 3.39569598e-01 -1.27095625e-01 -3.99028748e-01
9.10083473e-01 5.02811551e-01 1.02887189e+00 -7.24925518e-01
6.97849751e-01 1.57932734e+00 -5.89685813e-02 6.85228407e-01
-9.57571566e-01 -3.41282874e-01 7.19342887e-01 2.44092032e-01
-1.06762850e+00 -3.75723034e-01 8.33031058e-01 -2.57256389e-01
6.01374030e-01 6.65223420e-01 7.67429352e-01 1.24364698e+00
-4.79230195e-01 1.23164558e+00 9.03665125e-01 -4.68839467e-01
2.93282419e-01 6.39655709e-01 4.99138683e-02 8.11238408e-01
-4.82821822e-01 -4.58050728e-01 -4.96225446e-01 -4.32009697e-01
6.63286328e-01 1.02401718e-01 -2.70403236e-01 -5.75872242e-01
-1.25232077e+00 7.16322362e-01 1.39981046e-01 -2.75552105e-02
-3.17683011e-01 -2.33404651e-01 2.13542864e-01 4.75995280e-02
-2.14475691e-01 2.22518921e-01 -2.36202955e-01 -5.21108985e-01
-1.23254120e-01 4.34963465e-01 1.21485555e+00 1.21400106e+00
5.25069952e-01 -3.64777833e-01 -4.41094428e-01 1.09796333e+00
6.42923892e-01 7.05070615e-01 1.95219919e-01 -9.42230165e-01
7.64065385e-01 1.03615737e+00 7.62359440e-01 -1.05927002e+00
-5.38128436e-01 3.90619695e-01 -3.45786482e-01 2.03247830e-01
6.65571034e-01 -2.60918528e-01 -7.27296829e-01 1.70811689e+00
5.16695678e-01 -2.21720129e-01 3.63726050e-01 1.05331481e+00
1.23371112e+00 8.11211884e-01 1.45540193e-01 -1.68234780e-01
1.72473013e+00 -6.72964811e-01 -9.16412711e-01 -2.30195373e-01
4.91864204e-01 -8.55318725e-01 1.62932634e+00 3.35470289e-01
-9.36829686e-01 -4.50150937e-01 -8.78974974e-01 -1.18167102e-01
-4.88343120e-01 4.73656744e-01 7.35861421e-01 4.54163045e-01
-6.49639666e-01 -4.71926667e-02 -5.28704643e-01 -6.50660574e-01
-3.33618104e-01 2.90977120e-01 -3.19403931e-02 -2.45430246e-02
-1.25738680e+00 7.81160414e-01 1.62723348e-01 2.73442030e-01
-7.23202288e-01 -6.70980215e-02 -9.45007265e-01 1.01409473e-01
6.98698699e-01 -5.78392088e-01 1.51630402e+00 -6.38871789e-01
-1.72925699e+00 4.32760209e-01 -2.06575021e-01 1.19581550e-01
6.45116508e-01 -1.96847431e-02 -3.56929123e-01 5.74080087e-02
-4.70738143e-01 5.90583742e-01 6.05276942e-01 -1.74500406e+00
-9.99489784e-01 -5.44209957e-01 8.22816849e-01 8.20991874e-01
-3.10697079e-01 -2.50548869e-01 -1.05804682e+00 -5.79054393e-02
-2.49288023e-01 -1.14114141e+00 9.82335024e-03 -1.08790465e-01
-5.01451433e-01 -3.25601995e-01 7.99342990e-01 -6.62822127e-01
1.04709351e+00 -2.13510609e+00 2.06162333e-01 2.16466263e-01
1.03377081e-01 7.92247504e-02 -1.24264672e-01 5.34417272e-01
6.01235449e-01 -2.09744006e-01 1.36567414e-01 -5.30374229e-01
4.07302737e-01 2.43260324e-01 -8.71864557e-02 -1.90155566e-01
-1.95923582e-01 7.16639936e-01 -7.18307436e-01 -6.65536523e-01
5.02060115e-01 5.27223825e-01 -3.92930269e-01 8.15423310e-01
-5.57211399e-01 3.26581150e-01 -6.90471053e-01 4.98503119e-01
5.22345901e-01 -2.28316456e-01 6.39931679e-01 -3.68145287e-01
-1.85044575e-02 -3.46612334e-02 -1.45677066e+00 1.74709511e+00
-8.46694231e-01 4.17480528e-01 4.79208589e-01 -3.30185384e-01
9.40912962e-01 2.68029988e-01 1.75769255e-01 -6.34271085e-01
1.77729696e-01 -3.12268168e-01 -1.79735377e-01 -8.27172577e-01
8.85226667e-01 5.74769735e-01 -2.35781297e-01 4.31291342e-01
-5.00002921e-01 -7.03786090e-02 1.29065707e-01 5.16790152e-01
6.28852606e-01 1.00815393e-01 3.77317667e-02 3.74326110e-01
6.61136389e-01 8.36376026e-02 1.52606443e-01 6.98495030e-01
1.89084355e-02 3.06894779e-01 6.09066010e-01 -3.06739002e-01
-7.97680199e-01 -9.09102440e-01 4.17039603e-01 1.36623394e+00
5.33519387e-01 -3.82696182e-01 -7.34798789e-01 -7.08856106e-01
-9.31475963e-03 8.33132625e-01 -3.04338008e-01 1.84553176e-01
-3.95933598e-01 -5.19073725e-01 1.22116648e-01 2.24949613e-01
5.38060427e-01 -1.19447291e+00 -6.34062290e-01 9.74629298e-02
-5.61991930e-01 -1.31528974e+00 -4.96841162e-01 -3.66125107e-01
-3.50889772e-01 -1.23600554e+00 -6.16181850e-01 -8.35396349e-01
5.88645160e-01 1.86601922e-01 8.74846399e-01 2.28098884e-01
-9.61950943e-02 7.71404684e-01 -3.93779397e-01 -1.24558806e-01
-6.16220653e-01 -2.29425095e-02 -2.17300773e-01 2.80420154e-01
2.97653496e-01 -6.42756522e-02 -7.77006209e-01 5.89661717e-01
-8.46842349e-01 5.35498142e-01 3.68714571e-01 6.49367452e-01
1.17369637e-01 -3.32946777e-01 2.81326592e-01 -7.04011619e-01
1.17576492e+00 -2.04895660e-01 -5.57219565e-01 7.59828627e-01
-3.17414850e-01 1.03462204e-01 6.00539327e-01 -6.50997519e-01
-1.48773909e+00 2.02727273e-01 7.27612674e-02 7.57799447e-02
-4.20407534e-01 3.89782339e-01 -6.03252351e-01 3.48820478e-01
3.10443670e-01 1.80961385e-01 8.20217580e-02 -4.31898952e-01
6.63307965e-01 1.15362144e+00 4.48339731e-01 -7.93156505e-01
4.70768660e-01 3.95897925e-01 -3.94248635e-01 -8.95235837e-01
-9.19816345e-02 -4.37947243e-01 -2.93078095e-01 -5.16047418e-01
8.81436169e-01 -7.61223674e-01 -1.80859482e+00 3.72933060e-01
-1.40250075e+00 -3.03409725e-01 4.67602462e-01 5.14215052e-01
-5.59659183e-01 5.36296129e-01 -4.44014579e-01 -1.26934791e+00
-1.55555695e-01 -1.44400394e+00 1.00896823e+00 4.41870451e-01
-1.14556856e-01 -8.20032120e-01 -3.11772615e-01 5.96811295e-01
2.35951856e-01 3.12011465e-02 9.26561058e-01 -8.17387283e-01
-6.02126241e-01 -1.38323024e-01 -3.89633626e-01 -3.05554688e-01
2.05848500e-01 1.23291552e-01 -8.34784865e-01 -9.13334116e-02
-3.99111301e-01 -2.84601480e-01 1.40302524e-01 1.90355703e-02
8.15862596e-01 -2.84878761e-01 -2.86925733e-01 1.21731766e-01
8.72691035e-01 7.37957537e-01 3.75235736e-01 2.28678375e-01
7.86141753e-01 8.13323617e-01 1.01976490e+00 8.34852517e-01
9.49048042e-01 1.06348205e+00 5.32754362e-01 -1.74949870e-01
2.40647703e-01 -3.38407904e-01 1.05340771e-01 3.32097203e-01
-1.09092630e-01 -6.37206495e-01 -6.49638593e-01 2.40144894e-01
-2.15293217e+00 -6.89210594e-01 4.95152920e-02 2.14939880e+00
5.94123065e-01 7.22547695e-02 3.20675701e-01 -3.67283463e-01
7.81430781e-01 4.00349451e-03 -6.02858782e-01 -3.81615609e-01
1.29280373e-01 -5.90752006e-01 -2.55385842e-02 7.85001516e-01
-7.13692665e-01 1.01984513e+00 5.28353214e+00 4.66139406e-01
-7.80550659e-01 -4.03527081e-01 4.91130322e-01 -8.97846837e-03
-4.98160005e-01 -2.37401605e-01 -8.34975362e-01 2.81995803e-01
4.52196389e-01 8.55689421e-02 9.01618719e-01 8.40945423e-01
4.18690681e-01 -2.79209852e-01 -1.23003042e+00 1.27077031e+00
8.40542689e-02 -1.13292766e+00 7.71595612e-02 4.54250239e-02
2.12511718e-02 -7.06954241e-01 8.07362609e-04 5.25103748e-01
3.59740883e-01 -8.17693233e-01 3.77351910e-01 5.52589953e-01
5.25631309e-01 -7.55566061e-01 4.37320143e-01 3.80410880e-01
-1.21319556e+00 -1.46195889e-02 -4.00925800e-03 1.33569643e-01
1.37670860e-01 -4.10586149e-01 -1.36780906e+00 2.62454420e-01
4.66650426e-01 2.17718303e-01 -5.23770750e-01 6.83719039e-01
3.45756626e-03 -9.34338197e-02 -1.96620569e-01 -6.98151469e-01
5.37009947e-02 -3.10043514e-01 3.83421600e-01 9.85936880e-01
2.79012859e-01 3.00208956e-01 4.09573972e-01 6.68260992e-01
1.50022402e-01 5.23414552e-01 -4.06997800e-01 -1.99230373e-01
6.21161997e-01 1.41822469e+00 -5.12257516e-01 -4.64698881e-01
-5.39072454e-01 1.16886330e+00 2.13090017e-01 6.41034663e-01
-9.42953825e-01 -2.85206944e-01 8.18606794e-01 -1.41377211e-01
-5.85166086e-03 -2.86085844e-01 9.44399834e-02 -1.11268771e+00
1.05160348e-01 -1.18541229e+00 3.67979884e-01 -1.11014485e+00
-8.38550985e-01 7.86403894e-01 -8.93899500e-02 -1.09819996e+00
-4.56073880e-01 -6.87395096e-01 -4.02982533e-01 9.90439832e-01
-1.14244759e+00 -1.34796703e+00 -7.89715230e-01 8.16011310e-01
7.13310599e-01 -2.71814853e-01 9.53596830e-01 -7.68138766e-02
-4.63280737e-01 3.60739172e-01 -1.60820484e-01 1.32654905e-01
8.80519092e-01 -1.23924541e+00 2.54760772e-01 5.96753173e-02
-5.08149415e-02 7.13096738e-01 9.09695446e-01 -5.62962234e-01
-1.80550086e+00 -3.84399384e-01 4.07292426e-01 -4.98482168e-01
3.45055223e-01 -6.09553516e-01 -7.84981191e-01 4.25062031e-01
3.88934374e-01 -5.72204947e-01 5.66444993e-01 3.14066201e-01
-2.43926253e-02 -8.22200552e-02 -1.11983418e+00 1.10067916e+00
8.92675757e-01 -5.07423580e-01 -5.82602799e-01 1.96120784e-01
6.60603940e-01 -7.06746399e-01 -6.25168383e-01 1.63343638e-01
8.04531276e-01 -8.85470927e-01 9.68302011e-01 -6.95913732e-01
5.70028871e-02 -3.71935308e-01 -2.38722578e-01 -1.28081691e+00
2.28307620e-01 -6.43448234e-01 -3.73050161e-02 1.35115361e+00
5.94570577e-01 -2.19671547e-01 8.86098325e-01 1.26647770e+00
1.85242400e-01 -3.28964800e-01 -4.67703521e-01 -1.53034747e-01
-8.39125395e-01 -4.19185907e-01 1.02873802e+00 6.97347224e-01
4.89487052e-01 5.79231560e-01 -6.25193477e-01 4.28761005e-01
1.15048267e-01 2.84128368e-01 1.32880425e+00 -1.13984478e+00
-2.96893209e-01 -1.40816465e-01 1.73521303e-02 -1.50011706e+00
-4.97375801e-02 -2.90966511e-01 6.26742318e-02 -1.88956547e+00
2.45940387e-01 -4.09425825e-01 6.58144578e-02 1.41976073e-01
-2.26424143e-01 -3.29809904e-01 3.86574566e-01 -5.61549291e-02
-8.09947014e-01 4.57255036e-01 1.43417192e+00 -2.81597167e-01
-7.77800620e-01 2.18116343e-01 -4.86752510e-01 6.46558464e-01
5.16303658e-01 2.91651696e-01 -8.35930586e-01 -2.40600526e-01
2.45852381e-01 7.60313213e-01 3.71914059e-01 -4.43730861e-01
3.11643392e-01 -5.90899110e-01 3.62317324e-01 -6.87679827e-01
9.51557338e-01 -1.09706736e+00 -1.09277256e-01 5.94861321e-02
-5.81504166e-01 1.89860180e-01 1.42807946e-01 6.20093167e-01
2.69926116e-02 -4.88286242e-02 -1.84471924e-02 -1.03031367e-01
-8.85461271e-01 4.62745912e-02 -5.42560637e-01 -2.27468312e-01
1.06070232e+00 -1.27744796e-02 -5.50649464e-01 -1.01300657e+00
-9.54213440e-01 7.49297202e-01 5.13252914e-01 8.45665216e-01
8.01384449e-01 -1.19578874e+00 -2.83984065e-01 -6.85472367e-03
4.65642095e-01 -1.96474090e-01 4.26502556e-01 1.92449689e-01
-4.40258384e-01 3.48310530e-01 -9.50062945e-02 -5.59168577e-01
-1.76577055e+00 7.62700438e-01 1.09391250e-01 6.83115050e-02
-2.82873183e-01 3.31531227e-01 4.56063241e-01 -6.64597571e-01
6.79993212e-01 -4.37901765e-01 -7.62215436e-01 8.06898475e-02
6.73627555e-01 4.23221178e-02 -2.79584885e-01 -6.23393416e-01
-1.81527808e-01 1.93230078e-01 8.48418847e-02 -4.11362559e-01
8.52333844e-01 -5.01185000e-01 3.75002056e-01 5.29076040e-01
7.44502604e-01 9.79108736e-02 -1.31020963e+00 -1.45907193e-01
-2.21537158e-01 -5.52237928e-01 -3.81597281e-01 -1.08462954e+00
-5.66315234e-01 7.17692018e-01 7.65934646e-01 5.06234050e-01
8.20546448e-01 -5.98675385e-02 6.22434139e-01 6.91528320e-01
3.23421538e-01 -1.29375279e+00 2.47748882e-01 3.31916839e-01
9.15740192e-01 -1.54312789e+00 -3.47338378e-01 -5.45202076e-01
-1.34394979e+00 1.07941604e+00 9.65417922e-01 7.53945947e-01
9.58087742e-02 -7.16396645e-02 4.49448854e-01 -1.13515906e-01
-8.17930937e-01 -3.23340774e-01 2.78377384e-01 6.89495862e-01
3.64345908e-01 8.69871974e-02 -2.72122249e-02 7.37625062e-01
-6.07122630e-02 -3.03602397e-01 4.61188108e-01 7.74845541e-01
-4.07180130e-01 -1.03000641e+00 -4.64201212e-01 3.83732878e-02
1.25584871e-01 8.08164030e-02 -5.92810988e-01 8.53886902e-01
-3.11843336e-01 1.41285205e+00 6.40156269e-02 -4.65334296e-01
8.07502508e-01 1.14130162e-01 2.96558470e-01 -4.70062345e-01
-5.45491934e-01 9.71561596e-02 4.92110699e-01 -3.48660380e-01
-1.83022454e-01 -4.01495218e-01 -1.42303932e+00 -3.42525214e-01
-9.42595974e-02 3.21220607e-01 1.00596881e+00 7.65983343e-01
2.15532809e-01 3.04515123e-01 7.04119444e-01 -6.41997278e-01
-3.03961903e-01 -1.03130817e+00 -3.04063201e-01 8.15382957e-01
9.11353901e-02 -5.94754696e-01 -1.68494016e-01 -8.40578750e-02] | [10.990988731384277, 1.2674751281738281] |
c13c0f14-5ce9-47d5-a232-e4a5f1982901 | object-augmented-rgb-d-slam-for-wide | 2108.02522 | null | https://arxiv.org/abs/2108.02522v1 | https://arxiv.org/pdf/2108.02522v1.pdf | Object-Augmented RGB-D SLAM for Wide-Disparity Relocalisation | We propose a novel object-augmented RGB-D SLAM system that is capable of constructing a consistent object map and performing relocalisation based on centroids of objects in the map. The approach aims to overcome the view dependence of appearance-based relocalisation methods using point features or images. During the map construction, we use a pre-trained neural network to detect objects and estimate 6D poses from RGB-D data. An incremental probabilistic model is used to aggregate estimates over time to create the object map. Then in relocalisation, we use the same network to extract objects-of-interest in the `lost' frames. Pairwise geometric matching finds correspondences between map and frame objects, and probabilistic absolute orientation followed by application of iterative closest point to dense depth maps and object centroids gives relocalisation. Results of experiments in desktop environments demonstrate very high success rates even for frames with widely different viewpoints from those used to construct the map, significantly outperforming two appearance-based methods. | ['Andrew Calway', 'Xingrui Yang', 'Yuhang Ming'] | 2021-08-05 | null | null | null | null | ['geometric-matching'] | ['computer-vision'] | [ 2.78934449e-01 1.38723310e-02 2.29152054e-01 -6.53521955e-01
-6.69551849e-01 -4.93931115e-01 7.57821441e-01 2.35834956e-01
-6.63834393e-01 5.17070830e-01 -1.52679428e-01 3.08426738e-01
-2.05516055e-01 -7.78181434e-01 -8.55465233e-01 -4.70942914e-01
-3.23837847e-02 1.16739523e+00 7.88730145e-01 -1.07326142e-01
5.72170794e-01 1.11411607e+00 -1.76975358e+00 6.17329590e-02
1.17695257e-01 9.99165773e-01 4.83102709e-01 8.11211348e-01
-3.88252698e-02 7.09714353e-01 -4.53371644e-01 -3.69004614e-04
4.39932942e-01 -1.13924429e-01 -6.52076185e-01 1.40345350e-01
9.14421320e-01 -5.46439946e-01 -3.42827439e-01 7.97202051e-01
4.59475607e-01 2.90408164e-01 4.53012764e-01 -1.38849676e+00
-6.00479990e-02 -1.25399768e-01 -5.46404898e-01 -1.09438218e-01
9.50521708e-01 -2.47787356e-01 6.30469561e-01 -1.10398138e+00
9.54451621e-01 1.25442755e+00 9.48379695e-01 1.97141975e-01
-1.19702721e+00 -3.26677710e-01 -1.52826980e-01 2.60085613e-01
-1.80065525e+00 -6.78128898e-01 9.09064710e-01 -3.10624421e-01
1.05862987e+00 2.40467727e-01 7.11230993e-01 4.54044729e-01
2.00499490e-01 4.37472463e-01 8.94532859e-01 -7.27070212e-01
4.10785079e-01 9.05191451e-02 -3.73529851e-01 6.51899695e-01
6.22246638e-02 -6.67404830e-02 -8.40456843e-01 -3.75457466e-01
1.15715313e+00 2.79123962e-01 -2.18670174e-01 -1.26103640e+00
-1.53802490e+00 5.47522366e-01 6.92358315e-01 5.06231599e-02
-6.77338421e-01 5.87695539e-01 -2.08814573e-02 -1.31654203e-01
3.78244936e-01 1.15853354e-01 -3.74104619e-01 -7.79721281e-03
-9.57184315e-01 2.70760030e-01 4.73232031e-01 1.18657625e+00
1.26696885e+00 -4.39771950e-01 3.98420155e-01 3.70026410e-01
5.66261351e-01 7.37702489e-01 1.85598522e-01 -1.46548772e+00
1.41415119e-01 5.86232483e-01 3.13991934e-01 -1.34561157e+00
-5.01944780e-01 2.91008651e-01 -4.67828512e-01 6.33426726e-01
3.24423522e-01 5.00033140e-01 -9.23397064e-01 1.28259504e+00
7.36827552e-01 8.93711820e-02 -7.53358006e-02 9.18585956e-01
5.40051877e-01 3.57051283e-01 -4.70036954e-01 1.08786210e-01
1.11583185e+00 -4.19858515e-01 -3.61921161e-01 -1.54232621e-01
4.08039182e-01 -8.39145362e-01 1.84454769e-01 4.01667982e-01
-1.20505834e+00 -7.45916724e-01 -1.03978062e+00 -9.66968536e-02
-2.05614761e-01 -9.35169309e-02 3.99237812e-01 2.74017781e-01
-1.52570105e+00 7.04441845e-01 -1.00528193e+00 -5.72216213e-01
1.47551835e-01 6.30980074e-01 -8.80724192e-01 -1.04281768e-01
-6.27676845e-01 1.18361413e+00 9.18301105e-01 2.05599219e-01
-8.88143778e-01 -2.97175646e-01 -9.79263484e-01 -5.90645969e-01
1.65456086e-01 -7.84795582e-01 9.69841182e-01 -9.48962688e-01
-1.34754336e+00 1.05777860e+00 -4.56206053e-01 -2.21790612e-01
5.20889342e-01 -3.13343793e-01 1.15420774e-01 4.03034866e-01
2.06121370e-01 9.92679894e-01 7.76880026e-01 -1.72473001e+00
-1.12794268e+00 -5.55204749e-01 1.63585067e-01 5.61796904e-01
3.89858663e-01 -1.45565316e-01 -6.46388233e-01 1.71536773e-01
1.27967024e+00 -1.05231488e+00 -2.04115435e-01 4.03881639e-01
-1.38015568e-01 -2.25932583e-01 9.51174498e-01 -6.25824094e-01
2.36534923e-01 -1.85612929e+00 3.44835788e-01 4.92323756e-01
2.82042682e-01 -3.62615049e-01 1.32392257e-01 2.40992844e-01
-3.81444627e-03 -5.44483364e-01 1.10263146e-01 -7.10845649e-01
-1.46114975e-01 4.34852362e-01 5.00561893e-02 1.03700459e+00
-2.29720473e-02 6.48413897e-01 -1.12332225e+00 -4.90383327e-01
7.97419429e-01 6.65689409e-01 -1.72763675e-01 1.76399112e-01
-1.01013832e-01 2.79618263e-01 -1.67372897e-01 5.66368282e-01
8.83444726e-01 1.94491118e-01 -7.48005807e-02 -3.90053093e-01
-2.33575121e-01 2.14319512e-01 -1.55451155e+00 2.44002986e+00
-2.60971308e-01 7.30925143e-01 -1.69985861e-01 -5.29759824e-01
1.01371920e+00 1.77130640e-01 7.26649046e-01 -5.26367247e-01
8.64885896e-02 2.23658293e-01 -5.41781187e-01 -8.61950293e-02
8.74221802e-01 1.12332441e-01 2.07244813e-01 5.44081867e-01
2.36805350e-01 -6.24997795e-01 -2.46415406e-01 2.18696073e-01
9.63476121e-01 9.15906131e-01 3.64420086e-01 -1.53888704e-03
5.04292965e-01 1.38543531e-01 2.60311037e-01 6.99711144e-01
-1.11215688e-01 9.93227720e-01 -1.47352263e-01 -1.06750369e+00
-1.22385812e+00 -1.39195180e+00 -1.76674768e-01 6.85690284e-01
6.92563713e-01 -1.36687607e-01 -2.87453771e-01 -4.85077083e-01
4.47789095e-02 5.33889174e-01 -6.05033398e-01 9.29830968e-02
-6.73771203e-01 -2.72734463e-01 -3.79283950e-02 4.49875832e-01
3.31427157e-01 -8.81098330e-01 -1.10030031e+00 2.36943975e-01
-3.17340165e-01 -7.92731941e-01 1.36487648e-01 4.29808855e-01
-7.63333499e-01 -9.62535679e-01 -6.94026172e-01 -6.41110063e-01
1.00573075e+00 5.66189885e-01 1.03925312e+00 -2.02838495e-01
-9.68613848e-02 6.38919532e-01 -3.92948121e-01 -4.11968529e-01
-3.15808326e-01 -2.64646888e-01 4.25075501e-01 -1.72450319e-01
4.05663520e-01 -5.78287899e-01 -5.58577657e-01 5.91435254e-01
-4.30235445e-01 9.80104133e-02 3.86218995e-01 2.70902961e-01
7.79416025e-01 -1.88385129e-01 -2.32854933e-01 -1.02495290e-02
-1.82868659e-01 -2.19451878e-02 -8.22763681e-01 1.16484262e-01
-1.79128155e-01 -3.51795852e-02 -2.69811720e-01 -2.11442724e-01
-6.83302283e-01 7.47082651e-01 1.53522015e-01 -5.64106703e-01
-2.97208160e-01 -1.62397116e-01 5.46431541e-02 -4.73131716e-01
7.70863593e-01 1.77497402e-01 -5.10926992e-02 -1.87462956e-01
5.11621773e-01 4.48364854e-01 7.52242029e-01 -4.06718075e-01
8.96366298e-01 9.50999379e-01 1.15287416e-01 -4.78289574e-01
-4.45607156e-01 -7.55254984e-01 -1.41535306e+00 -5.61996818e-01
9.58209813e-01 -1.05495310e+00 -7.25516498e-01 2.66855359e-01
-1.71689665e+00 2.43571401e-03 -4.56016898e-01 7.98910737e-01
-1.02548075e+00 2.35459059e-01 -1.07048519e-01 -9.56223845e-01
-6.71206461e-03 -1.06208849e+00 1.47097957e+00 6.11821935e-02
-2.36786544e-01 -6.90866709e-01 4.23461944e-01 -4.74770069e-02
-3.21375695e-03 4.72271085e-01 6.60693645e-02 -3.50365549e-01
-1.01720142e+00 -4.10349399e-01 -2.29252636e-01 1.87156145e-02
1.17396869e-01 -9.54843238e-02 -1.29113019e+00 -1.99181855e-01
4.24381793e-02 1.34254145e-02 4.73291546e-01 4.73607600e-01
4.40440744e-01 2.65617341e-01 -6.60051167e-01 5.81551790e-01
1.63211882e+00 3.79874468e-01 6.23697579e-01 6.82815313e-01
7.76917934e-01 6.44312143e-01 6.80142999e-01 3.50480437e-01
5.07334948e-01 8.89990687e-01 9.23531473e-01 1.54095814e-02
1.47055030e-01 -1.06727324e-01 1.08199738e-01 5.61730623e-01
-4.65869129e-01 -7.11216480e-02 -1.04341114e+00 4.65046853e-01
-1.96935380e+00 -7.26128519e-01 -2.37082943e-01 2.51684642e+00
3.83313745e-01 1.45867258e-01 -1.74932815e-02 -5.88517562e-02
8.58109593e-01 -1.35148838e-01 -1.84135899e-01 -2.96350215e-02
5.90596022e-03 -1.23404242e-01 8.28362882e-01 7.63084710e-01
-9.35909867e-01 8.75929713e-01 5.99465656e+00 3.74427199e-01
-6.46698356e-01 2.15283796e-01 1.76508427e-02 -6.34025186e-02
1.24588601e-01 3.49775225e-01 -7.53091037e-01 1.43901296e-02
5.88211358e-01 4.03311074e-01 2.80001372e-01 9.07397330e-01
-1.00754946e-01 -7.21975505e-01 -1.30841732e+00 1.28407025e+00
5.22877753e-01 -1.26670885e+00 -3.21983904e-01 9.08258557e-02
7.21405327e-01 4.67616200e-01 -6.16517305e-01 -3.15363824e-01
3.48495632e-01 -4.30718839e-01 1.19439936e+00 8.81946027e-01
5.25353670e-01 -8.53572667e-01 8.64401102e-01 3.70862275e-01
-1.26818395e+00 3.64895254e-01 -7.92079926e-01 -8.71668458e-02
4.14928973e-01 4.72354382e-01 -1.30581570e+00 6.56656623e-01
8.78469229e-01 5.21296740e-01 -5.11704206e-01 1.05438733e+00
-2.12969437e-01 -4.93310481e-01 -8.23769391e-01 3.14448953e-01
6.30867034e-02 -1.81943133e-01 4.54169452e-01 7.08750069e-01
4.44797516e-01 -2.12269738e-01 2.05143884e-01 6.68441772e-01
3.72997046e-01 -2.81622320e-01 -6.89416528e-01 8.42040241e-01
4.66471970e-01 1.29469073e+00 -1.24064398e+00 -4.63861465e-01
-6.22856244e-02 1.41556990e+00 1.95896298e-01 5.73145896e-02
-8.38778913e-01 -3.14216882e-01 2.37313688e-01 1.71443895e-01
2.87855983e-01 -6.05892718e-01 2.38910213e-01 -7.46604562e-01
1.66327938e-01 -1.74106196e-01 -1.01421297e-01 -1.49716938e+00
-6.86153412e-01 6.17613196e-01 2.59157598e-01 -1.36375642e+00
-3.66913557e-01 -4.17804122e-01 -1.17491923e-01 1.00002205e+00
-1.18873072e+00 -1.25699651e+00 -7.23100722e-01 5.10542095e-01
2.61529803e-01 2.59755671e-01 7.91691065e-01 -7.48612955e-02
4.11550164e-01 -2.69602656e-01 2.17238367e-01 -6.73511252e-03
7.22565591e-01 -1.35231555e+00 4.95133668e-01 5.32587588e-01
5.53861916e-01 3.66895378e-01 5.13443708e-01 -7.93978572e-01
-1.31764376e+00 -7.81019807e-01 8.70586872e-01 -1.20208144e+00
2.22923934e-01 -6.04466498e-01 -5.63042462e-01 8.86186421e-01
-1.03632249e-01 2.59304315e-01 -6.03875257e-02 -2.25445822e-01
-9.07263681e-02 -1.59821928e-01 -1.20263338e+00 1.57255828e-01
1.19353521e+00 -6.51910841e-01 -6.36705339e-01 5.19672811e-01
6.94432378e-01 -9.21287477e-01 -6.93374276e-01 2.78138995e-01
8.40699613e-01 -1.20654666e+00 1.36847484e+00 1.13472961e-01
-1.85284838e-01 -7.48458862e-01 -5.29814899e-01 -9.76306319e-01
-2.45855883e-01 -3.14200759e-01 8.58672708e-02 9.26892757e-01
-2.48874992e-01 -4.83653843e-01 9.91633892e-01 3.73061299e-01
2.90426202e-02 -1.09689578e-01 -1.34287655e+00 -6.09558582e-01
-7.39446640e-01 -5.59256911e-01 3.89430612e-01 8.22247922e-01
-3.93302411e-01 -3.96152176e-02 -1.23463117e-01 5.30166566e-01
9.78309035e-01 -9.75114033e-02 1.25184488e+00 -1.41861451e+00
9.22987014e-02 9.48790181e-03 -1.21197927e+00 -9.75583017e-01
2.83983126e-02 -6.54416800e-01 5.01535892e-01 -1.72409701e+00
7.98934978e-03 -8.12621713e-01 -1.55326709e-01 5.13943553e-01
3.99688959e-01 7.44519472e-01 9.74650830e-02 4.58262801e-01
-9.65327859e-01 3.04857790e-01 6.01307631e-01 8.84443223e-02
-2.64912009e-01 -2.77410358e-01 1.89214408e-01 1.04385269e+00
4.79808629e-01 -8.27081919e-01 -1.78897500e-01 -4.96390998e-01
3.34887058e-01 -5.81093691e-03 8.00906360e-01 -1.48712647e+00
4.09325719e-01 7.60485455e-02 8.92781734e-01 -1.39951289e+00
9.55114901e-01 -1.27466798e+00 6.81065321e-01 3.93268287e-01
5.62459454e-02 4.01010692e-01 3.38814557e-02 5.80485702e-01
1.00349255e-01 -4.31051642e-01 4.30803716e-01 -3.63319784e-01
-1.00434208e+00 1.83349326e-01 -1.22879088e-01 -6.03152752e-01
1.21268225e+00 -8.07498991e-01 3.23458254e-01 -3.73769462e-01
-8.12851429e-01 -2.97309965e-01 9.90173399e-01 3.31711590e-01
9.53046739e-01 -1.53768420e+00 -5.93839407e-01 3.49737018e-01
2.11780697e-01 7.43826985e-01 3.41158500e-03 9.66035903e-01
-1.17669618e+00 2.62124687e-01 -1.88922554e-01 -1.55447567e+00
-1.30695403e+00 5.42501509e-01 4.17209327e-01 4.40737754e-01
-7.17443824e-01 9.70682323e-01 7.30091333e-02 -6.00152075e-01
1.35681361e-01 -3.72566670e-01 3.02594602e-01 -8.53773430e-02
5.37014008e-01 4.47613984e-01 2.91425109e-01 -1.19614649e+00
-7.42915213e-01 9.32136357e-01 1.88097998e-01 -5.23628175e-01
1.30462158e+00 -4.71665114e-01 -2.85696566e-01 6.69092417e-01
1.17150056e+00 -1.93898350e-01 -1.51864278e+00 -5.10792375e-01
-6.31754994e-02 -8.05218935e-01 6.37962371e-02 -3.81913692e-01
-5.45759141e-01 6.21444345e-01 1.01337600e+00 -1.64736211e-01
7.93297708e-01 1.19580396e-01 8.81214738e-02 6.55949175e-01
1.06825054e+00 -8.91113281e-01 3.10420385e-03 3.77404064e-01
8.85123134e-01 -1.20945060e+00 5.10717809e-01 -6.46198466e-02
-4.48937267e-02 1.29275930e+00 3.93012643e-01 -3.05586696e-01
3.58357847e-01 1.03720129e-01 2.77932286e-02 -2.56475717e-01
-8.80253909e-04 -6.65966496e-02 1.92068189e-01 9.16725755e-01
-2.07726330e-01 -3.30095172e-01 4.06663984e-01 -7.28417337e-01
-5.76340668e-02 -1.43349051e-01 3.07221115e-01 1.36432433e+00
-8.08572531e-01 -8.22236776e-01 -1.04135585e+00 -6.30839840e-02
2.94878870e-01 1.90568641e-01 -3.83334160e-01 9.84379888e-01
4.47957844e-01 5.65081954e-01 6.40727222e-01 -2.94577360e-01
1.60269722e-01 6.36577979e-02 9.88582015e-01 -2.63486683e-01
-7.38152564e-02 2.04952091e-01 -1.58796042e-01 -1.04925203e+00
-1.05831730e+00 -7.56959677e-01 -1.29976118e+00 -1.05741829e-01
-7.10158825e-01 -9.82066542e-02 1.19137096e+00 8.51878583e-01
1.66543782e-01 1.10287562e-01 5.06060362e-01 -1.88154876e+00
1.87506869e-01 -7.56868124e-01 -4.73558038e-01 2.74368674e-01
4.12629098e-01 -8.54411185e-01 -2.17472330e-01 2.39569306e-01] | [7.390466213226318, -2.3389484882354736] |
39d7a801-43e2-43ef-9956-12be8c6109c6 | robust-graph-learning-from-noisy-data | 1812.06673 | null | http://arxiv.org/abs/1812.06673v1 | http://arxiv.org/pdf/1812.06673v1.pdf | Robust Graph Learning from Noisy Data | Learning graphs from data automatically has shown encouraging performance on
clustering and semisupervised learning tasks. However, real data are often
corrupted, which may cause the learned graph to be inexact or unreliable. In
this paper, we propose a novel robust graph learning scheme to learn reliable
graphs from real-world noisy data by adaptively removing noise and errors in
the raw data. We show that our proposed model can also be viewed as a robust
version of manifold regularized robust PCA, where the quality of the graph
plays a critical role. The proposed model is able to boost the performance of
data clustering, semisupervised classification, and data recovery
significantly, primarily due to two key factors: 1) enhanced low-rank recovery
by exploiting the graph smoothness assumption, 2) improved graph construction
by exploiting clean data recovered by robust PCA. Thus, it boosts the
clustering, semi-supervised classification, and data recovery performance
overall. Extensive experiments on image/document clustering, object
recognition, image shadow removal, and video background subtraction reveal that
our model outperforms the previous state-of-the-art methods. | ['Haiqi Pan', 'Zenglin Xu', 'Zhao Kang', 'Steven C. H. Hoi'] | 2018-12-17 | null | null | null | null | ['shadow-removal', 'video-background-subtraction', 'imagedocument-clustering', 'image-shadow-removal'] | ['computer-vision', 'computer-vision', 'computer-vision', 'computer-vision'] | [ 3.26903760e-01 -2.22090259e-01 -1.26458360e-02 -2.05837324e-01
-8.62582803e-01 -3.64279062e-01 4.41274583e-01 1.63908258e-01
-5.22104912e-02 4.33898777e-01 1.95578769e-01 1.58590525e-01
-3.66015881e-01 -4.85861808e-01 -7.19841301e-01 -1.28130257e+00
1.11329347e-01 4.07049716e-01 1.61273420e-01 1.36829108e-01
1.30446970e-01 4.56293315e-01 -1.43236327e+00 -8.31896141e-02
9.91158366e-01 7.36573160e-01 2.56826818e-01 4.28884566e-01
-6.61657900e-02 9.19427574e-01 -3.17835808e-01 -5.48432814e-03
1.63185805e-01 -2.59364843e-01 -5.39921880e-01 9.17879045e-01
4.06197369e-01 1.05980784e-01 -5.20402372e-01 1.57168090e+00
2.59983569e-01 2.74808407e-01 4.41370010e-01 -1.26262009e+00
-6.19497240e-01 4.24501359e-01 -1.09764493e+00 -1.95783257e-01
1.67963848e-01 -1.98853999e-01 7.22282708e-01 -8.82110834e-01
6.13520384e-01 1.24171114e+00 4.04925674e-01 1.15420409e-01
-1.28162527e+00 -5.53436160e-01 2.61608690e-01 3.44896108e-01
-1.56382394e+00 -5.56008458e-01 1.17234552e+00 -3.41525674e-01
1.52859792e-01 2.58226335e-01 4.26107794e-01 6.72709703e-01
-2.49739513e-01 7.90242791e-01 1.30260932e+00 -5.58376551e-01
2.13670418e-01 -5.55670001e-02 3.97772998e-01 1.03623664e+00
5.30795515e-01 -2.65252352e-01 -5.10771871e-01 -3.18229288e-01
5.50294340e-01 2.26429582e-01 -7.25068271e-01 -8.19137871e-01
-1.10711467e+00 5.78876853e-01 3.81115019e-01 2.22745642e-01
-3.98478717e-01 -1.04291245e-01 1.39872715e-01 6.26720637e-02
5.24506807e-01 -1.52696976e-02 -3.36064361e-02 2.21999764e-01
-7.71896362e-01 -3.14127356e-01 7.48104513e-01 1.03656113e+00
8.84306014e-01 2.94932157e-01 1.95542470e-01 9.85594988e-01
3.66181493e-01 7.80088067e-01 3.42440546e-01 -8.33243906e-01
3.67049336e-01 7.12511420e-01 -2.31504515e-01 -1.46362448e+00
-2.64268994e-01 -3.42193961e-01 -1.19711328e+00 -3.26309539e-02
5.10573506e-01 2.25715041e-01 -9.74803090e-01 1.52080107e+00
4.38748419e-01 5.29488623e-01 -8.80413572e-04 9.41368163e-01
7.85462797e-01 6.09377801e-01 -4.02971655e-01 -6.57448530e-01
7.49860108e-01 -8.53930295e-01 -1.01887953e+00 -1.12888426e-01
2.49241471e-01 -7.41657138e-01 8.54735315e-01 7.48859406e-01
-6.49394751e-01 -4.98718053e-01 -9.50940192e-01 2.11153314e-01
-2.14010868e-02 2.37265661e-01 5.52676618e-01 5.49890995e-01
-8.33280623e-01 5.30310452e-01 -9.20533359e-01 -3.93047094e-01
2.77470648e-01 3.34650069e-01 -7.34336019e-01 -6.71110928e-01
-3.64748716e-01 3.08752298e-01 3.22719991e-01 2.79243082e-01
-6.10761583e-01 -1.65404510e-02 -8.28756213e-01 -1.05578452e-01
8.63459527e-01 -2.69594908e-01 3.43908131e-01 -8.83433282e-01
-1.18059492e+00 6.96385562e-01 -2.61497110e-01 -1.95369720e-01
2.43889317e-01 -2.36896113e-01 -4.11302060e-01 3.60335082e-01
2.07436718e-02 -8.50096345e-02 1.48299384e+00 -1.59684670e+00
-2.79218704e-01 -7.39241004e-01 -5.82507610e-01 2.01326191e-01
-4.09268826e-01 -2.52437115e-01 -1.02551699e+00 -7.49126375e-01
6.78116560e-01 -1.03698838e+00 -3.47423434e-01 -2.22137392e-01
-3.72089237e-01 1.20367475e-01 1.14148021e+00 -1.03314066e+00
9.81297791e-01 -2.44117570e+00 4.36905742e-01 6.03418589e-01
4.33657706e-01 7.67874196e-02 -2.16031939e-01 1.17438383e-01
-1.53448775e-01 -2.61480540e-01 -4.29608315e-01 -2.67037541e-01
-4.44435000e-01 4.84005988e-01 -9.16209072e-02 9.56655622e-01
-8.96826461e-02 4.92605150e-01 -9.55259979e-01 -5.14003992e-01
3.54413062e-01 4.71793354e-01 -3.58046629e-02 2.01348767e-01
2.09400252e-01 6.10317111e-01 -3.70958269e-01 7.06801295e-01
9.22423065e-01 -3.18149239e-01 4.07640815e-01 -4.10638660e-01
3.35212678e-01 -2.90016592e-01 -1.80186570e+00 1.64211023e+00
6.82396293e-02 4.71279413e-01 6.81406260e-01 -1.39292598e+00
9.46611166e-01 6.22249618e-02 6.66621327e-01 -3.99741918e-01
-7.60017149e-03 -7.34939501e-02 -2.89266169e-01 -3.05412859e-01
3.54498744e-01 2.98314333e-01 3.47027153e-01 2.15939507e-01
8.59530494e-02 -7.00753508e-03 2.61015356e-01 6.11436367e-01
1.16507864e+00 -8.05162489e-02 1.85308844e-01 -3.15215886e-01
6.17211401e-01 -1.65915012e-01 7.01159000e-01 7.11264014e-01
-9.89886746e-02 5.86917758e-01 2.47027695e-01 8.97150710e-02
-5.25022924e-01 -9.44540262e-01 1.40249565e-01 7.96863854e-01
3.49513501e-01 -4.10100609e-01 -8.12733948e-01 -9.07181621e-01
-1.65193111e-01 2.45366350e-01 -1.98494270e-01 -2.01844782e-01
-4.63102549e-01 -1.11107194e+00 -9.36074648e-03 3.00209969e-01
3.76536131e-01 -4.98043805e-01 5.66883147e-01 6.96095601e-02
-4.18100983e-01 -1.41186917e+00 -5.82474172e-01 3.22085656e-02
-1.16963446e+00 -1.55082905e+00 -3.42999339e-01 -8.57065558e-01
1.25563717e+00 1.15656829e+00 7.14408755e-01 3.86494756e-01
-8.64321217e-02 7.89302349e-01 -5.97045600e-01 -3.51216178e-03
-4.25464481e-01 -4.27662015e-01 3.13146859e-01 6.08338237e-01
3.77931073e-02 -5.31090915e-01 -1.52555734e-01 3.52429926e-01
-1.01569343e+00 -2.31270269e-02 6.44456148e-01 1.02401125e+00
9.27515924e-01 8.16504240e-01 2.47909263e-01 -1.18047404e+00
4.44062471e-01 -1.85185403e-01 -7.52332985e-01 2.68651038e-01
-9.21590447e-01 2.25799352e-01 5.64963162e-01 -4.05389518e-01
-1.06673121e+00 4.82006788e-01 3.62343550e-01 -5.77800035e-01
-1.88851669e-01 4.94766682e-01 -5.66624880e-01 -4.80509251e-01
4.95269656e-01 2.81212717e-01 2.46270582e-01 -6.39232755e-01
5.38744509e-01 6.19571328e-01 7.49024570e-01 -4.48196352e-01
1.31673813e+00 8.19018066e-01 2.46860251e-01 -1.28808057e+00
-5.85595131e-01 -1.00772762e+00 -8.49604547e-01 -2.55184710e-01
4.98326838e-01 -1.11245549e+00 -1.98965102e-01 6.74019396e-01
-6.05974913e-01 -1.18781388e-01 3.73035707e-02 5.51355958e-01
-8.92702565e-02 1.14636743e+00 -6.40726328e-01 -8.28125834e-01
-2.65011758e-01 -8.41388285e-01 9.35428083e-01 2.67044529e-02
4.74604577e-01 -9.39331055e-01 -3.06074768e-01 6.45095468e-01
-7.95525536e-02 4.59581107e-01 6.45361662e-01 -4.73096907e-01
-6.00761294e-01 -4.50110316e-01 -3.90301973e-01 6.66108251e-01
4.50087219e-01 1.18882246e-01 -7.89867759e-01 -6.42031610e-01
2.07711115e-01 2.67757140e-02 9.27786648e-01 3.77960414e-01
1.10046852e+00 -4.08159882e-01 -3.78168523e-01 5.95737338e-01
1.28679037e+00 -2.48935670e-01 4.50581014e-01 -3.36727537e-02
1.35729384e+00 6.22358263e-01 7.32113123e-01 3.35674703e-01
2.07920045e-01 3.90394270e-01 4.32349086e-01 -2.87937433e-01
-1.30208328e-01 -2.45423231e-04 4.95175540e-01 1.41258526e+00
-1.53502196e-01 1.86297730e-01 -7.67944038e-01 4.65531975e-01
-2.27096748e+00 -7.64614165e-01 -7.33753979e-01 2.38257742e+00
4.56042171e-01 6.79974928e-02 -5.00261970e-02 4.76344287e-01
9.92661655e-01 1.72203794e-01 -2.49300316e-01 4.74196076e-01
-3.50306094e-01 -4.51000892e-02 6.16779208e-01 4.64142203e-01
-1.18362725e+00 7.77249873e-01 5.27830791e+00 8.98922026e-01
-9.25638914e-01 -7.54915364e-03 2.56989986e-01 3.28229904e-01
1.06188603e-01 1.20296851e-01 -2.90722579e-01 2.60159373e-01
3.71135771e-01 -2.49093566e-02 7.42301524e-01 9.41762626e-01
1.74640045e-01 -6.87627122e-02 -8.14985752e-01 1.23603106e+00
5.47662973e-01 -9.90559161e-01 1.45272672e-01 2.04469427e-01
8.37252021e-01 -1.86006904e-01 -2.57011652e-01 -6.58410341e-02
3.51246238e-01 -5.24589479e-01 2.96419650e-01 5.11029065e-01
2.82362789e-01 -6.96065485e-01 6.57353818e-01 4.48205620e-01
-1.21286547e+00 6.05124794e-02 -3.53015065e-01 2.27715597e-01
-1.01147667e-01 1.07867348e+00 -6.76053643e-01 8.46157134e-01
6.38243496e-01 1.08278084e+00 -8.99325848e-01 1.00583696e+00
-4.30548280e-01 8.03555906e-01 -2.75590450e-01 3.80391240e-01
-1.50463238e-01 -7.86256671e-01 7.06701338e-01 1.09681714e+00
1.66119058e-02 2.54967749e-01 5.42089701e-01 3.49000007e-01
-2.05873072e-01 2.58924752e-01 -5.73697209e-01 -4.60570119e-02
2.21834108e-01 1.60856962e+00 -9.32871401e-01 -2.72506773e-01
-5.54183960e-01 1.10648239e+00 2.87520468e-01 6.26158655e-01
-4.87662703e-01 -1.88289151e-01 1.52807653e-01 5.20806238e-02
2.00896814e-01 -5.56567669e-01 -4.49838340e-02 -1.36012411e+00
1.09510355e-01 -9.75314975e-01 6.03977501e-01 -6.52420938e-01
-1.41624153e+00 1.99713051e-01 -2.40894854e-01 -1.09559548e+00
1.70598313e-01 -5.64189613e-01 -3.76383096e-01 2.26368010e-01
-1.44026077e+00 -1.12228179e+00 -6.55741096e-01 9.76670146e-01
2.24521026e-01 -1.12646878e-01 4.99598682e-01 3.47706854e-01
-8.07752848e-01 1.31038159e-01 5.93425870e-01 2.63401955e-01
9.39417958e-01 -1.27762604e+00 -1.55615270e-01 1.47782958e+00
6.40458286e-01 4.76201296e-01 5.45545518e-01 -7.20686436e-01
-1.94767332e+00 -1.28894746e+00 -1.53861688e-02 -1.80375353e-01
6.87252879e-01 -3.16129059e-01 -1.27503347e+00 5.11088610e-01
-2.40635186e-01 2.10979849e-01 5.84372401e-01 5.97620942e-02
-3.57403338e-01 -3.95366877e-01 -1.02491391e+00 3.79988283e-01
8.75081599e-01 -4.06209111e-01 -2.61092395e-01 6.26489639e-01
3.94233346e-01 -3.23433757e-01 -6.89200997e-01 2.56535888e-01
2.90216729e-02 -6.30353391e-01 8.77934337e-01 -2.01368466e-01
-2.67594665e-01 -5.93259513e-01 -1.82127565e-01 -1.14240539e+00
-4.50182199e-01 -7.65678525e-01 -5.34709096e-01 1.48676121e+00
-1.26402646e-01 -4.16363955e-01 6.82743788e-01 5.36186814e-01
5.15559688e-02 -1.45611474e-02 -7.03853607e-01 -8.36339235e-01
-6.38659716e-01 -3.23571056e-01 1.42472327e-01 1.28764069e+00
-1.48307830e-01 4.26011831e-01 -6.12903178e-01 6.34285271e-01
1.32377791e+00 1.84054986e-01 1.13934374e+00 -1.36865914e+00
-3.53624761e-01 -4.95795347e-02 -5.42864561e-01 -8.92697096e-01
1.79739133e-01 -7.71852851e-01 1.30257607e-01 -1.51378047e+00
3.61584276e-01 -2.20369622e-01 -2.55099177e-01 3.17433476e-01
-5.44608891e-01 1.63053006e-01 1.16014153e-01 4.84423190e-01
-9.13808823e-01 4.26625848e-01 9.07774746e-01 -1.94462508e-01
-2.93148011e-01 1.27732977e-02 -6.47809684e-01 8.70098233e-01
4.95660931e-01 -5.14722645e-01 -2.28279471e-01 -1.80118844e-01
-1.97222367e-01 -9.78947058e-02 1.92154795e-01 -9.46129143e-01
3.25797021e-01 -1.70029521e-01 2.60883957e-01 -3.21770340e-01
2.17066124e-01 -1.10227084e+00 1.41869321e-01 1.36907399e-01
2.28813156e-01 -3.78766745e-01 -1.60907477e-01 1.26602244e+00
-3.50735217e-01 6.55355975e-02 8.37658048e-01 4.45207283e-02
-8.21630955e-01 3.68741363e-01 -1.80833116e-01 -1.47351086e-01
7.05413461e-01 -7.62831569e-02 -2.92893946e-01 -6.87707126e-01
-7.01215088e-01 1.37918130e-01 5.93479693e-01 3.26153159e-01
7.27754116e-01 -1.16985154e+00 -6.39250517e-01 1.67133123e-01
1.44850627e-01 1.06710486e-01 -1.57864171e-03 1.05361283e+00
-2.11138263e-01 -1.71054080e-01 2.23496899e-01 -8.47833157e-01
-1.68671167e+00 8.40607941e-01 -2.48591557e-01 -2.00932086e-01
-6.80995047e-01 4.29708511e-01 -8.05303454e-02 -2.25227609e-01
4.22270805e-01 -7.84288272e-02 -5.49866110e-02 -1.56038344e-01
4.16858703e-01 6.96574748e-01 9.69506204e-02 -1.00049222e+00
-2.82939494e-01 7.31983185e-01 4.13852707e-02 2.17832342e-01
1.32532990e+00 -3.97943646e-01 -4.63123083e-01 2.20604196e-01
1.01256514e+00 3.50683987e-01 -1.05177367e+00 -6.10767484e-01
6.96037933e-02 -7.07429945e-01 3.42798799e-01 -1.63149193e-01
-1.38074577e+00 5.45850694e-01 4.36471432e-01 2.23206297e-01
1.45917761e+00 -7.91941509e-02 4.94253010e-01 5.17387211e-01
4.77404207e-01 -1.17549860e+00 3.58694524e-01 1.34563908e-01
6.65982962e-01 -1.40224206e+00 4.69692439e-01 -8.80051553e-01
-5.75996697e-01 1.00093794e+00 3.21358562e-01 -1.79144397e-01
6.20707333e-01 6.65953755e-02 -6.38194457e-02 -2.18674585e-01
-1.90381750e-01 -4.18786079e-01 4.92588252e-01 8.05895984e-01
2.14108382e-03 2.01394334e-02 2.84206029e-02 4.65692401e-01
3.57266545e-01 -4.55480903e-01 6.27510190e-01 8.55926216e-01
-5.04121900e-01 -1.01908088e+00 -8.79685342e-01 4.81156945e-01
-3.04612339e-01 1.33399755e-01 -6.01000845e-01 6.41668379e-01
-3.67445052e-01 1.34152985e+00 -4.41015989e-01 -2.85780072e-01
2.62490034e-01 -1.21879905e-01 3.31587911e-01 -4.88782316e-01
1.05795294e-01 5.33812821e-01 -1.34851471e-01 -7.33444214e-01
-8.04200232e-01 -6.60439789e-01 -1.28064787e+00 -1.72519222e-01
-8.34932446e-01 2.90243924e-01 6.64126456e-01 8.17349732e-01
4.09835875e-01 3.13074708e-01 8.28539968e-01 -7.71174967e-01
-2.66012788e-01 -7.44387746e-01 -1.05676341e+00 8.16252887e-01
2.59550244e-01 -5.52728772e-01 -6.34742320e-01 4.37203020e-01] | [7.9842329025268555, 4.5394368171691895] |
3dbf81d5-d1f7-4453-9e2f-7b7390b0940b | bi-pointflownet-bidirectional-learning-for | 2207.07522 | null | https://arxiv.org/abs/2207.07522v1 | https://arxiv.org/pdf/2207.07522v1.pdf | Bi-PointFlowNet: Bidirectional Learning for Point Cloud Based Scene Flow Estimation | Scene flow estimation, which extracts point-wise motion between scenes, is becoming a crucial task in many computer vision tasks. However, all of the existing estimation methods utilize only the unidirectional features, restricting the accuracy and generality. This paper presents a novel scene flow estimation architecture using bidirectional flow embedding layers. The proposed bidirectional layer learns features along both forward and backward directions, enhancing the estimation performance. In addition, hierarchical feature extraction and warping improve the performance and reduce computational overhead. Experimental results show that the proposed architecture achieved a new state-of-the-art record by outperforming other approaches with large margin in both FlyingThings3D and KITTI benchmarks. Codes are available at https://github.com/cwc1260/BiFlow. | ['Jong Hwan Ko', 'Wencan Cheng'] | 2022-07-15 | null | null | null | null | ['scene-flow-estimation'] | ['computer-vision'] | [-3.61118525e-01 -7.25099802e-01 -4.04926777e-01 -3.07880849e-01
7.87030719e-03 -4.48521584e-01 6.36959732e-01 -4.15568024e-01
-4.20040816e-01 6.87197506e-01 5.91679931e-01 -1.71765178e-01
2.88965758e-02 -7.68253028e-01 -3.62256378e-01 -6.38614774e-01
-1.96533889e-01 -3.84500474e-01 5.78347802e-01 -5.56881726e-02
3.79479527e-01 4.92233276e-01 -1.37547493e+00 7.51143768e-02
6.25573754e-01 1.05796564e+00 1.00476727e-01 9.97294486e-01
-1.18445411e-01 1.41395450e+00 -2.28942841e-01 -3.64536613e-01
4.38690275e-01 -2.79513180e-01 -9.21477854e-01 -1.74889252e-01
7.43287981e-01 -8.27403367e-01 -1.19216001e+00 6.89223826e-01
4.24585015e-01 1.68862537e-01 3.60990912e-01 -1.38409531e+00
-6.41453743e-01 7.14351237e-02 -7.08399057e-01 6.51927769e-01
3.61237824e-01 4.96527463e-01 1.01783979e+00 -9.75359797e-01
6.23245418e-01 1.00337064e+00 5.20605326e-01 4.42297608e-01
-7.84061134e-01 -8.68982732e-01 2.14772195e-01 7.55054593e-01
-1.16654062e+00 -5.20542502e-01 7.99038827e-01 -6.12118363e-01
1.08581781e+00 1.24210820e-01 9.48899686e-01 9.04575467e-01
1.89084083e-01 1.09666359e+00 7.41448879e-01 1.53865099e-01
-1.68307960e-01 -2.67257839e-01 1.15485109e-01 9.87068772e-01
3.06653142e-01 2.37981558e-01 -7.44559646e-01 3.34160388e-01
9.47325706e-01 -5.31218611e-02 -6.45513594e-01 -4.69689190e-01
-1.57567787e+00 8.23637366e-01 9.54145789e-01 1.02919728e-01
-5.89749627e-02 5.53739667e-01 6.16347611e-01 -3.00611984e-02
3.93528014e-01 1.68355301e-01 -2.40910903e-01 -4.85806495e-01
-7.89763629e-01 2.36315265e-01 6.69239700e-01 1.08683968e+00
8.11908603e-01 3.48423183e-01 -6.89531937e-02 2.73995250e-01
3.42280328e-01 3.25606793e-01 3.77270222e-01 -1.24360061e+00
6.44320190e-01 6.32152677e-01 1.53484732e-01 -1.37326026e+00
-3.40066135e-01 -3.27613473e-01 -9.01001394e-01 1.34203061e-01
4.76314783e-01 -8.48306790e-02 -6.03941381e-01 1.48234534e+00
5.85027516e-01 6.91707611e-01 -8.41617361e-02 1.32088375e+00
1.10218060e+00 9.77751732e-01 -1.55261546e-01 1.85133651e-01
1.09862888e+00 -1.46732140e+00 -8.34142566e-01 -3.00421089e-01
5.57948947e-01 -9.42669809e-01 8.73284161e-01 3.76465358e-02
-9.30263400e-01 -8.02512944e-01 -1.19838357e+00 -6.47440732e-01
-2.36073270e-01 2.59678364e-01 1.16727602e+00 2.66568363e-01
-9.27312195e-01 4.36032355e-01 -1.07036436e+00 -2.03809440e-01
7.29475558e-01 1.51012450e-01 -5.80916107e-01 -1.81591183e-01
-1.13022482e+00 5.41709721e-01 2.24236906e-01 3.12203765e-01
-7.95168936e-01 -9.07926917e-01 -1.21114480e+00 -1.85598925e-04
-1.07369917e-02 -9.59005713e-01 1.08940244e+00 -3.52193415e-01
-1.61878288e+00 4.46303934e-01 -3.64982218e-01 -3.65371674e-01
8.07010591e-01 -5.72524071e-01 -1.82355911e-01 2.55148232e-01
4.70750146e-02 9.38616753e-01 6.88103914e-01 -6.57159925e-01
-7.89400756e-01 -9.40180346e-02 3.21473271e-01 2.96629876e-01
-3.40543985e-01 -3.50913465e-01 -4.94911134e-01 -5.03303409e-01
-4.65878174e-02 -8.08671832e-01 -1.16903327e-01 5.45454144e-01
-2.34760404e-01 -1.79759964e-01 1.18907464e+00 -5.50740361e-01
1.32682252e+00 -2.03950095e+00 1.88084990e-01 -6.33560538e-01
3.80127728e-01 3.25738162e-01 4.06721197e-02 3.95106643e-01
2.84457535e-01 -1.58252880e-01 -2.34316364e-01 -2.09707066e-01
-1.27019078e-01 -5.54829761e-02 -3.51350188e-01 9.12673652e-01
3.10986638e-01 8.87882113e-01 -1.13137293e+00 -6.01155221e-01
1.04549050e+00 7.36150265e-01 -7.87310064e-01 3.64290595e-01
3.27605128e-01 5.91253400e-01 -4.67785478e-01 2.94803321e-01
7.74126947e-01 -2.68791646e-01 -3.10192615e-01 -4.79424506e-01
-4.23269808e-01 3.77904445e-01 -1.19273460e+00 2.11978316e+00
-6.19527578e-01 1.27083993e+00 -3.86955738e-01 -7.54070818e-01
6.23355448e-01 -2.58159172e-03 7.15595365e-01 -4.76987541e-01
2.42178619e-01 -6.28442615e-02 5.89307323e-02 -6.82506323e-01
6.54854059e-01 4.42349166e-01 1.15040705e-01 -1.77593268e-02
3.56576592e-01 -2.00755313e-01 5.46917379e-01 2.19390109e-01
1.05373824e+00 4.90363330e-01 5.42447627e-01 -2.50934005e-01
8.83905172e-01 7.39863282e-03 7.85537541e-01 2.94211835e-01
-7.90718615e-01 4.78022933e-01 2.51012325e-01 -8.72989237e-01
-7.28671432e-01 -1.15049779e+00 -1.42796293e-01 5.40193617e-01
6.42765701e-01 -6.92431748e-01 -2.96580285e-01 -6.93111002e-01
1.69323925e-02 2.43365258e-01 -6.10056758e-01 -1.53869525e-01
-7.66716123e-01 -4.36692655e-01 4.18181628e-01 6.78729296e-01
1.12604225e+00 -8.81641567e-01 -1.09333825e+00 7.10222051e-02
-4.80879158e-01 -1.47641671e+00 -7.39663661e-01 -5.23073912e-01
-8.21497798e-01 -1.20625615e+00 -5.83878160e-01 -7.66242504e-01
3.32656085e-01 7.09036887e-01 9.18477058e-01 5.09697571e-02
-6.02183342e-01 1.46131575e-01 -2.79347360e-01 4.39652689e-02
3.12610090e-01 2.41756469e-01 -1.47372410e-01 -6.69308305e-02
3.99336755e-01 -4.70552117e-01 -1.14671028e+00 3.16617519e-01
-6.30402088e-01 3.28792185e-01 3.19258213e-01 8.30788732e-01
1.98932722e-01 -2.91570365e-01 1.68689236e-01 -4.12746429e-01
1.06853485e-01 -2.77706236e-01 -7.22915173e-01 -3.08296442e-01
-2.57161558e-01 5.88953868e-02 8.30190599e-01 -9.14050490e-02
-1.21398509e+00 3.43919732e-02 -1.46444082e-01 -4.65205669e-01
-1.73267230e-01 2.33747438e-02 3.64112295e-02 -2.43217364e-01
3.67408723e-01 2.62380272e-01 -2.89019257e-01 -9.26038548e-02
5.00772536e-01 4.84190613e-01 4.66930598e-01 -7.89402127e-02
8.54848146e-01 8.25780451e-01 8.75668675e-02 -9.62763667e-01
-7.97465146e-01 -8.94480407e-01 -7.57413983e-01 -4.56293732e-01
9.36034203e-01 -1.07492805e+00 -9.85259056e-01 5.60097158e-01
-1.13194919e+00 -2.48945430e-01 -5.43101802e-02 1.04301155e+00
-6.16712630e-01 3.49944174e-01 -8.26677382e-01 -3.25114787e-01
-3.28959078e-01 -1.30787325e+00 8.43465865e-01 4.73306865e-01
-7.95286223e-02 -1.07459080e+00 3.12772006e-01 3.65812868e-01
5.22468090e-01 2.11041525e-01 1.09336928e-01 1.86972991e-01
-9.19107497e-01 -1.91033743e-02 -5.43191195e-01 1.45980522e-01
3.17150176e-01 1.80408373e-01 -1.06580698e+00 -2.18079194e-01
-2.49140695e-01 -2.97297806e-01 1.11379862e+00 4.92630899e-01
1.08987617e+00 -1.42866716e-01 -5.72492667e-02 1.34261513e+00
1.45133531e+00 -4.35471050e-02 6.56830370e-01 3.87899190e-01
1.10575378e+00 5.65005720e-01 4.73714530e-01 5.64027727e-01
4.90791917e-01 5.84091067e-01 5.07192075e-01 6.03577718e-02
-5.68401814e-01 -3.79530787e-01 4.06827241e-01 9.80030715e-01
-1.11268193e-01 -1.77847475e-01 -8.25045168e-01 7.41449296e-01
-1.92915154e+00 -1.18901575e+00 -3.27530831e-01 1.71191525e+00
5.31357467e-01 1.07357770e-01 -1.98868141e-01 8.51944610e-02
3.71336728e-01 8.18768144e-01 -6.38915718e-01 -2.48711437e-01
1.68969467e-01 -1.00469597e-01 6.61995888e-01 7.03039944e-01
-1.42192698e+00 1.16491485e+00 5.41434050e+00 4.54209685e-01
-1.18217313e+00 -8.42308849e-02 4.08173621e-01 -2.52398551e-01
2.64674518e-02 -3.87416817e-02 -7.75939047e-01 4.51293319e-01
5.25642276e-01 -1.93301350e-01 1.89484492e-01 7.21993446e-01
1.80429235e-01 -3.86703759e-02 -8.52507353e-01 1.10832191e+00
-3.04948911e-02 -1.54558170e+00 -4.17463519e-02 -1.10256501e-01
7.55113661e-01 2.28036389e-01 -9.04470310e-02 2.50203639e-01
2.23745495e-01 -7.30139017e-01 5.42409778e-01 2.36048728e-01
6.45988584e-01 -5.47797024e-01 7.77790725e-01 -8.26752000e-03
-1.74504125e+00 5.80876060e-02 -3.83077383e-01 -4.64481831e-01
4.71884012e-01 5.22212923e-01 -4.84928071e-01 5.31456411e-01
9.93372619e-01 1.53364694e+00 -5.69540441e-01 1.22991765e+00
-3.56752813e-01 7.00755477e-01 -1.95547789e-01 4.64359298e-02
4.65759754e-01 -2.70312369e-01 5.94050407e-01 1.47072899e+00
2.81556129e-01 -2.10387841e-01 9.39290971e-02 6.07245207e-01
-1.25479117e-01 -7.66924471e-02 -7.80596375e-01 1.98574558e-01
2.42027968e-01 1.42249322e+00 -5.37313521e-01 -1.64443597e-01
-7.60216117e-01 8.81439984e-01 3.33479553e-01 3.32945764e-01
-1.12950373e+00 -6.61289692e-01 1.17292809e+00 -2.12704971e-01
2.77214289e-01 -5.81669986e-01 -9.09561142e-02 -1.55747414e+00
1.24378853e-01 -1.59438297e-01 2.68868327e-01 -4.93721306e-01
-1.00246859e+00 3.84050876e-01 -7.13604689e-02 -1.59050834e+00
-1.01127565e-01 -7.48646677e-01 -5.83746850e-01 5.71120799e-01
-2.00451040e+00 -1.09725463e+00 -9.42793250e-01 6.19094431e-01
8.08682799e-01 1.16719246e-01 3.76192153e-01 4.77608263e-01
-6.18789792e-01 3.17075878e-01 -1.08751260e-01 4.63255614e-01
7.93100476e-01 -1.16484654e+00 6.40868962e-01 1.20191264e+00
1.76459134e-01 4.02623683e-01 3.80581975e-01 -1.52402863e-01
-1.38875389e+00 -1.16470897e+00 7.15428352e-01 -2.10303232e-01
7.23123610e-01 -3.42917502e-01 -6.06415451e-01 7.20829308e-01
4.44934636e-01 6.55833542e-01 4.87173229e-01 -3.65123689e-01
-2.88304031e-01 -3.30467165e-01 -7.46351659e-01 7.03158498e-01
1.37237215e+00 -3.81736815e-01 -2.29929611e-01 -4.52305302e-02
7.06802368e-01 -5.56887507e-01 -8.03269506e-01 3.19694459e-01
8.98828924e-01 -1.23021162e+00 1.19292200e+00 -4.11284029e-01
6.95009112e-01 -5.88939309e-01 -9.21553746e-02 -1.08938682e+00
-5.68889499e-01 -5.64641833e-01 -5.99572539e-01 9.36375618e-01
-1.26177669e-01 -6.71433151e-01 7.99385965e-01 1.42691687e-01
-1.90767422e-01 -7.54409909e-01 -8.11485410e-01 -6.46624207e-01
-1.08836837e-01 -2.85854876e-01 4.83527929e-01 8.54236126e-01
-2.12835461e-01 4.06373322e-01 -5.80963254e-01 9.36545208e-02
6.10836029e-01 3.86782706e-01 9.67835844e-01 -6.95368469e-01
-3.13275270e-02 -5.32842875e-01 -9.91871357e-01 -1.53813052e+00
2.55741030e-01 -8.36055279e-01 -9.16233212e-02 -1.68751252e+00
-1.59907654e-01 -8.25310946e-02 -2.08819047e-01 7.61504173e-02
-3.03740293e-01 2.90646970e-01 3.79187644e-01 9.03634503e-02
-5.97657800e-01 9.32627082e-01 1.57526696e+00 -1.87350988e-01
-3.77386771e-02 -2.23531872e-01 -2.22258672e-01 7.76581228e-01
1.20243812e+00 -2.05312386e-01 -6.52001679e-01 -9.61179972e-01
-1.88363120e-01 -5.41390255e-02 4.25007641e-01 -1.40492690e+00
3.89410615e-01 -2.43300274e-01 5.89936733e-01 -5.07200956e-01
4.56733614e-01 -7.66365647e-01 -2.41642460e-01 7.29771316e-01
-2.11776659e-01 2.33824611e-01 3.57925117e-01 6.18142426e-01
-4.74079460e-01 1.92208216e-01 6.66929066e-01 9.10384804e-02
-1.29230940e+00 8.15795660e-01 -7.41662160e-02 1.77773058e-01
1.20568871e+00 -9.43479314e-02 -5.35323799e-01 -2.78991252e-01
1.06156236e-02 2.58048803e-01 2.29703143e-01 6.57481730e-01
8.12140107e-01 -1.39551473e+00 -6.46258652e-01 3.50879282e-01
1.81692243e-01 1.72298789e-01 3.22869480e-01 1.00160754e+00
-1.19390678e+00 6.59475744e-01 -4.82781827e-01 -7.39875317e-01
-9.62989032e-01 3.23162138e-01 2.59817362e-01 -1.48952127e-01
-8.00113857e-01 1.01199102e+00 5.05192518e-01 -2.70123392e-01
-2.18871906e-02 -3.18567067e-01 -2.59190917e-01 -2.10486099e-01
6.82937205e-01 7.39621460e-01 -3.00919354e-01 -8.91967595e-01
-4.94521081e-01 8.47279966e-01 3.84304337e-02 1.74559519e-01
1.03225923e+00 -2.07320243e-01 1.68216959e-01 3.22509557e-01
1.63668716e+00 -2.21140027e-01 -1.70260227e+00 -7.14446828e-02
-3.57414097e-01 -1.15616107e+00 4.12315965e-01 -1.17169552e-01
-1.49093091e+00 1.35614014e+00 5.30741036e-01 -2.29988888e-01
1.12631786e+00 -4.79632646e-01 1.07244086e+00 3.03783745e-01
3.47991467e-01 -5.94370842e-01 -7.27005750e-02 4.19906914e-01
6.76875412e-01 -1.54883277e+00 2.07122341e-01 -3.72155815e-01
-5.37759840e-01 1.34057415e+00 9.63552594e-01 -4.29359823e-01
7.68146634e-01 1.73385501e-01 2.01952755e-02 2.33100932e-02
-7.92661965e-01 -2.56221354e-01 3.25303853e-01 4.20515448e-01
6.50692105e-01 -9.07757804e-02 -2.27580786e-01 -2.02602684e-01
-2.23867118e-01 1.66924849e-01 5.13383269e-01 8.98570716e-01
-2.24044904e-01 -7.00657070e-01 2.34421730e-01 3.40946823e-01
-3.24963570e-01 2.84657814e-02 -2.66115181e-02 1.01235747e+00
-5.06513715e-02 9.03085649e-01 6.41780645e-02 -3.85907829e-01
2.25785628e-01 -4.55508024e-01 4.70652938e-01 -5.65456301e-02
-2.30599523e-01 -3.89026999e-01 -4.94244695e-02 -1.15700006e+00
-7.41126716e-01 -4.69668657e-01 -1.08698261e+00 -7.22263157e-01
-1.49609357e-01 -1.00104168e-01 3.16749632e-01 4.10307765e-01
4.59938496e-01 5.75800896e-01 7.24869609e-01 -8.88093412e-01
-8.33696127e-02 -8.04501832e-01 -1.22645572e-01 4.79551762e-01
6.74228251e-01 -8.33251357e-01 -4.58519757e-01 3.35857391e-01] | [8.69857120513916, -1.821067214012146] |
9d0e9268-195c-4fc2-ae6c-a3cf040bbdcd | seeing-what-you-miss-vision-language-pre | 2211.13437 | null | https://arxiv.org/abs/2211.13437v2 | https://arxiv.org/pdf/2211.13437v2.pdf | Seeing What You Miss: Vision-Language Pre-training with Semantic Completion Learning | Cross-modal alignment is essential for vision-language pre-training (VLP) models to learn the correct corresponding information across different modalities. For this purpose, inspired by the success of masked language modeling (MLM) tasks in the NLP pre-training area, numerous masked modeling tasks have been proposed for VLP to further promote cross-modal interactions. The core idea of previous masked modeling tasks is to focus on reconstructing the masked tokens based on visible context for learning local-to-local alignment. However, most of them pay little attention to the global semantic features generated for the masked data, resulting in a limited cross-modal alignment ability of global representations. Therefore, in this paper, we propose a novel Semantic Completion Learning (SCL) task, complementary to existing masked modeling tasks, to facilitate global-to-local alignment. Specifically, the SCL task complements the missing semantics of masked data by capturing the corresponding information from the other modality, promoting learning more representative global features which have a great impact on the performance of downstream tasks. Moreover, we present a flexible vision encoder, which enables our model to perform image-text and video-text multimodal tasks simultaneously. Experimental results show that our proposed method obtains state-of-the-art performance on various vision-language benchmarks, such as visual question answering, image-text retrieval, and video-text retrieval. | ['Wei Liu', 'Yujiu Yang', 'Hongfa Wang', 'Wenzhe Zhao', 'Chengfei Cai', 'Weijie Kong', 'Jie Jiang', 'RongCheng Tu', 'Yatai Ji'] | 2022-11-24 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Ji_Seeing_What_You_Miss_Vision-Language_Pre-Training_With_Semantic_Completion_Learning_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Ji_Seeing_What_You_Miss_Vision-Language_Pre-Training_With_Semantic_Completion_Learning_CVPR_2023_paper.pdf | cvpr-2023-1 | ['video-text-retrieval'] | ['computer-vision'] | [ 2.73925543e-01 -6.97868690e-02 -4.36409801e-01 -4.34423298e-01
-9.12212372e-01 -1.49971068e-01 9.32574809e-01 1.99513100e-02
-3.51532906e-01 3.48804206e-01 4.48773921e-01 -4.97954860e-02
1.69055790e-01 -3.86539549e-01 -8.47301126e-01 -7.68038750e-01
6.75418675e-01 2.82354891e-01 2.12562084e-01 -1.27594933e-01
1.14841230e-01 -1.34856984e-01 -1.48406172e+00 9.09732759e-01
7.62904584e-01 9.15532589e-01 8.19446504e-01 1.45208672e-01
-5.19901633e-01 1.04688025e+00 -7.75244012e-02 -4.34216559e-01
-1.06141128e-01 -5.17270446e-01 -6.38631344e-01 2.15728983e-01
6.28512561e-01 -3.89371306e-01 -5.42878449e-01 1.06246877e+00
1.85334861e-01 -4.10976671e-02 6.05142713e-01 -1.26986563e+00
-9.13886845e-01 5.24458945e-01 -6.61162376e-01 -1.41739666e-01
2.56342262e-01 8.48285332e-02 1.15731525e+00 -1.38597357e+00
5.09164214e-01 1.45752335e+00 1.16167329e-01 5.80440164e-01
-1.08831263e+00 -4.98434603e-01 5.06983042e-01 4.45083678e-01
-1.42124379e+00 -5.77557921e-01 8.21571648e-01 -4.54554915e-01
7.11032689e-01 1.78932175e-02 2.86972046e-01 1.20202863e+00
1.19425021e-02 1.14957702e+00 1.02787232e+00 -5.39044440e-01
-1.83149025e-01 1.91884652e-01 -1.05251364e-01 8.23864937e-01
-7.07709715e-02 -1.23323634e-01 -8.57429147e-01 2.35196009e-01
5.03151536e-01 2.00854242e-01 -4.31651235e-01 -3.71624261e-01
-1.52774835e+00 6.14027977e-01 4.52633768e-01 3.60277504e-01
-2.79945433e-01 1.56302750e-01 1.54946327e-01 1.04906283e-01
2.66935378e-01 -1.96352318e-01 -2.52834111e-01 2.70051599e-01
-8.95943701e-01 -1.05255798e-01 1.72190726e-01 9.50006068e-01
1.10585439e+00 -4.87313047e-02 -4.97124314e-01 7.77244210e-01
7.74110854e-01 5.59606671e-01 5.67385912e-01 -6.74516380e-01
9.54502583e-01 7.53109634e-01 -3.31106223e-02 -9.91128147e-01
-1.94354862e-01 -1.87468439e-01 -1.08499396e+00 -2.15337351e-01
1.45065114e-01 4.77432579e-01 -1.05812955e+00 1.89342427e+00
4.66143228e-02 1.17680334e-01 3.30885261e-01 8.84864986e-01
1.10055077e+00 7.54605293e-01 1.63558438e-01 -2.71773517e-01
1.37612867e+00 -1.40361989e+00 -7.45743334e-01 -5.46623290e-01
4.44816709e-01 -1.07369685e+00 1.27595150e+00 1.44095076e-02
-9.96158540e-01 -9.71740484e-01 -6.73038721e-01 -5.07264078e-01
-1.89258844e-01 2.92457938e-01 2.60358453e-01 6.46891668e-02
-8.07980895e-01 4.94856248e-03 -6.72399282e-01 -3.79999548e-01
4.23288345e-01 -4.59815888e-03 -4.46617484e-01 -5.62536776e-01
-1.09992337e+00 7.33836234e-01 4.81369883e-01 3.20760071e-01
-1.11772084e+00 -4.02206808e-01 -8.90061975e-01 1.04022123e-01
4.55373555e-01 -8.67991149e-01 8.55554938e-01 -1.01209295e+00
-1.14555597e+00 1.06660628e+00 -6.57128513e-01 -1.76020473e-01
3.52837533e-01 -6.36089817e-02 -2.07972988e-01 3.96516740e-01
3.34069610e-01 1.13951516e+00 1.29600072e+00 -1.56783557e+00
-4.51768279e-01 -4.39493179e-01 4.19358537e-03 4.54878747e-01
-4.33975399e-01 -3.65335084e-02 -1.07323587e+00 -8.15365732e-01
1.96263298e-01 -6.53529406e-01 -1.49577074e-02 -3.01266126e-02
-2.91907251e-01 -2.69492328e-01 5.96085012e-01 -8.83857012e-01
8.04528117e-01 -2.17609787e+00 4.71979856e-01 -2.63754934e-01
2.20257595e-01 1.65188223e-01 -4.40572441e-01 5.93505442e-01
1.81873348e-02 -2.46772662e-01 -2.13983625e-01 -8.75855267e-01
-2.77883634e-02 4.09162194e-01 -6.46288872e-01 1.43199369e-01
2.36481175e-01 1.27422249e+00 -6.45724535e-01 -8.21256697e-01
4.05672759e-01 3.90694380e-01 -4.51631606e-01 3.44571829e-01
-3.77268910e-01 7.14967847e-01 -3.57535809e-01 7.55842984e-01
5.65265656e-01 -4.74418432e-01 -8.88794139e-02 -5.27010560e-01
7.74344057e-02 -1.00251190e-01 -6.88273013e-01 2.17553234e+00
-5.36867440e-01 4.76544172e-01 7.50339702e-02 -1.16209841e+00
7.04575777e-01 2.67490476e-01 3.99630219e-01 -9.92061019e-01
-7.59324878e-02 5.07378280e-02 -4.43400949e-01 -7.24866629e-01
4.30093467e-01 -1.12293996e-01 5.05240187e-02 2.25096762e-01
1.52337924e-01 1.41672641e-01 -6.17699586e-02 4.53285158e-01
3.32577288e-01 1.32919222e-01 9.69702899e-02 1.28313243e-01
8.93547237e-01 -7.89090022e-02 3.10082704e-01 6.29437983e-01
-8.40344280e-03 7.22763300e-01 6.67718947e-02 -8.14936683e-02
-6.69909775e-01 -9.58177328e-01 1.66104868e-01 1.13758087e+00
6.12709582e-01 -5.21196425e-01 -4.63416278e-01 -5.00984550e-01
-2.81163454e-01 6.11976445e-01 -4.51860785e-01 -3.02682340e-01
-3.79720718e-01 -4.64444935e-01 3.10324639e-01 4.97867823e-01
7.81343758e-01 -1.10780752e+00 3.93900648e-02 -1.62450790e-01
-8.14616561e-01 -1.71914828e+00 -8.23643446e-01 -1.89076349e-01
-7.44108438e-01 -9.83808339e-01 -7.26146758e-01 -1.22615528e+00
8.22372258e-01 8.78910661e-01 8.31759036e-01 6.14128672e-02
-8.66879225e-02 6.71025336e-01 -3.72618705e-01 -7.96475336e-02
-2.96214551e-01 -2.40463868e-01 -1.84195995e-01 5.56474030e-01
2.98757195e-01 -2.76017368e-01 -5.64058065e-01 3.28105688e-01
-1.12147510e+00 6.25368893e-01 9.22097385e-01 9.84443843e-01
7.71525383e-01 -2.70042986e-01 4.56579030e-01 -3.88165236e-01
2.82979041e-01 -3.19818676e-01 -3.39781582e-01 7.82404482e-01
-3.90056670e-01 9.03335959e-02 4.06117290e-01 -5.51511228e-01
-1.21533632e+00 5.65400198e-02 7.77240619e-02 -8.72826636e-01
-7.77788162e-02 6.83762848e-01 -5.78694403e-01 8.06162879e-02
1.55727476e-01 7.88436890e-01 1.67636186e-01 -5.84569812e-01
6.31733716e-01 6.28400266e-01 6.03802800e-01 -5.12529016e-01
7.44121969e-01 6.97846472e-01 -6.27928600e-02 -7.72054434e-01
-1.07885909e+00 -6.68373823e-01 -5.65084636e-01 9.94134918e-02
1.22992897e+00 -1.41441453e+00 -4.79166597e-01 6.65983081e-01
-1.23810661e+00 -1.31643951e-01 -7.40621313e-02 4.52242792e-01
-6.03890479e-01 9.64045882e-01 -2.41211712e-01 -4.71514642e-01
-2.80834109e-01 -1.40507972e+00 1.48159313e+00 9.49026421e-02
3.34206879e-01 -8.37928593e-01 -3.16221297e-01 1.02008450e+00
2.63651192e-01 -4.67181087e-01 1.06521809e+00 -3.61788839e-01
-8.57373357e-01 1.10444285e-01 -6.00618899e-01 5.23470759e-01
1.60182625e-01 -5.19302130e-01 -1.06500995e+00 -3.24497908e-01
1.23605423e-01 -5.01297593e-01 1.32707560e+00 2.34261513e-01
1.11281717e+00 -1.57692507e-01 -3.06997120e-01 5.05620480e-01
1.22000647e+00 -3.16608787e-01 5.51375568e-01 9.96567532e-02
1.15801990e+00 7.83718646e-01 7.10999370e-01 4.98199500e-02
8.37987661e-01 8.67951632e-01 5.82655430e-01 -1.83132201e-01
-4.52715933e-01 -6.19110227e-01 6.66657269e-01 8.94141138e-01
3.33897859e-01 -2.75705438e-02 -7.17631042e-01 5.22052765e-01
-2.20464230e+00 -8.43293428e-01 4.93751280e-02 2.07930422e+00
8.54846358e-01 -2.60103434e-01 -2.79143751e-01 -3.69443595e-01
7.51577377e-01 3.86805683e-01 -3.69303793e-01 3.91758651e-01
-3.31689805e-01 -1.77500963e-01 -4.13871892e-02 5.72618008e-01
-9.78908718e-01 1.21543753e+00 4.67521811e+00 1.06164289e+00
-9.68928277e-01 3.09142947e-01 3.88582915e-01 6.81272149e-02
-4.61479366e-01 2.54719257e-01 -7.90265977e-01 4.41450685e-01
2.62806445e-01 3.81471030e-02 3.99580657e-01 5.10237873e-01
2.90271401e-01 -1.97527539e-02 -1.12634182e+00 1.39855492e+00
3.59736204e-01 -1.32084775e+00 7.74614513e-01 -9.38853770e-02
7.51454949e-01 -6.83265273e-03 5.39369844e-02 3.73869866e-01
-3.57770920e-01 -9.82674599e-01 6.79075420e-01 6.51275098e-01
7.76810110e-01 -3.49570900e-01 5.62369227e-01 5.66719949e-01
-1.25781453e+00 -6.08237423e-02 -4.53694791e-01 2.66364783e-01
2.12735817e-01 3.70736718e-01 -5.91272414e-01 7.42527604e-01
6.21688008e-01 9.28579211e-01 -6.80432439e-01 5.92342794e-01
-2.22320452e-01 2.73346066e-01 6.01553060e-02 3.70792180e-01
1.98282778e-01 -1.37409136e-01 3.55559438e-01 9.19659555e-01
6.15460016e-02 -1.89162880e-01 3.94545794e-01 9.71535087e-01
-1.93677187e-01 2.74621159e-01 -4.67511684e-01 -7.96450302e-02
2.89463788e-01 1.16815913e+00 -3.02825153e-01 -3.15276235e-01
-8.91942978e-01 1.10381961e+00 1.76277220e-01 6.71005785e-01
-6.41742170e-01 1.57724649e-01 4.25234020e-01 -1.42162949e-01
3.35831493e-01 -3.27649951e-01 2.55638659e-02 -1.55771780e+00
3.09611499e-01 -8.83259535e-01 4.44384307e-01 -1.07932043e+00
-1.41147685e+00 3.83207232e-01 5.69279976e-02 -1.28315210e+00
-1.29776180e-01 -5.01457930e-01 -3.76884401e-01 9.59028065e-01
-1.90975976e+00 -1.98689723e+00 -6.61609650e-01 1.17194104e+00
8.33648324e-01 -2.52757609e-01 4.86403823e-01 2.71204203e-01
-3.23292255e-01 4.59270656e-01 -6.75551295e-02 4.64632176e-02
1.00887358e+00 -6.54884636e-01 -1.32092591e-02 8.61178994e-01
5.67322552e-01 6.02410376e-01 1.27650455e-01 -5.01675248e-01
-1.53111756e+00 -1.17377555e+00 8.22759390e-01 -4.11151707e-01
5.86724281e-01 -1.81316838e-01 -8.88819933e-01 7.19946802e-01
3.55011821e-01 1.23625528e-02 4.02845323e-01 -2.16900900e-01
-5.08459747e-01 -1.91301346e-01 -4.67107207e-01 6.42576933e-01
8.55509818e-01 -1.06262600e+00 -7.25761592e-01 3.90850723e-01
9.22129452e-01 -1.36710957e-01 -3.95323962e-01 5.36742449e-01
3.19277644e-01 -8.42640281e-01 1.09437633e+00 -5.60907662e-01
5.18399358e-01 -4.57589895e-01 -5.41904569e-01 -8.28530073e-01
9.57739875e-02 -2.72721529e-01 -1.34147733e-01 1.39955699e+00
-6.94899932e-02 -4.18578058e-01 4.49804902e-01 1.69970363e-01
-1.70175225e-01 -5.66693842e-01 -9.34272170e-01 -4.07952905e-01
-9.60346162e-02 -5.00549853e-01 1.41266942e-01 9.71286893e-01
-2.50739515e-01 5.80478370e-01 -7.17341006e-01 3.37014705e-01
7.92707384e-01 5.19836009e-01 7.85527527e-01 -9.17950153e-01
-4.45599973e-01 -4.36321944e-01 -1.86510623e-01 -1.60989070e+00
4.32276279e-01 -1.13097477e+00 1.12893343e-01 -1.66163886e+00
6.04763687e-01 6.54408112e-02 -2.93437868e-01 6.21873558e-01
-5.14136970e-01 3.58434230e-01 3.26158494e-01 6.61788583e-01
-9.07579601e-01 1.00903666e+00 1.44982004e+00 -4.44226503e-01
1.05466060e-01 -2.27135658e-01 -7.52315283e-01 6.19361460e-01
4.44150299e-01 -2.39384085e-01 -5.52375317e-01 -8.05450082e-01
1.12994254e-01 9.10372287e-02 7.62904346e-01 -6.04368389e-01
3.66275758e-01 -2.24408075e-01 1.96234196e-01 -7.86999702e-01
5.89922249e-01 -8.79951119e-01 -3.60754341e-01 1.20365553e-01
-3.26049447e-01 -2.50035733e-01 5.48054054e-02 8.32524836e-01
-7.63444841e-01 -1.40068278e-01 4.45279241e-01 -1.49035394e-01
-9.45592999e-01 3.46177429e-01 -7.71225095e-02 3.60252969e-02
7.15823829e-01 -2.74915583e-02 -5.24994195e-01 -4.53813851e-01
-7.56364524e-01 5.53142190e-01 4.17252332e-01 8.21793675e-01
9.67362463e-01 -1.31951928e+00 -6.58969820e-01 2.46927992e-01
5.87083101e-01 2.60422677e-01 6.60494685e-01 1.15484095e+00
-9.36858878e-02 7.23779798e-01 -3.93517241e-02 -9.42467988e-01
-1.36674547e+00 8.23411345e-01 2.05947876e-01 -1.18069999e-01
-5.93989074e-01 7.51773179e-01 9.33674634e-01 -2.27223068e-01
4.15458471e-01 -1.64238364e-01 -1.26416132e-01 1.19402274e-01
5.02969205e-01 -9.29408446e-02 -1.27120391e-01 -8.47056031e-01
-2.79108405e-01 9.21838999e-01 -2.02857777e-01 -1.27607450e-01
9.77198422e-01 -7.03808248e-01 -4.02195960e-01 4.76017028e-01
1.17406464e+00 2.66045611e-02 -1.08527946e+00 -8.84562016e-01
-2.28574887e-01 -3.40398341e-01 3.91147584e-02 -4.85345811e-01
-9.76712227e-01 1.34884536e+00 4.79182571e-01 -3.78469676e-01
1.20899689e+00 3.93353462e-01 6.54697895e-01 4.34463769e-01
2.58868307e-01 -7.77630985e-01 6.14705980e-01 3.03946346e-01
8.94928396e-01 -1.63125229e+00 -1.11649312e-01 -4.09039497e-01
-8.99228990e-01 9.65844035e-01 6.98940992e-01 4.59589571e-01
5.00498593e-01 -4.39833105e-01 1.13482192e-01 -1.26315847e-01
-6.95225358e-01 -5.03629208e-01 5.57014167e-01 4.60620463e-01
1.56702757e-01 -2.71109164e-01 5.85043915e-02 5.79583228e-01
4.55075860e-01 -1.17215358e-01 6.27170503e-02 6.05832756e-01
-3.80457550e-01 -1.15625143e+00 -2.51461983e-01 -3.68780755e-02
-7.22490251e-02 -4.75822717e-01 -3.88269216e-01 6.43607914e-01
2.92386729e-02 1.10035324e+00 -9.27157328e-02 -2.39008814e-01
-1.12059914e-01 2.45904773e-01 6.25335932e-01 -5.94874203e-01
2.75946148e-02 2.51815438e-01 -2.63134748e-01 -5.04924953e-01
-8.29834878e-01 -3.61330569e-01 -1.28537607e+00 2.21606568e-01
-4.94825207e-02 -4.80644293e-02 5.10271370e-01 1.26188982e+00
4.77835625e-01 3.75030130e-01 4.66696709e-01 -8.20845664e-01
-4.28259641e-01 -1.01284337e+00 -1.89940274e-01 5.69059849e-01
3.75134379e-01 -5.87032378e-01 -1.64819539e-01 2.31201828e-01] | [10.782820701599121, 1.3792363405227661] |
57ba0ae8-b732-41b2-9e55-1033ed0b81ee | underwater-robotics-semantic-parser-assistant | 2301.12134 | null | https://arxiv.org/abs/2301.12134v1 | https://arxiv.org/pdf/2301.12134v1.pdf | Underwater Robotics Semantic Parser Assistant | Semantic parsing is a means of taking natural language and putting it in a form that a computer can understand. There has been a multitude of approaches that take natural language utterances and form them into lambda calculus expressions -- mathematical functions to describe logic. Here, we experiment with a sequence to sequence model to take natural language utterances, convert those to lambda calculus expressions, when can then be parsed, and place them in an XML format that can be used by a finite state machine. Experimental results show that we can have a high accuracy model such that we can bridge the gap between technical and nontechnical individuals in the robotics field. | ['Jake Imyak', 'Cedric McGuire', 'Parth Parekh'] | 2023-01-28 | null | null | null | null | ['semantic-parsing'] | ['natural-language-processing'] | [ 2.11477369e-01 6.03756368e-01 -6.67415559e-02 -7.06364334e-01
-3.78950924e-01 -9.07917142e-01 5.28058946e-01 -4.19326499e-02
-1.59157410e-01 6.03819013e-01 -6.04052059e-02 -1.07895696e+00
1.17983572e-01 -1.19949806e+00 -7.98693061e-01 1.83283523e-01
5.13778217e-02 4.21480089e-01 6.41137719e-01 -5.28652668e-01
7.41266683e-02 1.77536964e-01 -1.43127239e+00 2.78011262e-01
5.99948883e-01 6.50582850e-01 2.38127664e-01 9.11177516e-01
-9.62854564e-01 1.25581050e+00 -4.76614028e-01 -4.08684760e-01
1.37740105e-01 -3.63341004e-01 -1.59243727e+00 -1.41329408e-01
-3.53798389e-01 -2.77324438e-01 1.79144874e-01 1.29729557e+00
-4.68166649e-01 -3.39953035e-01 3.84350717e-01 -1.30007017e+00
-4.70275760e-01 8.81471097e-01 4.20769781e-01 -6.15781248e-01
1.13927209e+00 1.46770760e-01 7.75913417e-01 -1.21938162e-01
1.01252341e+00 1.79603291e+00 5.61313093e-01 8.02268744e-01
-1.04443216e+00 -1.34410039e-01 -1.10432208e-01 -2.25871295e-01
-9.53242123e-01 -2.03006625e-01 3.56991321e-01 -6.00514054e-01
1.27027428e+00 2.25595891e-01 6.10007405e-01 4.49053049e-01
3.17658871e-01 5.10150075e-01 9.24103856e-01 -9.87295866e-01
2.88645446e-01 1.28089070e-01 7.87592113e-01 9.97921824e-01
1.05371960e-01 -1.16279192e-01 3.83035690e-02 -7.35844672e-02
9.10860300e-01 -1.88346431e-01 1.88272625e-01 -1.05137661e-01
-1.00772870e+00 8.24218273e-01 1.58451572e-01 3.05130839e-01
-2.22933814e-01 5.92083991e-01 5.57738483e-01 5.71481347e-01
-2.43407965e-01 4.13633794e-01 -4.55096453e-01 -3.72827590e-01
-1.89714059e-02 4.87039685e-01 1.51116550e+00 1.07982934e+00
7.93040097e-01 -2.56241679e-01 4.44508046e-01 5.24448752e-01
5.94742775e-01 3.82778347e-01 3.16852331e-01 -1.50533414e+00
6.34866357e-02 7.50860989e-01 2.83384055e-01 -6.45108342e-01
-9.00641009e-02 7.28491485e-01 1.82198793e-01 4.40145224e-01
5.78482807e-01 -2.64650762e-01 -6.33821130e-01 1.65609968e+00
2.03967690e-01 -2.15903655e-01 7.26547360e-01 7.97575772e-01
6.27568126e-01 9.38041091e-01 3.57604176e-01 -6.73367009e-02
1.85390997e+00 -5.75281680e-01 -6.06264293e-01 -2.84986317e-01
9.76577044e-01 -5.77459753e-01 1.23911154e+00 6.06134057e-01
-1.24108231e+00 -2.38437250e-01 -1.16257846e+00 -3.05937141e-01
-5.40282965e-01 -4.75013733e-01 9.41377997e-01 6.47380531e-01
-1.10818064e+00 4.51955169e-01 -8.89737785e-01 -6.02296591e-01
-2.73576409e-01 3.07067335e-01 -2.45231375e-01 -9.04030446e-03
-1.39193678e+00 1.22132969e+00 9.49584305e-01 -1.06374711e-01
-4.67244714e-01 -1.35674909e-01 -1.03649902e+00 -1.06733486e-01
3.82994205e-01 -8.79045069e-01 2.17941689e+00 -1.04273522e+00
-1.86298048e+00 6.84694886e-01 -9.73094776e-02 -6.37587726e-01
3.64865512e-02 8.08877051e-02 -2.98679918e-01 2.21521594e-02
8.55165198e-02 5.34550965e-01 2.35447451e-01 -1.01931775e+00
-8.08296442e-01 -3.37703139e-01 9.98149216e-01 -6.66759089e-02
2.76548505e-01 6.14805937e-01 -4.55142617e-01 1.34965613e-01
4.05414164e-01 -9.40439641e-01 -3.08411181e-01 -1.64590981e-02
-1.23437151e-01 -6.06369376e-01 3.92361522e-01 -4.27616239e-01
9.70277011e-01 -1.98989189e+00 9.72006544e-02 9.76776704e-02
2.74021067e-02 1.87352479e-01 2.06674457e-01 5.27284265e-01
1.51660010e-01 4.04222041e-01 -1.71618775e-01 3.06241244e-01
5.30649483e-01 7.63204634e-01 -3.75553071e-01 -1.71868250e-01
2.81676590e-01 8.17496598e-01 -1.03132367e+00 -4.30026531e-01
2.28930309e-01 1.17425948e-01 -6.63936317e-01 3.53938758e-01
-1.01258171e+00 1.02696856e-02 -1.05386364e+00 1.36615142e-01
3.44657332e-01 9.58874524e-02 6.15907490e-01 4.66645807e-01
-1.85883075e-01 7.81848192e-01 -1.21720493e+00 1.59787130e+00
-8.03197324e-01 5.37980080e-01 3.91335249e-01 -9.91238058e-01
9.36228812e-01 2.69706488e-01 -1.18597155e-03 -6.01759851e-02
3.62301111e-01 3.68782878e-01 -1.35803401e-01 -6.96997941e-01
3.43908221e-01 -6.53635085e-01 -6.12026691e-01 6.08065188e-01
-1.05450295e-01 -8.29469323e-01 3.29193830e-01 1.81748822e-01
1.14840984e+00 3.50534827e-01 2.80679554e-01 -2.35968724e-01
6.58039391e-01 5.61841547e-01 3.34872037e-01 6.25672698e-01
1.49447858e-01 -6.09270185e-02 8.96392882e-01 -7.40538120e-01
-1.12193787e+00 -9.58894610e-01 1.17812119e-01 1.04613543e+00
6.06145822e-02 -5.93823433e-01 -1.29761851e+00 -1.64472282e-01
-1.84972733e-01 9.80115473e-01 1.65011123e-01 -9.40100774e-02
-5.73990941e-01 2.71774620e-01 9.21245039e-01 5.04154801e-01
3.56716335e-01 -1.36054361e+00 -9.59017754e-01 5.43871343e-01
-1.17678091e-01 -1.38031876e+00 1.55424759e-01 2.44052947e-01
-8.96415830e-01 -9.31488991e-01 3.18080366e-01 -1.25313044e+00
4.11850363e-01 -1.33323923e-01 8.17189038e-01 3.40193868e-01
1.02217561e-02 4.76112366e-01 -5.36335289e-01 -7.60228097e-01
-1.27215934e+00 -2.36916602e-01 -1.15540035e-01 -5.71786523e-01
5.89762270e-01 -4.18508857e-01 2.61365235e-01 -2.76389360e-01
-1.19320047e+00 1.50978297e-01 1.48874208e-01 3.25294048e-01
1.05376400e-01 1.27109423e-01 2.00783968e-01 -1.03599274e+00
9.04602766e-01 -5.34227677e-02 -7.41127968e-01 3.30084264e-01
-4.21088755e-01 5.97641945e-01 8.60306680e-01 -7.92706832e-02
-1.08979607e+00 4.24882323e-01 -3.50870639e-01 1.25205159e-01
-3.83964300e-01 8.98241699e-01 -1.12686276e-01 2.01728627e-01
6.23639643e-01 1.05797611e-01 4.21959460e-01 -2.68042386e-01
7.58521259e-01 1.00590777e+00 8.72971296e-01 -1.21771526e+00
3.79950881e-01 -3.65477167e-02 -6.22296706e-02 -6.84639394e-01
-4.74828303e-01 -6.91580325e-02 -4.46269274e-01 6.12887591e-02
9.58670557e-01 -4.64248747e-01 -8.91323805e-01 2.20870271e-01
-1.58918035e+00 -4.78891969e-01 -2.72133172e-01 2.72788018e-01
-9.92954254e-01 5.21969318e-01 -8.55554461e-01 -1.05491877e+00
-1.28871202e-01 -1.09686887e+00 8.81192386e-01 2.40286410e-01
-5.83989084e-01 -9.27506566e-01 -1.61189824e-01 8.37424491e-03
1.94028124e-01 -1.28656477e-01 1.26997733e+00 -5.50356448e-01
-5.54869354e-01 -3.07381034e-01 -9.05771106e-02 4.12795842e-01
-1.38254687e-01 2.69212484e-01 -4.74499971e-01 4.11609858e-01
-3.32944952e-02 -2.03227177e-01 -1.47776708e-01 1.22246847e-01
6.16325736e-01 -4.05659139e-01 -2.55110174e-01 7.24744797e-02
1.39854622e+00 5.72086692e-01 9.76100862e-01 4.63841945e-01
6.98858723e-02 6.49561942e-01 4.43507940e-01 -1.93126321e-01
7.39310622e-01 4.75226164e-01 -8.79365802e-02 4.40876156e-01
3.20506692e-01 -5.77981472e-01 3.54857415e-01 6.93304837e-01
2.09208980e-01 3.75022590e-01 -1.20684445e+00 3.77826631e-01
-2.03947949e+00 -8.48938644e-01 -2.30498090e-01 1.75257409e+00
1.17934442e+00 3.31123948e-01 -2.15186272e-02 1.41586177e-02
5.64226329e-01 -4.94949996e-01 1.84757840e-02 -1.30026531e+00
5.69767296e-01 2.70624608e-01 3.03293586e-01 9.80765224e-01
-6.79812133e-01 1.39342093e+00 7.09291601e+00 1.07551441e-01
-1.20110095e+00 -6.43428937e-02 -2.32349977e-01 8.01455617e-01
-2.77238309e-01 6.99845970e-01 -4.93058145e-01 1.48831710e-01
1.55506063e+00 -5.73787153e-01 8.19429994e-01 9.39273179e-01
1.47849023e-01 -2.61582702e-01 -1.37228906e+00 7.05390096e-01
-5.75230718e-01 -1.17529786e+00 -1.06226206e-02 -3.83413762e-01
-8.81596506e-02 -2.87784606e-01 -6.90849662e-01 4.95263785e-01
9.82189834e-01 -8.69245112e-01 9.74888265e-01 6.69406176e-01
3.95489752e-01 -2.24846452e-01 5.55060148e-01 6.77538753e-01
-1.08961976e+00 -3.58631730e-01 -3.81433278e-01 -6.85271680e-01
3.49580437e-01 1.18214861e-01 -1.02346897e+00 5.86029887e-01
3.01086247e-01 6.29778132e-02 1.17108226e-01 5.67550898e-01
-3.73138428e-01 3.74670297e-01 -5.94909191e-01 -6.97987974e-01
1.99126840e-01 -3.70723546e-01 2.46711969e-01 1.25982559e+00
1.76468685e-01 6.06873155e-01 4.53727752e-01 9.86155570e-01
4.04793620e-01 4.77590598e-03 -7.66976058e-01 -2.65699536e-01
3.34280461e-01 7.52443731e-01 -6.02598786e-01 -8.14341128e-01
-5.83640456e-01 6.23931289e-01 2.54197232e-02 1.49155423e-01
-5.92736363e-01 -7.26231992e-01 6.32161498e-01 -4.12223451e-02
-5.60830161e-02 -4.36416626e-01 -3.31712335e-01 -1.04762292e+00
2.19533339e-01 -8.72060657e-01 6.72957227e-02 -1.36125696e+00
-7.81356573e-01 5.54534316e-01 2.64058322e-01 -6.68110907e-01
-8.52950573e-01 -8.51267099e-01 -4.74735349e-01 8.68169010e-01
-9.54625905e-01 -1.05298698e+00 1.22895330e-01 2.37712309e-01
4.76422429e-01 2.15560704e-01 1.05837846e+00 4.17087562e-02
4.13456326e-03 -1.07473947e-01 -4.83400553e-01 2.89037645e-01
-1.91433892e-01 -1.05041480e+00 4.32372302e-01 7.19703555e-01
-1.29743502e-01 1.14453578e+00 9.26317275e-01 -3.77640873e-01
-1.96380031e+00 -7.59946406e-01 1.21541917e+00 -5.60400963e-01
8.32206130e-01 -2.35224038e-01 -8.55321586e-01 1.19140935e+00
-1.60826996e-01 -2.29485884e-01 1.76763535e-02 -2.27439940e-01
-4.60804284e-01 2.01855153e-01 -1.26400971e+00 5.95260382e-01
9.54240441e-01 -9.65685129e-01 -1.30643260e+00 7.71188438e-02
1.21398056e+00 -4.69460845e-01 -8.96042526e-01 -7.71679655e-02
6.05519772e-01 -4.91768658e-01 6.26002550e-01 -1.03159404e+00
4.13365304e-01 -6.22188032e-01 -2.54395425e-01 -1.12802076e+00
1.49379611e-01 -6.89220488e-01 4.47692901e-01 1.01792765e+00
5.76586783e-01 -9.14443135e-01 5.08083820e-01 1.41747761e+00
-1.41101807e-01 -2.03814492e-01 -6.21438265e-01 -8.00599337e-01
4.59463179e-01 -1.01577640e+00 9.63025510e-01 4.91633087e-01
1.08611858e+00 7.06427753e-01 4.52290237e-01 -2.18035933e-02
1.22401819e-01 1.73962846e-01 7.32298970e-01 -1.41915894e+00
-1.00053489e-01 -3.91457587e-01 -5.79315007e-01 -1.35070503e+00
5.43913186e-01 -9.72704709e-01 5.16810000e-01 -1.83556795e+00
-1.49280623e-01 -5.18978894e-01 5.16870856e-01 7.48888612e-01
4.75574434e-01 -5.50254703e-01 2.42362544e-01 -3.88935506e-02
-4.43613410e-01 -1.60838664e-01 9.41503465e-01 -4.70194295e-02
-2.18597129e-01 -3.06333482e-01 -7.35785484e-01 8.29231441e-01
5.68922997e-01 -4.11712468e-01 -3.30194026e-01 -4.42544669e-01
1.65250912e-01 6.98271632e-01 3.20781767e-01 -8.86915922e-01
2.59982675e-01 -6.35296404e-01 -6.02979004e-01 2.12843463e-01
-4.73803878e-02 -1.12280035e+00 3.93619120e-01 5.90721250e-01
-4.22189146e-01 -9.84307304e-02 5.95361553e-02 -4.75022271e-02
-2.00637519e-01 -9.64082301e-01 5.25985718e-01 -3.89290154e-01
-9.51409936e-01 -4.26921010e-01 -8.02563131e-01 -2.04210460e-01
9.91253495e-01 -3.79799865e-02 -1.71843097e-01 -4.94934320e-01
-8.34125698e-01 1.85881242e-01 6.14650249e-01 3.91847670e-01
2.32837856e-01 -9.09324884e-01 -1.53404921e-01 2.09815636e-01
-2.08543196e-01 1.18353181e-01 -5.73166966e-01 2.69400269e-01
-1.19492888e+00 7.15449512e-01 -2.68692166e-01 -4.21766877e-01
-9.35405970e-01 5.98706365e-01 4.61767256e-01 9.28488001e-02
-5.55865407e-01 2.39058107e-01 -2.29129583e-01 -1.02596760e+00
1.81837231e-02 -8.86710882e-01 -1.68390963e-02 -7.32166946e-01
7.03896105e-01 -1.63165495e-01 -2.57510513e-01 -4.56708938e-01
-4.89111513e-01 5.41221499e-01 5.93903780e-01 -6.04320228e-01
1.23401582e+00 -3.14516425e-02 -6.75108373e-01 4.14171427e-01
1.16617954e+00 -3.52420002e-01 -6.37431085e-01 -1.24550685e-01
4.79539782e-01 -7.17676729e-02 -3.65011096e-01 -3.95040065e-01
-2.87120223e-01 6.34097457e-01 4.92623858e-02 9.19015765e-01
6.90722108e-01 3.23464662e-01 8.61650646e-01 1.05127633e+00
1.08044255e+00 -9.68986392e-01 -4.96845424e-01 1.00922751e+00
5.31715631e-01 -5.66082060e-01 -4.43988234e-01 -7.97001064e-01
-4.39584255e-01 1.59327984e+00 1.96887761e-01 -2.26785466e-01
5.70253551e-01 8.40734601e-01 3.54805231e-01 -2.22811982e-01
-7.81643271e-01 -1.69394463e-01 -6.28158271e-01 7.16063857e-01
3.60120356e-01 3.15952152e-01 -7.04237461e-01 8.58087420e-01
-5.97126305e-01 8.01296592e-01 8.58953536e-01 1.46155822e+00
-9.17971075e-01 -1.49917519e+00 -4.78704244e-01 8.85430649e-02
-4.18651104e-01 2.52983660e-01 -4.84178275e-01 6.49924099e-01
-2.83785552e-01 1.18228173e+00 2.49639656e-02 -3.84842932e-01
6.58684373e-01 7.54440486e-01 5.88913918e-01 -9.17825580e-01
-4.14839149e-01 -4.36711878e-01 7.61380076e-01 -3.68159652e-01
-1.72312662e-01 -4.33862239e-01 -2.26195812e+00 -5.24271190e-01
-1.70727298e-02 4.71528411e-01 8.14534009e-01 1.21468961e+00
-2.36292239e-02 2.01376542e-01 2.75885761e-01 -4.02677387e-01
-7.87517011e-01 -5.62204003e-01 -2.42472842e-01 3.64217013e-01
1.03528053e-01 -2.66356707e-01 -1.18983813e-01 5.52274406e-01] | [8.90965747833252, 7.186615467071533] |
55196b6b-09f3-4566-8abd-f9700f8411b9 | parsing-line-segments-of-floor-plan-images | 2303.03851 | null | https://arxiv.org/abs/2303.03851v1 | https://arxiv.org/pdf/2303.03851v1.pdf | Parsing Line Segments of Floor Plan Images Using Graph Neural Networks | In this paper, we present a GNN-based Line Segment Parser (GLSP), which uses a junction heatmap to predict line segments' endpoints, and graph neural networks to extract line segments and their categories. Different from previous floor plan recognition methods, which rely on semantic segmentation, our proposed method is able to output vectorized line segment and requires less post-processing steps to be put into practical use. Our experiments show that the methods outperform state-of-the-art line segment detection models on multi-class line segment detection tasks with floor plan images. In the paper, we use our floor plan dataset named Large-scale Residential Floor Plan data (LRFP). The dataset contains a total of 271,035 floor plan images. The label corresponding to each picture contains the scale information, the categories and outlines of rooms, and the endpoint positions of line segments such as doors, windows, and walls. Our augmentation method makes the dataset adaptable to the drawing styles of as many countries and regions as possible. | ['Cihui Pan', 'Mingxiang Chen'] | 2023-03-07 | null | null | null | null | ['line-segment-detection'] | ['computer-vision'] | [ 3.66715848e-01 3.45400274e-01 -1.48764312e-01 -5.70612669e-01
-4.49943036e-01 -7.30581224e-01 3.48827362e-01 4.73173589e-01
-3.82282399e-02 3.54434431e-01 1.73415542e-01 -5.72328210e-01
1.35225236e-01 -1.34256339e+00 -5.97795367e-01 6.69883192e-02
-2.13357180e-01 5.84256470e-01 3.43409002e-01 -3.24982136e-01
4.67779279e-01 6.73013091e-01 -9.37629998e-01 3.92999619e-01
9.42847490e-01 9.88294482e-01 1.37935823e-03 6.45726740e-01
-6.75240219e-01 3.07093054e-01 -6.37512922e-01 -1.12049937e-01
2.50846028e-01 -4.61080909e-01 -7.82313406e-01 4.93791044e-01
7.48538435e-01 -9.06287730e-02 -2.00641051e-01 6.94795609e-01
2.65364289e-01 -3.71639878e-02 7.75690138e-01 -1.05705202e+00
-4.73357052e-01 8.74665618e-01 -7.21208096e-01 -1.58829316e-01
6.02440596e-01 -3.85753572e-01 1.22946525e+00 -4.57771629e-01
5.71657300e-01 9.45856273e-01 9.34466422e-01 9.88054127e-02
-1.13337028e+00 -3.27625453e-01 4.66498703e-01 -1.37754560e-01
-1.36719882e+00 1.27984256e-01 9.82549667e-01 -5.38694263e-01
8.61405790e-01 3.67466599e-01 9.00622427e-01 5.71178854e-01
-1.31585076e-01 9.29630756e-01 7.95302808e-01 -6.44989312e-01
3.77388835e-01 -2.27916151e-01 3.30774218e-01 1.06598544e+00
1.62741020e-01 -6.04642153e-01 1.43186981e-02 7.58685768e-02
1.24116242e+00 -6.33548200e-02 -4.33389634e-01 -6.78494394e-01
-1.25201285e+00 7.88314104e-01 9.22744393e-01 3.27186644e-01
-1.31326199e-01 -1.38045371e-01 3.12469542e-01 -2.34637409e-01
1.51382148e-01 4.77097660e-01 -4.39380020e-01 1.55538037e-01
-1.17848229e+00 3.40492129e-02 8.99294317e-01 1.33815777e+00
9.26628530e-01 -1.56389728e-01 -1.48043141e-01 1.00322378e+00
3.13217133e-01 4.07683522e-01 3.37317228e-01 -7.02701688e-01
1.09528601e+00 1.05216205e+00 -1.62908852e-01 -9.69018698e-01
-1.03552508e+00 -2.28909731e-01 -6.01553857e-01 -1.20822988e-01
3.01642418e-01 -1.51816517e-01 -1.30082858e+00 1.00201094e+00
4.24671620e-02 -4.00343359e-01 -3.37041885e-01 4.02208537e-01
9.73354638e-01 6.40789747e-01 -2.41014466e-01 4.08236176e-01
1.46445954e+00 -1.41387129e+00 -3.58250946e-01 -6.59150779e-01
7.26074338e-01 -5.87465823e-01 1.21216309e+00 7.78595954e-02
-6.00896060e-01 -6.37111485e-01 -1.31292868e+00 1.18608680e-02
-6.18967712e-01 2.75155962e-01 8.04309785e-01 8.91651273e-01
-7.35752523e-01 5.20477712e-01 -8.86615455e-01 -7.03548729e-01
6.09370887e-01 2.42164254e-01 -2.30473235e-01 -4.28043418e-02
-3.88378739e-01 3.43675405e-01 4.94193733e-01 1.89535916e-01
7.43877292e-02 -2.75055557e-01 -1.32723010e+00 1.85538217e-01
2.26777390e-01 -6.65094018e-01 1.17573655e+00 -5.66942692e-01
-1.33727741e+00 6.39917195e-01 1.30371690e-01 -2.75726438e-01
4.75462705e-01 -2.61719733e-01 -4.30976748e-01 1.51466310e-01
3.08100522e-01 8.38095546e-01 3.08255225e-01 -1.28028095e+00
-7.29539871e-01 -3.22379559e-01 -9.71674547e-02 2.77847499e-01
1.86770916e-01 -5.02250791e-01 -7.96118438e-01 -3.73048186e-01
7.46331632e-01 -8.83466959e-01 -4.13006812e-01 -2.05077231e-01
-1.22576666e+00 -8.45770314e-02 5.82252383e-01 -6.64542377e-01
1.39440453e+00 -2.03390861e+00 -5.47454476e-01 7.99426615e-01
-1.52200520e-01 -1.72095954e-01 7.16151148e-02 7.03068018e-01
-5.32576330e-02 2.77217835e-01 -7.41230965e-01 -3.24753255e-01
3.55366468e-01 1.35108277e-01 -1.54395238e-01 3.57290894e-01
-1.97511241e-01 8.44757497e-01 -3.81315500e-01 -5.41805327e-01
5.30797482e-01 4.52619083e-02 -3.32554042e-01 -5.93218058e-02
-1.91747576e-01 1.40703106e-02 -4.27770972e-01 7.66901195e-01
7.67967761e-01 -2.56734993e-02 7.14910403e-02 -9.35501084e-02
-2.24055886e-01 2.10715011e-01 -1.17206490e+00 1.89175284e+00
-2.06998095e-01 7.78179467e-01 -5.94664335e-01 -5.78835130e-01
1.36732936e+00 -3.50318849e-01 4.05388683e-01 -7.27985084e-01
-7.75042251e-02 5.32602370e-02 -3.52083266e-01 -2.71414280e-01
7.53441811e-01 5.27457654e-01 -6.08671665e-01 7.41482005e-02
-3.27940524e-01 -6.35874450e-01 5.43059349e-01 -1.15553632e-01
9.51483130e-01 3.91919374e-01 3.88453156e-01 -1.21355690e-02
3.89905483e-01 2.75428712e-01 4.21319336e-01 6.27328753e-01
2.67366748e-02 8.43383670e-01 5.43635070e-01 -6.69607282e-01
-1.12828159e+00 -1.18910825e+00 -2.44466104e-02 8.91350567e-01
1.80285498e-01 -5.45881212e-01 -1.07368219e+00 -6.57198310e-01
-2.18461752e-01 7.87539959e-01 -3.88415158e-01 6.85999990e-01
-9.20869350e-01 -4.81226206e-01 4.99209225e-01 8.01231146e-01
1.01311040e+00 -1.14596128e+00 -5.70372462e-01 1.83180287e-01
-1.01838350e-01 -1.20622432e+00 -5.84658265e-01 2.00319558e-01
-8.76765370e-01 -1.19718158e+00 -6.02908671e-01 -1.29522777e+00
9.81675625e-01 -3.63793271e-03 1.13035727e+00 -3.50293219e-01
-2.17176273e-01 2.79709816e-01 -2.23615587e-01 -3.20424914e-01
7.82072246e-02 5.06841838e-01 -7.46113956e-01 -3.86462569e-01
1.05754197e-01 -2.73625225e-01 -5.40066957e-01 2.94570267e-01
-4.67581898e-01 4.35526669e-01 3.41869175e-01 3.33598614e-01
9.11114931e-01 -3.94464470e-03 -8.14284086e-02 -1.30545199e+00
4.86800104e-01 1.55077159e-01 -6.58984244e-01 2.67606795e-01
-4.78887349e-01 -2.17191353e-01 6.20162606e-01 4.04223263e-01
-7.47063041e-01 2.87695467e-01 -4.27592009e-01 5.51853597e-01
-6.77954972e-01 4.46268141e-01 -1.35045871e-01 2.12044314e-01
5.69855928e-01 3.63106094e-02 -8.05030882e-01 -3.59996527e-01
4.59235698e-01 5.03535748e-01 7.47233152e-01 -1.56141222e-01
8.99068475e-01 4.22875494e-01 -1.31307095e-01 -1.14221108e+00
-6.70429766e-01 -7.54026532e-01 -1.13002467e+00 -4.27466705e-02
9.61986244e-01 -5.66511631e-01 -3.02663505e-01 4.91040707e-01
-9.60857689e-01 -6.80738926e-01 -2.81019926e-01 1.55409873e-01
-7.71556616e-01 3.34247917e-01 -7.29291081e-01 -6.14048958e-01
-2.68165290e-01 -7.18888342e-01 1.01873028e+00 5.83182395e-01
-2.28752792e-01 -8.37360084e-01 6.65634274e-02 3.07663292e-01
-4.91365418e-02 6.07028723e-01 1.12058687e+00 -4.78585392e-01
-5.90429187e-01 -2.99086809e-01 -2.86456764e-01 -6.91461414e-02
9.97923687e-02 1.66799039e-01 -6.77290976e-01 1.13111988e-01
-6.51281297e-01 1.51638418e-01 7.91749358e-01 5.82909763e-01
1.26572239e+00 -1.60475954e-01 -5.79747319e-01 8.38499725e-01
1.60770535e+00 4.97588396e-01 8.72184992e-01 8.49411786e-01
1.05168331e+00 5.45493007e-01 4.34418172e-01 3.24053824e-01
7.89225340e-01 3.88751090e-01 4.11719799e-01 -6.70452237e-01
9.04604048e-02 -5.94660223e-01 -1.44743398e-02 5.97113788e-01
-1.15403786e-01 -5.09615600e-01 -1.21368849e+00 3.76000971e-01
-1.65550148e+00 -7.44207323e-01 -5.07526100e-01 1.96187472e+00
5.75876050e-02 4.69061196e-01 2.15210631e-01 1.38154432e-01
6.79199338e-01 2.96976805e-01 -3.95816475e-01 -5.79637289e-01
-8.52321163e-02 4.36714664e-02 1.00275946e+00 3.02617371e-01
-1.62330842e+00 9.16799068e-01 6.39109373e+00 3.42879921e-01
-7.41222978e-01 -6.10141277e-01 6.99262619e-01 6.58037663e-01
-2.40076393e-01 -1.66899905e-01 -7.10915565e-01 9.07961503e-02
4.57015395e-01 5.14814675e-01 3.15591007e-01 1.12308490e+00
-7.18126297e-02 -1.72120571e-01 -9.43114221e-01 1.01817143e+00
1.79179311e-01 -1.27351725e+00 -3.11676651e-01 -3.30102220e-02
7.22074628e-01 2.65451789e-01 -2.20557407e-01 3.48261356e-01
3.23909134e-01 -9.27739620e-01 8.19720089e-01 2.71784633e-01
6.54914796e-01 -7.97621071e-01 6.73839629e-01 1.28759906e-01
-1.57211912e+00 -9.85946730e-02 -1.94025129e-01 1.16971888e-01
2.72057861e-01 1.50576890e-01 -1.22746122e+00 6.69482708e-01
5.27619839e-01 5.09373128e-01 -9.66952324e-01 1.45634449e+00
-5.86622775e-01 6.24858737e-01 -6.49098277e-01 3.38192061e-02
3.95901561e-01 -4.53785747e-01 -2.80205548e-01 1.36762464e+00
4.32554126e-01 -3.27662319e-01 6.41653419e-01 6.46706700e-01
-1.28688430e-02 4.56167608e-01 -6.65071547e-01 8.00997764e-02
4.51447904e-01 1.15760815e+00 -1.64101231e+00 -1.84563160e-01
-4.36751217e-01 1.27661324e+00 1.45874232e-01 4.59453285e-01
-7.26345122e-01 -8.66816759e-01 7.79046677e-03 7.87668079e-02
6.56018734e-01 -6.00219667e-01 -7.57703125e-01 -6.96861923e-01
-2.69216336e-02 -2.67304063e-01 3.44072491e-01 -6.65179133e-01
-8.62766325e-01 6.51367188e-01 -2.39281431e-01 -1.08834851e+00
-7.27117732e-02 -6.20269656e-01 -7.60437369e-01 3.28153908e-01
-1.49845994e+00 -1.55324149e+00 -5.44970512e-01 3.28241616e-01
7.16425240e-01 1.80559769e-01 9.92218971e-01 1.75180919e-02
-6.20139718e-01 3.89223188e-01 2.00668097e-01 1.07280171e+00
2.24837422e-01 -1.53804874e+00 1.02208710e+00 6.83014452e-01
3.63548994e-01 2.60043383e-01 3.16004604e-01 -6.43546999e-01
-6.87017262e-01 -1.11177599e+00 7.52134979e-01 -8.00769255e-02
3.07683587e-01 -6.88619256e-01 -5.16153574e-01 1.05544400e+00
7.68642947e-02 -2.89591759e-01 8.95844519e-01 2.33797818e-01
-1.83890209e-01 5.38601801e-02 -9.08364415e-01 6.61974490e-01
1.21095705e+00 -1.35547608e-01 -5.00426590e-01 2.94450015e-01
4.46149886e-01 -6.40781999e-01 -6.51988506e-01 1.94736734e-01
2.88433224e-01 -1.15602076e+00 1.01752996e+00 1.37994066e-01
3.88581097e-01 -3.09048742e-01 -1.67219251e-01 -1.25087714e+00
-4.70502138e-01 -1.29079714e-01 4.92091537e-01 1.29845846e+00
6.79400623e-01 -4.67707217e-01 1.00585246e+00 4.18847501e-01
-8.38301957e-01 -5.88258564e-01 -7.06854045e-01 -4.56942469e-01
-3.89987588e-01 -4.79950845e-01 9.11275387e-01 6.50182605e-01
-4.47697453e-02 3.58670801e-01 2.36556500e-01 1.65624350e-01
3.70764077e-01 7.68066406e-01 8.24583948e-01 -1.25978482e+00
5.38896918e-02 -7.33804882e-01 -5.12756705e-01 -1.26631522e+00
-1.74438164e-01 -8.13218474e-01 2.20115006e-01 -2.49952793e+00
-4.81108397e-01 -5.81770182e-01 9.92460996e-02 6.85079098e-01
3.36587280e-01 2.01348662e-01 2.48802796e-01 7.58548221e-03
-4.38875914e-01 3.27642977e-01 1.22795343e+00 -4.61515576e-01
-6.75975323e-01 1.26989335e-01 -3.53471458e-01 1.18873084e+00
9.26487505e-01 -2.21077055e-02 -2.32663870e-01 -3.80293846e-01
7.49921128e-02 -1.39982298e-01 -1.02941878e-01 -1.38648391e+00
1.43514574e-01 1.96237233e-03 8.21537971e-01 -1.22532785e+00
2.51134813e-01 -8.06425571e-01 -2.59226948e-01 5.68225622e-01
-2.13464975e-01 1.14175744e-01 1.76114798e-01 4.82459217e-01
1.06281668e-01 -2.72372752e-01 3.12992364e-01 -3.32299292e-01
-1.07076430e+00 2.55260915e-01 -3.83885473e-01 -2.73459524e-01
1.09408808e+00 -7.05247700e-01 -4.77106571e-01 -7.99111575e-02
-5.84100783e-01 3.04723501e-01 5.74633718e-01 2.35661685e-01
6.43633008e-01 -1.27849483e+00 -4.03381169e-01 4.84581590e-01
2.47253925e-01 5.17041087e-01 1.57434464e-01 1.20361701e-01
-1.32266891e+00 5.53105116e-01 -3.52983385e-01 -5.49484849e-01
-9.64504540e-01 3.18946332e-01 1.97211131e-01 -3.26179892e-01
-9.40621197e-01 7.28500366e-01 9.93935540e-02 -6.62125528e-01
3.66598256e-02 -1.00072443e+00 -5.43133020e-01 2.05140829e-01
-8.80191452e-04 2.52284646e-01 1.44036096e-02 -5.67374945e-01
-2.76313663e-01 9.01995122e-01 1.64423734e-01 5.60098141e-02
1.39178383e+00 -4.39704768e-02 1.23303048e-01 6.19770169e-01
8.44342530e-01 1.90244243e-01 -1.11281836e+00 1.33970544e-01
2.50898987e-01 -2.04272151e-01 -3.09803843e-01 -7.07595289e-01
-7.67362773e-01 5.12927890e-01 5.88968098e-01 3.58728349e-01
8.24511230e-01 7.83475786e-02 9.66281295e-01 3.97803992e-01
5.37594438e-01 -1.31255233e+00 -2.03523487e-01 5.23286462e-01
6.46504998e-01 -1.01099849e+00 6.71130372e-03 -9.65217531e-01
-4.56977338e-01 1.53666997e+00 5.71780682e-01 -1.40583172e-01
2.52401650e-01 1.11825103e-02 1.04966238e-01 -1.97863653e-01
2.83437669e-01 -3.16657960e-01 2.33939797e-01 7.89551198e-01
3.76382679e-01 6.22450590e-01 6.79623708e-02 2.53639489e-01
-9.02343750e-01 -4.57598627e-01 5.93184471e-01 1.07177973e+00
-6.52054429e-01 -9.85948861e-01 -6.40393674e-01 6.36701047e-01
9.68404114e-02 -1.94964092e-02 -2.74576336e-01 9.77339804e-01
1.15185268e-01 7.08919704e-01 4.55344856e-01 -2.57699788e-01
7.81730056e-01 -5.42608276e-02 3.13823313e-01 -7.51087010e-01
-3.72409314e-01 5.65573573e-02 2.67744422e-01 -4.31654662e-01
-1.43788144e-01 -6.14019632e-01 -1.77276933e+00 -2.13846710e-04
-1.80042922e-01 -1.43873990e-01 8.49226654e-01 7.35145628e-01
6.62750080e-02 7.35821843e-01 4.57932174e-01 -7.51549304e-01
3.01255822e-01 -7.77924001e-01 -7.52290845e-01 3.67117226e-01
1.98289305e-02 -2.09953308e-01 2.36890435e-01 2.40134913e-03] | [8.384824752807617, -1.6990293264389038] |
c088f82a-5b2e-4929-a3db-c447ce06de52 | goal-a-challenging-knowledge-grounded-video | 2303.14655 | null | https://arxiv.org/abs/2303.14655v1 | https://arxiv.org/pdf/2303.14655v1.pdf | GOAL: A Challenging Knowledge-grounded Video Captioning Benchmark for Real-time Soccer Commentary Generation | Despite the recent emergence of video captioning models, how to generate vivid, fine-grained video descriptions based on the background knowledge (i.e., long and informative commentary about the domain-specific scenes with appropriate reasoning) is still far from being solved, which however has great applications such as automatic sports narrative. In this paper, we present GOAL, a benchmark of over 8.9k soccer video clips, 22k sentences, and 42k knowledge triples for proposing a challenging new task setting as Knowledge-grounded Video Captioning (KGVC). Moreover, we conduct experimental adaption of existing methods to show the difficulty and potential directions for solving this valuable and applicable task. | ['Yu Xu', 'Hui Liu', 'Weidong Guo', 'Jie Tang', 'Juanzi Li', 'Lei Hou', 'Bin Xu', 'Yuxiao Dong', 'Xiaozhi Wang', 'Xinyu Guan', 'Yifan Xu', 'Kunyu Gao', 'Teng Tu', 'Jifan Yu', 'Ji Qi'] | 2023-03-26 | null | null | null | null | ['video-captioning'] | ['computer-vision'] | [ 2.66745239e-01 1.13827378e-01 -1.94444746e-01 -3.69120538e-01
-1.03071582e+00 -6.19649172e-01 6.01236165e-01 -1.94504559e-01
-1.30379528e-01 1.11796093e+00 7.95066953e-01 4.19889204e-02
-7.46079385e-02 -2.82880574e-01 -9.79892313e-01 -3.76790017e-01
-3.76464538e-02 3.11799198e-01 2.57509232e-01 -3.26735049e-01
2.57396936e-01 -6.84023201e-02 -1.51500976e+00 8.00601006e-01
4.51483339e-01 9.40245926e-01 1.48586437e-01 5.88950515e-01
1.22126602e-01 1.49087882e+00 -7.19447255e-01 -9.14363623e-01
-2.51204640e-01 -6.39586031e-01 -1.14771271e+00 2.55300939e-01
5.47757328e-01 -5.40883303e-01 -6.59134865e-01 8.74031782e-01
3.48376423e-01 4.54382718e-01 4.43386137e-01 -1.50772548e+00
-9.72897708e-01 8.45724106e-01 -1.67511359e-01 3.98857087e-01
8.24825883e-01 1.99572727e-01 7.75416970e-01 -7.46743202e-01
1.02133524e+00 1.07885993e+00 3.42258126e-01 8.64253700e-01
-4.56423521e-01 -5.28401852e-01 3.67697537e-01 8.76718700e-01
-1.56338954e+00 -5.45939744e-01 8.62703621e-01 -3.00707042e-01
6.10848784e-01 4.34965104e-01 8.43378305e-01 1.73529518e+00
-3.20296913e-01 1.01973319e+00 8.35656524e-01 -1.11262195e-01
1.58056572e-01 6.10274859e-02 -3.34265642e-02 3.44569236e-01
1.43159211e-01 -3.77156347e-01 -1.22501874e+00 1.53406635e-01
7.46493638e-01 -4.65792030e-01 -5.29854298e-01 -1.72010466e-01
-1.55743921e+00 6.86380267e-01 7.35142455e-02 -7.19060153e-02
-3.85302097e-01 2.72314429e-01 7.76164114e-01 -1.31620169e-01
2.82565951e-01 5.25229335e-01 1.19036533e-01 -8.71970236e-01
-8.79989743e-01 5.88490903e-01 7.02373564e-01 1.50641870e+00
3.11069012e-01 1.06809907e-01 -6.46702766e-01 4.79499012e-01
-4.48363833e-02 5.22496343e-01 1.27763927e-01 -1.19880223e+00
7.74913132e-01 3.79527211e-01 2.94989049e-01 -1.23085451e+00
1.91425104e-02 5.90205751e-02 -5.63902318e-01 -6.40265107e-01
1.94036931e-01 -9.71486047e-02 -5.42218387e-01 1.59964132e+00
9.24324468e-02 5.62386692e-01 3.84490997e-01 1.43522894e+00
1.42241263e+00 1.03929698e+00 3.15704793e-01 -2.46219397e-01
1.56850791e+00 -1.18513882e+00 -9.92384672e-01 -5.19215643e-01
2.04526067e-01 -3.95236075e-01 1.20117807e+00 2.62867332e-01
-1.25457489e+00 -2.49148712e-01 -9.04939473e-01 -3.33885610e-01
-7.25557432e-02 -2.23828241e-01 6.34336054e-01 2.33097881e-01
-9.39803541e-01 3.86191159e-02 -2.20936030e-01 -3.40209216e-01
4.61274356e-01 -2.81676173e-01 -4.72013742e-01 -4.84996408e-01
-1.59685647e+00 8.23269188e-01 7.34395385e-01 3.91517311e-01
-1.24984610e+00 -5.31104565e-01 -1.02417505e+00 -5.88525236e-02
8.83220673e-01 -5.64092517e-01 1.45487893e+00 -9.96373296e-01
-1.49790967e+00 9.81495440e-01 -4.82532382e-03 -4.67017740e-01
5.07923424e-01 -5.72749853e-01 -4.86548245e-01 7.98042357e-01
1.22841612e-01 7.01207519e-01 6.41858280e-01 -1.24734640e+00
-4.91929710e-01 6.80687279e-02 7.87952900e-01 6.13364816e-01
-4.01364803e-01 2.88263708e-01 -9.07404006e-01 -6.35196805e-01
-4.22619522e-01 -8.26323271e-01 -1.82413012e-01 -1.19633242e-01
-4.17419761e-01 1.07360616e-01 5.15121460e-01 -1.07051826e+00
1.26393390e+00 -2.29406691e+00 4.01399732e-01 -3.65094870e-01
1.42627105e-01 1.21630386e-01 -2.25929520e-03 6.39780879e-01
2.64401525e-01 9.08450335e-02 -1.13505289e-01 -1.33037905e-03
1.00899480e-01 1.93547681e-01 -5.45003831e-01 8.69816318e-02
1.81857854e-01 1.26129913e+00 -1.35927379e+00 -1.04418433e+00
1.74193561e-01 5.18636644e-01 -3.08473647e-01 3.23417664e-01
-3.29320639e-01 4.13151830e-01 -7.81955779e-01 5.11232793e-01
2.17396066e-01 -3.83161664e-01 1.05576284e-01 -2.31701672e-01
1.56110466e-01 -2.11804390e-01 -8.08896661e-01 2.05085731e+00
-1.46344513e-01 1.02835774e+00 -1.79095402e-01 -8.98943901e-01
8.55654657e-01 5.45350373e-01 1.45473659e-01 -8.15488577e-01
-2.56890114e-02 -3.18318695e-01 -6.27806008e-01 -1.08224440e+00
8.45094383e-01 -1.16996795e-01 -5.64397514e-01 7.49741420e-02
-1.10024810e-01 -5.15873954e-02 4.90892410e-01 6.70175493e-01
8.33700538e-01 4.96493369e-01 1.67659491e-01 5.17924577e-02
3.39325339e-01 5.38502395e-01 4.79570091e-01 7.53459811e-01
-4.20762390e-01 9.33676958e-01 6.13130391e-01 -4.88358438e-01
-1.02838695e+00 -8.50503802e-01 3.92836869e-01 9.64498222e-01
7.29173779e-01 -7.98946738e-01 -9.97465014e-01 -3.10883045e-01
-6.01026475e-01 7.62524664e-01 -5.41504145e-01 -2.63655603e-01
-8.80842388e-01 -2.83871621e-01 7.21595764e-01 7.94677377e-01
6.88905060e-01 -1.37880778e+00 -6.68418169e-01 1.20181344e-01
-9.42957103e-01 -1.88761079e+00 -3.81766111e-01 -7.51667202e-01
-3.30723733e-01 -1.04697335e+00 -9.09919381e-01 -8.41220796e-01
3.72509658e-01 3.79938900e-01 1.53195333e+00 -1.76332146e-01
6.45592734e-02 6.57609820e-01 -9.61990058e-01 -2.43540570e-01
-1.70265183e-01 -2.18489692e-01 -2.48341858e-01 8.97264704e-02
2.11387828e-01 -1.58787340e-01 -4.81970847e-01 1.53340980e-01
-9.27541673e-01 8.91997039e-01 5.08355498e-01 4.32973623e-01
5.42908370e-01 -3.01221311e-01 5.12940526e-01 -5.87820947e-01
6.31055832e-01 -4.95079845e-01 -9.07158032e-02 5.95105767e-01
3.12822849e-01 -2.29441345e-01 6.17639542e-01 -5.09145558e-01
-1.24768281e+00 -1.16518751e-01 9.17945206e-02 -5.85718036e-01
-9.84746218e-02 6.89965189e-01 -2.90021658e-01 2.84695655e-01
4.84598428e-01 5.89570403e-01 -4.23358083e-01 -1.33909985e-01
6.24509335e-01 5.82838178e-01 9.95775223e-01 -8.59973371e-01
7.70808101e-01 3.94336700e-01 -2.63824373e-01 -8.54740798e-01
-1.15372992e+00 -2.13566065e-01 -3.40625107e-01 -8.95893872e-01
1.15971088e+00 -1.15875137e+00 -9.01849091e-01 1.39698729e-01
-1.23572361e+00 -2.38449126e-01 -2.56006688e-01 3.21414024e-01
-1.02321661e+00 4.75080967e-01 -5.11873364e-01 -7.15322614e-01
-3.02720606e-01 -1.07188022e+00 1.02011645e+00 3.45333725e-01
-1.86655998e-01 -6.03786230e-01 -1.88540339e-01 8.70903194e-01
1.83632344e-01 6.24885738e-01 5.16559541e-01 -4.10600483e-01
-8.05460870e-01 -1.69167414e-01 -2.80334353e-01 -1.44478993e-03
-4.95632261e-01 -9.76136327e-02 -6.61332965e-01 7.88537413e-02
-3.39168161e-01 -8.62487555e-01 3.72646660e-01 8.37997273e-02
1.37248814e+00 -5.21066308e-01 -1.15503579e-01 3.85497838e-01
1.15172637e+00 2.18268871e-01 9.19721067e-01 3.74934435e-01
7.21330464e-01 5.66521287e-01 9.69881594e-01 4.53597218e-01
7.13344574e-01 7.95654178e-01 3.95474792e-01 2.92451501e-01
-2.96963483e-01 -6.51327431e-01 2.57236063e-01 8.47721219e-01
-3.76782179e-01 -3.90091419e-01 -9.12052512e-01 6.95764840e-01
-2.18696380e+00 -1.40388405e+00 -4.08114254e-04 1.55971825e+00
8.52714479e-01 5.07896729e-02 2.16635922e-03 -7.20027313e-02
9.46661890e-01 3.30075830e-01 -5.32916605e-01 -2.95845151e-01
-3.12640905e-01 -3.26594263e-01 4.67417985e-02 1.69896320e-01
-8.62812281e-01 1.08791554e+00 6.32607889e+00 1.04163897e+00
-6.89296484e-01 6.33756816e-02 6.83640182e-01 -2.52663881e-01
-2.24908739e-01 -7.68444762e-02 -6.58280909e-01 4.33513761e-01
9.27482784e-01 -5.37892938e-01 4.81797129e-01 9.03403640e-01
1.70360476e-01 -3.20479251e-03 -1.09214067e+00 1.42068768e+00
6.66532516e-01 -1.67059946e+00 5.06234288e-01 -5.13455033e-01
6.06379271e-01 -7.78477848e-01 -1.89877167e-01 6.43827081e-01
-3.88758853e-02 -1.06738067e+00 1.26697958e+00 5.48445523e-01
9.46769238e-01 -5.99447072e-01 6.49815977e-01 2.66092449e-01
-1.06592453e+00 8.14490914e-02 -1.79797590e-01 -2.27068916e-01
7.23453343e-01 8.22213814e-02 -4.54676330e-01 8.70539546e-01
8.87567401e-01 9.59517896e-01 -4.61568296e-01 9.80486035e-01
-3.12907279e-01 6.01062059e-01 2.10150912e-01 -2.85951167e-01
4.47687060e-01 1.81875795e-01 4.70545202e-01 1.25230491e+00
1.73416883e-01 7.19471514e-01 1.09160461e-01 4.42638367e-01
-9.09904465e-02 1.27928838e-01 -4.81076956e-01 -3.23902309e-01
4.17594433e-01 9.74047244e-01 -6.68968439e-01 -5.31046748e-01
-4.25157845e-01 9.62710619e-01 4.09151852e-01 5.04576445e-01
-1.27035964e+00 -5.76411411e-02 4.68281448e-01 9.41381529e-02
1.90978631e-01 -5.71601093e-02 9.82261226e-02 -1.43094611e+00
3.80520105e-01 -9.76553619e-01 4.48717564e-01 -1.49437022e+00
-1.12234938e+00 6.56755924e-01 4.92566824e-01 -1.34548116e+00
-2.12025270e-01 -4.76822853e-01 -3.53966296e-01 2.59596556e-01
-1.42234945e+00 -1.24289668e+00 -7.01948345e-01 6.33747458e-01
9.23479438e-01 1.48521513e-01 5.67469478e-01 2.24884197e-01
-4.92121011e-01 2.54931480e-01 -2.04688460e-01 2.38501489e-01
6.87184215e-01 -7.20604718e-01 2.75646776e-01 6.70942128e-01
-6.74128234e-02 8.69277716e-02 1.05233359e+00 -5.42259037e-01
-1.67229795e+00 -9.99016941e-01 8.22111309e-01 -6.35877252e-01
6.74300194e-01 -4.61385220e-01 -8.32203925e-01 7.44901121e-01
3.58540714e-01 -1.95369661e-01 6.16598606e-01 -2.85075486e-01
-3.28892946e-01 -2.61941040e-03 -6.94782317e-01 9.15251315e-01
1.47977257e+00 -4.59914356e-01 -7.85174131e-01 5.64013064e-01
9.76276100e-01 -6.66079462e-01 -7.39456475e-01 2.38871247e-01
4.53471214e-01 -7.30870485e-01 1.10925615e+00 -8.83978724e-01
1.09360158e+00 -1.51937157e-01 -2.58398265e-01 -8.54597926e-01
-2.06420854e-01 -6.67799413e-01 -2.35557050e-01 1.32242608e+00
1.39517546e-01 3.77485007e-01 7.83079803e-01 8.69930327e-01
-5.01741409e-01 -6.59498870e-01 -8.47692609e-01 -6.55893683e-01
-3.26444417e-01 -5.16241014e-01 4.84227210e-01 9.79561329e-01
4.01195258e-01 3.24374527e-01 -1.05848646e+00 1.75724993e-03
3.75838131e-01 1.46083906e-01 7.25133002e-01 -7.43476033e-01
1.25860469e-02 -1.63836181e-01 -6.58135414e-01 -9.52843845e-01
3.37044954e-01 -5.05042136e-01 1.78986445e-01 -1.89058721e+00
6.58643007e-01 3.09693933e-01 -5.29639870e-02 3.87619108e-01
-2.71992683e-01 4.64397132e-01 4.59423363e-01 8.72048140e-02
-1.57004261e+00 6.56524181e-01 1.44535172e+00 -1.72625348e-01
1.99152142e-01 -7.44814038e-01 -8.22811604e-01 5.11504352e-01
5.38037896e-01 -1.40527308e-01 -6.44261479e-01 -5.68197846e-01
5.39230108e-01 6.51603341e-01 6.24198139e-01 -1.03125083e+00
1.77047670e-01 -5.01894474e-01 1.04004718e-01 -4.19809163e-01
7.53482282e-01 -4.00960118e-01 3.64337623e-01 -2.90310606e-02
-5.99102736e-01 1.08170480e-01 3.40114683e-01 6.17000997e-01
-5.83155274e-01 -9.73057896e-02 1.44983917e-01 -3.11654538e-01
-1.50802493e+00 3.32644850e-01 -4.05416906e-01 6.49283767e-01
1.43035519e+00 -5.67301154e-01 -5.24748147e-01 -8.33409667e-01
-7.55750597e-01 4.46805179e-01 3.29512954e-01 7.67762184e-01
1.06398821e+00 -1.56459200e+00 -1.08267689e+00 -5.49893260e-01
5.56067526e-01 -9.34551880e-02 1.01627898e+00 1.80143297e-01
-5.99775314e-01 5.11731863e-01 -5.21739781e-01 -2.85461426e-01
-1.04116845e+00 7.69654930e-01 -1.68898329e-01 1.76477507e-01
-9.56037402e-01 7.69451082e-01 3.42686832e-01 4.36025470e-01
2.45638385e-01 1.91015571e-01 -6.46932960e-01 -6.87246770e-02
9.40957546e-01 1.35492951e-01 -3.32913488e-01 -9.45171416e-01
-4.81041551e-01 4.46190357e-01 6.25155419e-02 -1.21460199e-01
1.08700037e+00 -3.90945852e-01 2.82190442e-01 4.99279141e-01
7.87794828e-01 -5.82671583e-01 -1.42466056e+00 -1.55994326e-01
-1.22199290e-01 -5.41747332e-01 -2.88266867e-01 -8.58464599e-01
-8.77203107e-01 7.49845743e-01 -5.91494739e-02 -2.32516900e-01
1.05828929e+00 2.32225776e-01 8.49358857e-01 3.77850384e-01
4.91411448e-01 -9.80578423e-01 2.91198164e-01 3.08355838e-01
1.21748471e+00 -1.07284331e+00 -6.97908849e-02 -4.72820371e-01
-1.42214775e+00 9.38717902e-01 8.73368502e-01 1.41801387e-01
4.99555580e-02 -1.38143897e-01 8.28712732e-02 -2.67255276e-01
-8.68847907e-01 1.77949518e-01 1.23287484e-01 7.42518067e-01
7.77072161e-02 3.07387230e-03 -1.93233520e-01 9.79306519e-01
-2.43753552e-01 2.34500065e-01 9.66194928e-01 8.93438816e-01
-2.92468041e-01 -1.96984589e-01 -2.76164502e-01 8.19875970e-02
-3.63835841e-01 -8.62097889e-02 -3.18500072e-01 7.56118536e-01
-2.64363527e-01 9.90791082e-01 -8.11969414e-02 -2.58653015e-01
4.67122972e-01 -1.30309211e-02 3.73538673e-01 -4.75288540e-01
-3.53945583e-01 -5.04293919e-01 5.10261118e-01 -6.64793253e-01
-7.61012793e-01 -4.34842706e-01 -1.01015401e+00 -3.52242202e-01
3.18001956e-01 4.37383503e-01 3.85941982e-01 1.04530478e+00
2.81791449e-01 3.05196732e-01 8.81453007e-02 -8.66654456e-01
-3.35791707e-02 -6.38001680e-01 -3.47337842e-01 7.75458932e-01
4.60906141e-03 -6.63892269e-01 -3.84611785e-02 5.75514317e-01] | [10.486340522766113, 0.8389461040496826] |
7601bc8c-bc7c-42dc-8c69-0a20a270b96c | towards-solving-text-based-games-by-producing | 1812.00855 | null | http://arxiv.org/abs/1812.00855v1 | http://arxiv.org/pdf/1812.00855v1.pdf | Towards Solving Text-based Games by Producing Adaptive Action Spaces | To solve a text-based game, an agent needs to formulate valid text commands
for a given context and find the ones that lead to success. Recent attempts at
solving text-based games with deep reinforcement learning have focused on the
latter, i.e., learning to act optimally when valid actions are known in
advance. In this work, we propose to tackle the first task and train a model
that generates the set of all valid commands for a given context. We try three
generative models on a dataset generated with Textworld. The best model can
generate valid commands which were unseen at training and achieve high $F_1$
score on the test set. | ['Layla El Asri', 'Marc-Alexandre Côté', 'Xingdi Yuan', 'Ruo Yu Tao'] | 2018-12-03 | null | null | null | null | ['text-based-games'] | ['playing-games'] | [ 1.59666225e-01 4.78801504e-02 2.04992667e-01 -2.98016459e-01
-7.46683836e-01 -5.38370430e-01 6.53126180e-01 -6.58437461e-02
-7.38169312e-01 1.05565679e+00 1.16388716e-01 -2.90946275e-01
-1.18476026e-01 -1.23839700e+00 -6.64243698e-01 -4.83824551e-01
2.17279240e-01 1.12331927e+00 2.10924506e-01 -6.26758039e-01
5.50667405e-01 -7.04731867e-02 -1.53150034e+00 2.00778961e-01
1.04460645e+00 5.55183530e-01 6.47056103e-01 1.09940279e+00
-3.00168395e-01 1.24072993e+00 -1.22469091e+00 -5.04928112e-01
4.57146503e-02 -1.15799201e+00 -1.01965404e+00 -2.19361782e-01
-2.32830837e-01 -5.20879030e-01 -2.30028346e-01 1.08941031e+00
4.75779861e-01 4.28423673e-01 5.02144933e-01 -1.03461623e+00
-3.75821173e-01 8.34223390e-01 2.39863753e-01 2.14158893e-01
5.66461265e-01 5.04104495e-01 1.27112722e+00 -3.00050944e-01
7.20182419e-01 8.99118662e-01 2.72509232e-02 1.16376722e+00
-1.13050270e+00 -3.25879782e-01 8.75372365e-02 2.60593474e-01
-1.10932481e+00 -8.29673484e-02 5.69175541e-01 -1.43833145e-01
1.32416975e+00 2.58644909e-01 8.39645147e-01 1.41574943e+00
5.78473687e-01 8.04888725e-01 9.68836546e-01 -3.67543906e-01
7.94024885e-01 -5.17321706e-01 -3.12052429e-01 4.95595098e-01
-6.44418821e-02 3.90453398e-01 -6.09095037e-01 9.15201306e-02
7.67111838e-01 -4.45961148e-01 2.16842405e-02 -4.07990664e-02
-8.50722492e-01 1.23961365e+00 3.51532712e-03 3.60065490e-01
-4.80813384e-01 3.64135116e-01 2.30851725e-01 3.59835088e-01
8.50901008e-02 1.24560928e+00 -3.17105353e-01 -8.32878530e-01
-6.05236948e-01 7.95188367e-01 9.16935444e-01 7.29083598e-01
6.01537287e-01 3.71855557e-01 -3.76414508e-01 6.86582506e-01
1.14114650e-01 4.38011050e-01 7.05127954e-01 -6.05292261e-01
5.38639545e-01 3.20559084e-01 5.98040968e-02 -2.56827354e-01
-2.45140791e-01 -2.21067324e-01 -1.71250224e-01 2.77268708e-01
3.82312328e-01 -5.82090497e-01 -1.10003817e+00 1.95299172e+00
3.23676988e-02 2.87442058e-01 4.11118835e-01 6.74406052e-01
7.54639268e-01 8.34134281e-01 1.07992828e-01 -1.44036248e-01
7.32421577e-01 -6.49736881e-01 -5.82829654e-01 -7.61231244e-01
6.43116057e-01 -4.41463411e-01 1.09462702e+00 6.32533729e-01
-1.05326414e+00 -4.15883034e-01 -9.33965683e-01 5.30382156e-01
-2.14198679e-01 -2.52129555e-01 4.69845325e-01 6.80139661e-01
-1.08199918e+00 6.15415871e-01 -6.34754181e-01 -1.53759390e-01
5.90773709e-02 4.38423544e-01 9.15224031e-02 -1.45318955e-01
-1.37688422e+00 8.75885069e-01 8.77275586e-01 -1.99060932e-01
-1.64683211e+00 1.26781100e-02 -8.45753372e-01 1.25785872e-01
9.14982080e-01 -5.71206748e-01 1.65310442e+00 -8.06722045e-01
-1.85034049e+00 3.37715536e-01 1.16875647e-02 -6.36133432e-01
4.65571433e-01 -1.74726114e-01 7.32517540e-02 -1.73177913e-01
2.58693516e-01 5.36475360e-01 8.95498335e-01 -9.76777256e-01
-8.43272865e-01 -7.68714026e-02 4.61128920e-01 5.32511115e-01
1.07211834e-02 -1.36514589e-01 -3.34613681e-01 -4.37321097e-01
-1.84159949e-01 -9.50337529e-01 -4.43319052e-01 -1.02385604e+00
-4.00655895e-01 -5.87802172e-01 2.83467710e-01 -3.72714758e-01
1.08000278e+00 -1.63536417e+00 4.46092159e-01 9.34101045e-02
1.48512358e-02 4.20233428e-01 -3.77783477e-01 7.07845330e-01
9.17310119e-02 2.41469085e-01 1.17721163e-01 -5.79458773e-02
4.58330885e-02 7.50491142e-01 -3.06876957e-01 -2.89530814e-01
3.22203636e-01 1.18365097e+00 -1.09106672e+00 -2.93610156e-01
2.59303182e-01 -1.72076061e-01 -9.24637973e-01 8.34814072e-01
-9.55525160e-01 3.60760063e-01 -6.88811362e-01 1.15564026e-01
9.01437849e-02 1.68903936e-02 4.62755382e-01 8.30141187e-01
1.70371681e-01 4.23005462e-01 -1.02260613e+00 1.53607607e+00
-4.93434668e-01 2.99669385e-01 -4.97248411e-01 -1.05231607e+00
1.03019607e+00 3.25599790e-01 3.33500713e-01 -1.05866098e+00
2.15811864e-01 8.76335055e-02 3.55159849e-01 -6.29593909e-01
6.79678261e-01 -3.62612724e-01 -2.80781239e-01 4.41378295e-01
2.78299004e-01 -5.62672138e-01 4.97601509e-01 1.05695181e-01
1.30948400e+00 3.31916988e-01 1.45959139e-01 1.22301213e-01
2.98048198e-01 9.26314369e-02 5.17453671e-01 1.39109004e+00
6.39973730e-02 4.21232164e-01 7.63638258e-01 -5.86800516e-01
-9.27160859e-01 -9.01764333e-01 6.19455099e-01 1.27865982e+00
-2.95387711e-02 -6.13973916e-01 -9.95175958e-01 -8.84852886e-01
-5.73800981e-01 1.14595139e+00 -7.10119903e-01 -3.43643546e-01
-7.14100063e-01 -4.70608324e-01 4.25046712e-01 4.46985424e-01
5.13285041e-01 -1.64245725e+00 -9.68529522e-01 5.12453973e-01
-5.25130868e-01 -8.40362787e-01 -1.36032104e-01 5.12705922e-01
-5.86821556e-01 -1.20667899e+00 -3.80560577e-01 -7.03459501e-01
3.84602606e-01 -4.28780854e-01 1.47354960e+00 3.14880937e-01
-7.64571726e-02 4.24109064e-02 -8.54797781e-01 -5.18932223e-01
-7.16289818e-01 2.15826228e-01 -2.62454212e-01 -4.35754806e-01
2.49462247e-01 -1.86488599e-01 -5.96050583e-02 1.57441616e-01
-8.91273379e-01 2.17446491e-01 5.53632259e-01 1.26212728e+00
4.41821158e-01 3.29984128e-01 5.68895340e-01 -7.82583058e-01
1.19511855e+00 -3.52366209e-01 -5.23218274e-01 2.68243074e-01
-5.04966915e-01 3.54127914e-01 1.09784639e+00 -2.32768714e-01
-8.45603943e-01 -6.86563319e-03 -7.77723551e-01 -6.48753792e-02
-3.47006619e-01 7.16970205e-01 -8.32824782e-02 3.95678341e-01
8.41437876e-01 9.17680562e-01 -4.15775180e-01 -2.29093991e-02
1.16775677e-01 2.70283759e-01 2.99730301e-01 -1.08109331e+00
5.97464621e-01 -3.94707590e-01 -1.49176329e-01 -6.17756307e-01
-7.73779452e-01 -1.05671529e-02 -3.12366188e-01 -3.76934499e-01
9.75337148e-01 -3.21305275e-01 -7.85867095e-01 4.98302221e-01
-1.09239769e+00 -9.06267226e-01 -4.33327466e-01 2.83310503e-01
-1.07563663e+00 -5.61323389e-02 -1.11428060e-01 -8.21267843e-01
-2.81264037e-01 -1.36001205e+00 7.25788772e-01 4.99202222e-01
-2.92568564e-01 -8.73154402e-01 4.30126250e-01 1.57771841e-01
3.38482767e-01 1.72113195e-01 8.74966383e-01 -1.10149276e+00
-5.16009629e-01 -1.41680211e-01 4.75676835e-01 1.93617612e-01
9.57954153e-02 -2.10228056e-01 -5.80622077e-01 -3.38507921e-01
1.09499827e-01 -7.72809803e-01 3.23537052e-01 4.70324099e-01
1.08829486e+00 -3.21186602e-01 1.78907499e-01 3.50685894e-01
1.31528234e+00 9.33012068e-01 9.11240637e-01 3.89535278e-01
2.11281061e-01 2.05772832e-01 8.06195319e-01 6.10174894e-01
1.48732990e-01 7.25423872e-01 7.32168138e-01 5.38304925e-01
3.06610286e-01 -6.64861679e-01 1.64792702e-01 2.20012754e-01
-1.58056766e-01 -6.97555542e-01 -1.02936196e+00 4.75696117e-01
-1.95014000e+00 -1.16241026e+00 4.60819185e-01 1.91401768e+00
8.62416446e-01 7.00655997e-01 1.30541250e-01 2.13744957e-02
3.40944588e-01 1.97459877e-01 -6.75153375e-01 -6.06949806e-01
8.80461410e-02 1.00738049e+00 -1.04842626e-01 6.09124243e-01
-8.22797418e-01 1.38383555e+00 6.69813538e+00 9.15486455e-01
-1.14479959e+00 -1.56679630e-01 6.17379129e-01 -1.86805353e-01
-2.26851568e-01 -4.55612727e-02 -7.18844950e-01 4.16619956e-01
1.08441317e+00 -4.70723867e-01 9.08971846e-01 9.58097994e-01
2.54678279e-01 -3.67892444e-01 -1.07866037e+00 6.69760406e-01
-3.99025269e-02 -1.34139752e+00 2.37195626e-01 2.78375857e-02
7.09498227e-01 -6.13992177e-02 -2.50837707e-04 8.91615033e-01
1.00700927e+00 -1.45425797e+00 7.06610501e-01 2.60128468e-01
4.45432335e-01 -9.74419415e-01 7.66384482e-01 8.63328934e-01
-8.57574701e-01 -2.00349301e-01 -3.22391897e-01 -3.20968449e-01
-1.24847330e-01 -1.35814667e-01 -1.28922713e+00 5.20562768e-01
2.27146432e-01 3.58525097e-01 -4.65892732e-01 8.95397365e-01
-8.43805730e-01 7.92168260e-01 -2.89023425e-02 -8.05248559e-01
6.14357412e-01 -4.62765582e-02 4.21138287e-01 7.26774573e-01
3.44351172e-01 2.69131273e-01 3.89872819e-01 1.03609157e+00
1.37842774e-01 7.86661804e-02 -6.71532691e-01 -1.68202415e-01
2.41549224e-01 8.21722567e-01 -6.98212683e-01 -1.32712051e-01
4.79853712e-02 1.02766597e+00 4.74069118e-01 1.75781325e-01
-9.44728434e-01 -3.81271482e-01 4.99619663e-01 -1.68284550e-01
2.52660900e-01 -3.22237551e-01 -2.43238881e-01 -1.06521702e+00
-2.31519416e-01 -1.23000824e+00 4.15082246e-01 -7.47854292e-01
-7.51496732e-01 7.71907985e-01 -1.13512829e-01 -8.45187426e-01
-1.06767571e+00 -5.55617929e-01 -1.03319645e+00 9.12237346e-01
-1.08485639e+00 -7.85019517e-01 -1.97062045e-01 6.25221431e-01
1.07110918e+00 -4.20637846e-01 8.79670441e-01 -3.20629746e-01
-5.49862862e-01 5.23200452e-01 -1.54006928e-01 3.87960464e-01
9.74209756e-02 -1.58581424e+00 8.33949089e-01 1.03445888e+00
5.66188931e-01 5.43134987e-01 9.57413852e-01 -7.74799407e-01
-1.34321153e+00 -8.85683656e-01 5.83592713e-01 -4.10082012e-01
3.11343133e-01 -2.67890632e-01 -5.00859857e-01 6.06771529e-01
1.44411266e-01 -5.77556849e-01 3.72301400e-01 -9.63147171e-03
1.75618157e-01 3.49870712e-01 -9.17948484e-01 7.90037870e-01
8.82117987e-01 -2.19278485e-01 -7.18918443e-01 1.63018048e-01
3.73549759e-01 -8.26005459e-01 -4.62644488e-01 -8.98404121e-02
6.20837845e-02 -9.33742464e-01 6.37276292e-01 -1.13063848e+00
7.55945683e-01 2.75620027e-03 -2.02334914e-02 -1.97721732e+00
-2.16419995e-01 -8.43628824e-01 1.91828936e-01 7.00999737e-01
4.07099038e-01 -3.85368794e-01 1.14199507e+00 5.26491582e-01
-1.96770787e-01 -8.36284459e-01 -1.08946800e+00 -8.41074646e-01
3.94933432e-01 -5.48457086e-01 7.34022319e-01 4.74035203e-01
1.23872869e-01 4.74421054e-01 -6.40159786e-01 -2.50991732e-01
2.21214101e-01 6.39426410e-02 9.88791823e-01 -1.05974269e+00
-6.62921250e-01 -3.20448279e-01 -8.94632190e-02 -1.15101111e+00
2.87063420e-01 -6.83127224e-01 5.70579648e-01 -1.61360550e+00
6.84041753e-02 -3.56806248e-01 1.76196456e-01 5.38850486e-01
-4.19539064e-01 -1.80599302e-01 2.77506143e-01 -2.03408375e-01
-7.73651540e-01 6.54463887e-01 1.27960002e+00 -8.51808786e-02
-2.74796337e-01 1.77638575e-01 -7.38542676e-01 3.18004489e-01
1.16397047e+00 -4.42267627e-01 -6.76873744e-01 -4.26929355e-01
5.10315716e-01 4.46727067e-01 6.00217795e-03 -1.17244959e+00
1.72035843e-01 -1.01415992e+00 3.79357398e-01 -2.17837051e-01
3.07310373e-01 -3.42510164e-01 -5.01942188e-02 6.54423475e-01
-6.13401353e-01 2.26697221e-01 2.10550532e-01 4.13800657e-01
-7.03692213e-02 -8.58444273e-01 6.35745227e-01 -4.15580392e-01
-9.73880947e-01 1.11992240e-01 -8.25919390e-01 5.97776949e-01
1.23268139e+00 5.96201830e-02 -2.03984246e-01 -8.40960920e-01
-8.38874280e-01 2.52433002e-01 1.60117358e-01 5.27228475e-01
9.34046626e-01 -9.97625411e-01 -8.25117409e-01 2.32002333e-01
-1.79360956e-02 -6.42640814e-02 1.89936603e-03 -1.84479013e-01
-5.26445508e-01 4.27451253e-01 -5.23548782e-01 -2.46541947e-01
-1.17665303e+00 2.88338095e-01 7.59343982e-01 -7.79687583e-01
-3.09414685e-01 8.79888594e-01 -3.45031172e-02 -3.12629044e-01
-7.01068044e-02 -1.84916660e-01 -4.04399067e-01 -4.24404800e-01
5.17366529e-01 -9.17882174e-02 1.38227627e-01 -3.31571102e-01
-6.85421228e-02 1.03934124e-01 -1.23479657e-01 -4.55551147e-01
1.34474981e+00 5.52789688e-01 3.57879698e-01 2.45292690e-02
5.62327981e-01 -3.73016655e-01 -1.18688428e+00 -2.65689958e-02
-9.06716362e-02 -6.01567090e-01 -1.81566775e-01 -1.07685673e+00
-8.75302792e-01 6.95830464e-01 2.22663939e-01 4.72865313e-01
9.51501846e-01 -1.55286252e-01 5.67762673e-01 7.30927169e-01
5.57070732e-01 -1.31037164e+00 7.82333493e-01 1.23986149e+00
6.39967382e-01 -9.25212443e-01 -5.05621374e-01 1.12861387e-01
-8.78691494e-01 1.13161802e+00 1.13759995e+00 -9.17199031e-02
2.86809299e-02 1.48303986e-01 -1.40934706e-01 -2.67833412e-01
-1.01709509e+00 -5.56005776e-01 -1.23876147e-01 9.07582819e-01
2.28831187e-01 1.29448369e-01 -3.76452535e-01 4.15838569e-01
-6.12846136e-01 5.14671914e-02 7.91377902e-01 1.26733172e+00
-7.53501594e-01 -1.43700647e+00 -2.07119122e-01 7.35861003e-01
-3.92564505e-01 -5.95664345e-02 -7.10084558e-01 4.14705127e-01
-1.54613275e-02 1.22700191e+00 -1.10714547e-01 -8.11938703e-01
4.54082310e-01 3.16801071e-01 4.90967423e-01 -1.02490735e+00
-7.38702655e-01 -1.37971774e-01 2.14616224e-01 -4.39538091e-01
1.77731723e-01 -6.95219040e-01 -1.33416617e+00 -3.06704879e-01
-1.48055181e-01 4.99063224e-01 4.12330866e-01 1.21516895e+00
-7.90561289e-02 6.71390295e-01 6.62566483e-01 -5.50676882e-01
-7.77844250e-01 -8.83030474e-01 -3.46606463e-01 4.21656221e-01
-1.72680885e-01 -4.63671863e-01 1.32015526e-01 -2.62263447e-01] | [3.7598073482513428, 1.516864538192749] |
b3948b90-342e-4ce5-a6ce-24c008e96d5c | clare-conservative-model-based-reward | 2302.04782 | null | https://arxiv.org/abs/2302.04782v2 | https://arxiv.org/pdf/2302.04782v2.pdf | CLARE: Conservative Model-Based Reward Learning for Offline Inverse Reinforcement Learning | This work aims to tackle a major challenge in offline Inverse Reinforcement Learning (IRL), namely the reward extrapolation error, where the learned reward function may fail to explain the task correctly and misguide the agent in unseen environments due to the intrinsic covariate shift. Leveraging both expert data and lower-quality diverse data, we devise a principled algorithm (namely CLARE) that solves offline IRL efficiently via integrating "conservatism" into a learned reward function and utilizing an estimated dynamics model. Our theoretical analysis provides an upper bound on the return gap between the learned policy and the expert policy, based on which we characterize the impact of covariate shift by examining subtle two-tier tradeoffs between the exploitation (on both expert and diverse data) and exploration (on the estimated dynamics model). We show that CLARE can provably alleviate the reward extrapolation error by striking the right exploitation-exploration balance therein. Extensive experiments corroborate the significant performance gains of CLARE over existing state-of-the-art algorithms on MuJoCo continuous control tasks (especially with a small offline dataset), and the learned reward is highly instructive for further learning. | ['Junshan Zhang', 'Ju Ren', 'Sen Lin', 'Zhaofeng Zhang', 'Wei Shao', 'Guanbo Wang', 'Sheng Yue'] | 2023-02-09 | null | null | null | null | ['continuous-control'] | ['playing-games'] | [ 4.53424044e-02 3.57651085e-01 -5.05587101e-01 1.29427537e-01
-7.08672166e-01 -6.12167060e-01 3.35101157e-01 -1.00605085e-01
-7.69206703e-01 1.07748830e+00 2.06603184e-02 -4.09931272e-01
-5.14145195e-01 -2.95751899e-01 -1.03416288e+00 -8.46032858e-01
-2.97358483e-01 3.30105782e-01 -2.19426960e-01 -3.17530513e-01
2.96140879e-01 8.10851008e-02 -1.22219682e+00 -3.46335441e-01
1.22861159e+00 1.13009644e+00 3.01767856e-01 4.89158362e-01
5.60548127e-01 8.95503402e-01 -4.90241587e-01 -2.08470598e-02
6.64938629e-01 -4.76331532e-01 -4.51802969e-01 -6.05417974e-02
-7.09515959e-02 -8.01539123e-01 -3.76834959e-01 1.04517663e+00
4.74922448e-01 3.36550593e-01 4.16380376e-01 -1.37055671e+00
-5.66374838e-01 7.20898747e-01 -6.64084613e-01 3.38186145e-01
-1.65318042e-01 6.48157537e-01 1.00725317e+00 -3.60858113e-01
5.24063349e-01 1.17557442e+00 3.96389306e-01 7.15623140e-01
-1.28083003e+00 -7.89404750e-01 4.90928978e-01 -1.23480214e-02
-7.63260126e-01 -1.80146083e-01 5.04735470e-01 -4.89077389e-01
4.68943506e-01 -7.14577213e-02 7.48737812e-01 1.34931445e+00
2.48583242e-01 9.07908142e-01 1.41519034e+00 1.72281526e-02
6.60301149e-01 9.52913463e-02 -2.24519759e-01 5.62761903e-01
3.94596428e-01 9.85065043e-01 -7.20266581e-01 -6.28014132e-02
1.14501166e+00 -1.24612600e-01 -4.49422717e-01 -7.96200097e-01
-1.05213392e+00 8.23443890e-01 4.40959692e-01 -3.69978726e-01
-5.48807859e-01 5.18618822e-01 4.12406385e-01 6.64834321e-01
1.51019424e-01 8.37199509e-01 -6.90843642e-01 -5.59123099e-01
-3.52697760e-01 5.40971518e-01 6.67080522e-01 8.45306814e-01
4.95131880e-01 2.23859370e-01 -2.73230761e-01 3.08794081e-01
7.95447379e-02 4.09031004e-01 3.96410704e-01 -1.36590421e+00
9.00165021e-01 2.55075723e-01 8.97580206e-01 -5.59582949e-01
-2.40114003e-01 -7.98845232e-01 -4.13144469e-01 5.83735585e-01
8.37986588e-01 -6.71103597e-01 -4.51133788e-01 2.23383760e+00
4.36643571e-01 -5.86740486e-03 1.14614695e-01 1.23299706e+00
-2.57964432e-01 2.98997551e-01 5.59952110e-02 -3.82772684e-01
8.62877846e-01 -1.01167738e+00 -6.57353699e-01 -4.30687517e-01
5.90896130e-01 4.22119629e-03 1.32465541e+00 3.91277671e-01
-1.16661680e+00 -2.54238755e-01 -1.07044303e+00 2.96965659e-01
2.94434488e-01 9.51276571e-02 5.42570055e-01 3.16437006e-01
-6.45008326e-01 1.12177372e+00 -9.19754148e-01 1.64840966e-02
4.86259460e-01 3.68267745e-01 2.18437873e-02 3.11065674e-01
-1.08394682e+00 8.62389863e-01 2.33203769e-01 1.32438779e-01
-1.48434603e+00 -1.15817165e+00 -5.03610969e-01 2.04911590e-01
1.23620272e+00 -6.16147816e-01 1.52104211e+00 -1.01022649e+00
-1.94265831e+00 1.84985012e-01 3.93137068e-01 -7.76461720e-01
1.08393013e+00 -5.46517730e-01 2.28617772e-01 -6.33194521e-02
7.02556223e-02 5.21849751e-01 1.09406090e+00 -1.20512402e+00
-6.68698192e-01 -4.95856315e-01 2.39063248e-01 5.27164042e-01
-3.27984005e-01 -6.66485071e-01 1.84480190e-01 -6.01409972e-01
-4.10836935e-01 -1.12734938e+00 -5.31960964e-01 1.24791034e-01
-5.70582822e-02 -1.58848926e-01 4.61502641e-01 -5.36879063e-01
1.20613456e+00 -1.97498524e+00 2.92100459e-01 1.34106288e-02
2.69837409e-01 -8.24312028e-03 -1.88360453e-01 2.99502283e-01
1.56027675e-01 -1.26908883e-01 -1.78316846e-01 -4.78764698e-02
1.91050112e-01 2.87698269e-01 -8.33550215e-01 6.22927666e-01
-1.42914401e-02 1.07281637e+00 -1.24013698e+00 1.65324703e-01
-1.40926585e-01 -1.14280567e-01 -7.36510992e-01 4.30490017e-01
-4.55407530e-01 9.15134847e-01 -8.92934799e-01 2.71322101e-01
3.37585539e-01 -2.27775112e-01 3.66145998e-01 3.59609812e-01
-4.05879878e-02 1.38361022e-01 -1.03036678e+00 1.56921113e+00
-5.48507512e-01 2.22814053e-01 3.99131864e-01 -9.96777236e-01
6.80388987e-01 1.33822626e-02 3.69179606e-01 -8.94566476e-01
6.77162856e-02 3.96663934e-01 6.53909892e-02 -5.03850102e-01
4.10743475e-01 -2.18205407e-01 -8.61658454e-02 6.34485245e-01
-1.90068826e-01 6.56814948e-02 -6.37259558e-02 5.57965972e-02
1.01153803e+00 6.30256593e-01 2.09907755e-01 -6.48662984e-01
1.15482055e-01 3.25218253e-02 6.75691247e-01 1.08230269e+00
-6.38155937e-01 -1.65168107e-01 1.03356278e+00 -2.41528884e-01
-1.13416350e+00 -8.54253292e-01 2.31268272e-01 1.24756932e+00
3.37194473e-01 3.86149995e-02 -5.86159706e-01 -9.19816911e-01
5.47819197e-01 7.25148380e-01 -9.51103389e-01 -3.99682134e-01
-5.03651738e-01 -5.34525037e-01 1.93586633e-01 6.42308891e-01
2.83071578e-01 -8.96360874e-01 -1.16352046e+00 2.34499425e-01
2.12840408e-01 -8.45120132e-01 -7.67223775e-01 3.32495242e-01
-9.90321219e-01 -9.56745386e-01 -7.49275684e-01 -7.73025081e-02
5.84562302e-01 1.81450561e-01 8.45323622e-01 -2.25278720e-01
-5.34636341e-02 6.21936381e-01 -6.11270294e-02 -3.80412787e-01
-2.18683392e-01 -6.26170337e-02 4.25043553e-01 -2.43869096e-01
-2.14298189e-01 -5.99818110e-01 -8.63748074e-01 4.07378882e-01
-5.73462009e-01 -3.37047167e-02 4.88849610e-01 1.19394386e+00
4.45663124e-01 -2.45631054e-01 9.89354968e-01 -6.10760331e-01
8.42630565e-01 -6.73080504e-01 -1.25292218e+00 1.05554499e-01
-1.05328953e+00 6.26661777e-01 8.44380856e-01 -7.77152777e-01
-1.07331371e+00 -1.13220803e-01 4.24699306e-01 -6.64486349e-01
5.52859187e-01 1.85846254e-01 2.05297649e-01 -7.79022183e-03
5.31407177e-01 1.13469802e-01 3.92911077e-01 -3.92072767e-01
3.94091249e-01 2.96933413e-01 3.67418766e-01 -1.16571355e+00
6.44779384e-01 4.07646716e-01 1.20681569e-01 -6.79764226e-02
-1.16164327e+00 5.12266196e-02 -5.47009632e-02 -2.36422330e-01
4.76573855e-01 -1.10558999e+00 -1.34703350e+00 5.06914258e-02
-6.67785227e-01 -8.73609483e-01 -7.60800958e-01 5.36424577e-01
-1.22705579e+00 4.23590764e-02 -3.03854108e-01 -1.24042106e+00
-1.06095873e-01 -1.04641032e+00 7.53826439e-01 3.24880809e-01
2.65488654e-01 -7.63583660e-01 1.51476115e-01 1.90881252e-01
5.05986571e-01 2.37386823e-01 7.16480374e-01 -3.41608942e-01
-6.25782132e-01 3.84869218e-01 -8.49674568e-02 9.18827727e-02
-3.26051444e-01 -6.22043431e-01 -6.41491473e-01 -6.95623279e-01
2.98638731e-01 -8.43725622e-01 8.37337375e-01 2.78635740e-01
1.22974098e+00 -6.47196531e-01 -7.31241629e-02 4.98747885e-01
1.28313398e+00 1.12707108e-01 1.54690281e-01 6.64265692e-01
3.29634845e-01 4.89586651e-01 9.92254555e-01 9.87939417e-01
3.37873548e-01 4.43751216e-01 7.31332242e-01 3.95790845e-01
4.19392914e-01 -7.24020302e-01 6.56497478e-01 3.61167699e-01
2.45753676e-02 1.01838246e-01 -4.32242215e-01 4.91829723e-01
-2.30567551e+00 -7.26063251e-01 5.16750574e-01 2.53185105e+00
1.10986173e+00 1.11603357e-01 3.56639653e-01 -3.75512451e-01
3.91300857e-01 -8.19838140e-03 -1.48085654e+00 -2.52793550e-01
1.87536880e-01 -1.35382816e-01 9.39707041e-01 4.15013373e-01
-6.43607616e-01 7.33390927e-01 6.57511520e+00 9.32659030e-01
-9.67322707e-01 3.87698784e-02 6.79894745e-01 -4.18727607e-01
-2.34473795e-01 2.19728146e-02 -6.86257243e-01 4.96441692e-01
9.63652432e-01 -5.68597257e-01 1.13953805e+00 1.17690921e+00
4.19648588e-01 -1.98457852e-01 -1.30965650e+00 6.65544033e-01
-5.75371742e-01 -1.18668044e+00 -5.10823965e-01 3.49425673e-01
8.54177833e-01 1.50990894e-03 3.82169932e-01 7.45208085e-01
6.27952635e-01 -8.40433776e-01 1.02380574e+00 4.47623432e-01
7.61173785e-01 -6.87396049e-01 2.47827709e-01 7.04561114e-01
-7.24583805e-01 -7.18622446e-01 -2.99103260e-01 -2.55320042e-01
-1.79378211e-01 9.58012119e-02 -5.55971384e-01 3.02613676e-01
3.85288805e-01 6.56824768e-01 -2.83735543e-02 5.71090281e-01
-3.73652071e-01 3.83229762e-01 -2.57552266e-01 -8.56686309e-02
4.76833194e-01 -5.20814836e-01 6.94554389e-01 3.89265925e-01
1.76565364e-01 -6.55057607e-03 2.69663364e-01 1.22556150e+00
-8.62486884e-02 -2.64918625e-01 -3.72874528e-01 -3.70561451e-01
4.42557842e-01 9.76441741e-01 -2.12464005e-01 -9.48249102e-02
1.71561331e-01 7.44512618e-01 6.81871593e-01 5.52304626e-01
-9.40125585e-01 3.33663411e-02 7.91200936e-01 -1.63443699e-01
4.51688379e-01 -3.21801752e-01 -3.19482237e-01 -1.12883925e+00
2.07974896e-01 -8.92189980e-01 2.76146501e-01 -4.06576484e-01
-1.20755529e+00 1.70372950e-03 -1.43152744e-01 -1.22851455e+00
-3.68949771e-01 -4.77820426e-01 -2.78320014e-01 6.80694461e-01
-1.70715272e+00 -3.92801017e-01 1.63388833e-01 3.48727822e-01
4.89744723e-01 -1.10789433e-01 2.26143152e-01 -3.73532847e-02
-7.14263022e-01 6.41686618e-01 5.28531253e-01 -5.93070507e-01
4.97161031e-01 -1.31073511e+00 -4.54631299e-02 4.01247144e-01
-4.17632520e-01 4.21043009e-01 8.74934375e-01 -5.62406778e-01
-2.04728031e+00 -8.77425253e-01 -1.27904475e-01 -4.82518137e-01
1.17867255e+00 -2.59014755e-01 -5.80777228e-01 6.38501525e-01
-2.22282246e-01 8.57077241e-02 3.64426710e-02 -3.24331224e-02
-2.33479604e-01 -4.55989987e-02 -1.12781858e+00 7.88221300e-01
1.15109956e+00 -2.19797641e-01 -3.32074910e-01 7.71945268e-02
1.11431563e+00 -6.12189531e-01 -8.64163339e-01 3.24889779e-01
7.20129430e-01 -7.85093904e-01 7.44021952e-01 -1.09626126e+00
6.20302916e-01 3.30569397e-04 -1.64678041e-02 -1.53946030e+00
3.73088084e-02 -1.19844604e+00 -6.78078175e-01 5.00078797e-01
2.51134664e-01 -7.23214507e-01 7.60865331e-01 6.69153214e-01
1.73352182e-01 -1.11464393e+00 -1.09581876e+00 -1.23793566e+00
3.07071984e-01 -1.13861576e-01 3.19294006e-01 6.47191226e-01
1.00814790e-01 6.03721738e-02 -8.15962255e-01 -4.15789410e-02
8.06646407e-01 1.51267022e-01 7.47146785e-01 -7.63260961e-01
-7.39803672e-01 -3.83942187e-01 3.49203944e-01 -1.46842074e+00
2.36439779e-01 -4.72986847e-01 1.78984776e-01 -8.86865079e-01
1.26662567e-01 -6.98815763e-01 -4.26538318e-01 2.85409838e-01
-3.67263019e-01 -5.40274858e-01 5.35798550e-01 4.22351092e-01
-7.85397887e-01 1.20443594e+00 1.81685877e+00 2.00308397e-01
-4.55750287e-01 7.29493797e-02 -1.02159286e+00 4.46990341e-01
6.88898861e-01 -6.13251925e-01 -6.05904341e-01 -3.68486911e-01
3.94252926e-01 5.10623276e-01 3.57887059e-01 -4.98608738e-01
1.03377640e-01 -5.82187772e-01 3.18482816e-02 1.58680268e-02
7.28991106e-02 -7.73957372e-01 -3.13972026e-01 8.51080000e-01
-8.41555893e-01 -6.52920306e-02 9.50420573e-02 1.26077902e+00
4.59229678e-01 4.03894596e-02 7.65625417e-01 -4.64633480e-02
-3.23349774e-01 4.32313532e-01 -8.15963373e-02 5.27759790e-01
1.09540784e+00 3.02330911e-01 -4.53933150e-01 -5.03309965e-01
-5.95759988e-01 7.39349246e-01 2.93819904e-01 3.84972781e-01
2.71351099e-01 -1.16671205e+00 -4.64150429e-01 4.28856015e-02
-1.45154938e-01 -2.10073709e-01 2.01156721e-01 9.96162891e-01
2.05862910e-01 3.25909466e-01 -2.73958862e-01 -2.39349991e-01
-4.07122523e-01 6.99514508e-01 5.52953720e-01 -7.21650600e-01
-7.63335645e-01 5.83082139e-01 3.02410275e-01 -3.18329096e-01
4.75234061e-01 -3.18823457e-01 2.72677809e-01 -1.81362286e-01
3.38456005e-01 5.60412228e-01 -3.97324622e-01 2.67791688e-01
-5.32719120e-02 1.32806018e-01 -9.85052809e-02 -4.11859632e-01
1.35008430e+00 -2.43156612e-01 5.65979302e-01 3.96848112e-01
6.41121328e-01 -3.92167658e-01 -2.35977721e+00 -2.98789263e-01
5.99660538e-02 -6.46137178e-01 1.43251628e-01 -9.76805270e-01
-7.88546026e-01 6.87288165e-01 5.77022254e-01 9.39115286e-02
8.56358588e-01 -2.95843333e-01 7.27113128e-01 5.40188372e-01
6.81343496e-01 -1.51013780e+00 4.31975067e-01 4.12234366e-01
1.00692809e+00 -1.31996512e+00 -1.96150132e-03 2.20088452e-01
-1.09541214e+00 8.22135806e-01 7.49400854e-01 -4.84128237e-01
2.56881863e-01 1.40012592e-01 -3.86877775e-01 5.54903075e-02
-1.18907678e+00 -3.70944440e-02 -1.45637736e-01 3.83155495e-01
-2.96943963e-01 1.30802542e-01 -3.99426460e-01 8.96455526e-01
6.61602318e-02 1.07376993e-01 6.20465279e-01 9.78577197e-01
-4.95880574e-01 -8.54174912e-01 -1.98557302e-01 3.14707518e-01
-2.49390006e-01 2.25640208e-01 -4.14315313e-02 8.88095260e-01
-3.77879620e-01 5.98652422e-01 -4.82809581e-02 -2.38712020e-02
2.22010225e-01 -4.02021170e-01 5.49373031e-01 -1.65063754e-01
-4.91149992e-01 2.64931358e-02 -1.83365509e-01 -1.02634966e+00
7.62694003e-03 -4.98613656e-01 -1.16239059e+00 -8.82912278e-02
-2.00613469e-01 1.06097423e-01 6.72715127e-01 1.02959526e+00
6.59831107e-01 5.32043576e-01 1.03001642e+00 -8.13479960e-01
-1.94183671e+00 -5.89836895e-01 -6.78172588e-01 1.27501473e-01
7.73126841e-01 -9.95381832e-01 -6.66543305e-01 -4.53784049e-01] | [4.1310715675354, 2.3046841621398926] |
be76c717-5d4c-43b6-ba6b-f35094a7f77e | real-time-speech-frequency-bandwidth | 2010.10677 | null | https://arxiv.org/abs/2010.10677v2 | https://arxiv.org/pdf/2010.10677v2.pdf | Real-time Speech Frequency Bandwidth Extension | In this paper we propose a lightweight model for frequency bandwidth extension of speech signals, increasing the sampling frequency from 8kHz to 16kHz while restoring the high frequency content to a level almost indistinguishable from the 16kHz ground truth. The model architecture is based on SEANet (Sound EnhAncement Network), a wave-to-wave fully convolutional model, which uses a combination of feature losses and adversarial losses to reconstruct an enhanced version of the input speech. In addition, we propose a variant of SEANet that can be deployed on-device in streaming mode, achieving an architectural latency of 16ms. When profiled on a single core of a mobile CPU, processing one 16ms frame takes only 1.5ms. The low latency makes it viable for bi-directional voice communication systems. | [] | 2020-10-21 | null | null | null | null | ['bandwidth-extension', 'bandwidth-extension'] | ['audio', 'speech'] | [ 2.51139313e-01 4.09109980e-01 2.68119961e-01 -2.02489749e-01
-8.95675182e-01 -4.59743381e-01 2.73216695e-01 -4.00569260e-01
-5.69383442e-01 4.39859688e-01 5.67446411e-01 -5.64364910e-01
1.30842119e-01 -8.22126091e-01 -7.54197299e-01 -6.99593961e-01
-4.47803527e-01 -2.09793955e-01 5.01822650e-01 -3.30791444e-01
-4.16008472e-01 4.92857337e-01 -1.76288092e+00 6.71087325e-01
8.08662474e-02 1.03529978e+00 -4.69972426e-03 1.56861591e+00
1.74842313e-01 5.67324400e-01 -9.67794836e-01 -2.70182759e-01
2.75478572e-01 3.73546667e-02 -7.90361822e-01 -4.94258821e-01
4.42092329e-01 -9.64584649e-01 -7.53623426e-01 8.92843366e-01
1.03725076e+00 6.60856739e-02 -1.10836633e-01 -1.16338015e+00
3.65017831e-01 7.31434464e-01 1.88485250e-01 2.67752349e-01
3.74267995e-01 2.55836010e-01 7.17911601e-01 -7.59869039e-01
2.23225549e-01 1.16712487e+00 1.23676133e+00 7.33070076e-01
-1.07728803e+00 -7.32092440e-01 -5.14773905e-01 -2.29844503e-04
-1.31456792e+00 -1.23722827e+00 5.51489890e-01 2.04655841e-01
1.40221655e+00 6.17099941e-01 5.03518999e-01 1.22027469e+00
1.43402502e-01 3.67961377e-01 6.31390333e-01 -3.10358346e-01
2.87930697e-01 -2.29726806e-01 -6.46860957e-01 2.49465898e-01
-2.72242367e-01 4.10551876e-01 -9.74009454e-01 -4.07231778e-01
4.92176801e-01 -6.47523165e-01 -7.50501275e-01 4.67050463e-01
-9.99164522e-01 3.84744167e-01 1.87363222e-01 7.88447633e-02
-3.63563806e-01 7.92900801e-01 6.98635578e-01 6.33050203e-01
4.85337108e-01 1.30690306e-01 -6.50569022e-01 -5.64756691e-01
-1.28578234e+00 2.54422098e-01 9.23786640e-01 6.19455457e-01
4.88878161e-01 8.34443867e-01 4.90551889e-01 3.26768309e-01
5.26303053e-01 7.14144945e-01 5.91233075e-01 -1.20985961e+00
2.70822823e-01 -5.80148458e-01 -5.20777851e-02 -7.41990089e-01
-5.47145069e-01 -4.98983234e-01 -8.52157891e-01 3.76360565e-01
2.42580175e-01 -5.64839900e-01 -5.13294041e-01 1.44626296e+00
3.95087183e-01 8.67608786e-01 2.85824060e-01 1.03999603e+00
8.50760221e-01 9.22914922e-01 -2.64384300e-01 6.62189201e-02
1.17824304e+00 -9.58923459e-01 -7.55471885e-01 -5.41697815e-02
3.28701913e-01 -9.49418724e-01 1.21544719e+00 6.28730714e-01
-1.24203670e+00 -5.49098551e-01 -1.28142548e+00 -4.20621373e-02
1.05163688e-02 -5.18795252e-01 3.94774765e-01 1.13979781e+00
-1.50740290e+00 8.60211849e-01 -9.50519562e-01 2.44097888e-01
1.23329042e-02 5.69285214e-01 -4.02104080e-01 3.43003184e-01
-1.47638702e+00 2.08470657e-01 -3.28463838e-02 5.08806556e-02
-9.55439270e-01 -1.33900118e+00 -6.60920799e-01 2.89968282e-01
-1.87829643e-01 -6.08116210e-01 1.73464465e+00 -9.27133918e-01
-2.11301684e+00 1.28510848e-01 2.21949816e-01 -8.19313467e-01
4.22697246e-01 -2.57021695e-01 -1.04815209e+00 5.86249173e-01
-5.19653976e-01 3.39912862e-01 1.29980564e+00 -8.26307952e-01
-4.74171489e-01 1.41315937e-01 -1.18509913e-02 -1.40946120e-01
-2.83874661e-01 1.19218096e-01 9.80807561e-03 -8.19584608e-01
-1.45241931e-01 -4.83217746e-01 -8.62480178e-02 1.91937268e-01
-9.80546698e-02 5.18942773e-01 1.00165534e+00 -7.44626462e-01
8.88916850e-01 -2.32465434e+00 -2.81505287e-01 1.19370013e-01
-4.28031124e-02 5.98104596e-01 -1.77012578e-01 5.03037453e-01
6.39220104e-02 1.35061830e-01 -2.78426200e-01 -5.31084478e-01
-2.01649778e-02 6.61855116e-02 -6.50185049e-01 4.65987861e-01
5.66224344e-02 3.82545173e-01 -7.94671714e-01 1.87828407e-01
6.31481633e-02 1.01231408e+00 -9.52979624e-01 2.30911836e-01
5.76805547e-02 2.07179621e-01 7.61632919e-02 3.09719145e-01
1.08088517e+00 5.75719416e-01 1.08523279e-01 -4.08601284e-01
-1.44875884e-01 9.65962708e-01 -1.27548194e+00 1.88518131e+00
-1.06195295e+00 8.62125874e-01 7.66004562e-01 -3.71116310e-01
7.07145691e-01 8.70027006e-01 1.80425614e-01 -3.38349491e-01
1.24392301e-01 4.42782611e-01 -7.53586441e-02 -4.79727685e-01
6.22709751e-01 -3.19869667e-01 1.89407617e-01 4.69409585e-01
2.01938421e-01 -5.83792150e-01 -6.59449816e-01 -3.31383524e-03
1.37726367e+00 -1.26221642e-01 -2.56091624e-01 -2.23923787e-01
3.67652804e-01 -5.68476796e-01 3.03364694e-01 6.18176818e-01
-1.87024683e-01 6.46445096e-01 4.22175564e-02 -4.40700978e-01
-1.18196607e+00 -1.20762861e+00 -1.99569881e-01 8.67453933e-01
-3.34748477e-01 -5.92536569e-01 -8.87608588e-01 -2.72548586e-01
-1.85994327e-01 5.37430406e-01 8.76027197e-02 -3.53687882e-01
-6.05499148e-01 -3.22270840e-01 1.57109714e+00 2.97234356e-01
7.60695636e-01 -9.07262743e-01 -8.68470192e-01 4.47617978e-01
-9.50942785e-02 -1.00218117e+00 -4.94334996e-01 1.73784897e-01
-5.97895205e-01 -5.54026365e-01 -3.27418357e-01 -6.58826053e-01
-2.27085158e-01 9.32993516e-02 1.18319201e+00 2.92391419e-01
-8.71457253e-03 2.18130723e-01 -4.41323817e-01 -3.37953389e-01
-6.96949184e-01 4.46730517e-02 4.10283893e-01 7.29966760e-02
-1.45184964e-01 -1.05695045e+00 -6.37248039e-01 -1.84860229e-01
-1.14746904e+00 -2.16568813e-01 5.27721494e-02 7.25526631e-01
1.73463404e-01 2.76956409e-01 8.24263871e-01 -3.04126143e-01
3.27142775e-01 -4.47409868e-01 -2.49654189e-01 -5.86325049e-01
-1.74916863e-01 -2.96598792e-01 9.54892933e-01 -4.26393390e-01
-1.08659160e+00 -6.72887266e-03 -1.18905795e+00 -2.65661836e-01
-1.32196888e-01 1.11015089e-01 -2.11861685e-01 -2.31730461e-01
5.67513645e-01 1.50817633e-01 2.38409460e-01 -6.71715796e-01
3.41370732e-01 1.32890403e+00 1.04964590e+00 -1.14668705e-01
6.96420491e-01 6.30143940e-01 -2.07077608e-01 -1.23587060e+00
-1.64851040e-01 -2.16721445e-01 -1.01442628e-01 -1.03962950e-01
3.85389268e-01 -1.12811387e+00 -8.81000638e-01 8.01489115e-01
-1.02482474e+00 -7.77409136e-01 -6.07211590e-01 7.40961194e-01
-6.61301434e-01 3.74920636e-01 -6.87484741e-01 -6.72897458e-01
-7.68764973e-01 -7.85305023e-01 1.16289389e+00 1.94351524e-02
1.41915884e-02 -7.26840913e-01 -1.80228632e-02 -8.72460231e-02
9.80578899e-01 -2.02910706e-01 2.83216655e-01 -3.48159969e-01
-1.92094386e-01 -7.90603682e-02 3.05025637e-01 5.81279397e-01
-1.18442163e-01 4.13227500e-03 -1.50571442e+00 -4.76501137e-01
2.55758047e-01 -3.06936383e-01 7.79665947e-01 2.42462680e-01
1.05833685e+00 -6.41172290e-01 2.54469126e-01 1.17846096e+00
1.06195927e+00 -1.94821358e-01 9.45741117e-01 1.24460198e-01
1.50240615e-01 2.78632849e-01 1.04326114e-01 6.49496675e-01
1.27487466e-01 6.28089905e-01 7.06308305e-01 -1.54721990e-01
-7.60826945e-01 -2.17229202e-01 5.93652010e-01 1.08316600e+00
2.76506394e-01 -3.81941706e-01 -6.05406344e-01 5.75569212e-01
-9.47911501e-01 -1.05298424e+00 -1.56485483e-01 2.19229317e+00
1.07667470e+00 1.65676981e-01 3.21722217e-02 5.59794307e-01
3.64645094e-01 4.63007420e-01 -9.57344249e-02 -9.09664989e-01
1.03566058e-01 7.35379755e-01 4.93903697e-01 1.00120568e+00
-9.22695339e-01 7.18256891e-01 6.81351566e+00 7.40433812e-01
-1.44737101e+00 1.84322327e-01 1.05351888e-01 -3.84499758e-01
-5.55415094e-01 -3.00338626e-01 -3.88259321e-01 3.99534255e-01
1.95933330e+00 -8.39022249e-02 7.64687538e-01 6.66497350e-01
3.51725847e-01 3.81014615e-01 -5.34221053e-01 7.00305760e-01
-2.03249589e-01 -1.34071767e+00 -1.71676263e-01 -3.80453058e-02
1.80134103e-01 3.99189144e-01 2.89468281e-02 1.81761682e-01
-5.83065078e-02 -9.23863232e-01 8.97838295e-01 2.62649298e-01
1.49518895e+00 -1.11823487e+00 8.35521102e-01 5.08513808e-01
-1.19353914e+00 2.86161929e-01 -3.78665507e-01 -2.99854606e-01
4.52349484e-01 5.54048121e-01 -1.11918664e+00 3.13066989e-01
1.18138921e+00 -1.71655715e-02 1.30357847e-01 7.92606771e-01
2.02184208e-04 1.18925905e+00 -7.51635551e-01 3.84663314e-01
6.56546801e-02 7.34207988e-01 8.21370423e-01 1.45788550e+00
6.02696240e-01 4.15029190e-02 -3.34094375e-01 1.28895581e-01
-2.94136584e-01 -3.07287544e-01 -5.06517053e-01 4.27001864e-01
6.83689117e-01 1.04112637e+00 -1.53796254e-02 -1.04262322e-01
-4.16505337e-01 9.76175010e-01 -4.22605604e-01 1.48393616e-01
-7.36347258e-01 -1.01834202e+00 1.17605424e+00 7.63297838e-04
3.97768557e-01 -2.26684779e-01 1.68865561e-01 -8.55652452e-01
-5.08036688e-02 -9.94156539e-01 -2.73847729e-01 -7.78640091e-01
-7.13677824e-01 9.34638858e-01 -4.96940196e-01 -1.05445373e+00
-4.28341955e-01 -2.76124358e-01 -5.14382064e-01 9.17196035e-01
-1.65228593e+00 -7.71051764e-01 -2.90411592e-01 6.52491808e-01
4.76022214e-01 -1.36181191e-01 1.37494516e+00 5.03792822e-01
8.95627365e-02 8.95262122e-01 1.15717232e-01 -1.13104597e-01
4.73777920e-01 -1.08986676e+00 1.27417064e+00 7.73098767e-01
2.51383968e-02 1.88983083e-01 1.00125766e+00 -2.92805493e-01
-1.46130466e+00 -1.28121448e+00 1.00472307e+00 2.93387175e-01
5.43581188e-01 -4.70813692e-01 -1.04430461e+00 2.90353864e-01
4.39730436e-01 4.01426435e-01 7.83355236e-01 -5.78604519e-01
-4.67222452e-01 -2.59544224e-01 -1.48212612e+00 4.37087059e-01
7.50253439e-01 -7.58749425e-01 -2.44720221e-01 1.46757662e-01
1.26741517e+00 -6.47519350e-01 -1.05875134e+00 9.72304195e-02
6.29742682e-01 -1.06346238e+00 1.07378209e+00 -4.02805060e-01
4.04017493e-02 -3.59324425e-01 -4.66904610e-01 -1.42577791e+00
3.43253836e-02 -1.45588052e+00 -2.86770701e-01 1.05789351e+00
1.88374743e-01 -7.24910736e-01 7.82680750e-01 9.77162179e-03
-5.33252239e-01 -4.26253706e-01 -1.66643667e+00 -6.82913125e-01
-4.57555354e-02 -1.03994834e+00 1.11604333e+00 3.14349562e-01
-1.05307236e-01 -9.59777385e-02 -5.97875893e-01 7.89018750e-01
4.56393540e-01 -6.17302477e-01 5.59195936e-01 -7.78384805e-01
-6.32820666e-01 4.80304286e-03 -4.73936558e-01 -1.07322681e+00
5.21622300e-02 -6.36671960e-01 2.56380320e-01 -8.93043399e-01
-7.04570770e-01 -2.34863430e-01 -7.48638669e-03 3.14542890e-01
5.58076859e-01 7.21113205e-01 6.81080520e-02 -1.92425549e-01
1.47744432e-01 5.76032639e-01 7.64083564e-01 -7.69074038e-02
-2.62950677e-02 2.82354236e-01 -2.04918250e-01 9.65431750e-01
8.31704199e-01 -5.79681575e-01 -3.52987826e-01 -6.14504755e-01
1.57963380e-01 2.70068616e-01 5.49039543e-01 -1.31057835e+00
1.64288342e-01 3.51693869e-01 -8.27368125e-02 -2.16422096e-01
6.92881286e-01 -9.71969962e-01 1.66545838e-01 5.92841744e-01
-2.51276821e-01 -1.99895382e-01 6.20868623e-01 4.67578053e-01
-2.23726183e-01 -2.53930956e-01 6.82113469e-01 2.67158031e-01
-3.34605753e-01 4.44741957e-02 -9.00654554e-01 -6.25657141e-02
3.88373196e-01 2.33497456e-01 -3.27296644e-01 -7.77174175e-01
-5.79897344e-01 -5.40618360e-01 1.56609669e-01 1.58176467e-01
6.26428008e-01 -1.13566267e+00 -8.35993528e-01 5.74158311e-01
-5.97378790e-01 -1.27845630e-01 6.56473279e-01 3.01759869e-01
-1.12406504e+00 1.63119659e-01 1.35016829e-01 -3.55831712e-01
-1.16381717e+00 1.14161484e-01 6.98728442e-01 3.26629788e-01
-1.02337682e+00 1.06464779e+00 -4.11458343e-01 -2.89722025e-01
3.74458253e-01 -3.16478133e-01 2.21331328e-01 -4.95366126e-01
1.13901567e+00 4.95451599e-01 7.06629634e-01 -4.90946800e-01
-5.50184429e-01 1.04988374e-01 5.25518000e-01 -5.79637945e-01
1.28285921e+00 -1.76814750e-01 1.20351516e-01 7.67486989e-02
1.43355906e+00 4.86059308e-01 -1.49388385e+00 -7.81030506e-02
-4.68671262e-01 -4.66783881e-01 7.36384630e-01 -7.03414917e-01
-1.10343254e+00 8.97738695e-01 7.01650560e-01 3.22153240e-01
1.55062521e+00 -4.02255923e-01 1.41368306e+00 2.98090041e-01
3.55470330e-01 -7.20009625e-01 -1.29796371e-01 4.73829538e-01
9.52879429e-01 -4.56358045e-01 -2.78589517e-01 -2.59516627e-01
-1.88941553e-01 1.31381631e+00 6.57869354e-02 -2.62274835e-02
9.38478649e-01 1.26589596e+00 3.01422745e-01 3.36865634e-01
-9.70085263e-01 1.91061437e-01 -9.89194512e-02 9.66485381e-01
1.95012599e-01 4.11395133e-02 2.19743833e-01 7.33523667e-01
-6.24101341e-01 -1.37797058e-01 8.40653241e-01 6.33207142e-01
-4.85730380e-01 -8.84637237e-01 -4.52290863e-01 -2.79136691e-02
-8.29822004e-01 -4.53587174e-01 2.09730923e-01 3.42088580e-01
-2.34697629e-02 1.33211899e+00 4.62177873e-01 -7.13668823e-01
3.98128867e-01 3.19101512e-02 1.44841895e-02 -3.11376810e-01
-1.12497604e+00 2.50046700e-01 4.53054935e-01 -7.42424130e-01
-6.82337061e-02 -4.27138418e-01 -1.44180644e+00 -8.14248443e-01
-1.90784201e-01 -6.74931193e-03 1.01017690e+00 4.34938341e-01
5.63668072e-01 8.34786355e-01 8.69402051e-01 -1.08952999e+00
-8.65472198e-01 -8.79686654e-01 -7.13420689e-01 -3.20457846e-01
1.00405538e+00 6.08254932e-02 -7.87253976e-01 4.08966132e-02] | [15.2560396194458, 5.879628658294678] |
768bd3ad-a358-4d26-b34a-a921d79cd35b | hardware-aware-graph-neural-network-automated | 2303.10875 | null | https://arxiv.org/abs/2303.10875v2 | https://arxiv.org/pdf/2303.10875v2.pdf | Hardware-Aware Graph Neural Network Automated Design for Edge Computing Platforms | Graph neural networks (GNNs) have emerged as a popular strategy for handling non-Euclidean data due to their state-of-the-art performance. However, most of the current GNN model designs mainly focus on task accuracy, lacking in considering hardware resources limitation and real-time requirements of edge application scenarios. Comprehensive profiling of typical GNN models indicates that their execution characteristics are significantly affected across different computing platforms, which demands hardware awareness for efficient GNN designs. In this work, HGNAS is proposed as the first Hardware-aware Graph Neural Architecture Search framework targeting resource constraint edge devices. By decoupling the GNN paradigm, HGNAS constructs a fine-grained design space and leverages an efficient multi-stage search strategy to explore optimal architectures within a few GPU hours. Moreover, HGNAS achieves hardware awareness during the GNN architecture design by leveraging a hardware performance predictor, which could balance the GNN model accuracy and efficiency corresponding to the characteristics of targeted devices. Experimental results show that HGNAS can achieve about $10.6\times$ speedup and $88.2\%$ peak memory reduction with a negligible accuracy loss compared to DGCNN on various edge devices, including Nvidia RTX3080, Jetson TX2, Intel i7-8700K and Raspberry Pi 3B+. | ['Chunming Hu', 'Weisheng Zhao', 'Tong Qiao', 'Yumeng Shi', 'Yingjie Qi', 'Jianlei Yang', 'Ao Zhou'] | 2023-03-20 | null | null | null | null | ['architecture-search'] | ['methodology'] | [-2.80751109e-01 -4.01997745e-01 -4.27803308e-01 7.94759840e-02
6.29421845e-02 -2.39782885e-01 -3.90498862e-02 -4.90774028e-02
-3.67136449e-01 2.58728147e-01 -3.92378956e-01 -7.78815448e-01
-3.34995866e-01 -1.06600845e+00 -5.28512061e-01 -3.17959428e-01
-7.72322640e-02 3.01157773e-01 1.42804474e-01 -1.81619644e-01
1.86465606e-01 4.85291988e-01 -1.60974944e+00 -2.10886583e-01
6.46744728e-01 1.47809267e+00 2.23707575e-02 7.88134038e-01
-1.21434055e-01 7.44843423e-01 -4.94475126e-01 -2.44789317e-01
5.80241859e-01 1.57891947e-03 -1.35082290e-01 -6.21772528e-01
3.51916313e-01 -2.91348100e-01 -7.09670842e-01 1.18583226e+00
7.14039981e-01 -9.39150807e-03 2.44476512e-01 -1.42071688e+00
-4.30999249e-01 6.62539661e-01 -6.60515368e-01 4.03033018e-01
-2.16724515e-01 2.15840250e-01 9.39661264e-01 -4.51352268e-01
3.21059793e-01 7.05341041e-01 9.10875201e-01 4.01365876e-01
-7.36059546e-01 -9.07147884e-01 3.18149440e-02 4.60381627e-01
-1.64473033e+00 -1.81342870e-01 7.92687237e-01 6.34810305e-04
1.71086550e+00 2.36953333e-01 9.76777613e-01 8.79004657e-01
4.15243924e-01 3.10311288e-01 5.72972298e-01 -3.44945163e-01
4.74748880e-01 -5.39498389e-01 3.85090709e-01 8.13892603e-01
9.49413598e-01 1.49991840e-01 -7.06880808e-01 -1.30034596e-01
9.37531292e-01 -4.35165204e-02 -2.68721521e-01 -1.75045341e-01
-5.90321600e-01 4.19928849e-01 7.73056686e-01 2.91004940e-03
-4.34286475e-01 7.58625448e-01 8.74212205e-01 -1.24905355e-01
-1.37593433e-01 2.16691062e-01 -5.20046294e-01 -4.23554599e-01
-6.19688630e-01 -6.86996058e-02 7.84578443e-01 1.28575194e+00
5.79629064e-01 6.83273911e-01 2.19258383e-01 4.19330865e-01
2.42930412e-01 3.85558337e-01 5.95808029e-01 -4.05743003e-01
4.33027714e-01 1.15170026e+00 -5.53848684e-01 -1.38177490e+00
-9.10340846e-01 -7.92307615e-01 -1.13741362e+00 1.61743015e-01
6.34201467e-02 -1.72906891e-01 -6.76129639e-01 1.43029261e+00
3.16433012e-01 3.64412338e-01 -7.79707506e-02 9.16644394e-01
9.10969019e-01 5.55603683e-01 8.09317008e-02 1.78194433e-01
1.42625022e+00 -1.16354370e+00 -3.12787652e-01 -5.08174300e-01
9.27848279e-01 -3.51780087e-01 8.89580011e-01 4.75283384e-01
-9.71026778e-01 -7.40932703e-01 -1.50163043e+00 -1.58995479e-01
-2.74764329e-01 2.95350015e-01 1.17342675e+00 1.13807082e+00
-1.18784809e+00 4.85742211e-01 -1.17952967e+00 -2.62515634e-01
1.81975886e-01 8.15827489e-01 1.97280794e-01 1.96422309e-01
-7.69682288e-01 2.82989204e-01 7.64271617e-01 3.44539464e-01
-3.89700502e-01 -8.46422255e-01 -6.05027318e-01 3.58150721e-01
3.62782538e-01 -7.68198133e-01 8.21823835e-01 -5.43082893e-01
-1.58646083e+00 3.23206514e-01 4.86127436e-01 -6.29440010e-01
2.45517138e-02 5.68684153e-02 -9.05705452e-01 -1.95902571e-01
-4.27947521e-01 3.45665127e-01 4.70771581e-01 -4.36487973e-01
-6.30189657e-01 -4.44780022e-01 -1.30722731e-01 2.25552872e-01
-7.70282149e-01 -2.66274333e-01 -8.10558438e-01 -3.87154907e-01
1.01836607e-01 -1.13131678e+00 -3.56796265e-01 -1.90575495e-01
-4.34365541e-01 5.06665967e-02 8.77179444e-01 -1.42597809e-01
1.67357898e+00 -1.80167770e+00 -1.98199645e-01 3.82133216e-01
4.47346061e-01 5.71363270e-01 1.06256753e-01 1.71866789e-02
2.26011083e-01 -2.18445599e-01 5.61335206e-01 1.13949150e-01
7.43336603e-02 -3.37122986e-03 2.21143328e-02 3.90998304e-01
-5.12362003e-01 1.02087533e+00 -4.61996734e-01 -7.07853064e-02
2.12279141e-01 5.32649100e-01 -6.70952916e-01 -3.12136233e-01
-4.07020524e-02 -2.28801340e-01 -5.02336800e-01 7.52614737e-01
7.19310999e-01 -6.39744520e-01 5.89367807e-01 -5.54842651e-01
-5.16362898e-02 -2.64831912e-02 -1.21637416e+00 1.56568038e+00
-5.68481326e-01 7.81395972e-01 1.34046897e-01 -7.00512946e-01
1.13618159e+00 -2.39594385e-01 1.57516688e-01 -1.15900695e+00
7.96068609e-01 3.56347591e-01 1.70906261e-01 -6.50221540e-04
8.69942844e-01 5.59159160e-01 -1.02027528e-01 2.99841195e-01
-3.07655126e-01 5.64853609e-01 7.78338909e-02 -1.18298113e-01
1.39092159e+00 -1.91633061e-01 2.60118574e-01 -6.06749833e-01
2.03475192e-01 2.42126584e-01 6.03130937e-01 8.09817553e-01
-1.58166260e-01 7.88997263e-02 1.32750139e-01 -8.66897225e-01
-9.20531929e-01 -6.00105047e-01 6.60167485e-02 1.07544649e+00
4.76938695e-01 -7.54813135e-01 -9.57669973e-01 -2.86226422e-01
-2.93801814e-01 5.56742907e-01 -3.18351269e-01 -3.19604099e-01
-7.73132980e-01 -1.04048169e+00 7.99837589e-01 8.85988057e-01
8.49479020e-01 -7.82668412e-01 -1.42481267e+00 3.73216420e-01
7.22275496e-01 -1.26817799e+00 -2.99887210e-01 2.31611595e-01
-1.09018183e+00 -9.36798275e-01 4.26794067e-02 -8.11238170e-01
5.33550143e-01 3.66835088e-01 1.32193804e+00 4.75151300e-01
-6.23507559e-01 1.35441154e-01 -1.17469780e-01 -1.31228536e-01
7.28245303e-02 4.71174896e-01 2.73019642e-01 -4.30437267e-01
5.15187979e-01 -7.30535507e-01 -9.47586298e-01 1.76791504e-01
-4.61889088e-01 3.40478480e-01 5.66637516e-01 6.56937063e-01
7.33353794e-01 4.49639142e-01 6.32891878e-02 -6.58757269e-01
6.03015423e-01 -2.27618143e-01 -1.20852637e+00 1.20368421e-01
-1.22886872e+00 -5.17339166e-03 1.09411395e+00 -4.74889040e-01
-5.76143384e-01 -6.23968691e-02 5.70913963e-02 -6.47252619e-01
2.91913629e-01 5.89046955e-01 -1.84680030e-01 -4.69881058e-01
7.92561293e-01 6.32081106e-02 -2.67348528e-01 -1.22704178e-01
-1.50943547e-01 3.73135120e-01 6.70428753e-01 -6.29739881e-01
3.05947244e-01 2.98777640e-01 5.11536539e-01 -7.95675159e-01
-1.33002996e-01 -8.91853422e-02 6.17428534e-02 -2.51698613e-01
5.64124346e-01 -9.24385011e-01 -1.49701881e+00 5.71193099e-01
-7.39867389e-01 -5.25632620e-01 2.94134200e-01 4.89564687e-01
-9.41191465e-02 1.15167156e-01 -7.74576426e-01 -5.50406635e-01
-1.15572166e+00 -1.40603995e+00 6.77294374e-01 8.36807072e-01
-1.50091341e-02 -9.15238023e-01 -4.02526438e-01 3.73302549e-02
8.51715863e-01 8.17223713e-02 9.52389061e-01 -4.07957375e-01
-8.88348520e-01 -2.67863721e-01 -6.61574364e-01 -2.26453200e-01
-2.99063742e-01 -5.40309660e-02 -5.09462237e-01 -5.05876839e-01
-1.07380591e-01 4.32166718e-02 3.90313238e-01 5.84600687e-01
1.29904413e+00 -1.21912368e-01 -5.00165403e-01 1.30549645e+00
2.00805211e+00 4.31327909e-01 5.37765980e-01 5.44668078e-01
1.25630641e+00 -2.49256283e-01 1.71418056e-01 5.08008242e-01
5.01233339e-01 6.59838378e-01 8.78216982e-01 1.35176927e-01
-1.12347463e-02 -1.50702983e-01 1.40346617e-01 1.05219758e+00
4.30643521e-02 -5.57824135e-01 -1.21445870e+00 1.27444133e-01
-1.79098594e+00 -2.45345950e-01 -4.02991652e-01 2.08055997e+00
3.27724740e-02 5.91342270e-01 -1.22568987e-01 -1.18041597e-02
6.42062604e-01 3.63023430e-02 -1.11686134e+00 -6.72872663e-01
4.36206944e-02 3.28259408e-01 1.13632536e+00 -1.04353175e-01
-6.43834472e-01 8.33087027e-01 5.31940031e+00 1.06565475e+00
-1.46279716e+00 -1.86232269e-01 6.16669297e-01 -2.41991237e-01
2.89935451e-02 -1.83760549e-03 -1.14000142e+00 4.47364479e-01
1.28214753e+00 -3.79065871e-01 5.67319989e-01 1.39478457e+00
-2.38165736e-01 1.95557579e-01 -8.47831130e-01 1.53325558e+00
-7.84754828e-02 -1.64839125e+00 -1.34285629e-01 4.05236363e-01
5.74505985e-01 3.32675725e-01 -4.01244909e-02 4.76764888e-01
2.76590705e-01 -1.01455808e+00 6.57426119e-01 -9.06676650e-02
8.48742545e-01 -1.10218394e+00 5.57461083e-01 1.63463593e-01
-1.57956123e+00 -2.49822766e-01 -4.56200093e-01 -3.42210889e-01
1.10186756e-01 3.44655663e-01 -6.75107598e-01 4.19778079e-01
1.01786435e+00 1.70468718e-01 -5.40910006e-01 7.85650074e-01
1.67123616e-01 4.63003844e-01 -5.57655334e-01 -5.92487514e-01
1.96221277e-01 -2.37672850e-01 1.64754346e-01 9.69946980e-01
6.76876783e-01 2.52146512e-01 9.98367090e-03 6.47754192e-01
-2.67353058e-01 1.40539810e-01 -2.65655041e-01 1.43992499e-01
7.31775165e-01 1.47235513e+00 -1.15631557e+00 -4.13052589e-02
-4.88611370e-01 7.47965574e-01 3.93759817e-01 -1.50060263e-02
-1.26318526e+00 -5.52340686e-01 8.72421145e-01 3.26958322e-03
2.68813282e-01 -4.17613029e-01 -6.50134146e-01 -7.17883050e-01
-1.27627011e-02 -7.05056369e-01 4.59710687e-01 -6.64527893e-01
-7.81319320e-01 1.01231027e+00 -4.60437953e-01 -1.03486788e+00
2.32065227e-02 -1.09439147e+00 -5.93561649e-01 5.35844326e-01
-1.10612822e+00 -9.95925367e-01 -8.35216165e-01 3.56564105e-01
2.28207707e-01 -3.53150845e-01 5.50750196e-01 4.13824022e-01
-1.12069011e+00 1.12702954e+00 -8.83040763e-03 3.20207253e-02
-1.37936920e-02 -7.78520048e-01 9.91412401e-01 9.32883441e-01
-9.43239257e-02 6.36213422e-01 4.98630345e-01 -7.58178651e-01
-2.38165212e+00 -1.05316436e+00 -7.64347473e-03 2.87241518e-01
6.95393324e-01 -2.66821206e-01 -7.84832895e-01 4.78159070e-01
-1.36658251e-01 4.88562047e-01 5.40370762e-01 9.55041721e-02
-2.39415333e-01 -2.47653499e-01 -8.07570398e-01 1.09228921e+00
1.51156223e+00 -2.32725069e-01 6.55188203e-01 1.03135534e-01
6.60858154e-01 -1.09166026e+00 -7.89409339e-01 5.11645019e-01
6.10738337e-01 -1.10676885e+00 9.40180302e-01 -2.49488041e-01
-1.31519213e-01 -3.19848120e-01 -3.70658010e-01 -7.09753513e-01
-4.54326361e-01 -5.36082983e-01 -4.56269860e-01 7.23180294e-01
1.95638940e-01 -6.50377154e-01 1.52935338e+00 7.69696891e-01
-3.65931690e-01 -1.06235588e+00 -8.59906614e-01 -9.40369248e-01
-4.80932415e-01 -8.53173554e-01 8.82413745e-01 5.61382294e-01
-2.57002085e-01 3.28720182e-01 -2.29743049e-01 4.68214989e-01
6.65686727e-01 -1.09794699e-02 8.60706568e-01 -1.05751967e+00
-5.23352206e-01 -7.68326283e-01 -9.08613980e-01 -1.11463535e+00
6.15996588e-03 -7.29551613e-01 -5.44775546e-01 -1.07204628e+00
-1.36369988e-01 -6.66368723e-01 -2.98650682e-01 2.86508262e-01
1.49094105e-01 2.71446019e-01 9.63569432e-03 -4.87273671e-02
-7.32352197e-01 3.27546179e-01 6.59788847e-01 1.97846703e-02
-3.47705543e-01 -3.99699092e-01 -7.14329600e-01 7.29700089e-01
7.73488760e-01 -1.70562759e-01 -7.39687383e-01 -1.00161159e+00
5.69568098e-01 -2.93433899e-03 2.08574057e-01 -1.64251137e+00
9.28274155e-01 7.72484243e-02 2.96038568e-01 -5.48105001e-01
1.55829087e-01 -8.79784703e-01 7.23626018e-01 6.97555482e-01
3.66358072e-01 7.56618202e-01 6.39694750e-01 4.85577434e-01
2.44013503e-01 5.99926226e-02 4.73170817e-01 3.33042622e-01
-1.14193928e+00 5.58782160e-01 1.94379315e-02 -1.47874370e-01
9.91104424e-01 -6.55478299e-01 -6.73148870e-01 6.53439611e-02
1.81397982e-02 2.05269590e-01 5.71272016e-01 3.31846714e-01
5.22172213e-01 -1.30809033e+00 -2.49173075e-01 3.27470541e-01
-3.47700007e-02 9.10336226e-02 8.33849132e-01 4.88911390e-01
-1.16704810e+00 6.07086182e-01 -4.21478719e-01 -5.77682674e-01
-1.13668716e+00 6.17455721e-01 3.48802865e-01 -3.97759020e-01
-6.88105166e-01 9.57545161e-01 -3.49075273e-02 -5.58829606e-02
1.86149165e-01 -2.15350091e-01 8.25665146e-02 -4.75450873e-01
4.71392035e-01 8.96342635e-01 5.61683297e-01 -3.53624552e-01
-4.86698657e-01 5.39460599e-01 -1.01056546e-01 4.96069252e-01
1.00836933e+00 5.46830446e-02 -6.15725480e-03 -3.06627184e-01
1.13075459e+00 -3.55651945e-01 -1.07282650e+00 7.43834302e-02
-1.05944976e-01 -7.05509111e-02 5.20127058e-01 -4.28563386e-01
-1.49905610e+00 3.91774893e-01 8.84760439e-01 5.59981959e-03
1.62393510e+00 -5.56288302e-01 1.21319830e+00 2.29988724e-01
8.40122700e-01 -1.19710124e+00 -2.15360627e-01 5.30672371e-01
1.10775948e-01 -7.62559295e-01 1.90688744e-01 -4.04072225e-01
-4.20997590e-02 1.17828178e+00 1.25482464e+00 -9.77970138e-02
6.28675938e-01 6.78808153e-01 -2.22083509e-01 -4.13811952e-01
-6.40526891e-01 2.13874429e-01 1.34742454e-01 4.04275596e-01
-7.16990530e-02 2.68837065e-01 -3.37962434e-02 8.88680637e-01
-3.69570285e-01 -8.77613500e-02 3.13816309e-01 9.17401075e-01
-3.85657251e-02 -8.42882037e-01 8.41390528e-03 6.26578629e-01
-2.84618258e-01 -3.25190574e-01 1.32549733e-01 8.38532567e-01
-2.08055928e-01 5.20900011e-01 3.10555428e-01 -1.00766635e+00
3.07532370e-01 -5.51797092e-01 3.20927262e-01 -1.00105897e-01
-8.72719824e-01 -1.96250066e-01 3.21618132e-02 -7.79573560e-01
5.73524058e-01 -1.15943767e-01 -1.45231140e+00 -7.86140800e-01
-3.66142005e-01 -2.30136350e-01 1.08249247e+00 3.74692202e-01
1.06018972e+00 7.64676034e-01 1.08235769e-01 -9.82222974e-01
-1.91915035e-01 -2.99253553e-01 -4.56731856e-01 -3.04793149e-01
-2.64460057e-01 -5.76008201e-01 -2.55314428e-02 -7.63359427e-01] | [7.107876300811768, 5.518571376800537] |
99626ab6-39ff-4097-a5a6-989d65dff4c4 | contrastive-learning-of-general-purpose-audio | 2010.10915 | null | https://arxiv.org/abs/2010.10915v1 | https://arxiv.org/pdf/2010.10915v1.pdf | Contrastive Learning of General-Purpose Audio Representations | We introduce COLA, a self-supervised pre-training approach for learning a general-purpose representation of audio. Our approach is based on contrastive learning: it learns a representation which assigns high similarity to audio segments extracted from the same recording while assigning lower similarity to segments from different recordings. We build on top of recent advances in contrastive learning for computer vision and reinforcement learning to design a lightweight, easy-to-implement self-supervised model of audio. We pre-train embeddings on the large-scale Audioset database and transfer these representations to 9 diverse classification tasks, including speech, music, animal sounds, and acoustic scenes. We show that despite its simplicity, our method significantly outperforms previous self-supervised systems. We furthermore conduct ablation studies to identify key design choices and release a library to pre-train and fine-tune COLA models. | ['Neil Zeghidour', 'David Grangier', 'Aaqib Saeed'] | 2020-10-21 | null | null | null | null | ['spoken-command-recognition'] | ['speech'] | [ 4.23474073e-01 -1.32306829e-01 1.64521173e-01 -5.44993043e-01
-1.14371514e+00 -7.22092450e-01 3.82164717e-01 2.71726429e-01
-4.55364525e-01 1.41812906e-01 5.54407716e-01 2.61702649e-02
1.65856462e-02 -4.92844015e-01 -8.71167183e-01 -2.99628764e-01
-5.80742836e-01 3.94762665e-01 1.00066520e-01 -1.52521282e-01
-2.05197651e-02 9.83363315e-02 -1.80969810e+00 6.50990367e-01
-2.49507707e-02 1.00692582e+00 9.75491926e-02 1.07688498e+00
2.46214643e-01 8.11845303e-01 -5.41673064e-01 7.48698507e-03
6.81648776e-02 -4.69541907e-01 -9.29677904e-01 -3.58892791e-02
6.86030567e-01 -8.84645581e-02 -3.96316350e-01 5.24692118e-01
8.33511889e-01 1.43849969e-01 7.11166978e-01 -1.30208945e+00
-7.10769296e-01 9.85652864e-01 -2.36851633e-01 4.75693822e-01
4.64273810e-01 2.03507051e-01 1.57162690e+00 -8.31865072e-01
2.58438051e-01 1.16188931e+00 9.93740082e-01 5.87261021e-01
-1.39012909e+00 -8.16982448e-01 -1.65164284e-02 2.67969012e-01
-1.09373176e+00 -9.21943128e-01 9.46078300e-01 -4.37309980e-01
1.05307627e+00 1.65475562e-01 7.04149485e-01 1.34550071e+00
-2.21489459e-01 9.23546731e-01 8.92124593e-01 -5.17245471e-01
3.60757560e-01 -2.18640745e-01 -3.07610370e-02 4.54014003e-01
-4.89076495e-01 2.15318635e-01 -8.44362199e-01 -4.51013029e-01
4.78242844e-01 -4.05451566e-01 2.52406509e-03 -5.10051787e-01
-1.21997750e+00 7.04136312e-01 5.17449319e-01 2.90238291e-01
-1.01005457e-01 5.35221875e-01 7.71599174e-01 6.64721906e-01
1.94022030e-01 8.99432361e-01 -5.43143153e-01 -4.12095994e-01
-7.86051452e-01 2.02430665e-01 5.38885415e-01 6.74080312e-01
6.47356093e-01 5.41275918e-01 -1.00598847e-02 1.11609781e+00
2.00059801e-01 2.04947248e-01 8.49953830e-01 -1.11138582e+00
1.25251487e-01 -1.95441768e-03 -3.23733717e-01 -6.19685173e-01
-3.73475403e-01 -5.41402221e-01 -4.78044629e-01 9.79619995e-02
-6.78994507e-03 -1.19377360e-01 -6.44335091e-01 1.81993937e+00
-7.80652836e-02 6.45640612e-01 1.88123375e-01 6.25246644e-01
8.61377478e-01 6.83308184e-01 2.41469033e-02 1.97427884e-01
9.57269788e-01 -1.13990068e+00 -2.50922292e-01 -3.66342455e-01
4.95513380e-01 -8.01101744e-01 1.37018633e+00 4.79074001e-01
-1.06398237e+00 -9.13453400e-01 -1.21778226e+00 6.06075116e-02
-1.89321116e-01 4.39285189e-02 5.95658302e-01 2.66645014e-01
-1.02167249e+00 9.27571118e-01 -6.90512359e-01 -2.21262664e-01
4.68842000e-01 2.54377693e-01 -3.95284951e-01 4.14308935e-01
-9.49221611e-01 3.80430877e-01 4.38876212e-01 -4.19656307e-01
-1.53617203e+00 -9.40403700e-01 -9.42428172e-01 5.31167835e-02
5.95002174e-02 -4.33141917e-01 1.76266301e+00 -1.18409395e+00
-1.66330492e+00 9.88076329e-01 2.37786576e-01 -8.89034510e-01
-1.56639159e-01 -4.52799112e-01 -5.10203362e-01 1.58393249e-01
1.49362311e-02 9.54049766e-01 1.16789317e+00 -1.02706039e+00
-4.63582605e-01 1.41202986e-01 2.93647442e-02 2.20201209e-01
-5.46587467e-01 7.87045583e-02 -1.12509906e-01 -7.97991931e-01
-2.88811117e-01 -8.98924708e-01 -6.23016506e-02 -8.73784162e-03
-2.65623301e-01 -1.81084827e-01 5.84359527e-01 -1.48984134e-01
8.98992121e-01 -2.67698693e+00 9.53454003e-02 -1.61708985e-02
1.45059302e-01 1.17568992e-01 -8.38424921e-01 6.00445330e-01
-4.25186366e-01 -2.21979722e-01 -2.91661680e-01 -4.21353668e-01
1.02449276e-01 1.84682325e-01 -7.17517316e-01 1.73280969e-01
3.14131349e-01 6.20351195e-01 -1.23905110e+00 -4.14259523e-01
8.92025381e-02 3.84963959e-01 -1.13259482e+00 6.54666185e-01
-2.83415526e-01 2.27178350e-01 7.50455931e-02 4.39727396e-01
1.42464370e-01 4.93086092e-02 9.67582092e-02 -2.73897558e-01
3.26039046e-02 7.89879918e-01 -1.11855447e+00 2.21200061e+00
-6.74252272e-01 7.44083762e-01 9.02275182e-03 -1.35977387e+00
8.58875036e-01 3.51714611e-01 5.92208803e-01 -5.21882713e-01
-5.22163622e-02 1.27105908e-02 6.75754398e-02 -4.27494079e-01
3.14831495e-01 -1.54559240e-01 -3.38858575e-01 6.59426332e-01
8.23923528e-01 -5.58086932e-01 -8.27987045e-02 1.65866330e-01
1.26908112e+00 2.18925655e-01 1.96141601e-01 -7.60861561e-02
1.12553552e-01 -4.43559170e-01 3.11229050e-01 7.39595890e-01
-2.56114870e-01 7.86453664e-01 1.25394687e-01 -3.57785821e-01
-8.77359450e-01 -1.37505770e+00 -1.10823512e-02 1.88314748e+00
-4.33165878e-01 -9.14162397e-01 -5.13347745e-01 -5.74347675e-01
1.25040382e-01 3.43465269e-01 -5.63530564e-01 -3.83578092e-01
-3.72728348e-01 -2.29138881e-01 9.09043252e-01 7.41060853e-01
3.13582532e-02 -1.41094434e+00 -4.18121606e-01 4.01105881e-01
1.52605474e-01 -8.92704427e-01 -4.70426261e-01 8.14864039e-01
-5.76010883e-01 -9.48260128e-01 -2.28452817e-01 -1.21967602e+00
-1.13999248e-02 1.74848095e-01 1.52486718e+00 -2.33342588e-01
-3.83350074e-01 7.31725097e-01 -3.63305271e-01 -6.31775081e-01
-6.52502775e-01 1.10240363e-01 4.97158051e-01 -9.37354416e-02
6.63458556e-02 -1.06413841e+00 -3.39080244e-01 -1.12686194e-02
-7.56894231e-01 -4.81889993e-01 2.38020435e-01 9.05826628e-01
6.63092613e-01 -2.17749178e-01 8.94674778e-01 -7.35332549e-01
6.86153173e-01 -5.00240564e-01 -2.18175977e-01 -4.38101441e-02
-2.48428017e-01 7.18972981e-02 7.00326681e-01 -6.62568808e-01
-1.94080666e-01 2.74988562e-01 -5.00171661e-01 -5.60861230e-01
-3.16413641e-01 2.98360378e-01 1.24335267e-01 1.00571968e-01
1.02225959e+00 1.15921311e-01 -3.50559987e-02 -5.72645664e-01
6.93840206e-01 8.90582681e-01 9.10659790e-01 -7.08855689e-01
8.79713535e-01 2.21530318e-01 -5.03888369e-01 -9.68458652e-01
-1.03917456e+00 -5.26945591e-01 -4.41782027e-01 -2.88108061e-03
3.76992702e-01 -1.16511726e+00 -5.43795168e-01 1.47121161e-01
-7.10217953e-01 -5.91194749e-01 -9.31762338e-01 5.21113753e-01
-1.10267127e+00 1.54834285e-01 -5.53370357e-01 -5.77237427e-01
-3.00822586e-01 -6.68870509e-01 1.12008202e+00 -5.00920191e-02
-6.82879686e-01 -8.29885602e-01 8.03903997e-01 1.06585592e-01
3.70316118e-01 -2.27846548e-01 7.98021972e-01 -8.67803156e-01
-2.45270338e-02 2.33417422e-01 1.87183648e-01 7.50021815e-01
2.93083966e-01 1.22277148e-01 -1.60104012e+00 -5.08520603e-01
-4.15351599e-01 -1.11472440e+00 1.04175723e+00 2.21929789e-01
1.48102272e+00 -1.98261753e-01 1.74744323e-01 7.21249759e-01
9.88539398e-01 -5.27522585e-04 3.53228897e-01 4.07792032e-01
4.01592910e-01 2.92141855e-01 3.58813703e-01 5.16280055e-01
2.10190222e-01 6.35910988e-01 4.18551773e-01 -1.63010538e-01
-3.28424066e-01 -6.65661097e-01 5.32708526e-01 1.32874537e+00
4.26471740e-01 1.86749041e-01 -5.48223138e-01 9.17722106e-01
-1.42398000e+00 -1.08502603e+00 6.65900946e-01 1.97057450e+00
1.30451226e+00 1.94285393e-01 5.29965401e-01 6.52343333e-01
3.03464144e-01 2.65860170e-01 -4.21366215e-01 -5.86603582e-01
5.89793511e-02 8.25195491e-01 -1.97403848e-01 2.38983288e-01
-1.49328291e+00 1.00950193e+00 7.89365387e+00 6.01580441e-01
-1.14090979e+00 -1.64437093e-04 7.23940730e-02 -2.33593345e-01
-2.12381944e-01 -1.62903666e-02 -3.34024161e-01 2.37760678e-01
1.36332309e+00 -8.33063498e-02 5.44857621e-01 9.64986026e-01
-2.17051610e-01 6.29947126e-01 -1.54702258e+00 1.14752662e+00
7.23783076e-02 -1.44398355e+00 -1.24775562e-02 -5.00194013e-01
5.62755644e-01 3.55186373e-01 2.64415354e-01 7.60854304e-01
5.38737655e-01 -9.90245581e-01 6.90452099e-01 1.28145799e-01
8.02200854e-01 -6.84734166e-01 2.54302502e-01 -7.36229047e-02
-1.19935000e+00 -2.02319339e-01 -4.52659220e-01 -3.07050914e-01
-1.69131741e-01 2.41066560e-01 -1.11676800e+00 1.14690207e-01
1.01855612e+00 9.97250915e-01 -6.36457324e-01 1.15101361e+00
-2.08815813e-01 1.18598318e+00 -2.63198793e-01 1.19336836e-01
2.45839208e-01 4.14795905e-01 4.53467399e-01 1.55423558e+00
4.25857045e-02 -6.16107166e-01 3.44157338e-01 5.81407428e-01
-4.25946981e-01 1.36032149e-01 -7.95437634e-01 -2.21724063e-01
7.16828644e-01 1.12041259e+00 -1.97597817e-01 -2.81523854e-01
-2.44563505e-01 8.39423597e-01 5.23996592e-01 2.08143648e-02
-5.99705696e-01 -6.77152872e-01 9.21273947e-01 -2.00009216e-02
6.29593313e-01 -1.26884267e-01 1.33509949e-01 -9.20398712e-01
-2.76401341e-01 -1.18750775e+00 4.35997814e-01 -9.76467073e-01
-1.43129849e+00 7.50223219e-01 -1.76435068e-01 -1.47414923e+00
-7.57185459e-01 -3.00658017e-01 -7.15012014e-01 2.46579066e-01
-1.44647515e+00 -9.47346866e-01 9.51165259e-02 6.98881030e-01
5.85002005e-01 -7.63657928e-01 1.46473110e+00 2.13692293e-01
-2.32716471e-01 9.15458918e-01 1.28942549e-01 3.21445465e-01
9.82441187e-01 -1.43285942e+00 7.03002334e-01 1.78441390e-01
1.19179904e+00 3.95630419e-01 5.75812757e-01 9.44146216e-02
-1.22313726e+00 -1.07370746e+00 4.91792113e-01 -4.02761847e-01
1.06608868e+00 -5.94712257e-01 -9.05735791e-01 8.01500618e-01
4.88253027e-01 2.11252302e-01 1.27658296e+00 6.59740031e-01
-9.98574913e-01 -4.41797704e-01 -7.07874119e-01 4.15320277e-01
1.00301695e+00 -1.24119139e+00 -1.01969862e+00 2.02102974e-01
8.61367702e-01 -9.56134573e-02 -9.05306101e-01 3.00909042e-01
5.93970001e-01 -6.65169477e-01 1.09414005e+00 -8.76271188e-01
2.78036922e-01 -1.33860603e-01 -3.86280686e-01 -1.68202889e+00
-3.95598024e-01 -9.26814854e-01 -8.15225542e-02 1.31722629e+00
2.48063326e-01 -1.39965117e-01 6.46618128e-01 -5.25313377e-01
-4.18106258e-01 -4.72498059e-01 -9.98266876e-01 -9.17750359e-01
1.32464036e-01 -7.47825861e-01 3.67803097e-01 1.02112842e+00
1.57642812e-01 8.19880128e-01 -3.34023863e-01 9.89622250e-03
5.15464842e-01 1.98028237e-01 9.60788667e-01 -1.44059026e+00
-8.22372675e-01 -1.65053219e-01 -7.29637444e-01 -9.77857888e-01
3.97809595e-01 -1.19038343e+00 2.48984486e-01 -9.88734305e-01
8.04089531e-02 -4.48601037e-01 -8.20417345e-01 8.74753356e-01
3.46903294e-01 6.42601788e-01 1.39204785e-01 1.22365199e-01
-8.02842617e-01 6.52368605e-01 6.28668547e-01 -5.40544927e-01
-3.19425493e-01 -5.76407872e-02 -8.73698950e-01 7.56419003e-01
8.89531791e-01 -6.41647875e-01 -5.78822315e-01 -5.73462009e-01
4.79931831e-02 -3.14083695e-01 3.34965110e-01 -1.45304739e+00
-1.12922624e-01 2.70401925e-01 2.11175248e-01 -3.10512334e-01
5.51962316e-01 -4.91884530e-01 -4.85287011e-01 8.18200111e-02
-9.77469146e-01 -6.89998195e-02 5.01764357e-01 5.81086934e-01
-4.77081597e-01 -2.55872250e-01 8.32472801e-01 4.86730523e-02
-7.03044474e-01 1.58767533e-02 -5.46595931e-01 3.34559560e-01
4.41911340e-01 1.53445646e-01 1.42524093e-01 -6.04530931e-01
-9.23057497e-01 -6.17364906e-02 1.50443062e-01 7.25086272e-01
7.47366250e-01 -1.60850155e+00 -8.49104881e-01 4.53406721e-01
4.56266105e-01 -2.59647220e-01 1.70341849e-01 5.49166761e-02
-1.31417170e-01 2.20434010e-01 -3.94748598e-01 -7.58132994e-01
-1.31234360e+00 4.90206510e-01 1.84655249e-01 3.59865017e-02
-5.95646977e-01 1.13106990e+00 9.08842385e-02 -7.32642293e-01
6.77629590e-01 -5.27730882e-01 -9.76630151e-02 -7.80497678e-03
6.43047750e-01 -1.22146998e-02 1.22982480e-01 -4.15067106e-01
-4.31297511e-01 5.56102157e-01 -3.77098024e-02 -3.71585280e-01
1.61866069e+00 3.56316060e-01 4.79227543e-01 1.04935646e+00
1.47495770e+00 1.11359484e-01 -1.31219661e+00 -4.04835850e-01
-1.65621266e-01 3.02607100e-02 1.12258904e-01 -4.94095832e-01
-8.56818914e-01 1.03904808e+00 8.27655017e-01 1.78913936e-01
1.06579816e+00 1.77662805e-01 5.50404668e-01 7.94750154e-01
1.43331870e-01 -1.24115145e+00 8.08384120e-01 5.93652964e-01
1.05632782e+00 -8.78685713e-01 -7.80064315e-02 2.65119046e-01
-7.87101209e-01 1.06610823e+00 4.01062369e-01 -6.24304056e-01
7.20261395e-01 5.17172337e-01 2.13793293e-01 -3.53288911e-02
-1.10467017e+00 -3.63777518e-01 2.69178391e-01 1.08657968e+00
6.82951927e-01 -1.94796221e-03 5.82754910e-01 7.42010832e-01
-6.72266185e-01 -8.30494463e-02 4.56525177e-01 1.05466521e+00
-3.88520449e-01 -1.16226161e+00 -4.80044596e-02 1.82065696e-01
-3.59853119e-01 -1.74368173e-01 -5.25583088e-01 3.83584052e-01
-2.61652023e-02 9.28144038e-01 3.32394809e-01 -8.47528696e-01
5.06781578e-01 2.58955479e-01 5.65445185e-01 -9.15110409e-01
-8.31543505e-01 1.33691058e-01 1.34714052e-01 -4.42841321e-01
-6.16747439e-01 -5.58272898e-01 -1.10112751e+00 3.45517844e-01
-4.06678505e-02 3.81155312e-01 5.61650872e-01 6.53435707e-01
5.31422734e-01 7.37087786e-01 1.02342212e+00 -1.12033892e+00
-7.20934927e-01 -1.03525531e+00 -5.02356946e-01 4.40829337e-01
6.03310347e-01 -5.37528455e-01 -3.98876369e-01 2.18427896e-01] | [15.34662914276123, 5.135128974914551] |
2efece1f-ec04-4995-a3f1-c9ce9eb07f61 | unsupervised-dual-cascade-learning-with | 1811.00436 | null | http://arxiv.org/abs/1811.00436v1 | http://arxiv.org/pdf/1811.00436v1.pdf | Unsupervised Dual-Cascade Learning with Pseudo-Feedback Distillation for Query-based Extractive Summarization | We propose Dual-CES -- a novel unsupervised, query-focused, multi-document
extractive summarizer. Dual-CES is designed to better handle the tradeoff
between saliency and focus in summarization. To this end, Dual-CES employs a
two-step dual-cascade optimization approach with saliency-based pseudo-feedback
distillation. Overall, Dual-CES significantly outperforms all other
state-of-the-art unsupervised alternatives. Dual-CES is even shown to be able
to outperform strong supervised summarizers. | ['Haggai Roitman', 'Doron Cohen', 'David Konopnicki', 'Odellia Boni', 'Guy Feigenblat'] | 2018-11-01 | null | null | null | null | ['query-based-extractive-summarization'] | ['natural-language-processing'] | [-3.18762264e-03 7.31563941e-02 -5.22455037e-01 -2.42740780e-01
-1.50328898e+00 -3.73179525e-01 7.49923468e-01 6.63533628e-01
-3.98311049e-01 6.04842424e-01 9.76293623e-01 9.51815024e-02
3.66286263e-02 -2.66634643e-01 -5.00890195e-01 -1.31398991e-01
1.33486599e-01 5.97332895e-01 2.98520386e-01 -5.82757950e-01
9.12132323e-01 1.64431378e-01 -1.34192526e+00 6.29161179e-01
1.22060525e+00 4.16568458e-01 4.05679286e-01 9.89579260e-01
-7.41519928e-02 9.04218912e-01 -7.90319920e-01 -4.67138797e-01
-4.33556944e-01 -6.03556275e-01 -8.88510227e-01 -2.16550499e-01
8.73803496e-01 -3.22398931e-01 -4.85594183e-01 8.38197827e-01
6.15018606e-01 2.97272712e-01 9.09612894e-01 -6.71970725e-01
-1.09248316e+00 9.24582660e-01 -6.40847743e-01 9.26462591e-01
2.71134645e-01 -9.03703794e-02 1.54342222e+00 -1.35365033e+00
6.64501548e-01 1.42964578e+00 3.60979825e-01 5.16119123e-01
-1.18099463e+00 -2.48683169e-01 2.17870638e-01 -3.71498279e-02
-9.80072021e-01 -6.45375788e-01 8.87317717e-01 -7.27146538e-03
1.22416484e+00 3.97846162e-01 4.14604723e-01 1.00268066e+00
3.13981891e-01 1.42406070e+00 7.42760122e-01 -3.37516844e-01
2.78171897e-01 7.88459089e-03 3.70493680e-01 7.55802512e-01
3.24120373e-01 -5.13609707e-01 -1.20539784e+00 7.86117911e-02
3.19919169e-01 -1.36273488e-01 -1.98800698e-01 1.47286996e-01
-1.16477013e+00 8.98281753e-01 6.17864728e-01 3.52544039e-01
-5.53893089e-01 3.19680363e-01 5.35180926e-01 1.27336353e-01
1.23635435e+00 1.09612656e+00 -1.63919047e-01 -2.26968914e-01
-1.85757077e+00 6.63681507e-01 7.07390845e-01 9.21464443e-01
3.72332096e-01 2.47284397e-01 -1.00334799e+00 6.20581269e-01
-1.11665418e-02 4.23547924e-01 6.36246562e-01 -7.72696078e-01
7.25797057e-01 4.68752027e-01 1.07832335e-01 -1.00121772e+00
-3.18623155e-01 -9.70609844e-01 -6.40240073e-01 -3.78022045e-01
-5.98549485e-01 6.08447604e-02 -9.49062824e-01 1.39344323e+00
-2.51844645e-01 -8.25825110e-02 2.34780222e-01 8.07934403e-01
1.29058015e+00 8.75091314e-01 1.67657390e-01 -2.43981525e-01
9.60686862e-01 -1.81825674e+00 -8.92958522e-01 -6.08919978e-01
3.36236119e-01 -4.88148600e-01 1.09300864e+00 -2.65838560e-02
-1.63761926e+00 -3.99435103e-01 -1.30268395e+00 -5.99396169e-01
-2.67119139e-01 5.72067916e-01 3.85783851e-01 5.83529882e-02
-1.25041270e+00 6.55756593e-01 -7.39064693e-01 -2.47332498e-01
5.36959648e-01 -8.82161967e-03 -7.94332996e-02 1.61962539e-01
-8.90374064e-01 9.96347845e-01 5.67521811e-01 -2.60344744e-01
-1.04983664e+00 -9.35054302e-01 -9.12351310e-01 5.57903171e-01
4.46801901e-01 -1.12637031e+00 1.74344361e+00 -6.64619505e-01
-1.38117898e+00 7.48351634e-01 -6.23936474e-01 -8.35456073e-01
2.71318585e-01 -8.27975929e-01 -2.88995802e-01 7.01212406e-01
6.66274607e-01 7.88937092e-01 9.27104890e-01 -1.32608747e+00
-5.11828303e-01 -1.21609688e-01 -1.12545229e-01 6.90939903e-01
-5.03787100e-01 8.65803510e-02 -5.94394267e-01 -1.03712595e+00
-3.02662194e-01 -4.15985256e-01 -1.88798472e-01 -5.39606512e-01
-8.64154696e-01 -3.41702044e-01 8.62516820e-01 -6.20685339e-01
1.83634531e+00 -1.89570904e+00 4.97426927e-01 -5.15522778e-01
5.94424248e-01 5.54499447e-01 -3.50665987e-01 7.02770591e-01
2.11639032e-01 2.66024798e-01 -1.77801669e-01 -1.06562185e+00
-2.11238489e-01 -4.08635229e-01 -4.40842927e-01 3.11761089e-02
5.67706108e-01 1.30930746e+00 -1.30930543e+00 -6.55582547e-01
-3.31760980e-02 5.22153452e-02 -5.09450257e-01 1.00263253e-01
-5.29496372e-01 -1.00398205e-01 -6.71905935e-01 5.47876477e-01
1.75760344e-01 -5.03676414e-01 -2.99464434e-01 -7.62018468e-03
-2.03753144e-01 6.69411957e-01 -2.86655635e-01 2.02822065e+00
-2.59177417e-01 8.53432834e-01 -1.17860489e-01 -7.77820230e-01
7.48554587e-01 6.48045819e-03 3.75297926e-02 -6.39815927e-01
3.04277167e-02 3.81607533e-01 -6.66827500e-01 -2.37744495e-01
1.53236175e+00 8.12742934e-02 -1.92857549e-01 4.06151742e-01
6.46024168e-01 -4.05351251e-01 6.13622010e-01 1.23651469e+00
1.15046430e+00 -2.91187614e-01 3.24561566e-01 -6.27862394e-01
3.12911689e-01 1.15893364e-01 2.51392514e-01 1.18056297e+00
-3.27512361e-02 8.66448283e-01 3.78223926e-01 2.91823775e-01
-8.68586421e-01 -1.24428129e+00 5.24619579e-01 1.35447955e+00
2.66927958e-01 -9.46910501e-01 -7.70343125e-01 -8.59296262e-01
1.45810008e-01 1.12298846e+00 -6.05256736e-01 -3.62305969e-01
-4.41478014e-01 -1.73957288e-01 4.99529004e-01 5.74003875e-01
5.39604425e-01 -1.03109407e+00 -7.32511699e-01 4.42683071e-01
-2.65634030e-01 -7.76871026e-01 -1.04342723e+00 3.29782456e-01
-1.10262620e+00 -4.43057328e-01 -1.05216980e+00 -8.91573429e-01
1.95744172e-01 7.90345669e-01 1.57946181e+00 -1.00435749e-01
4.69061881e-02 4.00743663e-01 -4.41268712e-01 -6.69361889e-01
-2.67837673e-01 8.77995610e-01 -2.14631498e-01 -3.53129059e-01
2.17619821e-01 -5.77128291e-01 -7.52091944e-01 -4.71748680e-01
-9.78798330e-01 2.81553328e-01 8.60739350e-01 8.55613768e-01
4.58440453e-01 -6.02500677e-01 1.15260506e+00 -9.66845214e-01
1.46947742e+00 -4.94652301e-01 4.22134995e-02 6.11366391e-01
-6.89278126e-01 5.01202703e-01 5.07059515e-01 -1.96878687e-01
-1.04101205e+00 -4.49638218e-01 7.08597824e-02 -3.37145239e-01
3.58011782e-01 9.31416810e-01 3.01284641e-01 2.36723781e-01
7.65580714e-01 5.40930390e-01 -4.51043248e-01 -5.34534037e-01
7.40711808e-01 7.27937520e-01 9.65870976e-01 -1.99384749e-01
4.68598694e-01 2.38504350e-01 -3.69060874e-01 -8.79878402e-01
-1.47245717e+00 -8.84930372e-01 -3.66290540e-01 -1.06866568e-01
7.72218585e-01 -1.14903355e+00 8.95532966e-02 3.35729063e-01
-1.30938625e+00 -1.04780979e-01 -3.63330632e-01 1.30892426e-01
-4.13642824e-01 2.94492871e-01 -9.09648478e-01 -5.83180964e-01
-1.12591863e+00 -7.33031213e-01 1.44405568e+00 5.93874931e-01
-1.38956815e-01 -9.70271826e-01 2.57508516e-01 2.12833345e-01
7.69609988e-01 -8.70587304e-02 4.91010845e-01 -9.83518481e-01
-2.86390215e-01 6.74168020e-03 -3.20499718e-01 2.12811530e-01
4.73330282e-02 -2.45028064e-02 -6.45023286e-01 -3.11945587e-01
-2.81703591e-01 -7.07769990e-01 1.75418150e+00 6.10685945e-01
8.75066936e-01 -6.01440370e-01 -3.37702721e-01 2.08777174e-01
1.31285882e+00 -1.85268074e-01 3.09498996e-01 2.56845087e-01
5.74799895e-01 2.68920034e-01 7.82876730e-01 3.04600477e-01
6.29855633e-01 2.75930732e-01 1.43839732e-01 -2.28568077e-01
-3.22049469e-01 -6.07044280e-01 3.07554930e-01 1.36566448e+00
4.17757720e-01 -5.17266929e-01 -6.03547096e-01 8.65292072e-01
-2.08273315e+00 -1.06653333e+00 3.01122777e-02 1.39852965e+00
1.01971972e+00 3.32230985e-01 1.44094571e-01 -4.61552739e-01
6.19970202e-01 8.63758504e-01 -6.09857082e-01 -4.90260720e-01
-4.29397225e-01 3.27665538e-01 2.57566094e-01 4.38824415e-01
-1.26303720e+00 1.37558138e+00 7.25678349e+00 9.19624209e-01
-1.03890967e+00 1.20439753e-01 5.45124471e-01 -3.29621643e-01
-4.37013805e-01 -2.07825407e-01 -8.23397458e-01 2.61714131e-01
9.39587712e-01 -6.92381680e-01 -2.53862794e-02 9.37615514e-01
8.16571489e-02 -3.34485769e-01 -9.16123152e-01 8.92635345e-01
7.12698579e-01 -1.92087781e+00 4.84568536e-01 -5.03277481e-01
1.08053696e+00 2.78186768e-01 2.36175254e-01 4.25097793e-01
2.02692896e-01 -7.44087279e-01 1.03786111e+00 4.73038882e-01
7.42500961e-01 -7.44830132e-01 4.81084168e-01 3.69882584e-01
-9.08974111e-01 1.15535289e-01 -1.23587683e-01 2.56075114e-01
3.70546371e-01 6.53172493e-01 -3.79923224e-01 6.99372470e-01
4.05337572e-01 9.72323656e-01 -8.99128199e-01 1.08395779e+00
-4.16833192e-01 6.99661255e-01 1.08938739e-01 -4.03630972e-01
6.00844860e-01 2.85984010e-01 1.05084062e+00 1.85104537e+00
-1.15658222e-02 2.72419583e-02 2.12474212e-01 9.38856184e-01
-4.10415828e-01 -6.14375025e-02 -3.71902108e-01 -3.61791462e-01
5.08521438e-01 1.12371588e+00 -5.69293737e-01 -7.78848052e-01
3.78007293e-02 1.32808900e+00 6.13491297e-01 3.49487096e-01
-4.67491269e-01 -7.96761155e-01 1.77605361e-01 -3.41496348e-01
7.03622639e-01 -2.95655489e-01 -3.63516480e-01 -1.45003486e+00
-1.24757871e-01 -8.02637219e-01 2.74354637e-01 -8.85778368e-01
-9.54337776e-01 6.28084600e-01 1.51224047e-01 -6.61269724e-01
-3.13660741e-01 -9.27384049e-02 -9.20931101e-01 6.72586262e-01
-1.72676861e+00 -1.16586232e+00 -5.40322848e-02 1.88930586e-01
1.31106663e+00 -1.72686026e-01 6.52485132e-01 -8.51860419e-02
-4.84068722e-01 4.29129839e-01 2.95948386e-01 -3.47629964e-01
7.69865632e-01 -1.62316203e+00 8.73142481e-01 1.15771127e+00
2.95538098e-01 6.14015281e-01 9.96002913e-01 -7.30316997e-01
-1.10193276e+00 -1.17557240e+00 1.07596278e+00 -3.31007481e-01
6.03690684e-01 -2.29070500e-01 -7.51541853e-01 5.63710749e-01
7.12570906e-01 -4.59643692e-01 5.03057003e-01 1.00450918e-01
-4.03144270e-01 -8.70769247e-02 -6.91146791e-01 6.93657041e-01
7.36081958e-01 -7.18555152e-01 -1.33532739e+00 2.57687628e-01
1.12389147e+00 -3.19047183e-01 -4.20406282e-01 1.68666407e-01
2.16188163e-01 -8.81539762e-01 9.03351963e-01 -5.75466514e-01
1.09221673e+00 -7.94816986e-02 2.54743040e-01 -1.91142273e+00
-6.22319221e-01 -7.02578783e-01 -8.51201177e-01 1.38173139e+00
5.86702764e-01 -1.16341546e-01 5.22023737e-01 -1.11782342e-01
-6.04483426e-01 -8.72306824e-01 -5.23298264e-01 -6.37478232e-01
3.88529263e-02 2.91082114e-01 4.94472496e-02 4.29176509e-01
2.38152415e-01 1.06672573e+00 -4.31768119e-01 -3.24479371e-01
4.82093483e-01 1.34692639e-01 4.46710885e-01 -9.69154418e-01
-1.85825720e-01 -7.99974978e-01 1.77969187e-01 -1.50538909e+00
1.48473322e-01 -7.61800647e-01 1.55756339e-01 -2.10295653e+00
4.50382888e-01 2.25064620e-01 -4.35528219e-01 1.79077268e-01
-7.10575759e-01 6.59783836e-03 3.10082227e-01 2.70751953e-01
-1.43503714e+00 1.02536440e+00 1.12666941e+00 -5.35734475e-01
-4.22135055e-01 -2.15191305e-01 -1.23803806e+00 1.87286049e-01
9.34317827e-01 -5.95005274e-01 -4.15365458e-01 -5.93853831e-01
-1.18759833e-02 1.20513774e-02 3.39648463e-02 -8.56514633e-01
6.58913791e-01 5.32060787e-02 3.45264040e-02 -1.21215963e+00
2.63248593e-01 3.76822740e-01 -7.43430257e-01 2.15237647e-01
-1.00529814e+00 3.55556071e-01 2.10472748e-01 7.21623123e-01
-6.27637386e-01 -2.82463878e-01 3.04913372e-01 -3.52221370e-01
-5.25378346e-01 4.96146083e-02 -4.02344167e-01 4.89182949e-01
3.78166407e-01 2.92765528e-01 -6.01412892e-01 -5.87589264e-01
-1.55813366e-01 5.97917020e-01 1.70642212e-01 5.60404778e-01
8.42119157e-01 -1.03449380e+00 -1.08794510e+00 -4.48545814e-01
3.66526902e-01 2.98239775e-02 2.00700834e-01 6.31431878e-01
-4.09533083e-01 7.80004025e-01 1.56726256e-01 -2.89771408e-01
-1.24838209e+00 5.80634296e-01 -2.28530690e-01 -7.88943589e-01
-7.00349212e-01 9.29569483e-01 1.91849582e-02 1.10229798e-01
2.45197490e-01 -2.89966792e-01 -2.51395285e-01 1.19545847e-01
7.50903726e-01 5.04100919e-01 1.75725147e-01 -3.11341405e-01
-2.10004374e-01 5.73153272e-02 -6.56894207e-01 -3.95904839e-01
1.40056562e+00 -1.84303313e-01 -1.60935283e-01 5.89268923e-01
1.09825552e+00 5.88637032e-02 -9.45967317e-01 -2.41663218e-01
3.63190711e-01 -1.36707917e-01 4.25812274e-01 -9.68636155e-01
-6.81340337e-01 7.67227530e-01 -1.96636632e-01 2.14435682e-01
1.11587453e+00 1.56815112e-01 8.08510661e-01 6.21644497e-01
-2.08633870e-01 -1.26815712e+00 6.53713822e-01 6.25190616e-01
1.24506736e+00 -1.27581286e+00 2.96757936e-01 -1.35473534e-01
-9.15100515e-01 1.12991834e+00 6.42075419e-01 -5.03143787e-01
1.74025610e-01 -8.86570364e-02 -2.55420804e-01 -5.13488531e-01
-1.11946213e+00 -1.62966371e-01 7.82105803e-01 1.88228458e-01
3.98258001e-01 -1.09136641e-01 -4.77303624e-01 5.38721204e-01
-1.62581682e-01 -2.30989859e-01 4.28366095e-01 1.06918693e+00
-7.44403839e-01 -5.20488203e-01 9.49752554e-02 7.13299453e-01
-5.79527676e-01 -4.39285427e-01 -7.93270826e-01 5.10708213e-01
-7.49276936e-01 1.06414306e+00 -1.05646208e-01 -1.22983843e-01
2.29627937e-01 -2.78536767e-01 2.12890446e-01 -8.85002375e-01
-1.11028421e+00 1.80580318e-01 1.48427188e-01 -4.12061572e-01
-4.81692553e-01 -3.70844543e-01 -8.63325715e-01 -1.35097012e-01
-5.06560683e-01 4.75613832e-01 5.34043968e-01 8.46214235e-01
8.29910278e-01 8.45628858e-01 6.53012812e-01 -1.18998313e+00
-8.06191981e-01 -1.18988287e+00 -4.25515383e-01 7.50503540e-02
9.27608073e-01 -3.50355506e-01 -3.71406347e-01 -2.64190435e-01] | [12.451024055480957, 9.436189651489258] |
2b642b98-63fd-4fea-90cf-6f613291c807 | comparative-study-and-optimization-of-feature | 1208.6335 | null | https://arxiv.org/abs/1208.6335v2 | https://arxiv.org/pdf/1208.6335v2.pdf | Comparative Study and Optimization of Feature-Extraction Techniques for Content based Image Retrieval | The aim of a Content-Based Image Retrieval (CBIR) system, also known as Query by Image Content (QBIC), is to help users to retrieve relevant images based on their contents. CBIR technologies provide a method to find images in large databases by using unique descriptors from a trained image. The image descriptors include texture, color, intensity and shape of the object inside an image. Several feature-extraction techniques viz., Average RGB, Color Moments, Co-occurrence, Local Color Histogram, Global Color Histogram and Geometric Moment have been critically compared in this paper. However, individually these techniques result in poor performance. So, combinations of these techniques have also been evaluated and results for the most efficient combination of techniques have been presented and optimized for each class of image query. We also propose an improvement in image retrieval performance by introducing the idea of Query modification through image cropping. It enables the user to identify a region of interest and modify the initial query to refine and personalize the image retrieval results. | ['Ravdeep Johar', 'Aman Chadha', 'Sushmit Mallik'] | 2012-08-30 | null | null | null | null | ['content-based-image-retrieval', 'image-cropping'] | ['computer-vision', 'computer-vision'] | [ 1.73632622e-01 -6.33507669e-01 -1.88439429e-01 -1.96398929e-01
-8.71524811e-01 -7.94655442e-01 6.55372798e-01 7.02513456e-01
-5.71072400e-01 3.39798361e-01 -5.70431612e-02 1.91946983e-01
-6.92320526e-01 -8.03272784e-01 -5.55095486e-02 -7.52270997e-01
-4.77592275e-03 3.25833201e-01 5.26756465e-01 -3.70995402e-01
9.29578722e-01 1.14817441e+00 -2.16261458e+00 2.90462673e-01
3.17543298e-01 1.03856766e+00 4.88270998e-01 7.61201382e-01
-3.62233490e-01 6.07258797e-01 -7.23055243e-01 2.66617805e-01
1.39202163e-01 -3.29081178e-01 -9.68104780e-01 2.67604589e-01
1.80136070e-01 -2.58593112e-01 -1.47242639e-02 9.52852190e-01
6.71986759e-01 3.69733483e-01 9.48335171e-01 -1.05150259e+00
-8.95757556e-01 -1.04513727e-01 -4.42422420e-01 4.75654304e-01
7.00992227e-01 -5.27432919e-01 5.01756132e-01 -8.76368642e-01
8.38478863e-01 1.22495735e+00 -4.88549396e-02 8.89948979e-02
-9.27408099e-01 -2.35792503e-01 -3.99706304e-01 7.92532802e-01
-1.87175441e+00 -3.81181128e-02 9.37744439e-01 -2.10163787e-01
6.09937191e-01 7.94924617e-01 5.23374677e-01 -2.34522358e-01
1.72286198e-01 6.17844701e-01 1.14704585e+00 -1.03131461e+00
1.68691859e-01 5.39300501e-01 1.03368320e-01 4.94391233e-01
-1.23933055e-01 -1.56526759e-01 -4.18313563e-01 -2.53181636e-01
5.38662016e-01 3.81137460e-01 3.12128924e-02 -4.53521043e-01
-8.63637984e-01 7.38432407e-01 6.46692038e-01 9.56965446e-01
-5.77084124e-01 -1.08293118e-02 2.79766917e-01 2.40038753e-01
9.32084471e-02 3.32799137e-01 -1.96345419e-01 3.57464522e-01
-7.50848234e-01 3.53592694e-01 4.73718166e-01 8.03962350e-01
1.19761848e+00 -6.86326563e-01 -4.41705734e-01 9.38217640e-01
3.98822397e-01 5.86871326e-01 7.70167649e-01 -7.88398504e-01
-3.93240303e-01 9.43048239e-01 3.68398651e-02 -1.56413150e+00
-1.28578141e-01 3.18765789e-01 -3.31549406e-01 1.28758863e-01
-2.29238078e-01 6.65837765e-01 -1.04825473e+00 7.99135864e-01
3.57411951e-01 -4.29685563e-01 -1.40637964e-01 8.78175199e-01
9.84662354e-01 8.00433874e-01 1.38118431e-01 -1.30219325e-01
1.54879069e+00 -6.94384098e-01 -7.55418301e-01 4.40155774e-01
1.30240902e-01 -1.42353916e+00 6.85581207e-01 2.53271133e-01
-8.78618777e-01 -6.52829349e-01 -7.79844522e-01 -6.37991503e-02
-1.13367105e+00 1.44842207e-01 2.30571136e-01 5.72752357e-01
-1.42482889e+00 1.99546397e-01 -8.22625160e-02 -5.73141277e-01
-3.53565753e-01 4.43144739e-01 -4.16986346e-01 -2.75584489e-01
-8.64623666e-01 7.93759942e-01 6.21854067e-01 -3.39448601e-01
-5.31244397e-01 -3.20349813e-01 -5.28932929e-01 -6.10490404e-02
9.56379548e-02 -1.78518295e-01 7.35923529e-01 -1.06518984e+00
-1.28475225e+00 1.16802812e+00 -3.38243514e-01 1.92972496e-02
-7.62049109e-02 2.15862423e-01 -4.15915519e-01 8.64785731e-01
3.14332038e-01 8.10930431e-01 7.84390330e-01 -1.51545751e+00
-7.61266351e-01 -5.14342904e-01 -1.77862585e-01 3.83464813e-01
-2.83635825e-01 4.24000919e-01 -1.02504504e+00 -5.08248866e-01
3.01273942e-01 -7.50198722e-01 -1.06558263e-01 -7.27154501e-03
-3.32067199e-02 -4.66018558e-01 1.24293005e+00 -4.79023606e-01
1.23934436e+00 -2.24370432e+00 -2.91135520e-01 8.24653029e-01
-2.76835918e-01 2.96030283e-01 -3.53829227e-02 7.49062657e-01
1.60186782e-01 3.89998078e-01 3.73147041e-01 4.54810262e-01
-3.60514998e-01 1.55544296e-01 2.37168819e-01 3.06818843e-01
-1.02654487e-01 6.43122494e-01 -5.21492302e-01 -1.13617575e+00
6.65405095e-01 7.31575906e-01 -1.96766987e-01 6.52262643e-02
1.47206306e-01 2.02071428e-01 -7.83293247e-01 9.21627343e-01
7.43158817e-01 -1.09724976e-01 -1.80110425e-01 -3.93323481e-01
-1.99291185e-01 -5.85458517e-01 -1.31280243e+00 1.34123862e+00
-2.59946197e-01 4.45726067e-01 -1.41821668e-01 -9.39088941e-01
1.03716397e+00 4.46925670e-01 9.53066230e-01 -1.18144310e+00
2.71225162e-02 2.41604298e-01 -5.23554683e-01 -8.77083004e-01
7.78796911e-01 4.77629900e-01 2.26166889e-01 4.93577838e-01
-1.43164635e-01 -1.05122857e-01 4.99432445e-01 1.19627506e-01
4.87450659e-01 -1.27766520e-01 4.21768576e-01 -3.34185928e-01
1.12692451e+00 2.46264890e-01 -3.70900840e-01 8.30565810e-01
-2.59985447e-01 5.84149361e-01 -2.52556622e-01 -5.08166492e-01
-1.06830013e+00 -7.75050223e-01 -3.27352077e-01 1.23920548e+00
6.80206239e-01 -9.84768793e-02 -5.66268861e-01 1.47771249e-02
-5.39553687e-02 2.10863743e-02 -4.83800262e-01 1.23234749e-01
-3.31580818e-01 -4.29725200e-01 -9.70057026e-02 -3.05247426e-01
6.29168510e-01 -1.28557813e+00 -8.48765492e-01 1.45761862e-01
-1.53403310e-02 -5.47577322e-01 -3.66797715e-01 -3.53116512e-01
-7.73287594e-01 -9.31661725e-01 -1.09884667e+00 -1.13476992e+00
8.84799957e-01 7.62644529e-01 1.05146396e+00 6.98015749e-01
-9.88272846e-01 1.06904006e+00 -9.48107719e-01 -2.92668104e-01
-1.36358351e-01 -2.34269630e-03 -5.36773145e-01 -4.80684340e-02
3.85491252e-01 9.57828388e-03 -1.16322792e+00 4.87627149e-01
-1.56535244e+00 -5.70755780e-01 6.26456857e-01 4.55541492e-01
9.89921570e-01 3.15564036e-01 1.11885846e-01 -4.97663260e-01
7.86256313e-01 -1.90528795e-01 -4.76421595e-01 9.09703493e-01
-8.11805427e-01 9.52135772e-03 -1.49092808e-01 -2.49994963e-01
-7.58201599e-01 6.37153834e-02 3.23080540e-01 -2.34192349e-02
-2.81284392e-01 6.21196926e-01 2.09818199e-01 -6.43581629e-01
6.97305441e-01 6.44017339e-01 7.74307102e-02 -4.73734528e-01
3.47752512e-01 8.91248763e-01 2.46576726e-01 -3.69527042e-01
3.96770328e-01 5.06491423e-01 1.21729843e-01 -9.43413079e-01
-1.44226149e-01 -1.33786595e+00 -7.28885293e-01 -4.48592573e-01
8.75620544e-01 -6.97796524e-01 -7.39647686e-01 3.80562872e-01
-8.64647269e-01 4.85244066e-01 1.89136386e-01 2.57441819e-01
-1.17885783e-01 4.91559088e-01 -1.90026999e-01 -9.48308289e-01
-6.22240305e-01 -1.11234832e+00 1.17257190e+00 6.02623165e-01
3.33463520e-01 -8.08432877e-01 3.97681119e-03 2.98770875e-01
7.59910882e-01 5.56588508e-02 9.29842234e-01 -5.36132514e-01
-6.60053790e-01 -6.26157939e-01 -4.52352524e-01 5.87115586e-02
5.00640750e-01 2.02019244e-01 -5.53674340e-01 -1.25928387e-01
-4.76463914e-01 3.81546021e-02 5.20237625e-01 3.56786251e-01
1.27418971e+00 -3.01347703e-01 -5.02506554e-01 9.38070863e-02
2.04187727e+00 6.79121256e-01 9.24906135e-01 6.93401098e-01
2.07839951e-01 5.28831005e-01 8.56476903e-01 4.63261336e-01
6.07143492e-02 8.62286389e-01 1.48850143e-01 -1.46512926e-01
-7.83288479e-02 3.07204157e-01 -2.69122005e-01 3.40960860e-01
-1.28266662e-01 -9.91883650e-02 -7.76703119e-01 5.08779287e-01
-1.45461690e+00 -9.39039469e-01 8.79967287e-02 2.25711632e+00
6.03815556e-01 -5.15225887e-01 1.31499961e-01 1.07219882e-01
7.53843725e-01 -2.56131560e-01 -1.22979321e-01 -4.37614739e-01
-5.88263758e-02 3.56518179e-01 7.37084985e-01 4.32206571e-01
-1.09104764e+00 8.31382155e-01 6.82870865e+00 1.28838074e+00
-1.40618718e+00 -9.71254483e-02 6.91634536e-01 4.81437296e-01
-2.75656879e-01 -6.05057813e-02 -5.62323153e-01 3.22784841e-01
4.89452541e-01 -3.86675060e-01 5.70122600e-01 8.72659862e-01
1.04676954e-01 -8.60838950e-01 -3.63804787e-01 1.17518735e+00
4.65923309e-01 -1.03445041e+00 4.88787025e-01 -6.14562631e-02
7.26439238e-01 -3.27350438e-01 6.87847584e-02 -1.59708545e-01
-4.34447408e-01 -7.63168633e-01 4.03445423e-01 8.35747242e-01
4.60358560e-01 -9.21761930e-01 6.80212855e-01 -9.79900584e-02
-1.31064439e+00 6.39255741e-04 -4.03347909e-01 4.89669532e-01
-3.38772565e-01 5.75888380e-02 -7.83097506e-01 6.22494400e-01
1.02391851e+00 8.80887210e-02 -1.18555105e+00 1.44476020e+00
4.17986244e-01 -1.35884807e-01 -2.87062556e-01 -2.93925315e-01
1.62733242e-01 -2.94290423e-01 1.58442408e-01 1.15199232e+00
4.03520435e-01 2.17589036e-01 1.99357942e-01 4.41466689e-01
3.55672091e-01 9.47416544e-01 -5.25105000e-01 8.37378204e-03
4.86654311e-01 1.42510688e+00 -1.12898350e+00 -5.27540982e-01
-9.43110809e-02 9.93880153e-01 -4.22462165e-01 3.65659028e-01
-8.23295787e-02 -6.87080741e-01 2.51600333e-03 1.30407721e-01
2.43478835e-01 -5.13000004e-02 2.87861973e-01 -4.42561120e-01
-1.63655430e-01 -7.03356862e-01 4.89330620e-01 -9.78302777e-01
-8.83981645e-01 4.95414197e-01 4.84312445e-01 -1.23912585e+00
-3.97976249e-01 -6.47067666e-01 -1.12243988e-01 9.45123434e-01
-1.41828668e+00 -1.23693120e+00 -4.07560021e-01 1.12431467e+00
3.86572301e-01 -2.53040642e-01 8.93910110e-01 4.92859811e-01
3.29219624e-02 2.26381496e-01 3.81984323e-01 -6.47084564e-02
6.73550785e-01 -1.10728943e+00 -6.57614291e-01 3.62046629e-01
1.43461525e-01 5.33542931e-01 5.79990089e-01 -4.31883484e-01
-1.44222331e+00 -4.76629376e-01 8.88324201e-01 -8.28363076e-02
1.25748500e-01 3.46529752e-01 -6.45926058e-01 -9.76004675e-02
2.76104689e-01 -4.68553007e-02 5.82052350e-01 -3.92091036e-01
-3.19965035e-02 -5.26330471e-01 -1.39410996e+00 3.56805533e-01
-4.71846163e-02 -6.64657295e-01 -1.31445929e-01 5.46940625e-01
2.52175599e-01 3.93993333e-02 -1.05234873e+00 2.18255252e-01
6.51312888e-01 -1.05393100e+00 1.42084622e+00 1.25388816e-01
-2.14074060e-01 -5.75364053e-01 -1.96568057e-01 -5.80028355e-01
-1.52466550e-01 -4.21478078e-02 6.80582821e-01 1.14241421e+00
1.96892351e-01 -3.39247167e-01 5.39226592e-01 6.86394095e-01
4.21571642e-01 -2.01611519e-01 -6.70627356e-01 -4.30025488e-01
-3.72147530e-01 6.36873841e-02 5.48179865e-01 5.46655536e-01
-1.61782935e-01 -1.73828334e-01 -7.28949830e-02 -6.30095676e-02
5.44712782e-01 2.00369045e-01 5.18232465e-01 -1.14752197e+00
1.36362419e-01 -5.19821584e-01 -7.92473495e-01 -3.05387646e-01
-5.03717721e-01 -5.31551361e-01 -1.68605506e-01 -1.58862293e+00
6.27476990e-01 -5.69662154e-01 -6.38986051e-01 2.35462680e-01
-7.89652467e-02 6.88525617e-01 3.10157746e-01 7.22323418e-01
-7.31066167e-01 -2.43420810e-01 1.31208348e+00 -2.82105803e-01
-2.46169105e-01 -1.24113292e-01 -3.20060968e-01 -2.14220900e-02
6.85074329e-01 -4.49068457e-01 -3.64612252e-01 -2.29751598e-02
1.73996776e-01 -4.80071791e-02 4.50631261e-01 -9.36953545e-01
4.25129771e-01 -1.67894587e-01 5.44037104e-01 -9.98697996e-01
2.66803563e-01 -1.10615671e+00 3.69666636e-01 4.34033930e-01
-3.47875476e-01 4.33720708e-01 4.92045470e-03 3.56688470e-01
-7.63373733e-01 -7.18041003e-01 6.52920127e-01 -5.11582553e-01
-1.34313250e+00 6.62453324e-02 -5.35667181e-01 -7.20705867e-01
1.24123979e+00 -4.12993520e-01 4.71286476e-03 -5.39840400e-01
-5.92252553e-01 -2.14073867e-01 3.79363686e-01 2.77135134e-01
7.02935994e-01 -1.42162633e+00 -3.72534811e-01 6.25079125e-02
5.51596344e-01 -5.54659128e-01 4.04019535e-01 4.60962832e-01
-1.22076321e+00 6.98863149e-01 -4.38444465e-01 -6.07619047e-01
-1.74497628e+00 9.80324388e-01 1.29871383e-01 -1.47747666e-01
-8.33072066e-02 4.34204370e-01 -3.02147061e-01 1.90562848e-02
7.68965483e-02 1.79814398e-01 -8.64869058e-01 2.98871249e-01
6.61038697e-01 3.43891233e-01 2.67011046e-01 -1.02349484e+00
-4.02119428e-01 1.15130270e+00 -6.79483637e-02 -2.76627719e-01
9.86312270e-01 -5.06746113e-01 -6.65696681e-01 1.41835675e-01
1.65613747e+00 -1.90810144e-01 -2.23049656e-01 -3.83442640e-01
2.57376701e-01 -7.80695140e-01 2.20369980e-01 -8.36295784e-01
-1.04123092e+00 3.91415358e-01 1.53270018e+00 5.07763982e-01
1.59733677e+00 1.39473692e-01 2.33240277e-01 4.73155588e-01
2.93410957e-01 -1.34345829e+00 2.84955233e-01 9.43557024e-02
8.82654548e-01 -1.37787879e+00 3.87928694e-01 -1.75162449e-01
-3.81828994e-01 1.26069546e+00 3.04133222e-02 -2.09532022e-01
1.14537096e+00 -2.91335285e-01 3.43043387e-01 -5.04192412e-01
-1.72308788e-01 -5.54514110e-01 7.30412781e-01 5.40239155e-01
4.58135515e-01 -8.63301307e-02 -7.73703933e-01 -4.50603157e-01
3.59688103e-01 5.20761125e-04 -6.57018572e-02 1.28384936e+00
-7.25470304e-01 -1.44309819e+00 -9.94102657e-01 3.03916842e-01
-7.44317532e-01 2.24686209e-02 -4.29964334e-01 7.89214551e-01
-3.84248979e-02 1.21200395e+00 1.22155264e-01 -1.10491753e-01
1.30178943e-01 -2.69329280e-01 4.57372785e-01 -2.08823636e-01
-5.96014380e-01 4.05445576e-01 -5.98844707e-01 -3.63245010e-01
-1.00798512e+00 -3.56230408e-01 -8.44936728e-01 -1.33163203e-02
-3.00348133e-01 4.36867595e-01 1.30952847e+00 6.40426934e-01
3.76215100e-01 -1.74894594e-02 7.97304988e-01 -9.00171697e-01
1.34659201e-01 -8.38934302e-01 -7.63534725e-01 6.96920931e-01
1.61386266e-01 -3.86262149e-01 -7.84818903e-02 4.06673551e-01] | [10.766709327697754, 0.09267441183328629] |
f1ac4b2f-1d3f-4262-934b-a7274e6e7908 | deep-texture-and-structure-aware-filtering | 1712.02893 | null | http://arxiv.org/abs/1712.02893v2 | http://arxiv.org/pdf/1712.02893v2.pdf | Deep Texture and Structure Aware Filtering Network for Image Smoothing | Image smoothing is a fundamental task in computer vision, that aims to retain
salient structures and remove insignificant textures. In this paper, we aim to
address the fundamental shortcomings of existing image smoothing methods, which
cannot properly distinguish textures and structures with similar low-level
appearance. While deep learning approaches have started to explore the
preservation of structure through image smoothing, existing work does not yet
properly address textures. To this end, we generate a large dataset by blending
natural textures with clean structure-only images, and then build a texture
prediction network (TPN) that predicts the location and magnitude of textures.
We then combine the TPN with a semantic structure prediction network (SPN) so
that the final texture and structure aware filtering network (TSAFN) is able to
identify the textures to remove ("texture-awareness") and the structures to
preserve ("structure-awareness"). The proposed model is easy to understand and
implement, and shows excellent performance on real images in the wild as well
as our generated dataset. | ['Nick Barnes', 'ShaoDi You', 'Kaiyue Lu'] | 2017-12-07 | deep-texture-and-structure-aware-filtering-1 | http://openaccess.thecvf.com/content_ECCV_2018/html/Kaiyue_Lu_Deep_Texture_and_ECCV_2018_paper.html | http://openaccess.thecvf.com/content_ECCV_2018/papers/Kaiyue_Lu_Deep_Texture_and_ECCV_2018_paper.pdf | eccv-2018-9 | ['image-smoothing'] | ['computer-vision'] | [ 7.71984398e-01 2.18936607e-01 1.84599414e-01 -4.19694483e-01
-2.87359208e-01 -1.27596125e-01 7.51058757e-01 1.53953522e-01
1.34114558e-02 3.67800564e-01 3.42022479e-01 -8.99441657e-04
-2.06277310e-03 -9.86616910e-01 -6.53892994e-01 -1.03031719e+00
2.26214454e-01 -6.16245084e-02 6.09492123e-01 -2.42679626e-01
5.26886344e-01 5.35210013e-01 -1.85377049e+00 5.90132713e-01
9.14564788e-01 1.20849180e+00 4.31735516e-01 4.65321451e-01
-3.52598101e-01 1.05633140e+00 -1.77079052e-01 -1.99144706e-01
1.84288010e-01 -1.87974215e-01 -7.44120538e-01 3.72494578e-01
9.68846619e-01 -1.48027480e-01 -1.41961679e-01 1.50616586e+00
3.07321876e-01 1.04104832e-01 4.92572546e-01 -7.36080766e-01
-9.43152010e-01 3.19865525e-01 -8.41147900e-01 1.76449399e-02
-2.55391985e-01 2.66775917e-02 5.74331403e-01 -8.23736072e-01
6.23447478e-01 1.46773994e+00 8.05616558e-01 5.70378661e-01
-1.35734308e+00 -1.82818979e-01 2.99178302e-01 9.80990529e-02
-1.02181840e+00 -5.78465283e-01 1.04922485e+00 -2.99722373e-01
4.89072889e-01 4.24163669e-01 5.04119933e-01 9.71507072e-01
4.78024334e-01 8.79874051e-01 1.32848930e+00 -2.50935465e-01
2.47686848e-01 -1.45922359e-02 4.59350199e-02 6.53662384e-01
1.16038345e-01 7.05740452e-02 -4.43150967e-01 2.47509509e-01
9.86391068e-01 1.08994760e-01 -2.07917690e-01 -4.87569898e-01
-1.07830584e+00 3.85782897e-01 4.09188479e-01 1.82706028e-01
-4.29112762e-01 -1.05604574e-01 3.49550277e-01 1.04639590e-01
8.72734725e-01 1.47308171e-01 -3.88238162e-01 3.61744463e-01
-9.33374107e-01 2.60696203e-01 4.61216390e-01 6.45395994e-01
1.07198358e+00 1.66731238e-01 -3.54344934e-01 1.12144876e+00
-1.40190823e-02 4.49913740e-01 2.18794703e-01 -8.73587251e-01
-1.08598813e-01 6.50363743e-01 -4.11301479e-02 -1.39756608e+00
-2.37012580e-01 -3.89747977e-01 -1.26885605e+00 6.74730420e-01
3.33561420e-01 1.96789220e-01 -1.12731087e+00 1.43984199e+00
3.15210342e-01 2.89716423e-01 -1.59534365e-01 9.56529140e-01
9.52749014e-01 4.95205313e-01 1.29452780e-01 1.79894015e-01
1.46864247e+00 -1.14451158e+00 -6.00595474e-01 -4.65474635e-01
5.67774512e-02 -1.11872840e+00 1.18894696e+00 4.52367455e-01
-1.11990106e+00 -8.25640857e-01 -8.22860479e-01 -3.93435031e-01
-3.64895582e-01 1.89434186e-01 7.29041338e-01 4.56869632e-01
-1.23428667e+00 8.37221920e-01 -7.90576637e-01 -4.13614929e-01
5.91281474e-01 9.04178545e-02 -4.58081424e-01 5.36013804e-02
-6.81386828e-01 7.42478251e-01 3.60234916e-01 3.88573140e-01
-8.06070030e-01 -7.12015808e-01 -9.36647296e-01 2.65235692e-01
4.86578733e-01 -6.59141302e-01 7.80395389e-01 -1.54348195e+00
-1.47019041e+00 9.33992386e-01 -4.11060393e-01 -3.54674250e-01
3.57568800e-01 -1.92347959e-01 -3.31721872e-01 2.72264588e-03
-1.09587787e-02 6.91023827e-01 1.27245533e+00 -1.73698747e+00
-9.51256454e-01 -2.73564517e-01 -2.79354066e-01 8.72197747e-02
-2.35116988e-01 -1.77813441e-01 -5.01152992e-01 -1.02983081e+00
3.08886111e-01 -4.29280907e-01 -2.27227151e-01 2.76374966e-01
-4.68031228e-01 1.05450548e-01 9.76278305e-01 -9.13848460e-01
9.62520540e-01 -2.22498918e+00 3.78373452e-02 2.16486976e-01
4.51099217e-01 4.10655558e-01 -4.76911783e-01 6.00067601e-02
-5.18471710e-02 -1.47452608e-01 -2.73564309e-01 -5.48608959e-01
-1.02096207e-01 1.99435085e-01 -5.28315425e-01 2.75572121e-01
4.32367474e-01 7.44140506e-01 -6.72085047e-01 -4.07782286e-01
5.21733999e-01 6.56414032e-01 -6.08804524e-01 2.90781111e-02
-4.30781662e-01 3.63934636e-01 -3.95212263e-01 7.74671018e-01
1.23180616e+00 -1.18618786e-01 1.63337573e-01 -6.34577990e-01
-3.85675281e-01 -3.41603346e-02 -1.17428505e+00 1.07186472e+00
-2.21470192e-01 5.52988768e-01 4.33644354e-01 -9.67038095e-01
1.19222224e+00 -2.48207927e-01 3.35494667e-01 -1.10515976e+00
9.76593569e-02 1.41048402e-01 -2.71086901e-01 -3.81750554e-01
6.43862009e-01 1.48571566e-01 3.88971835e-01 1.29268721e-01
-1.79358616e-01 7.66512938e-03 -2.72474699e-02 -1.18383564e-01
7.66206920e-01 1.39233977e-01 -7.11134216e-03 -6.70278430e-01
8.48143518e-01 -2.47833222e-01 6.58994853e-01 7.50755787e-01
-1.86696604e-01 8.22731256e-01 4.98193741e-01 -9.61273074e-01
-1.25110459e+00 -1.14519918e+00 -1.17156636e-02 1.21564555e+00
6.04216099e-01 -1.17401384e-01 -1.07212424e+00 -4.59671825e-01
-4.82547702e-03 3.27350676e-01 -9.80401158e-01 -1.13137931e-01
-6.27387047e-01 -4.74421978e-01 5.06968349e-02 2.66562313e-01
8.77793968e-01 -1.36887848e+00 -1.89602330e-01 1.55862838e-01
-2.50429034e-01 -8.51070940e-01 -5.05068839e-01 4.46369722e-02
-6.33044660e-01 -9.65387464e-01 -5.83779931e-01 -1.14753950e+00
9.07433629e-01 5.23919106e-01 1.09980750e+00 3.77542406e-01
-2.45548114e-01 -1.35867968e-01 -3.16811912e-02 -3.66713583e-01
-3.08876961e-01 -6.87742382e-02 -3.59428525e-01 6.04944587e-01
1.52527660e-01 -5.00456393e-01 -7.82924950e-01 2.85696715e-01
-1.09705353e+00 3.95155549e-01 7.19855607e-01 8.76405597e-01
1.03815258e+00 4.42589760e-01 2.34885275e-01 -1.08047307e+00
4.90954399e-01 8.10697973e-02 -6.08738720e-01 4.21099395e-01
-4.22874779e-01 1.53063545e-02 6.41512394e-01 -3.48778665e-01
-1.58413386e+00 1.78512976e-01 -3.53322476e-01 -1.52467817e-01
-4.24116671e-01 1.55726090e-01 -2.54477888e-01 -3.70773375e-01
3.44275206e-01 6.09110296e-01 1.88086703e-01 -8.74596119e-01
2.05103770e-01 2.21622601e-01 8.21458817e-01 -4.03757155e-01
6.42319679e-01 8.96150470e-01 -2.78400909e-02 -8.79294038e-01
-1.30240178e+00 -1.58012688e-01 -6.61082327e-01 -5.04971482e-02
8.25166881e-01 -7.04562962e-01 -6.27596617e-01 8.91508520e-01
-8.04293752e-01 -5.47400713e-01 -2.55023301e-01 -1.36233598e-01
-5.69064915e-01 6.95496202e-01 -5.28300166e-01 -6.26755476e-01
-4.89600271e-01 -1.05131221e+00 1.34050322e+00 3.61500204e-01
1.24604091e-01 -1.09437513e+00 -2.32809782e-01 1.13127418e-01
8.14358234e-01 2.96786249e-01 1.09881449e+00 2.12157980e-01
-7.68122733e-01 1.80529296e-01 -7.54912913e-01 4.69537944e-01
3.94276351e-01 1.45247474e-01 -1.03912544e+00 -1.20302677e-01
1.21718787e-01 8.33051205e-02 1.37618136e+00 8.16145360e-01
1.58970380e+00 -1.83745801e-01 -1.50647908e-01 8.67419720e-01
1.44751537e+00 -1.08699694e-01 1.15857947e+00 4.97247040e-01
7.53100574e-01 8.01081836e-01 3.73707235e-01 1.68948174e-01
2.43971363e-01 4.62777138e-01 4.85950440e-01 -8.70620608e-01
-6.49499893e-01 -3.29915643e-01 9.18068662e-02 4.68075931e-01
-6.80981800e-02 -3.49789262e-02 -4.95230585e-01 5.02554119e-01
-1.86356604e+00 -8.94580841e-01 -3.04715782e-01 1.87503338e+00
7.83404529e-01 3.02375019e-01 -2.43316427e-01 -1.01076692e-01
8.45985532e-01 3.87668520e-01 -4.25205439e-01 -3.05426061e-01
-5.93603134e-01 2.89287090e-01 5.74156702e-01 5.54034293e-01
-1.51021254e+00 1.29787779e+00 6.23927307e+00 1.20396197e+00
-1.43703818e+00 -2.72796452e-01 9.31375861e-01 3.78660798e-01
-2.72567809e-01 8.65027979e-02 -5.98418057e-01 5.20370185e-01
1.59405425e-01 1.80375859e-01 4.37109709e-01 8.66880238e-01
4.38874781e-01 -1.87310413e-01 -6.40434325e-01 8.64038229e-01
-3.06151602e-02 -1.57861888e+00 3.31873178e-01 -1.42640591e-01
7.86868811e-01 -2.23878771e-01 2.53127694e-01 -5.13773933e-02
5.61649725e-02 -1.04615402e+00 9.29738939e-01 1.01852548e+00
5.64266324e-01 -4.73552793e-01 7.43069112e-01 1.50455073e-01
-1.33339572e+00 -8.51617754e-02 -7.92165995e-01 1.17102258e-01
-2.55850405e-02 8.87297451e-01 -1.12874694e-01 2.53875434e-01
9.37192440e-01 9.25662458e-01 -9.23192143e-01 1.02079725e+00
-1.91416696e-01 4.63121027e-01 -2.65391986e-03 2.24554330e-01
6.30994812e-02 -4.20227200e-01 1.67940423e-01 1.08788753e+00
1.42346397e-01 -2.88508207e-01 2.72356927e-01 9.66632843e-01
1.78895295e-01 5.78942932e-02 -2.11810753e-01 1.90595567e-01
1.35184124e-01 1.45397091e+00 -1.10539496e+00 -5.45584977e-01
-2.68243194e-01 1.09764707e+00 4.59451318e-01 2.58253574e-01
-4.56018448e-01 -2.61419684e-01 7.81826615e-01 1.74918815e-01
4.29120332e-01 -2.61129886e-02 -8.53520453e-01 -1.20282531e+00
4.54484709e-02 -1.02361298e+00 5.52441403e-02 -7.68628359e-01
-1.63003016e+00 6.07938588e-01 -6.49596274e-01 -1.06912506e+00
6.62908614e-01 -6.10904813e-01 -7.78515339e-01 9.89782274e-01
-1.80132258e+00 -1.51762772e+00 -5.58423698e-01 5.74697196e-01
6.36270940e-01 1.22449510e-01 5.35283506e-01 3.82768363e-01
-4.67693955e-01 3.30225497e-01 9.65652466e-02 -7.49192387e-02
8.21608365e-01 -1.28117466e+00 6.08692050e-01 8.80074322e-01
-1.97740301e-01 6.08354390e-01 7.64733732e-01 -9.23519373e-01
-1.23183298e+00 -1.36352110e+00 8.31347823e-01 -2.09793091e-01
5.25257170e-01 -4.67619061e-01 -1.36175776e+00 2.16264188e-01
1.30496427e-01 1.57859158e-02 -1.52996048e-01 -9.04985592e-02
-3.04632187e-01 -2.68076152e-01 -9.99123335e-01 8.06837261e-01
1.06379926e+00 -4.56669360e-01 -4.01486754e-01 1.23326555e-01
3.24623674e-01 -2.48646691e-01 -6.34761095e-01 5.24235845e-01
6.29302442e-01 -1.28217828e+00 1.06996322e+00 -1.40189141e-01
5.86527288e-01 -4.53392893e-01 9.91297290e-02 -1.10353827e+00
-6.80889428e-01 -3.39323938e-01 2.72449911e-01 1.30158305e+00
8.87356848e-02 -6.35785162e-01 1.12313688e+00 2.84824252e-01
-2.18863472e-01 -4.88735825e-01 -5.21418512e-01 -5.23796320e-01
-6.96511865e-02 9.68167838e-03 7.15890288e-01 8.62295032e-01
-7.45246530e-01 -2.13597491e-02 -7.38517523e-01 2.86200821e-01
8.32305491e-01 3.59822124e-01 7.62806058e-01 -1.40215468e+00
4.63948697e-02 -7.62634695e-01 -1.27105817e-01 -1.04963577e+00
4.03594878e-03 -5.55904567e-01 2.30291247e-01 -1.53620780e+00
3.78293455e-01 -3.76234889e-01 -2.98189044e-01 5.51394701e-01
-4.39719796e-01 5.06804109e-01 -1.40277058e-01 6.84755296e-02
-5.26014805e-01 7.46298015e-01 1.47234190e+00 -2.59909183e-01
-3.06477156e-02 -1.58329293e-01 -8.87816846e-01 1.03851032e+00
7.03415811e-01 -6.68015257e-02 -3.05486709e-01 -5.03241658e-01
-1.63625449e-01 -4.97193187e-01 6.95146501e-01 -8.43036234e-01
5.56322709e-02 -3.11310589e-01 7.17528701e-01 -7.69714773e-01
1.43004404e-02 -6.96094573e-01 4.58438843e-02 1.87477097e-01
-2.81613469e-01 -4.22864765e-01 2.20848620e-01 4.95587468e-01
-3.89375269e-01 1.06690645e-01 1.11328471e+00 -4.37103920e-02
-1.15510094e+00 4.23961461e-01 -3.42191845e-01 -2.84482926e-01
6.72969639e-01 -4.48898613e-01 -6.31167293e-01 -1.34375215e-01
-7.36757755e-01 -5.13370745e-02 7.38371909e-01 4.56124425e-01
7.48856723e-01 -1.21557307e+00 -4.95662928e-01 6.54677331e-01
-1.29949441e-02 -8.47098082e-02 7.83553243e-01 7.04626322e-01
-7.19255567e-01 -1.16765000e-01 -4.80984122e-01 -5.10012388e-01
-1.26336813e+00 6.41352415e-01 4.12161857e-01 -9.20458660e-02
-9.48877096e-01 7.82414138e-01 8.21860552e-01 -4.05680329e-01
3.35816622e-01 -4.51152414e-01 -3.89339864e-01 -2.02620432e-01
7.29576766e-01 8.07133242e-02 1.49378449e-01 -5.65720916e-01
-1.88434925e-02 8.55510414e-01 -9.77695212e-02 5.49200594e-01
1.56069267e+00 -5.07649481e-01 -5.99372923e-01 -1.83461651e-01
8.82176518e-01 1.97544336e-01 -1.77778673e+00 -4.24609959e-01
9.93190482e-02 -7.08992779e-01 3.16433161e-01 -6.25755191e-01
-1.36439717e+00 9.25536513e-01 8.27349782e-01 3.74984235e-01
1.39832807e+00 -2.72396058e-01 9.23862457e-01 1.09769970e-01
9.69246253e-02 -1.21176994e+00 9.58338380e-02 6.34623230e-01
9.45699334e-01 -1.18568075e+00 -9.36068818e-02 -9.48233247e-01
-5.85771799e-01 1.13326418e+00 6.93067014e-01 -3.62241864e-01
7.90073752e-01 3.63490850e-01 1.34279236e-01 -3.08234692e-01
-7.12616861e-01 -4.63337332e-01 5.07941246e-01 8.66031706e-01
2.04031929e-01 1.29741346e-02 4.80863601e-02 2.89450914e-01
-9.30761471e-02 -1.13514766e-01 3.53417993e-01 6.83279097e-01
-8.27018619e-01 -1.11102641e+00 -3.76370162e-01 4.60902005e-01
-4.04512703e-01 -2.76317686e-01 -4.38234359e-01 3.61372530e-01
4.01906669e-01 7.43275881e-01 2.34808534e-01 -2.50719905e-01
3.26481283e-01 -3.26975316e-01 1.98293880e-01 -3.46602947e-01
-3.74329686e-01 3.87487352e-01 -4.66260687e-02 -6.89324379e-01
-4.59946632e-01 -3.49006861e-01 -7.18411684e-01 -4.69169647e-01
1.79438312e-02 -3.72797161e-01 4.24742997e-01 7.35547006e-01
3.22811514e-01 5.97612321e-01 4.79909062e-01 -1.07326329e+00
-1.08296201e-01 -6.49479210e-01 -9.08463061e-01 8.18548083e-01
4.51237410e-01 -6.13725007e-01 -2.33294755e-01 4.88058537e-01] | [10.89677906036377, -1.4455939531326294] |
4bddfb93-98ac-4fad-8474-14afabe5278b | language-model-cascades | 2207.10342 | null | https://arxiv.org/abs/2207.10342v2 | https://arxiv.org/pdf/2207.10342v2.pdf | Language Model Cascades | Prompted models have demonstrated impressive few-shot learning abilities. Repeated interactions at test-time with a single model, or the composition of multiple models together, further expands capabilities. These compositions are probabilistic models, and may be expressed in the language of graphical models with random variables whose values are complex data types such as strings. Cases with control flow and dynamic structure require techniques from probabilistic programming, which allow implementing disparate model structures and inference strategies in a unified language. We formalize several existing techniques from this perspective, including scratchpads / chain of thought, verifiers, STaR, selection-inference, and tool use. We refer to the resulting programs as language model cascades. | ['Charles Sutton', 'Kevin Murphy', 'Jascha Sohl-Dickstein', 'Rif A. Saurous', 'Henryk Michalewski', 'Yuhuai Wu', 'Raphael Gontijo Lopes', 'David Bieber', 'Jacob Austin', 'Aitor Lewkowycz', 'Winnie Xu', 'David Dohan'] | 2022-07-21 | null | null | null | null | ['probabilistic-programming'] | ['methodology'] | [ 1.25567600e-01 -4.16844748e-02 -5.42791724e-01 -1.19215488e-01
-6.74122810e-01 -8.79544139e-01 7.05312729e-01 -2.47340128e-02
2.08378360e-01 6.06973767e-01 -1.57247156e-01 -9.02449906e-01
-4.05550331e-01 -1.21960557e+00 -6.76136672e-01 -1.59240395e-01
-2.20665127e-01 6.69001281e-01 6.25350296e-01 -1.37821913e-01
4.87734795e-01 6.13182001e-02 -1.57376993e+00 2.40023226e-01
6.28715456e-01 2.52121091e-01 -1.99362040e-02 1.21110487e+00
-6.85486317e-01 1.35644126e+00 -7.08674371e-01 -5.76231062e-01
-3.62782151e-01 -1.32449865e-01 -4.46037740e-01 -4.05430406e-01
-1.32691264e-01 -3.15118372e-01 -1.88593000e-01 9.87723410e-01
1.85839429e-01 -2.32976489e-02 4.09942687e-01 -1.65618646e+00
-6.27091467e-01 1.33928168e+00 -5.05035400e-01 -6.02417216e-02
7.82323956e-01 4.21284407e-01 7.78262973e-01 -6.48028076e-01
4.97316808e-01 1.40919363e+00 7.52039075e-01 4.76017267e-01
-1.94958317e+00 -5.18972635e-01 4.63533513e-02 4.06328999e-02
-1.37940288e+00 -1.85286492e-01 5.42024493e-01 -8.46750975e-01
1.22448468e+00 3.19590777e-01 6.65911496e-01 1.34260821e+00
3.86676550e-01 9.33625102e-01 9.99420047e-01 -8.66019845e-01
5.25822163e-01 2.32442692e-01 7.90538847e-01 9.93652582e-01
3.68077874e-01 3.81036848e-01 -6.27998114e-01 -7.76797354e-01
7.86530972e-01 1.59842581e-01 3.13325703e-01 -6.98224127e-01
-8.60436261e-01 8.54197264e-01 -5.64706624e-01 6.99643185e-03
2.98618168e-01 5.43837667e-01 2.69127786e-01 3.59740019e-01
-2.13845640e-01 6.82619691e-01 -4.84587550e-01 -7.41591156e-01
-8.40789139e-01 4.39964771e-01 1.33920503e+00 1.51781952e+00
5.01366019e-01 1.91495538e-01 -2.98614770e-01 4.13486898e-01
4.31783438e-01 3.65627706e-01 2.43868545e-01 -7.88396597e-01
2.39668205e-01 5.44814706e-01 3.67406279e-01 -3.94229650e-01
-4.38084304e-02 -5.12828901e-02 -2.08110034e-01 5.14439702e-01
2.14771256e-01 -4.57885504e-01 -9.78570282e-01 1.62696743e+00
-9.99594666e-03 3.05055350e-01 -1.73721351e-02 -7.49751627e-02
3.24956149e-01 5.86209714e-01 2.54960775e-01 6.50607422e-02
1.05900168e+00 -7.16545105e-01 -5.88320911e-01 -1.82166800e-01
7.68548071e-01 -3.41776848e-01 1.17691064e+00 7.16650248e-01
-1.26637900e+00 -3.88065875e-01 -1.26068544e+00 2.33230725e-01
-4.58989710e-01 -4.76901680e-01 8.85965347e-01 1.00868344e+00
-1.09261680e+00 6.53760016e-01 -1.11458874e+00 1.41835466e-01
1.25113711e-01 3.77315611e-01 2.49428555e-01 -3.69501524e-02
-9.92512107e-01 7.50079632e-01 1.30179003e-01 -3.18585634e-01
-1.27615893e+00 -8.46004486e-01 -1.03413188e+00 2.79621065e-01
5.71996689e-01 -6.22935593e-01 1.73078060e+00 -3.36066633e-01
-1.54707062e+00 3.36746544e-01 -2.41983771e-01 -2.58451998e-01
4.11142468e-01 -2.69466966e-01 -3.20650667e-01 -6.00029886e-01
-1.68440625e-01 -1.17426127e-01 7.21460342e-01 -1.24046731e+00
-4.50072438e-01 -5.03449887e-02 3.41494709e-01 -5.58176160e-01
2.88405746e-01 5.26596665e-01 -1.94061846e-01 -1.57726645e-01
-2.50270784e-01 -6.76471710e-01 -5.37392557e-01 -3.92911196e-01
-6.20054483e-01 -3.41604412e-01 4.07271385e-01 -2.74725080e-01
1.90498126e+00 -2.03032231e+00 2.52868414e-01 3.00067276e-01
2.22233176e-01 -1.88004375e-01 3.62381265e-02 7.34637737e-01
-1.55602455e-01 7.08083868e-01 1.33780874e-02 9.41592380e-02
4.97345865e-01 8.36725757e-02 -3.41156334e-01 -4.31686863e-02
3.90070409e-01 1.07989430e+00 -9.93809104e-01 -6.57908618e-01
3.28066438e-01 -6.62568435e-02 -5.96335173e-01 3.15599203e-01
-7.29874790e-01 -3.24017853e-01 -6.50635064e-01 9.39686298e-01
2.94409037e-01 -4.15136814e-02 3.86363119e-01 6.19061887e-01
-2.32989907e-01 8.42685029e-02 -1.36858201e+00 1.25028896e+00
-5.80578804e-01 3.73443156e-01 -2.62601495e-01 -4.18013245e-01
7.46231914e-01 1.47076666e-01 -2.05506667e-01 5.25201485e-02
-2.47587204e-01 -1.90314859e-01 2.22236335e-01 -7.70804226e-01
3.21910679e-01 -4.17193435e-02 -6.69749439e-01 7.45385051e-01
1.58310443e-01 -6.47646427e-01 5.13696134e-01 3.10481787e-01
1.58241034e+00 4.47223395e-01 4.79796171e-01 1.80286765e-01
-2.89655417e-01 1.14368707e-01 3.42600048e-01 1.64435887e+00
2.40781903e-01 2.98737496e-01 1.22929490e+00 -1.06924042e-01
-1.04164445e+00 -1.53646028e+00 8.73434693e-02 1.44599605e+00
-2.25356787e-01 -8.42646301e-01 -3.39812636e-01 -5.08445382e-01
7.37719014e-02 1.42751157e+00 -4.81232315e-01 -2.50392467e-01
-1.99087784e-01 -3.98373663e-01 5.60895920e-01 9.37793136e-01
-3.72535467e-01 -8.58535409e-01 -5.67870200e-01 4.60735172e-01
6.17869735e-01 -3.43366921e-01 6.41630739e-02 6.32102489e-01
-7.37415850e-01 -1.01746154e+00 -1.61970422e-01 -5.58544517e-01
1.89823791e-01 -1.28614545e-01 1.46472418e+00 -7.59938359e-03
-3.76144916e-01 5.26960254e-01 -1.18886136e-01 -6.31728232e-01
-6.22321606e-01 -3.46709341e-01 -2.08063290e-01 -8.02901089e-01
4.76459831e-01 -7.75671124e-01 3.37208509e-01 7.34804869e-02
-6.87619090e-01 -3.14015709e-02 3.95426154e-01 1.15561020e+00
2.59772688e-01 1.91650793e-01 4.32520686e-03 -1.24146307e+00
7.85592675e-01 -7.33047962e-01 -8.83531690e-01 7.60681272e-01
-4.36748236e-01 2.93847293e-01 3.19642395e-01 -9.20699239e-01
-1.03348517e+00 -1.87018365e-01 2.74906397e-01 -4.26675677e-01
-2.34500423e-01 9.65914547e-01 -1.80277824e-01 1.47076577e-01
7.75353909e-01 4.38557714e-01 -1.79856136e-01 -6.94782585e-02
3.98763448e-01 2.87892818e-01 1.86163813e-01 -1.21396053e+00
7.58863151e-01 -2.86920309e-01 -1.46508917e-01 -4.14181113e-01
-2.47347996e-01 7.31415898e-02 -3.72383833e-01 2.87601277e-02
4.64612275e-01 -6.18223071e-01 -6.56422198e-01 4.43513483e-01
-1.06050920e+00 -7.26903915e-01 -3.58947992e-01 1.23355217e-01
-7.26563454e-01 -9.47125349e-03 -6.48366511e-01 -1.41018200e+00
2.62154132e-01 -1.33541846e+00 7.13725328e-01 3.60850573e-01
-7.84205377e-01 -9.62107599e-01 2.13287830e-01 -2.97427416e-01
3.99791598e-01 -2.18003392e-02 1.69961023e+00 -1.05538082e+00
-6.97842896e-01 -7.32380450e-01 4.20783579e-01 2.29567215e-02
-3.56375873e-01 7.31619895e-01 -3.44308048e-01 2.73893118e-01
-2.15021521e-01 -2.86738425e-01 7.88160563e-02 2.42397115e-01
1.33626044e+00 -3.77700254e-02 -6.47068381e-01 6.45274669e-02
1.39461207e+00 4.87072498e-01 5.90843201e-01 -2.76098363e-02
4.15297210e-01 3.89941067e-01 1.94957241e-01 6.88993454e-01
1.63134262e-01 3.51615310e-01 -1.14282548e-01 4.62338954e-01
2.66100317e-01 -6.58785105e-01 4.94578481e-01 5.14289320e-01
1.63585603e-01 -1.16201431e-01 -1.55227292e+00 4.46129411e-01
-1.87164044e+00 -1.31177819e+00 -2.63140678e-01 2.19261384e+00
1.04409695e+00 6.17615700e-01 -8.33180100e-02 -1.16773389e-01
7.30830133e-01 -1.92028612e-01 -4.15911674e-01 -5.11262894e-01
3.99114698e-01 7.13144004e-01 1.77450672e-01 6.03107691e-01
-5.95635533e-01 7.70977676e-01 8.43692017e+00 8.93190742e-01
-3.84535402e-01 4.55783494e-02 1.86379075e-01 -1.93576083e-01
-7.19575286e-01 4.74733621e-01 -8.83897662e-01 4.51874435e-01
1.15803862e+00 -8.15045178e-01 5.64243972e-01 1.33084655e+00
-1.65651277e-01 -3.77624273e-01 -1.58380818e+00 5.48917294e-01
-2.26765648e-01 -1.33981597e+00 -6.56897575e-02 -2.25802451e-01
8.54821444e-01 -2.89584726e-01 4.05381694e-02 1.02061212e+00
1.42332006e+00 -1.13794279e+00 8.41397583e-01 7.96261907e-01
4.48051721e-01 -6.66828632e-01 4.88766402e-01 3.39245915e-01
-8.42324376e-01 -3.30766916e-01 2.39919145e-02 -1.97029620e-01
-1.76850371e-02 1.78967118e-01 -7.31650352e-01 1.35979101e-01
3.19079071e-01 1.66044891e-01 -5.27929127e-01 1.22133005e+00
-4.78291065e-01 8.86984169e-01 -1.85689911e-01 -6.48813844e-01
-2.41818294e-01 -7.81340897e-02 5.31458795e-01 1.17724133e+00
2.88591266e-01 -2.91838735e-01 3.35771322e-01 1.31063056e+00
6.30622208e-01 -5.60635090e-01 -9.28142786e-01 -3.88534516e-01
7.96597958e-01 8.27126205e-01 -4.57688570e-01 -6.79576099e-01
-6.14152610e-01 -3.99661884e-02 5.79140522e-02 5.68527699e-01
-9.86061096e-01 -5.51487565e-01 5.18909037e-01 1.91002384e-01
1.95796892e-01 -5.06655931e-01 -5.65580845e-01 -9.94040310e-01
-2.71052033e-01 -1.08776999e+00 2.16092229e-01 -9.37065423e-01
-1.21517134e+00 -3.66477519e-02 5.55967867e-01 -7.36952066e-01
-6.69361472e-01 -5.70800304e-01 -1.19207168e+00 9.22084987e-01
-4.12875146e-01 -9.36884999e-01 1.32705018e-01 2.90742725e-01
6.32951260e-01 -1.52799696e-01 1.06026125e+00 -4.29062843e-01
-6.71012700e-01 5.62106550e-01 -6.50634468e-02 -1.59262404e-01
3.17577980e-02 -1.50809884e+00 5.78973174e-01 8.12143326e-01
-7.50931278e-02 1.21407735e+00 8.53479922e-01 -7.74297357e-01
-2.00764108e+00 -5.52846849e-01 4.72184688e-01 -8.84790123e-01
1.35789275e+00 -5.80667675e-01 -8.70796442e-01 9.75475550e-01
1.89608859e-03 -4.61823314e-01 8.60871375e-01 9.05604661e-01
-5.17200947e-01 4.47344959e-01 -7.78630376e-01 1.02769339e+00
8.84348691e-01 -7.64543533e-01 -7.90588498e-01 1.72207534e-01
7.70331323e-01 -2.40643725e-01 -7.40852475e-01 4.88206223e-02
7.86578298e-01 -7.76095688e-01 7.46130586e-01 -1.14952040e+00
5.38488090e-01 -2.12870687e-01 -1.83628291e-01 -9.80811417e-01
-3.97519886e-01 -1.06360793e+00 -7.41020203e-01 1.33156240e+00
6.93522453e-01 -4.88394499e-01 6.11405194e-01 1.15482223e+00
8.04069042e-02 -5.74657798e-01 -5.98497272e-01 -7.56042421e-01
1.21753998e-01 -9.90324795e-01 6.56996071e-01 6.55089855e-01
7.44855583e-01 2.00252473e-01 -1.51936412e-01 -6.62692562e-02
5.83057225e-01 9.77447033e-02 8.89480114e-01 -1.28638268e+00
-1.02420115e+00 -6.29381061e-01 -3.79933476e-01 -5.27132511e-01
2.72669345e-01 -6.33888781e-01 2.41996378e-01 -1.20969617e+00
4.09682512e-01 -5.23079574e-01 -1.61884665e-01 5.58495879e-01
-1.44615233e-01 -8.15169632e-01 -7.30055049e-02 -2.34160304e-01
-6.15615904e-01 4.20819335e-02 6.49454236e-01 -2.18187839e-01
-4.22821999e-01 3.53694081e-01 -6.55217648e-01 7.67158747e-01
6.48118496e-01 -5.18921375e-01 -5.87979794e-01 -2.03489080e-01
5.47758043e-01 6.27187848e-01 3.53225529e-01 -1.10723543e+00
4.04867619e-01 -5.87137163e-01 3.00476253e-01 -2.43519902e-01
2.17548922e-01 -2.54945129e-01 5.46262622e-01 4.26389277e-01
-6.43876493e-01 8.02164674e-02 2.18666971e-01 5.69485903e-01
3.55132930e-02 -8.46463442e-01 2.08273575e-01 -2.51797676e-01
-7.38124192e-01 -1.48806706e-01 -7.63369262e-01 -1.46984667e-01
1.20950282e+00 -2.52489358e-01 -3.01102459e-01 -1.97380230e-01
-1.18240047e+00 3.48081499e-01 5.25243163e-01 4.54966307e-01
5.69170654e-01 -7.99277008e-01 -3.42988044e-01 2.01508462e-01
1.18729524e-01 -1.58689424e-01 1.51471779e-01 6.23956323e-01
-1.38002619e-01 -1.10573359e-01 2.29807454e-03 -5.01329839e-01
-1.12181783e+00 8.65069568e-01 1.48713380e-01 -3.93223286e-01
-9.60749760e-02 8.74638379e-01 -9.26649421e-02 -4.41080302e-01
1.45953879e-01 -3.48291159e-01 2.74066746e-01 -4.88785282e-02
7.58890331e-01 2.97847569e-01 -4.17628109e-01 6.61925912e-01
-2.48706937e-01 6.08216189e-02 -1.61308885e-01 -4.01481748e-01
1.19145095e+00 2.41025090e-01 -2.70040125e-01 1.26478493e+00
3.02579552e-01 1.10722564e-01 -1.04924297e+00 -1.06233358e-03
2.91373193e-01 -2.48470306e-01 -1.98611379e-01 -9.85471368e-01
-1.06873019e-02 8.57735336e-01 1.84481278e-01 5.59964538e-01
4.27333593e-01 2.11717546e-01 -2.59071887e-01 3.92682344e-01
8.99452686e-01 -7.21988201e-01 -1.18563287e-01 5.84378839e-01
4.54267532e-01 -7.47890234e-01 -2.37606883e-01 -3.58026266e-01
-4.61428672e-01 1.22686827e+00 7.75900126e-01 -1.66259870e-01
6.74910367e-01 1.20232868e+00 -5.74444294e-01 2.57740486e-02
-1.43810678e+00 -7.34373704e-02 -3.59430492e-01 9.46703970e-01
3.95798564e-01 3.31781417e-01 1.30129457e-01 1.25786197e+00
-1.14987358e-01 5.71091115e-01 7.53591001e-01 1.60431516e+00
-2.91388422e-01 -1.13380909e+00 -4.42309290e-01 7.48665333e-01
-2.19631359e-01 -3.72987032e-01 -6.20907471e-02 9.36753690e-01
-1.24739252e-01 9.52954888e-01 1.06307715e-01 -7.39827216e-01
2.55499929e-01 6.26154840e-01 7.61764765e-01 -1.09713185e+00
-5.01206100e-01 -3.11925918e-01 2.40862235e-01 -4.03535664e-01
4.42096472e-01 -7.26152658e-01 -8.58782351e-01 -3.13372165e-01
-4.48430926e-01 1.27135813e-01 3.68086755e-01 7.70855606e-01
1.58669814e-01 6.50696576e-01 3.59997749e-01 -5.69712222e-01
-9.79944468e-01 -8.66121113e-01 -5.33724368e-01 -4.09131110e-01
1.43358828e-02 -7.10423887e-01 -1.52432635e-01 -4.83378954e-03] | [8.47161865234375, 6.867863178253174] |
df377099-0a5e-4148-9aa1-1d6144e3cd13 | the-npu-aslp-system-for-audio-visual-speech | 2303.06341 | null | https://arxiv.org/abs/2303.06341v1 | https://arxiv.org/pdf/2303.06341v1.pdf | The NPU-ASLP System for Audio-Visual Speech Recognition in MISP 2022 Challenge | This paper describes our NPU-ASLP system for the Audio-Visual Diarization and Recognition (AVDR) task in the Multi-modal Information based Speech Processing (MISP) 2022 Challenge. Specifically, the weighted prediction error (WPE) and guided source separation (GSS) techniques are used to reduce reverberation and generate clean signals for each single speaker first. Then, we explore the effectiveness of Branchformer and E-Branchformer based ASR systems. To better make use of the visual modality, a cross-attention based multi-modal fusion module is proposed, which explicitly learns the contextual relationship between different modalities. Experiments show that our system achieves a concatenated minimum-permutation character error rate (cpCER) of 28.13\% and 31.21\% on the Dev and Eval set, and obtains second place in the challenge. | ['Peikun Chen', 'Ao Zhang', 'Bingshen Mu', 'He Wang', 'Pengcheng Guo'] | 2023-03-11 | null | null | null | null | ['audio-visual-speech-recognition'] | ['speech'] | [ 2.04079881e-01 -7.20832497e-02 2.89046258e-01 -2.16607735e-01
-1.56918800e+00 -3.45740348e-01 6.97732687e-01 -2.01284438e-01
-2.20909640e-01 3.23985219e-01 7.43297279e-01 -2.99557328e-01
2.08304301e-01 8.19893647e-03 -5.94394982e-01 -8.36693883e-01
2.09388986e-01 -9.64339525e-02 -5.12012169e-02 -1.07161656e-01
-5.57395443e-02 2.79312015e-01 -1.53426802e+00 6.65873706e-01
8.20827007e-01 1.03173792e+00 4.00128543e-01 1.07260180e+00
1.19637765e-01 8.15378129e-01 -7.62236714e-01 -2.23993629e-01
5.26397824e-02 -4.30411696e-01 -3.27598780e-01 -1.70685336e-01
3.98347706e-01 -1.30092084e-01 -5.45523286e-01 8.68260980e-01
1.34339893e+00 5.49672544e-01 7.20479906e-01 -1.17572343e+00
-5.90201914e-01 8.93268466e-01 -7.98672795e-01 4.16444153e-01
4.86709088e-01 9.87970829e-02 8.76792014e-01 -1.53182018e+00
2.40958199e-01 1.47087109e+00 3.99130613e-01 6.72385275e-01
-1.03819370e+00 -7.47243226e-01 1.56565711e-01 6.75416529e-01
-1.57427704e+00 -1.26843309e+00 7.60436177e-01 -1.89993814e-01
1.35363972e+00 5.21574795e-01 6.83284774e-02 1.35206747e+00
-3.21377456e-01 1.07104731e+00 8.65214050e-01 -6.21630728e-01
8.41443241e-02 -9.18602105e-03 1.65358514e-01 1.71206325e-01
-5.18768728e-01 2.79384255e-01 -1.12749016e+00 1.73109040e-01
1.78581268e-01 -5.70648670e-01 -6.02144063e-01 4.76471603e-01
-1.10160875e+00 3.90967131e-01 1.87240854e-01 4.09886330e-01
-4.09109563e-01 7.02356398e-02 1.60010040e-01 5.39510250e-02
4.75613534e-01 1.61746159e-01 -3.20213884e-01 -1.95922762e-01
-1.15999722e+00 -1.03358082e-01 4.80123907e-01 8.96206617e-01
3.97854783e-02 6.80392861e-01 -8.73291731e-01 1.43015170e+00
8.16158891e-01 1.00314128e+00 3.58583152e-01 -8.23124230e-01
7.30605841e-01 -3.53826433e-01 -1.20417237e-01 -8.35271776e-01
-7.88832530e-02 -4.57875580e-01 -9.00839448e-01 4.39638346e-02
-1.88463077e-01 -2.97487557e-01 -1.24701715e+00 1.80142820e+00
2.46859059e-01 5.55727780e-01 2.60267794e-01 9.66207743e-01
1.23424280e+00 1.24050426e+00 -7.23804464e-04 -5.21260381e-01
1.19796503e+00 -1.03375852e+00 -9.71867085e-01 -2.55844086e-01
-1.10063873e-01 -9.15462613e-01 6.13133848e-01 5.47010481e-01
-1.29541254e+00 -7.75445223e-01 -1.04531050e+00 -4.65816585e-03
1.23346061e-01 2.03808889e-01 -3.18915308e-01 6.94686115e-01
-1.15942788e+00 2.87816655e-02 -7.04433918e-01 1.12268865e-01
2.33085483e-01 -1.58522964e-01 -8.24584886e-02 -6.14954643e-02
-1.10794759e+00 7.24586427e-01 2.46715061e-02 1.67715728e-01
-1.28260362e+00 -8.27817440e-01 -9.49806273e-01 1.38880461e-01
9.74344388e-02 -2.71597743e-01 1.14009607e+00 -7.31876373e-01
-1.75565934e+00 4.24903482e-01 -6.16005301e-01 -5.78460991e-01
-1.15031740e-02 -3.61958355e-01 -8.54040444e-01 2.17533782e-01
-3.35575700e-01 6.77169085e-01 1.03527379e+00 -1.46146786e+00
-4.92319793e-01 -1.66112244e-01 -6.53690279e-01 5.87466240e-01
-1.54276982e-01 5.17711461e-01 -6.80559874e-01 -1.04868734e+00
1.20272554e-01 -6.13637865e-01 1.90325856e-01 -6.75607085e-01
-7.09456384e-01 -1.94173828e-01 5.14537394e-01 -1.47552478e+00
1.30083632e+00 -2.65733528e+00 3.72246087e-01 2.49592453e-01
-3.50625627e-02 5.61425745e-01 -6.24798596e-01 2.16746360e-01
-2.12492704e-01 -1.26760244e-01 1.37568056e-03 -9.05913472e-01
1.93910420e-01 -2.63853937e-01 -4.47499067e-01 1.24040030e-01
5.86093627e-02 5.45160770e-01 -4.06116754e-01 -2.59148091e-01
2.08518729e-01 1.10395491e+00 -5.32002211e-01 3.07499915e-01
1.97610676e-01 3.11968118e-01 2.83874691e-01 4.81087029e-01
7.69832611e-01 3.45822752e-01 -1.03730768e-01 -1.77558213e-01
-1.59220546e-02 6.06999993e-01 -1.28116393e+00 1.69629323e+00
-5.85207105e-01 8.36025536e-01 4.47850019e-01 -5.24606586e-01
8.27155530e-01 8.19179475e-01 9.12197456e-02 -8.66534472e-01
5.68302460e-02 6.63774237e-02 -8.74669179e-02 -2.24163815e-01
3.98619086e-01 -1.07972976e-02 2.41292387e-01 9.51732602e-03
3.80893499e-01 -3.30046862e-02 -1.12438850e-01 2.37829328e-01
9.36147749e-01 -2.15042755e-01 -6.98841140e-02 9.30004194e-02
6.27025187e-01 -7.97058880e-01 6.12371743e-01 6.11606658e-01
-2.36914232e-01 1.10018897e+00 2.26814244e-02 4.76365000e-01
-7.41146624e-01 -1.39008117e+00 1.00019075e-01 1.12872458e+00
-1.04870781e-01 -5.13466060e-01 -6.78699017e-01 -2.02998191e-01
-3.99841219e-01 1.21989703e+00 -1.22442149e-01 -7.91346803e-02
-4.00258541e-01 -5.11021733e-01 7.36627519e-01 5.56409180e-01
2.43543163e-01 -8.90505493e-01 4.48581800e-02 6.10108338e-02
-6.87816918e-01 -1.12950075e+00 -7.44517446e-01 -2.53761690e-02
1.99676659e-02 -4.70831275e-01 -9.32896614e-01 -6.47432566e-01
1.91441149e-01 3.94753933e-01 6.76830590e-01 -5.43656707e-01
1.38307512e-01 5.61303675e-01 -4.74256426e-01 -4.09449905e-01
-5.23289979e-01 -4.27693039e-01 2.76208818e-01 3.91422987e-01
1.16630271e-01 -4.81479883e-01 -3.64102006e-01 1.41135603e-01
-3.89727533e-01 1.96506649e-01 4.63056743e-01 9.56719697e-01
6.49715006e-01 -5.66195808e-02 6.28082812e-01 -2.22940519e-02
6.72236800e-01 -2.91466832e-01 -3.19810301e-01 3.28367800e-01
-3.18932444e-01 -2.95554340e-01 2.74053931e-01 -5.88040471e-01
-1.62141466e+00 -1.67682916e-01 -4.83035147e-01 -7.34299123e-01
-1.98082894e-01 4.29014593e-01 -7.48639286e-01 4.04600322e-01
5.50894499e-01 5.45867503e-01 -4.86759990e-01 -5.85234463e-01
5.23531914e-01 1.25032222e+00 7.91127205e-01 -1.77688599e-01
5.30680776e-01 -7.86861032e-02 -5.52282333e-01 -1.26487494e+00
-4.31415021e-01 -4.53163534e-01 1.94791015e-02 -1.87969446e-01
9.28936303e-01 -1.43540537e+00 -5.24371862e-01 6.30205572e-01
-1.41336143e+00 -2.01510891e-01 -8.98565426e-02 6.71824396e-01
-2.53905922e-01 4.42122996e-01 -5.87061644e-01 -1.30921340e+00
-5.87644160e-01 -1.22723591e+00 9.55022097e-01 1.50138259e-01
-4.10026014e-02 -2.99504936e-01 9.72183347e-02 7.85887301e-01
5.75383306e-01 -3.24039489e-01 3.85820478e-01 -8.15538466e-01
-3.37962985e-01 1.95740044e-01 -7.19008297e-02 7.47405291e-01
-2.13026181e-01 -2.56085485e-01 -1.58292818e+00 -3.45489651e-01
3.25435512e-02 1.14707559e-01 9.35232878e-01 4.63685125e-01
7.89322495e-01 -3.49166989e-01 -9.83868688e-02 5.43350697e-01
8.69248211e-01 6.08946562e-01 8.08123052e-01 -3.70948195e-01
7.80335009e-01 4.41637367e-01 1.86705440e-01 5.33830047e-01
3.77207011e-01 9.22742665e-01 1.27458140e-01 6.88522905e-02
-8.45208526e-01 -1.27768010e-01 7.14614809e-01 1.23616600e+00
2.85921782e-01 -6.72287524e-01 -9.54317272e-01 7.08678663e-01
-1.45316529e+00 -1.07606208e+00 -9.43623632e-02 2.09786129e+00
9.65744734e-01 -2.04309836e-01 5.25895879e-03 3.47341299e-01
8.01753044e-01 1.13782227e-01 -1.71823874e-01 -4.00530964e-01
-6.41421795e-01 2.28659242e-01 7.72603005e-02 8.24004710e-01
-9.89061058e-01 6.06054604e-01 5.71108389e+00 1.37752128e+00
-8.89167070e-01 4.80891138e-01 6.78384364e-01 -4.01215285e-01
-2.94557154e-01 -5.42121470e-01 -7.36725628e-01 5.97183466e-01
1.35025382e+00 -9.01522115e-03 7.85244524e-01 3.90788287e-01
3.09928328e-01 -4.00141664e-02 -8.21908474e-01 1.50755525e+00
5.08454382e-01 -8.71307731e-01 -1.97952554e-01 -3.04383308e-01
5.87617457e-01 2.20539808e-01 3.01099211e-01 2.49932364e-01
6.67275488e-02 -1.04697645e+00 9.77341115e-01 5.91303170e-01
8.39978993e-01 -9.39483881e-01 3.20101321e-01 1.71653301e-01
-1.20725870e+00 -1.42900497e-01 9.48790014e-02 5.85615575e-01
3.31568122e-01 5.86864769e-01 -9.03852284e-01 6.03358746e-01
6.87316716e-01 2.18588829e-01 -1.46916449e-01 1.19107258e+00
-5.13098180e-01 9.92227495e-01 -4.32701468e-01 2.72726148e-01
-3.47713262e-01 3.75668198e-01 1.21644211e+00 1.51132524e+00
4.41655397e-01 1.86099634e-01 -3.32605481e-01 5.21147013e-01
-1.25612393e-01 -6.72450289e-02 -1.70082882e-01 5.24070859e-02
7.84396946e-01 9.64854121e-01 -2.44144015e-02 -2.05515504e-01
-1.78826392e-01 1.14560568e+00 -7.65494555e-02 7.09557116e-01
-1.08063507e+00 -5.41179359e-01 8.20333123e-01 -4.41495478e-01
6.36346877e-01 -1.92796931e-01 -2.91498095e-01 -1.14046431e+00
-2.71197930e-02 -1.07301736e+00 8.23678747e-02 -1.14317822e+00
-1.08788824e+00 8.36328983e-01 -2.27717817e-01 -9.99486208e-01
-2.81145811e-01 -2.59915709e-01 -4.82042789e-01 1.25296295e+00
-1.48037648e+00 -1.21277821e+00 1.83916315e-01 6.91296518e-01
7.84473479e-01 -4.57679093e-01 6.04900897e-01 7.25864828e-01
-7.14069843e-01 1.09683096e+00 1.65750176e-01 1.13516604e-03
5.87178826e-01 -9.37351406e-01 3.06231409e-01 1.48474050e+00
4.55054551e-01 2.14734048e-01 8.17865670e-01 -3.87639523e-01
-1.33032739e+00 -1.11315417e+00 1.13244987e+00 -2.89747328e-01
2.32411921e-01 -4.39643383e-01 -8.00038218e-01 3.34285676e-01
5.35021245e-01 -2.75493979e-01 7.69577444e-01 -6.91385716e-02
-6.54605508e-01 -1.01831235e-01 -7.61338890e-01 6.34010851e-01
7.41397858e-01 -8.78838599e-01 -6.02190375e-01 2.61689973e-04
1.01186836e+00 -4.14047748e-01 -4.80691612e-01 2.92159081e-01
2.12679327e-01 -7.81457067e-01 1.34889317e+00 -2.35731885e-01
2.16119200e-01 -4.97984886e-01 -6.44764423e-01 -1.67795610e+00
-2.18687713e-01 -9.93296802e-01 -3.15502882e-01 1.94081008e+00
6.18467748e-01 -1.95291206e-01 -5.34617603e-02 4.13719743e-01
-4.36630279e-01 -1.60027832e-01 -1.33714628e+00 -7.35726595e-01
-3.63039017e-01 -1.01166928e+00 3.49124581e-01 7.01015234e-01
2.96778698e-02 5.38523674e-01 -7.37148643e-01 5.81978858e-01
4.83161628e-01 -3.14622998e-01 4.25429404e-01 -5.97959638e-01
-6.32716179e-01 -2.90952921e-01 -3.86439227e-02 -1.08534241e+00
1.09899230e-01 -7.80829310e-01 3.69505227e-01 -1.42610991e+00
-2.45394498e-01 2.89193124e-01 -5.75011551e-01 3.21138144e-01
-2.47139975e-01 8.08207169e-02 6.36724114e-01 -1.20380379e-01
-6.50241017e-01 9.01175141e-01 5.95832169e-01 -3.40521961e-01
-3.62899095e-01 -7.58380210e-03 -5.96891582e-01 4.53570396e-01
5.09288371e-01 -3.62616330e-01 -2.68242061e-01 -6.78710222e-01
-1.73096925e-01 3.69499236e-01 2.12444112e-01 -1.00465524e+00
3.81941020e-01 1.93593606e-01 2.02451766e-01 -8.41650844e-01
9.28805590e-01 -5.57416797e-01 1.30100846e-01 -3.14160958e-02
-3.83579135e-01 -4.47526485e-01 3.87094527e-01 4.96547639e-01
-3.73414248e-01 3.18766952e-01 8.74534190e-01 4.52470809e-01
-4.08795506e-01 -8.04135203e-02 -4.77404237e-01 4.38211933e-02
6.43547833e-01 1.38388753e-01 -3.24530095e-01 -5.57866335e-01
-7.56758690e-01 3.06656271e-01 -3.42434049e-01 4.70519155e-01
1.08294761e+00 -1.45202804e+00 -1.26830971e+00 3.19366455e-01
2.94779167e-02 -4.19278681e-01 8.02193224e-01 8.23736012e-01
7.57742599e-02 2.60531873e-01 1.30260140e-01 -3.47964704e-01
-1.53806114e+00 3.39799643e-01 3.63461912e-01 1.59936205e-01
-3.06589514e-01 1.49871063e+00 1.49533138e-01 -3.11810207e-02
6.32271111e-01 1.34440586e-01 -3.85536879e-01 2.22893357e-01
1.04911423e+00 6.14516020e-01 2.02317283e-01 -1.01381516e+00
-6.11656666e-01 5.70678115e-02 1.54399857e-01 -8.09954643e-01
1.26223826e+00 -5.64045966e-01 2.04976931e-01 3.73333693e-01
1.12207329e+00 3.47339064e-01 -1.15016806e+00 -4.86507207e-01
-3.07297826e-01 -3.95983815e-01 4.00089413e-01 -1.23413634e+00
-9.57953393e-01 1.22583759e+00 7.64225006e-01 1.72824591e-01
1.41813099e+00 -7.16023371e-02 7.32533753e-01 -6.54451409e-03
-2.29804590e-01 -1.08420622e+00 -2.81562544e-02 7.95819402e-01
1.40255702e+00 -9.47337449e-01 -5.33289790e-01 -1.69641614e-01
-9.75835085e-01 7.41771996e-01 4.02532160e-01 4.54483509e-01
5.54916978e-01 4.99020725e-01 2.16947541e-01 3.21370363e-01
-8.55155647e-01 -2.16510713e-01 6.68253899e-01 8.12662005e-01
3.25520903e-01 8.12455863e-02 2.30830044e-01 1.04492307e+00
-8.72772411e-02 -6.46959484e-01 2.47307599e-01 4.80388790e-01
-4.26533341e-01 -6.67155862e-01 -7.21761107e-01 -1.65527731e-01
-4.69097167e-01 -6.92397535e-01 -1.97684214e-01 -2.28185534e-01
-1.55065596e-01 1.74921310e+00 -1.17368914e-01 -6.55805171e-01
3.91378969e-01 3.72162849e-01 2.65351534e-01 -3.13736588e-01
-5.45082033e-01 7.78799474e-01 2.14035615e-01 -5.09438872e-01
-2.50901788e-01 -6.25435889e-01 -9.98075426e-01 -6.93851486e-02
-2.61435688e-01 -1.61526725e-02 9.05301690e-01 7.72231221e-01
7.58947611e-01 9.13425326e-01 7.74131536e-01 -9.37955618e-01
-3.61678213e-01 -1.25105393e+00 -5.73872507e-01 1.77355796e-01
6.18992925e-01 -3.20305049e-01 -5.51131845e-01 2.04985261e-01] | [14.641359329223633, 5.618353366851807] |
0b0a7ccb-e3ac-4853-9ba8-7a19abbeb657 | precise-and-generalized-robustness | 2306.06747 | null | https://arxiv.org/abs/2306.06747v1 | https://arxiv.org/pdf/2306.06747v1.pdf | Precise and Generalized Robustness Certification for Neural Networks | The objective of neural network (NN) robustness certification is to determine if a NN changes its predictions when mutations are made to its inputs. While most certification research studies pixel-level or a few geometrical-level and blurring operations over images, this paper proposes a novel framework, GCERT, which certifies NN robustness under a precise and unified form of diverse semantic-level image mutations. We formulate a comprehensive set of semantic-level image mutations uniformly as certain directions in the latent space of generative models. We identify two key properties, independence and continuity, that convert the latent space into a precise and analysis-friendly input space representation for certification. GCERT can be smoothly integrated with de facto complete, incomplete, or quantitative certification frameworks. With its precise input space representation, GCERT enables for the first time complete NN robustness certification with moderate cost under diverse semantic-level input mutations, such as weather-filter, style transfer, and perceptual changes (e.g., opening/closing eyes). We show that GCERT enables certifying NN robustness under various common and security-sensitive scenarios like autonomous driving. | ['Zhendong Su', 'Shuai Wang', 'Yuanyuan Yuan'] | 2023-06-11 | null | null | null | null | ['style-transfer'] | ['computer-vision'] | [ 3.46193731e-01 -3.03890884e-01 1.43636823e-01 -3.48278761e-01
-2.28336334e-01 -9.95750368e-01 6.46168411e-01 -5.77433884e-01
-2.00442716e-01 3.00492018e-01 -1.63968787e-01 -5.22521436e-01
-4.03863072e-01 -8.07047069e-01 -1.10246861e+00 -7.55980015e-01
3.15073460e-01 -4.26241338e-01 2.32640952e-01 -3.69029999e-01
2.32006758e-01 8.41398239e-01 -1.92191577e+00 -3.49404011e-03
1.15477216e+00 9.40894067e-01 6.03070594e-02 9.95452046e-01
2.82476842e-01 7.05290377e-01 -8.93309057e-01 -8.01515281e-01
6.91373825e-01 3.93315889e-02 -3.86521637e-01 1.35878017e-02
7.41001606e-01 -2.48308033e-01 -5.48846126e-01 1.68803120e+00
2.82488942e-01 3.11350916e-02 5.60015380e-01 -1.79297256e+00
-1.53245831e+00 2.44083643e-01 -9.48539376e-02 -1.47905901e-01
1.51855564e-02 4.75202799e-01 5.11747599e-01 -5.53447127e-01
3.88690650e-01 1.23100317e+00 7.12350905e-01 6.40838206e-01
-1.11836612e+00 -6.13615453e-01 2.36443445e-01 2.56924689e-01
-1.22424746e+00 -1.94779426e-01 8.40325952e-01 -4.12031978e-01
5.13328433e-01 6.97620749e-01 1.55223623e-01 1.08273435e+00
6.29261196e-01 4.95664448e-01 1.19028497e+00 -3.25932950e-02
2.70337373e-01 1.02056608e-01 1.03359878e-01 4.75430250e-01
3.78692687e-01 4.69402611e-01 -3.57918739e-01 1.08637162e-01
8.25997114e-01 -2.70293672e-02 -4.35131848e-01 -4.81628209e-01
-1.06507957e+00 4.33601975e-01 3.20145071e-01 -1.64525002e-01
7.68498331e-02 2.74727166e-01 1.94367439e-01 4.82075155e-01
1.12312295e-01 2.97979504e-01 -5.41319609e-01 -5.67757990e-03
-7.25216866e-01 9.33513511e-03 3.95148695e-01 1.31928062e+00
7.79900432e-01 6.55413985e-01 -2.02901006e-01 5.54076850e-01
1.77518830e-01 8.49895775e-01 3.54008168e-01 -1.21222150e+00
2.62240618e-01 3.74041259e-01 1.33241400e-01 -1.23612809e+00
-3.81005369e-02 -7.48427749e-01 -9.36246336e-01 7.35793114e-01
-1.61156002e-02 -1.63759142e-02 -1.12257957e+00 2.04732513e+00
-7.51393009e-03 2.18537167e-01 4.29986656e-01 7.54813313e-01
6.69340193e-01 3.91896546e-01 -2.36987844e-01 -1.33041358e-02
1.13363707e+00 -6.94078565e-01 -8.18918288e-01 -3.00505050e-02
2.02928081e-01 -7.80983925e-01 1.29688728e+00 6.27008080e-01
-8.47717166e-01 -1.06417668e+00 -1.28734660e+00 1.24832369e-01
-7.39451051e-01 2.11990610e-01 2.74657428e-01 1.23506069e+00
-1.37187994e+00 5.16576827e-01 -4.98157263e-01 -6.25263378e-02
-2.48801578e-02 2.50590920e-01 -3.75002503e-01 1.69665404e-02
-1.56583023e+00 9.47399199e-01 1.89637497e-01 4.79988158e-01
-9.67744291e-01 -6.65679216e-01 -1.00569963e+00 -2.30765492e-01
1.80360109e-01 -7.23246276e-01 1.07872272e+00 -1.03368163e+00
-1.58545113e+00 5.36270499e-01 1.34471223e-01 -5.60479164e-01
8.78561795e-01 -4.33521450e-01 -1.04756367e+00 -4.43921126e-02
-9.67539921e-02 8.06400239e-01 1.29289770e+00 -1.42155123e+00
-4.18633491e-01 -1.06745698e-01 3.19424778e-01 -2.41335537e-02
-4.77843970e-01 -2.82228682e-02 -4.52240020e-01 -7.95991719e-01
1.32231927e-02 -9.34339404e-01 -8.97393376e-03 1.26964197e-01
-5.00368893e-01 1.93039954e-01 1.36454809e+00 -5.02385259e-01
9.10685837e-01 -2.28384805e+00 -9.95965153e-02 1.44228309e-01
1.42381921e-01 2.84763396e-01 -1.92505464e-01 -1.65975261e-02
-4.38917905e-01 2.07534447e-01 -2.30693549e-01 1.00772995e-04
2.40214124e-01 3.81819248e-01 -6.92542851e-01 6.18924856e-01
2.95147091e-01 9.81076121e-01 -5.84092796e-01 -6.67692125e-02
4.05607760e-01 4.95020717e-01 -4.17216748e-01 -1.80059642e-01
5.37895337e-02 6.71773180e-02 -1.43116778e-02 6.18950009e-01
1.05100524e+00 2.63243079e-01 -5.28128684e-01 -5.38271487e-01
-2.01415882e-01 -5.52107930e-01 -1.24944782e+00 1.34869981e+00
-2.24511743e-01 9.08353269e-01 -9.17699337e-02 -5.24392247e-01
1.08155704e+00 1.16524249e-01 -8.89180675e-02 -4.66443896e-01
4.15828340e-02 2.35215351e-02 -1.71354175e-01 -3.86903673e-01
8.61647367e-01 2.09276482e-01 -4.01179701e-01 -1.07596286e-01
-1.27444519e-02 -1.68756306e-01 -2.38256324e-02 2.73848534e-01
9.11213040e-01 1.13615677e-01 -1.85666949e-01 -2.47237623e-01
4.89635289e-01 -3.43749851e-01 7.13308692e-01 1.05873239e+00
-4.47929919e-01 7.41286814e-01 4.99311030e-01 -1.71017423e-01
-1.06185734e+00 -1.38183939e+00 -7.28028640e-02 6.11819923e-01
4.20448601e-01 8.07747394e-02 -1.05799770e+00 -4.85154629e-01
7.24260462e-03 8.28870237e-01 -6.30376339e-01 -7.58415163e-01
-6.16937988e-02 -4.06037241e-01 9.69870627e-01 6.18137121e-01
1.10142529e+00 -6.75961673e-01 -3.79324377e-01 -3.60472977e-01
1.58871233e-01 -1.20518959e+00 -4.30689454e-01 1.08461007e-01
-6.46859944e-01 -8.79515409e-01 -6.22116089e-01 -4.71968919e-01
8.52226675e-01 2.82655895e-01 6.81053221e-01 -1.79364681e-01
-2.06648797e-01 4.92844373e-01 -1.19259082e-01 -2.37810358e-01
-5.11449099e-01 -6.36709094e-01 5.30143380e-01 1.45910978e-01
-9.28255692e-02 -3.24363917e-01 -3.80932927e-01 6.82234287e-01
-1.31315815e+00 -6.03782460e-02 5.03593981e-01 6.47297084e-01
3.70770842e-01 5.83472192e-01 1.74578711e-01 -4.99453306e-01
7.20899880e-01 1.46819457e-01 -9.59165037e-01 5.67786455e-01
-7.80659437e-01 2.07952023e-01 6.19768143e-01 -4.82838333e-01
-1.02940345e+00 7.27236792e-02 7.36903474e-02 -9.59368050e-01
-3.27477306e-01 2.27670625e-01 -7.03011155e-01 -3.71844202e-01
8.38616490e-01 3.70443434e-01 -1.33392707e-01 -1.60861477e-01
8.18179667e-01 4.61709350e-01 1.29641891e+00 -4.94874269e-01
1.62954330e+00 3.87290388e-01 -1.93680748e-02 -5.47243416e-01
-4.67213273e-01 -3.94140035e-02 -6.08505666e-01 -5.00779927e-01
8.46261203e-01 -7.85468400e-01 -6.18415236e-01 1.01262760e+00
-1.02961266e+00 -9.44118127e-02 -3.17063540e-01 9.80213732e-02
-5.25923848e-01 5.74215829e-01 -4.92274523e-01 -6.49141550e-01
1.04681827e-01 -1.37487447e+00 7.76117623e-01 2.42140889e-01
1.22597054e-01 -7.82206178e-01 -3.33745539e-01 2.53871560e-01
3.92744362e-01 1.99817792e-01 7.54487813e-01 -1.85936212e-03
-6.58061683e-01 -3.30008447e-01 -2.67523974e-01 9.50496018e-01
2.05573142e-01 7.47636259e-01 -1.17009747e+00 -1.91723540e-01
2.20591068e-01 1.98351845e-01 8.82152438e-01 4.05915797e-01
1.36716115e+00 -3.86661053e-01 1.78450882e-01 9.61007059e-01
1.34519565e+00 2.76036382e-01 8.79197359e-01 3.26036394e-01
7.14606524e-01 4.21557099e-01 5.40440083e-01 -6.14273958e-02
-8.45692754e-02 4.87463295e-01 7.51236141e-01 -2.80594587e-01
-6.47296757e-02 -1.59071341e-01 7.46459544e-01 6.23743892e-01
1.32013741e-03 -3.42059106e-01 -6.14396274e-01 1.79080829e-01
-1.58236563e+00 -1.03637588e+00 -1.30676568e-01 1.91079652e+00
5.21021724e-01 5.06739974e-01 -4.55029786e-01 3.92305940e-01
1.13142490e+00 6.18980080e-02 -8.49137843e-01 -5.32832205e-01
-5.50227940e-01 -3.30192029e-01 1.02877617e+00 3.68906915e-01
-1.19294858e+00 8.36739480e-01 6.96525145e+00 1.02778280e+00
-1.13781726e+00 -9.16344929e-04 5.34273565e-01 2.48150617e-01
-6.85169160e-01 -2.17598453e-02 -5.74886620e-01 4.29613233e-01
7.63678193e-01 1.21791266e-01 2.26671487e-01 9.26054537e-01
1.95900738e-01 1.95299417e-01 -7.10466623e-01 8.96764576e-01
1.42353848e-01 -1.33736098e+00 3.28452647e-01 -8.23128782e-03
9.40761209e-01 -2.73188412e-01 8.53620946e-01 8.17767456e-02
4.82517242e-01 -8.14888895e-01 1.31285393e+00 8.41523588e-01
1.23779869e+00 -9.36156154e-01 6.73319995e-01 5.45926094e-02
-1.04698348e+00 -3.48243713e-01 -3.96893740e-01 1.49691939e-01
-1.12625800e-01 5.31876922e-01 -1.04791254e-01 7.20999002e-01
8.02522063e-01 5.14751256e-01 -8.08112264e-01 8.25623095e-01
-5.06241858e-01 4.49313283e-01 3.65104154e-02 3.68425161e-01
3.89669323e-04 -8.87863711e-02 7.27865756e-01 9.10766184e-01
4.99056399e-01 -3.24894160e-01 -2.84601122e-01 1.09520674e+00
5.91869317e-02 -7.14634299e-01 -5.79101682e-01 1.93147331e-01
4.44299400e-01 1.06159592e+00 -6.75347388e-01 -1.14867523e-01
-9.18121412e-02 1.14149070e+00 -2.90606886e-01 8.80219162e-01
-1.34957778e+00 -6.48564994e-01 9.42480028e-01 -3.07760388e-01
2.11777225e-01 -3.00027370e-01 -5.80575585e-01 -1.00229096e+00
4.99471985e-02 -6.97953880e-01 6.99584372e-03 -1.39368212e+00
-1.04086328e+00 6.04634643e-01 7.97316025e-04 -1.65144479e+00
1.60019815e-01 -8.75044286e-01 -6.30500197e-01 7.09929526e-01
-1.50558674e+00 -1.04878902e+00 -4.52657551e-01 8.60449314e-01
3.68740439e-01 -3.97948951e-01 5.12458742e-01 1.25626147e-01
-9.20899808e-01 9.77808058e-01 1.60880744e-01 2.14318365e-01
6.07256889e-01 -1.24469936e+00 6.03420854e-01 1.69735408e+00
-1.37469666e-02 9.26351309e-01 1.05875981e+00 -6.42946303e-01
-1.25638115e+00 -1.52056968e+00 2.99743861e-01 -5.74617565e-01
6.09661400e-01 -3.14755857e-01 -6.71811998e-01 4.53406543e-01
-1.35717653e-02 3.53457965e-02 1.32253632e-01 -3.04621607e-01
-6.58189893e-01 -4.21598047e-01 -1.17421436e+00 8.18207800e-01
9.38984275e-01 -9.84577835e-01 -5.23583829e-01 -2.41368972e-02
1.19501996e+00 -5.25303006e-01 -6.80660486e-01 6.73381329e-01
4.13265258e-01 -9.70810056e-01 1.09245801e+00 -2.05975801e-01
2.42978528e-01 -8.67560029e-01 -2.81831175e-01 -9.98087525e-01
-4.07996595e-01 -8.67124498e-01 4.55080234e-02 1.22536123e+00
3.91167372e-01 -8.21998596e-01 4.59835202e-01 6.76209390e-01
-6.35663867e-01 -3.42669904e-01 -8.39667559e-01 -1.27563679e+00
-1.04269892e-01 -8.40357542e-01 8.19544077e-01 8.35509956e-01
-6.85324192e-01 -4.99137133e-01 -5.14980614e-01 8.10516000e-01
6.31215394e-01 -7.89196566e-02 6.33749902e-01 -9.73010302e-01
-7.35676661e-02 -5.33019066e-01 -9.26952302e-01 -8.75387132e-01
-7.67019317e-02 -4.71214890e-01 1.51233405e-01 -1.23395169e+00
-1.25574633e-01 -1.49546536e-02 -5.72541595e-01 4.79799241e-01
1.71123799e-02 2.33402431e-01 5.61138391e-02 2.24999338e-01
-3.20907414e-01 5.04996896e-01 1.06914687e+00 -2.55060107e-01
1.28297538e-01 1.11774066e-02 -6.70606792e-01 6.33094192e-01
9.08409536e-01 -2.26978570e-01 -7.19998896e-01 -4.78439331e-01
3.55665565e-01 -3.99780333e-01 9.97480571e-01 -1.36811280e+00
3.33862156e-01 -3.60162050e-01 3.94779891e-01 -6.49615943e-01
1.34424359e-01 -7.75850058e-01 5.40800035e-01 5.39381266e-01
-3.21406603e-01 1.97633848e-01 3.03409606e-01 6.69199049e-01
-3.41115266e-01 -7.02036396e-02 7.99101114e-01 3.01654011e-01
-1.23604202e+00 2.72614151e-01 -2.99557149e-01 -3.24797213e-01
9.86296773e-01 -9.64376628e-01 -7.41047740e-01 -3.40569824e-01
-5.58032513e-01 -1.25462517e-01 7.05591440e-01 7.42910445e-01
8.92995834e-01 -1.43262935e+00 -3.61768037e-01 5.07722080e-01
1.03149779e-01 -2.95197219e-01 6.33375704e-01 6.23697996e-01
-5.23362815e-01 4.11196083e-01 -4.89416659e-01 -7.42049754e-01
-1.30543506e+00 7.13097930e-01 5.19703507e-01 4.17534232e-01
-4.26627100e-01 1.07208419e+00 3.74238759e-01 -5.14438033e-01
1.98361769e-01 -6.47934675e-01 -2.34904494e-02 -3.48678052e-01
1.84736654e-01 2.77661592e-01 1.80550233e-01 -6.76275432e-01
-1.26840249e-01 6.59064531e-01 3.17876756e-01 -1.03911392e-01
7.08109856e-01 -2.18021393e-01 7.82279223e-02 2.15959013e-01
1.01600587e+00 -5.48366271e-02 -1.81882024e+00 1.35330826e-01
-4.22229946e-01 -3.82008553e-01 1.31483406e-01 -8.00481558e-01
-1.25475836e+00 8.21932852e-01 9.78411257e-01 7.83661380e-02
1.41076446e+00 -4.03682053e-01 4.65995938e-01 4.41488206e-01
2.46024340e-01 -1.25778913e+00 -1.15055501e-01 4.90722716e-01
9.60101247e-01 -7.42917240e-01 -2.01041266e-01 -2.67711103e-01
-5.72489262e-01 9.44215715e-01 7.25498438e-01 1.11129349e-02
4.59090322e-01 4.25629735e-01 1.54177994e-01 3.85726020e-02
-6.14444554e-01 2.16813207e-01 5.05097091e-01 8.00727785e-01
-2.64780819e-01 1.34954110e-01 4.27701414e-01 3.13693196e-01
-5.05727530e-01 -3.12142015e-01 6.70566320e-01 6.71436608e-01
-4.98182923e-01 -7.61117339e-01 -7.03721583e-01 7.30458125e-02
-8.80538747e-02 -4.38655876e-02 -2.55107641e-01 6.77491188e-01
6.49749696e-01 9.32028174e-01 -1.76331624e-02 -8.84775102e-01
4.84764159e-01 -2.09881037e-01 1.58748522e-01 -1.58438131e-01
-1.69468969e-01 -2.54628062e-01 -2.96629757e-01 -6.93983614e-01
-3.81524235e-01 -6.21136129e-01 -9.68230009e-01 -3.20293993e-01
-4.19765979e-01 -1.55129328e-01 7.75456429e-01 8.12208116e-01
2.19849139e-01 7.85485983e-01 5.60037315e-01 -3.87153983e-01
-6.17512107e-01 -4.82518286e-01 -5.23896873e-01 3.18366051e-01
4.87676233e-01 -6.49761558e-01 -6.45752788e-01 3.70549232e-01] | [5.471975803375244, 7.972486972808838] |
78400780-2c03-421c-aeff-a2753ff1c36b | temporal-aware-mixed-attention-based | 2305.18234 | null | https://arxiv.org/abs/2305.18234v1 | https://arxiv.org/pdf/2305.18234v1.pdf | Temporal Aware Mixed Attention-based Convolution and Transformer Network (MACTN) for EEG Emotion Recognition | Emotion recognition plays a crucial role in human-computer interaction, and electroencephalography (EEG) is advantageous for reflecting human emotional states. In this study, we propose MACTN, a hierarchical hybrid model for jointly modeling local and global temporal information. The model is inspired by neuroscience research on the temporal dynamics of emotions. MACTN extracts local emotional features through a convolutional neural network (CNN) and integrates sparse global emotional features through a transformer. Moreover, we employ channel attention mechanisms to identify the most task-relevant channels. Through extensive experimentation on two publicly available datasets, namely THU-EP and DEAP, our proposed method, MACTN, consistently achieves superior classification accuracy and F1 scores compared to other existing methods in most experimental settings. Furthermore, ablation studies have shown that the integration of both self-attention mechanisms and channel attention mechanisms leads to improved classification performance. Finally, an earlier version of this method, which shares the same ideas, won the Emotional BCI Competition's final championship in the 2022 World Robot Contest. | ['Dong Ming', 'Yulin Sun', 'Dong Huang', 'Xiaopeng Si'] | 2023-05-18 | null | null | null | null | ['eeg-emotion-recognition'] | ['miscellaneous'] | [-1.34192511e-01 -5.62361293e-02 -4.60530557e-02 -2.47457653e-01
-3.27631295e-01 3.82064208e-02 1.43224865e-01 -2.28013411e-01
-5.71029067e-01 8.81366849e-01 1.98795378e-01 4.05639380e-01
-4.07630438e-03 -1.80901915e-01 -4.72122699e-01 -5.86847246e-01
-5.38008332e-01 -3.19329441e-01 -4.03896928e-01 -2.15448543e-01
1.95669279e-01 2.34205112e-01 -1.48388803e+00 2.66873002e-01
9.18185294e-01 1.69752479e+00 1.20696940e-01 3.22868645e-01
4.19916958e-01 9.60633934e-01 -5.51821053e-01 -2.55223483e-01
-1.51291564e-01 -6.15461648e-01 -9.07132983e-01 -4.25785363e-01
-4.27062988e-01 2.75754750e-01 -3.05296361e-01 1.10137165e+00
6.99607730e-01 1.34355247e-01 4.64124560e-01 -1.59491789e+00
-5.63970983e-01 4.21420157e-01 -3.95826131e-01 2.67154217e-01
3.39843303e-01 -6.65973052e-02 8.84496272e-01 -7.55039871e-01
4.68686759e-01 7.26204216e-01 8.97363961e-01 5.90945303e-01
-1.05572939e+00 -9.09382880e-01 1.56801775e-01 6.81245863e-01
-1.45251679e+00 -1.41935587e-01 8.36248398e-01 -1.69253722e-01
1.17299664e+00 -8.41157790e-03 1.12354481e+00 1.62769771e+00
8.78868520e-01 1.02430713e+00 1.26691747e+00 -4.57098428e-03
3.52801800e-01 1.48397610e-01 6.03968538e-02 3.08498442e-01
-3.89313102e-01 1.62179217e-01 -8.91517282e-01 -1.95904493e-01
5.09064436e-01 -2.81479359e-01 -5.99377275e-01 -4.30665016e-02
-1.17308533e+00 6.36324346e-01 5.90664387e-01 5.61812341e-01
-1.00537050e+00 3.29189360e-01 7.20715106e-01 2.72667646e-01
3.28689575e-01 7.93151617e-01 -5.38917959e-01 -5.89912474e-01
-5.51639795e-01 -2.36880079e-01 4.86470491e-01 5.78352869e-01
4.16897923e-01 3.86576504e-02 -1.86818257e-01 9.05452013e-01
-9.81435850e-02 8.43431707e-03 8.18364501e-01 -9.99017775e-01
-1.91330403e-01 3.40321749e-01 -3.27657610e-02 -1.11499870e+00
-8.88060331e-01 -5.01379788e-01 -1.09066999e+00 -4.42505330e-02
-3.63505512e-01 -4.68181074e-01 -4.14163530e-01 2.16865015e+00
-1.92728832e-01 2.38813162e-01 1.44887000e-01 1.14064407e+00
6.13802910e-01 4.37965870e-01 3.85727048e-01 -2.17052519e-01
1.30030882e+00 -8.36794198e-01 -1.10214555e+00 -2.41484866e-01
1.83408365e-01 -1.80870622e-01 7.13516951e-01 8.50206792e-01
-9.72080529e-01 -4.48344767e-01 -1.12478304e+00 2.12662548e-01
-4.75163251e-01 1.95231423e-01 1.19043493e+00 3.38314205e-01
-1.02448404e+00 5.42118847e-01 -6.69257879e-01 -3.85100424e-01
5.39649248e-01 6.03994429e-01 -3.97390872e-01 5.08457601e-01
-1.73261499e+00 1.13754487e+00 1.74554840e-01 4.56225753e-01
-6.67656362e-01 -3.06076258e-01 -7.17744529e-01 1.54382557e-01
-1.34219930e-01 -3.99478912e-01 1.02841282e+00 -1.45871437e+00
-1.85413694e+00 4.13701802e-01 -1.74622685e-01 -5.33388078e-01
-1.25148907e-01 -3.81958514e-01 -5.07017970e-01 2.77842551e-01
-1.73275575e-01 1.00974119e+00 6.66147232e-01 -7.37430155e-01
-3.60946506e-01 -3.63917559e-01 -3.10816944e-01 1.04083829e-01
-6.57228887e-01 2.24483564e-01 -4.16231215e-01 -5.79936624e-01
-2.05717266e-01 -8.97825241e-01 -3.61266762e-01 -4.68080044e-01
-8.10151920e-02 -3.73263657e-01 4.82340962e-01 -5.96671939e-01
1.16453731e+00 -2.28664494e+00 5.14984310e-01 3.03697050e-01
2.44581595e-01 -4.55808640e-02 -1.70147344e-01 1.68388590e-01
-4.93813068e-01 4.21302654e-02 -1.01174630e-01 -9.37741324e-02
1.67165428e-01 -1.62606776e-01 -6.90608397e-02 4.14388955e-01
3.61967713e-01 1.27200115e+00 -7.17165709e-01 -2.10731998e-01
-5.76991811e-02 5.94929159e-01 -4.96083915e-01 5.82646951e-02
3.32583070e-01 6.31487012e-01 -2.23062187e-01 4.41611230e-01
3.19966316e-01 -1.73784181e-01 2.07063675e-01 -2.45119303e-01
-2.14180667e-02 8.96178931e-02 -5.82554638e-01 2.02282953e+00
-5.78159451e-01 9.82888162e-01 5.79936691e-02 -1.25019574e+00
7.71697104e-01 7.45347083e-01 8.50073278e-01 -1.31597221e+00
8.72961164e-01 -6.21832423e-02 5.11138365e-02 -5.51657796e-01
3.64545375e-01 -1.06153101e-01 -4.69344884e-01 2.32342795e-01
3.92681062e-01 2.13094860e-01 -2.80325472e-01 1.45861447e-01
1.06376541e+00 8.90654251e-02 5.05905807e-01 -3.71840477e-01
1.49566159e-01 -4.59014773e-01 8.09000611e-01 4.09923702e-01
-6.06426120e-01 2.64051080e-01 7.18644500e-01 -2.41512999e-01
-4.79514629e-01 -4.75346118e-01 -1.74433529e-01 9.19119596e-01
4.40718651e-01 -4.93135512e-01 -6.79861367e-01 -3.85116369e-01
-3.89196426e-01 5.57127953e-01 -1.13646698e+00 -7.93734968e-01
3.05753257e-02 -9.08218443e-01 6.10447824e-01 8.35775673e-01
5.83495557e-01 -1.72417772e+00 -1.24836528e+00 3.71476620e-01
-6.47120237e-01 -1.17649972e+00 -1.16938606e-01 7.52027094e-01
-4.25839394e-01 -8.16833496e-01 -6.23363614e-01 -7.19393313e-01
2.23543867e-01 -3.75433505e-01 7.30442643e-01 -2.80718029e-01
-4.96887177e-01 4.00168866e-01 -4.90069449e-01 -6.56054735e-01
4.96907681e-01 1.17495775e-01 1.06676541e-01 2.35936865e-01
5.09281158e-01 -7.11486280e-01 -5.40305555e-01 1.41682759e-01
-4.87985462e-01 -2.16898490e-02 7.65146255e-01 1.03564203e+00
3.62256467e-01 1.51358649e-01 1.07432401e+00 -2.71142602e-01
9.66363907e-01 -4.90574628e-01 -3.73394936e-02 -8.56395960e-02
-3.66113216e-01 -7.38318861e-02 4.35828358e-01 -4.87152845e-01
-1.14454472e+00 1.49431475e-03 -2.15356380e-01 -4.49438125e-01
-8.13244954e-02 6.40550852e-01 -7.01334924e-02 -1.56554997e-01
4.18391585e-01 1.96056768e-01 -3.20222855e-01 1.11815400e-01
-8.55778009e-02 7.29625404e-01 6.41789019e-01 -4.17910904e-01
-1.39433756e-01 2.56820798e-01 -4.76088017e-01 -6.84011459e-01
-4.87905651e-01 -3.00162464e-01 -2.94540465e-01 -4.21405762e-01
1.03165138e+00 -9.66832101e-01 -1.23151481e+00 6.74023271e-01
-1.12266028e+00 -4.61854935e-01 1.03395887e-01 7.83964396e-01
-7.67724454e-01 -4.63540927e-02 -7.73400366e-01 -9.58229065e-01
-4.36508238e-01 -9.15686369e-01 9.40051913e-01 3.20313305e-01
-6.43011928e-01 -5.85684776e-01 3.19366194e-02 -2.21895039e-01
6.14172280e-01 2.35486984e-01 5.48475623e-01 -2.83259064e-01
1.21958055e-01 -2.04757124e-01 -2.62136459e-01 3.25901389e-01
-3.13967258e-01 -3.87743145e-01 -1.08687067e+00 1.72584161e-01
3.83848101e-02 -7.54077911e-01 7.95681715e-01 4.78631079e-01
1.63796616e+00 2.14421764e-01 -3.53170186e-01 4.83977944e-01
1.07373345e+00 7.08056688e-01 1.15336049e+00 1.81652993e-01
1.42990559e-01 5.51435888e-01 4.50787246e-01 6.73707902e-01
4.17660266e-01 5.93627930e-01 5.22209406e-01 -2.63372749e-01
4.47757035e-01 1.71283096e-01 4.39651072e-01 8.66015494e-01
-2.45380431e-01 1.72436506e-01 -6.31107926e-01 6.58701479e-01
-1.93666732e+00 -9.15706873e-01 2.67526895e-01 1.77430809e+00
6.91752791e-01 -4.62751053e-02 -1.16742536e-01 1.56457163e-02
4.49373126e-01 -2.10390776e-01 -7.47903526e-01 -7.28586555e-01
-3.21539849e-01 5.57264268e-01 7.75586069e-02 -1.07680455e-01
-1.10933733e+00 9.53877211e-01 6.63717985e+00 6.35214448e-01
-1.42894447e+00 3.05080414e-01 6.08129203e-01 -2.01360047e-01
2.68922359e-01 -5.56612372e-01 -1.68247879e-01 4.51919764e-01
1.27033222e+00 -1.65454134e-01 6.52289987e-01 6.55400634e-01
4.44579795e-02 -2.93979615e-01 -7.85694540e-01 1.37413812e+00
2.54642278e-01 -8.76953781e-01 -7.95635819e-01 1.84122790e-02
5.58952689e-01 6.61935359e-02 1.75130829e-01 7.17254937e-01
1.08152002e-01 -1.25874627e+00 8.28300059e-01 6.68712080e-01
6.50283277e-01 -1.11960471e+00 1.00901008e+00 7.47858658e-02
-1.11850643e+00 -4.32568908e-01 -1.40452951e-01 -2.39909410e-01
-5.47729284e-02 4.01342988e-01 -5.91961071e-02 4.71619397e-01
1.30989277e+00 9.97677267e-01 -2.43227154e-01 8.73518288e-01
-3.53993565e-01 4.19011056e-01 3.64890806e-02 -3.12955022e-01
1.82800040e-01 7.33473822e-02 2.58293808e-01 1.25873733e+00
1.91393510e-01 4.36423242e-01 -3.03267092e-01 8.97780061e-01
-8.69165584e-02 7.04171434e-02 -3.52428436e-01 -1.95152462e-01
1.39950380e-01 1.53092980e+00 -5.65540016e-01 -1.75707519e-01
-1.06008016e-01 1.57490075e+00 4.99390960e-01 5.90224266e-01
-1.35435784e+00 -9.57960784e-01 7.17514634e-01 -9.77991223e-01
1.85416684e-01 -4.13470622e-03 -3.22120607e-01 -1.12370872e+00
-1.36292428e-01 -7.97470331e-01 -1.07566896e-03 -1.18381727e+00
-1.02649832e+00 1.11238778e+00 -4.31681454e-01 -1.03432870e+00
-5.94084822e-02 -5.01226425e-01 -6.27187610e-01 6.43690705e-01
-1.28856528e+00 -5.77424645e-01 -3.82998168e-01 7.58217752e-01
2.29381427e-01 2.72915632e-01 1.01036429e+00 3.17639709e-01
-8.13336372e-01 5.58509707e-01 -2.93910980e-01 2.86622941e-02
7.35137343e-01 -9.80000854e-01 -2.04459190e-01 4.55187798e-01
-1.74449235e-01 5.72185218e-01 4.07892078e-01 -2.21495926e-01
-1.21905136e+00 -9.76271689e-01 6.71123385e-01 -9.49623715e-03
5.06375849e-01 -4.80944306e-01 -6.89483941e-01 6.11605823e-01
7.29737937e-01 -2.87673652e-01 8.28138351e-01 3.65188181e-01
-1.10461324e-01 9.30125639e-02 -9.14001584e-01 4.71168578e-01
9.53769565e-01 -5.52592576e-01 -4.56482083e-01 -1.38154596e-01
3.47269714e-01 -6.12322539e-02 -1.03471291e+00 5.16272128e-01
8.87082636e-01 -9.95204031e-01 5.09521723e-01 -4.79403853e-01
5.22624671e-01 2.06235304e-01 -1.94696248e-01 -1.88232529e+00
-4.74623322e-01 -5.62698722e-01 1.18772224e-01 8.16829562e-01
3.37024331e-01 -6.63032115e-01 3.05726111e-01 4.11929667e-01
-1.86696053e-01 -1.11600566e+00 -9.67585206e-01 -6.99716806e-01
-2.35830873e-01 -7.01784134e-01 3.58699143e-01 8.36851120e-01
9.24649775e-01 4.91933763e-01 -6.33335650e-01 -1.52358741e-01
1.01823494e-01 -2.95403004e-02 2.86422431e-01 -1.23983979e+00
-1.04829306e-02 -7.34011531e-01 -5.05652964e-01 -6.23899341e-01
6.45460248e-01 -8.64111841e-01 3.93841565e-01 -1.29229999e+00
3.96293670e-01 -1.56270317e-03 -8.97041023e-01 7.25909114e-01
-7.24866614e-02 5.00670493e-01 -8.88072047e-03 -2.81734258e-01
-1.01870143e+00 1.18011844e+00 8.70871544e-01 -6.05477206e-03
-2.72987068e-01 -5.69979072e-01 -9.61068690e-01 6.01166606e-01
9.67056990e-01 -2.55888045e-01 -5.23223877e-02 1.07143978e-02
3.18140566e-01 1.22445859e-01 3.01757276e-01 -1.23760283e+00
3.21719170e-01 2.17267692e-01 6.87492728e-01 -1.00570127e-01
6.32182777e-01 -8.05567980e-01 3.00623253e-02 4.20169532e-01
-4.13357884e-01 -7.79678300e-02 5.75385451e-01 4.38018769e-01
-5.35444200e-01 3.05387735e-01 7.97217786e-01 1.97637811e-01
-9.46499467e-01 3.79207075e-01 -9.76881742e-01 -2.60554373e-01
1.08729970e+00 -1.19584091e-01 -1.60584915e-02 -5.32934189e-01
-9.32690382e-01 3.67750376e-01 -1.83219984e-01 6.50938749e-01
6.25562370e-01 -1.52470970e+00 -3.54881287e-01 3.08603644e-01
1.62399143e-01 -8.66924524e-01 5.13226867e-01 1.54789400e+00
2.65612006e-01 5.76384544e-01 -6.60650373e-01 -3.63112390e-01
-1.07980549e+00 4.80177879e-01 4.57590610e-01 -1.24000870e-01
-4.11693722e-01 1.03493381e+00 3.58812064e-01 -2.28063077e-01
3.15823555e-01 -2.27742925e-01 -4.32409227e-01 2.10727856e-01
5.06976724e-01 1.27882794e-01 5.66735454e-02 -5.49664557e-01
-5.26017427e-01 3.36582929e-01 2.95359433e-01 -2.05124199e-01
1.38728154e+00 -1.59835234e-01 -2.84373134e-01 5.61711729e-01
1.28688633e+00 -3.96782190e-01 -9.70700026e-01 1.10833257e-01
-3.04187275e-02 1.52833343e-01 3.05472761e-01 -1.13850451e+00
-1.29775834e+00 1.11518168e+00 6.94510698e-01 -8.16437751e-02
1.67254496e+00 -1.93060085e-01 8.57940376e-01 2.70945996e-01
6.14808500e-01 -1.33109999e+00 2.28807554e-01 6.05695188e-01
1.03999543e+00 -1.06692755e+00 -4.86084729e-01 8.97944160e-03
-1.14392257e+00 7.48841345e-01 8.34210932e-01 -1.52314603e-01
7.27847099e-01 2.60953069e-01 -1.17950983e-01 -3.89774382e-01
-1.14076495e+00 -7.09428862e-02 1.99259609e-01 3.80743623e-01
3.70128036e-01 2.12348446e-01 -3.93939763e-01 1.59959519e+00
-1.56825036e-01 3.07119042e-01 1.08502023e-01 7.08611190e-01
-2.06090540e-01 -6.12075269e-01 -9.51267332e-02 3.26248974e-01
-6.51297033e-01 -1.26499757e-01 -5.44637680e-01 7.27711916e-01
1.25788271e-01 8.62356842e-01 1.51595175e-02 -9.10748839e-01
2.47026131e-01 3.80007207e-01 5.19670248e-01 -5.65333180e-02
-8.80424738e-01 1.76415890e-01 -7.32003301e-02 -1.11291361e+00
-4.74691540e-01 -5.88807642e-01 -1.52738166e+00 2.27248818e-01
-2.52418190e-01 3.91059995e-01 8.06286871e-01 7.75720716e-01
8.24865937e-01 1.02339721e+00 6.50470316e-01 -1.08051312e+00
1.75220549e-01 -1.01361251e+00 -8.41826081e-01 3.34480286e-01
1.36296451e-03 -9.76844907e-01 -2.37957925e-01 -1.80521503e-01] | [13.174604415893555, 3.520907402038574] |
15f1c45d-8902-4b49-ad5b-2ef329194a5e | using-orthophoto-for-building-boundary | 1905.09150 | null | https://arxiv.org/abs/1905.09150v1 | https://arxiv.org/pdf/1905.09150v1.pdf | Using Orthophoto for Building Boundary Sharpening in the Digital Surface Model | Nowadays dense stereo matching has become one of the dominant tools in 3D reconstruction of urban regions for its low cost and high flexibility in generating dense 3D points. However, state-of-the-art stereo matching algorithms usually apply a semi-global matching (SGM) strategy. This strategy normally assumes the surface geometry pieceswise planar, where a smooth penalty is imposed to deal with non-texture or repeating-texture areas. This on one hand, generates much smooth surface models, while on the other hand, may partially leads to smoothing on depth discontinuities, particularly for fence-shaped regions or densely built areas with narrow streets. To solve this problem, in this work, we propose to use the line segment information extracted from the corresponding orthophoto as a pose-processing tool to sharpen the building boundary of the Digital Surface Model (DSM) generated by SGM. Two methods which are based on graph-cut and plane fitting are proposed and compared. Experimental results on several satellite datasets with ground truth show the robustness and effectiveness of the proposed DSM sharpening method. | ['Rong-Jun Qin', 'Xiaohu Lu', 'Xu Huang'] | 2019-05-22 | null | null | null | null | ['stereo-matching'] | ['computer-vision'] | [ 5.13923585e-01 9.02778208e-02 3.89360726e-01 -4.07560587e-01
-3.43519509e-01 -7.57331327e-02 5.81396103e-01 2.51093566e-01
-8.71748626e-02 7.82458246e-01 -1.35809064e-01 -4.15503643e-02
-1.37231603e-01 -1.35423362e+00 -4.32436705e-01 -5.50388515e-01
2.61715323e-01 6.91393733e-01 6.87600434e-01 -3.95069331e-01
4.01151806e-01 6.15700662e-01 -1.74498200e+00 -1.01291567e-01
1.51455832e+00 7.65228808e-01 4.66421485e-01 6.43315539e-02
-6.17397904e-01 -8.79147723e-02 4.73779440e-02 -3.14603060e-01
4.95869219e-01 -1.68793187e-01 -4.24667954e-01 6.80868983e-01
3.56989175e-01 -3.65201056e-01 -1.01695798e-01 1.35745776e+00
1.36293694e-01 -4.74464521e-02 7.34839559e-01 -8.29684317e-01
3.56380999e-01 -1.38655201e-01 -9.85894978e-01 -3.65118057e-01
3.77042443e-01 -1.15941083e-02 5.04877687e-01 -8.20184469e-01
6.59414232e-01 1.32124543e+00 8.51379752e-01 -8.07991624e-02
-1.33895957e+00 -4.56124991e-01 -6.22103810e-02 1.28236175e-01
-1.56697547e+00 -2.56673664e-01 1.15383661e+00 -5.47892392e-01
3.69785994e-01 3.30765337e-01 7.36492097e-01 1.45328894e-01
1.54020250e-01 3.07597667e-01 1.29584455e+00 -3.23959142e-01
2.17053682e-01 -8.08340982e-02 7.65025616e-02 4.55935866e-01
4.12097692e-01 4.94154468e-02 2.20013540e-02 -2.77836174e-01
8.41026068e-01 1.48764268e-01 -4.71979409e-01 -4.87464339e-01
-6.78363204e-01 6.03353381e-01 3.76354277e-01 2.58599907e-01
-5.46778798e-01 -3.47869396e-01 1.48726851e-01 1.24215066e-01
8.33923459e-01 -2.22103462e-01 7.48265982e-02 2.48889729e-01
-1.27765238e+00 6.49943054e-01 3.88321579e-01 9.10239518e-01
1.34605539e+00 -3.45606282e-02 4.64046806e-01 8.75672340e-01
4.01869684e-01 5.89966357e-01 7.42592709e-03 -6.96260810e-01
6.85697317e-01 1.09523368e+00 1.73323765e-01 -1.38056326e+00
-2.95613974e-01 -1.36187017e-01 -1.02655101e+00 5.01112401e-01
2.47423008e-01 2.93942064e-01 -9.99182761e-01 1.01487350e+00
6.50933981e-01 1.35858923e-01 -2.32038870e-01 8.12183678e-01
5.83433092e-01 7.56594658e-01 -3.24526459e-01 -1.54247746e-01
9.43659067e-01 -2.91126519e-01 -5.84759235e-01 -2.67576456e-01
4.09325510e-01 -9.17319775e-01 7.83022821e-01 2.36685976e-01
-9.39773560e-01 -5.33748507e-01 -9.33440030e-01 1.15947530e-01
-1.20902658e-01 -1.90121368e-01 2.59338140e-01 4.71858919e-01
-8.83036971e-01 8.04430127e-01 -6.55018568e-01 -5.01477361e-01
1.24297656e-01 4.73641381e-02 -2.97388434e-01 -3.57242703e-01
-9.63007510e-01 6.93552077e-01 3.34366202e-01 3.02892625e-01
-3.01447004e-01 -4.87734169e-01 -7.18455911e-01 -3.21698003e-02
2.11247072e-01 -6.88890100e-01 6.28661692e-01 -9.51604307e-01
-1.34592605e+00 1.10825562e+00 -1.31252035e-01 -2.47324720e-01
8.88776720e-01 9.57612097e-02 -2.84743577e-01 2.54942407e-03
2.17568114e-01 1.87169641e-01 7.62413383e-01 -1.49247670e+00
-6.77806199e-01 -6.98139369e-01 -3.49672794e-01 2.37273827e-01
3.36830586e-01 -3.81699234e-01 -3.25670987e-01 -4.46355760e-01
9.04261053e-01 -6.89489305e-01 -3.81392032e-01 5.11939861e-02
-3.91856819e-01 1.85369775e-01 9.95989025e-01 -9.01203811e-01
1.02022994e+00 -2.24946928e+00 -4.00575250e-03 7.01725602e-01
-1.15422502e-01 8.46142545e-02 2.44902551e-01 5.18445134e-01
6.84348345e-02 -9.46107954e-02 -8.77186418e-01 -2.58833438e-01
-4.39329177e-01 1.67134106e-01 3.91772091e-02 6.14916503e-01
-8.44720379e-02 4.90674973e-02 -5.52927852e-01 -6.70981467e-01
5.89927614e-01 4.07944858e-01 -3.42345625e-01 1.37097523e-01
-2.49096125e-01 5.00090480e-01 -5.03099978e-01 4.99296546e-01
1.50582576e+00 3.74140948e-01 5.23197465e-02 -1.18480772e-01
-6.72107160e-01 1.86296292e-02 -1.77124357e+00 1.47149992e+00
-3.00767809e-01 2.44276941e-01 6.00367844e-01 -7.26227641e-01
1.42684567e+00 1.47800967e-01 2.99199134e-01 -7.22359061e-01
-2.77927071e-02 6.96370602e-01 -5.05286932e-01 -3.30118358e-01
5.34463584e-01 -2.33435079e-01 4.43556279e-01 -3.49807531e-01
-8.72484207e-01 -8.78992260e-01 -2.24914849e-01 -1.76665783e-01
5.31892359e-01 1.91522971e-01 1.42062694e-01 -5.36475837e-01
9.37665999e-01 2.17543826e-01 7.24047780e-01 1.51671246e-01
3.57038230e-01 8.71013105e-01 2.50506818e-01 -4.34546083e-01
-1.28592288e+00 -7.49302745e-01 -3.89387071e-01 -1.97988138e-01
4.90147859e-01 -3.78054008e-02 -8.27322185e-01 1.66999493e-02
1.19764090e-01 7.12681711e-01 -1.41700119e-01 2.80070394e-01
-6.86693132e-01 -4.73247409e-01 -4.48375046e-02 -1.51240215e-01
1.10273206e+00 -5.59343994e-01 -4.41443741e-01 4.23171431e-01
-2.33762249e-01 -8.62963736e-01 -3.11454013e-02 -3.49904388e-01
-1.52169049e+00 -1.08106136e+00 -8.66193771e-01 -6.35867059e-01
8.41384172e-01 6.31249666e-01 8.52016985e-01 3.16202283e-01
-2.30076388e-02 -1.83391944e-01 -1.65135697e-01 1.22135967e-01
-3.74265343e-01 -1.61363274e-01 -5.19454837e-01 3.79505366e-01
-2.56895144e-02 -7.62912452e-01 -4.60483223e-01 6.22506797e-01
-8.92463446e-01 5.71774781e-01 3.11668813e-01 4.57067251e-01
6.90036416e-01 3.51565927e-01 -4.59829904e-02 -8.39771807e-01
9.81066376e-02 -3.29919040e-01 -1.12967134e+00 -1.13969713e-01
-4.47673202e-01 -1.44909278e-01 2.98879772e-01 3.15285563e-01
-1.27933776e+00 1.76190943e-01 -3.97643059e-01 -1.26356721e-01
-3.57418984e-01 4.57156479e-01 -5.16595900e-01 -2.27221400e-01
3.76702338e-01 4.61649626e-01 5.05094556e-03 -7.93347716e-01
-8.52895528e-02 7.02741563e-01 4.19086337e-01 -3.59488368e-01
1.08238804e+00 9.57383811e-01 4.78215426e-01 -1.42784536e+00
-1.36435181e-01 -6.04996026e-01 -5.66384196e-01 -4.01469827e-01
7.65972257e-01 -6.67955637e-01 -1.04420558e-01 7.91565716e-01
-1.19496953e+00 -1.81375399e-01 7.37818629e-02 2.24820778e-01
-4.62696731e-01 8.85150731e-01 -2.34743446e-01 -9.46560621e-01
-2.31518373e-01 -1.08220577e+00 1.05847442e+00 1.99983224e-01
2.90299077e-02 -7.75082350e-01 8.65602642e-02 3.63855004e-01
3.41158748e-01 5.95616043e-01 9.99067247e-01 3.22586030e-01
-9.39323783e-01 -2.20712051e-01 -2.92324007e-01 2.42919028e-01
1.65001854e-01 1.83231577e-01 -9.06794012e-01 5.55218793e-02
1.83059067e-01 2.86333799e-01 6.75140500e-01 5.09889662e-01
6.50806308e-01 4.90040965e-02 -3.71041894e-01 7.62318194e-01
1.82075644e+00 4.76661474e-02 1.17634690e+00 5.93353510e-01
5.04179001e-01 8.93164754e-01 1.02223504e+00 5.13998151e-01
2.68724561e-01 8.92383397e-01 7.35570788e-01 -1.63184047e-01
8.56426079e-03 -2.63281554e-01 8.08564387e-03 5.68001747e-01
-4.26048905e-01 1.18683986e-01 -9.83328283e-01 5.06639481e-01
-1.81797576e+00 -9.46044803e-01 -1.04083085e+00 2.66555691e+00
4.68436211e-01 3.49585027e-01 -1.46450996e-01 5.36247015e-01
1.05634141e+00 1.75581276e-01 4.17975597e-02 -1.73375517e-01
-3.77445281e-01 -3.44689167e-03 8.27019513e-01 9.79136765e-01
-7.40240276e-01 8.36333275e-01 4.22291088e+00 1.05009830e+00
-1.04218984e+00 -1.27606645e-01 2.22718269e-01 7.15184093e-01
-6.18909061e-01 3.41397077e-01 -6.95520580e-01 5.17645895e-01
3.25914234e-01 1.05212517e-01 8.88995305e-02 7.87799001e-01
7.97175646e-01 -5.89997053e-01 -4.12895143e-01 1.23093772e+00
-4.07287389e-01 -1.19177961e+00 -2.38610059e-02 2.38497600e-01
9.33393478e-01 -1.74695551e-01 -4.54022914e-01 -2.12828055e-01
-2.07625359e-01 -4.50977206e-01 7.89678335e-01 6.27217770e-01
6.50545061e-01 -7.80485511e-01 7.71329522e-01 5.52129447e-01
-1.49107015e+00 5.31284034e-01 -4.63519722e-01 -1.05377831e-01
4.38225925e-01 1.16195452e+00 -5.37731707e-01 9.36220706e-01
4.10704404e-01 5.22887349e-01 -4.30495650e-01 1.47896945e+00
-1.00901693e-01 1.80455551e-01 -6.23562455e-01 4.76326048e-01
2.82287419e-01 -9.23037589e-01 8.13822567e-01 9.11281765e-01
5.67246318e-01 2.19749153e-01 1.42582551e-01 9.02356148e-01
4.05279994e-01 3.69099885e-01 -9.36834514e-01 5.13381660e-01
3.68848950e-01 9.95038629e-01 -7.46079326e-01 -3.38057786e-01
-3.22247475e-01 9.36650038e-01 -2.39331782e-01 1.28227323e-01
-5.16951859e-01 -2.55833894e-01 4.05346155e-01 8.64823341e-01
-6.86463565e-02 -5.36494553e-01 -6.50743246e-01 -9.79955673e-01
5.71640953e-02 -5.77245176e-01 5.06002791e-02 -6.19802058e-01
-8.74300897e-01 1.43843263e-01 1.71008566e-03 -1.45380700e+00
9.18328688e-02 -1.21953003e-01 -6.49283946e-01 1.04536092e+00
-1.60232913e+00 -1.06671894e+00 -7.25116432e-01 6.29500091e-01
7.42411137e-01 3.64181250e-01 3.28173161e-01 5.15431046e-01
-2.10186318e-01 -2.90840775e-01 3.86375278e-01 -4.46294606e-01
2.92930454e-01 -7.61451364e-01 3.56361747e-01 9.26437438e-01
-6.77316666e-01 8.21921825e-02 9.55698967e-01 -1.10502195e+00
-9.98693347e-01 -9.43728507e-01 1.14005029e+00 4.80344534e-01
1.40556857e-01 -1.49158418e-01 -1.20606852e+00 2.51548499e-01
-2.24079326e-01 -4.39331830e-01 -1.13347664e-01 -2.99980909e-01
2.64714837e-01 -2.64376312e-01 -1.41163683e+00 5.58016300e-01
8.88253331e-01 -1.76404655e-01 -6.59414291e-01 9.81051400e-02
4.44878563e-02 -4.00804371e-01 -5.55010080e-01 6.43861592e-01
2.66832888e-01 -1.52396333e+00 8.35396349e-01 3.80287796e-01
2.34600797e-01 -6.34965599e-01 -2.01185524e-01 -1.19421780e+00
-2.35144451e-01 -4.32882875e-01 6.78444982e-01 1.37821782e+00
8.24169740e-02 -7.08511591e-01 9.46184814e-01 4.57229823e-01
-4.19462442e-01 -2.28024021e-01 -1.07077122e+00 -7.70892620e-01
-3.91389728e-01 -3.11717778e-01 6.76820874e-01 9.01123106e-01
-4.39799935e-01 -1.07750289e-01 -8.67633671e-02 3.71484727e-01
9.91145730e-01 3.04011643e-01 1.00059581e+00 -1.65034866e+00
1.00084059e-01 -4.16677713e-01 -5.55681467e-01 -9.37612712e-01
-2.17250749e-01 -4.60413724e-01 1.75099801e-02 -1.74916089e+00
-2.10484073e-01 -8.50551128e-01 6.46754384e-01 -2.10969388e-01
6.93723634e-02 -8.13330561e-02 -1.11191437e-01 4.26891446e-01
3.48658085e-01 7.95494139e-01 1.11192703e+00 -4.31780480e-02
-4.27545309e-01 2.43406832e-01 2.05485106e-01 1.05137372e+00
6.14893138e-01 -4.31725919e-01 -2.66976655e-01 -4.32279289e-01
1.73359782e-01 1.00403778e-01 3.02733898e-01 -1.12382603e+00
-5.32248542e-02 -4.29583907e-01 -6.74007460e-02 -1.08674026e+00
5.08019209e-01 -1.29183376e+00 7.78550565e-01 5.87470233e-01
4.05346662e-01 -2.46586367e-01 7.47209713e-02 4.72474724e-01
-4.50514495e-01 -3.88546228e-01 1.34240639e+00 -3.37770402e-01
-7.41668582e-01 2.14497834e-01 -9.23760235e-02 -2.42315963e-01
1.00338554e+00 -7.65596092e-01 9.45672169e-02 -2.53297180e-01
-3.45964402e-01 1.08588122e-01 1.08573163e+00 -1.82631105e-01
6.94656849e-01 -1.21581411e+00 -7.84593642e-01 4.06217396e-01
6.40969500e-02 4.55502927e-01 5.09732723e-01 6.51811719e-01
-1.13755429e+00 2.94898510e-01 -1.71491995e-01 -8.13924551e-01
-1.23902345e+00 2.02281252e-01 4.08355922e-01 3.69903911e-03
-9.46557760e-01 3.06927204e-01 1.48134112e-01 -2.67458946e-01
-2.86006093e-01 -4.33907807e-01 -1.08714163e-01 2.32050698e-02
6.24055527e-02 7.61315465e-01 2.58628100e-01 -8.80035639e-01
-2.28803799e-01 1.22834837e+00 4.36153024e-01 -9.59415808e-02
1.28335798e+00 -3.11156064e-01 -1.90838844e-01 2.29535341e-01
7.84163475e-01 2.32853562e-01 -1.07955587e+00 -9.38263908e-02
5.18014096e-02 -1.01113701e+00 3.43175262e-01 1.02059823e-02
-9.46379840e-01 8.66510808e-01 4.79529649e-01 2.02670112e-01
8.65192056e-01 -4.01438981e-01 7.82720625e-01 -1.01992540e-01
7.03655601e-01 -1.21451402e+00 -7.66201079e-01 2.33715758e-01
8.71164978e-01 -8.57731462e-01 1.63143083e-01 -1.07710850e+00
-1.90271273e-01 1.36173213e+00 2.74438679e-01 -2.11945310e-01
6.39274478e-01 -1.47406220e-01 -3.33965957e-01 -1.83413789e-01
1.23777926e-01 -3.37786227e-01 -5.49856992e-03 4.07784194e-01
1.46626696e-01 -1.79464687e-02 -7.57293284e-01 -8.26535970e-02
-5.46682859e-04 1.33205354e-01 4.16643769e-01 8.93459499e-01
-9.25983667e-01 -1.10017908e+00 -1.02511823e+00 3.13989073e-01
1.62426323e-01 6.86767772e-02 -2.00059619e-02 8.50453496e-01
2.57316619e-01 8.13986778e-01 1.60839021e-01 -8.99100900e-02
6.44791961e-01 -2.20697559e-02 3.24246407e-01 -4.42599505e-01
-1.96360379e-01 2.64931053e-01 1.94776058e-01 -3.38545501e-01
-3.43565762e-01 -1.00844347e+00 -1.21201122e+00 -5.17910838e-01
-3.58191967e-01 -7.30761513e-02 8.13860595e-01 8.11823964e-01
8.06229785e-02 -1.75731733e-01 7.27604985e-01 -1.06836534e+00
-2.18763173e-01 -8.09831381e-01 -9.13073599e-01 6.03822351e-01
6.13324456e-02 -9.19961572e-01 -5.57715833e-01 -1.08440854e-01] | [8.406064987182617, -2.610987901687622] |
01b66095-7c7b-4189-900a-50c95069548b | meta-hallucinator-towards-few-shot-cross | 2305.06978 | null | https://arxiv.org/abs/2305.06978v1 | https://arxiv.org/pdf/2305.06978v1.pdf | Meta-hallucinator: Towards Few-Shot Cross-Modality Cardiac Image Segmentation | Domain shift and label scarcity heavily limit deep learning applications to various medical image analysis tasks. Unsupervised domain adaptation (UDA) techniques have recently achieved promising cross-modality medical image segmentation by transferring knowledge from a label-rich source domain to an unlabeled target domain. However, it is also difficult to collect annotations from the source domain in many clinical applications, rendering most prior works suboptimal with the label-scarce source domain, particularly for few-shot scenarios, where only a few source labels are accessible. To achieve efficient few-shot cross-modality segmentation, we propose a novel transformation-consistent meta-hallucination framework, meta-hallucinator, with the goal of learning to diversify data distributions and generate useful examples for enhancing cross-modality performance. In our framework, hallucination and segmentation models are jointly trained with the gradient-based meta-learning strategy to synthesize examples that lead to good segmentation performance on the target domain. To further facilitate data hallucination and cross-domain knowledge transfer, we develop a self-ensembling model with a hallucination-consistent property. Our meta-hallucinator can seamlessly collaborate with the meta-segmenter for learning to hallucinate with mutual benefits from a combined view of meta-learning and self-ensembling learning. Extensive studies on MM-WHS 2017 dataset for cross-modality cardiac segmentation demonstrate that our method performs favorably against various approaches by a lot in the few-shot UDA scenario. | ['S. Kevin Zhou', 'Cuntai Guan', 'Zeng Zeng', 'Fangcheng Zhou', 'Ziyuan Zhao'] | 2023-05-11 | null | null | null | null | ['cardiac-segmentation', 'unsupervised-domain-adaptation'] | ['medical', 'methodology'] | [ 5.06095409e-01 3.13568085e-01 -4.10562754e-01 -3.16892892e-01
-1.35495865e+00 -2.34797120e-01 3.16363752e-01 -5.76661378e-02
-2.14888811e-01 7.90968418e-01 2.69654006e-01 3.73613164e-02
1.15323298e-01 -5.48638344e-01 -5.89410126e-01 -8.57304573e-01
5.10719359e-01 5.96784532e-01 1.27236918e-01 -1.90152466e-01
-3.47819805e-01 -9.75223631e-02 -9.95657444e-01 5.78428686e-01
1.34274805e+00 5.23478031e-01 5.16466320e-01 3.03358883e-01
-1.94427237e-01 5.72122633e-01 -4.29739654e-01 -2.49360636e-01
1.40167177e-01 -8.96346271e-01 -1.00171447e+00 6.28970385e-01
7.54377022e-02 -1.45098627e-01 -2.08081394e-01 1.11747742e+00
1.08203554e+00 9.16594043e-02 7.32119679e-01 -1.08861840e+00
-8.32961917e-01 5.10363758e-01 -7.27816522e-01 7.77045190e-02
2.57479884e-02 4.17559534e-01 4.98155892e-01 -7.52685785e-01
9.71916854e-01 8.20950627e-01 7.03888178e-01 1.01535082e+00
-1.22633553e+00 -5.77475965e-01 -1.04927920e-01 -2.65943054e-02
-1.05353200e+00 -1.17153309e-01 8.83006454e-01 -4.17623460e-01
4.81474251e-01 -8.12076628e-02 4.61501718e-01 1.39682186e+00
8.43085274e-02 1.14901805e+00 1.32716000e+00 -1.57118484e-01
3.45045716e-01 2.92121083e-01 -9.34029520e-02 6.23292267e-01
-1.97558179e-01 -1.31836578e-01 -4.68271583e-01 -2.00436518e-01
6.86815798e-01 2.25579500e-01 -2.52443135e-01 -6.34525120e-01
-1.65762115e+00 8.09035063e-01 5.24304509e-01 3.21592510e-01
-5.81588328e-01 -3.38573426e-01 6.82057559e-01 2.42771477e-01
5.16706765e-01 6.45797372e-01 -3.78549069e-01 2.58032292e-01
-1.09141457e+00 -3.69818881e-02 3.66244346e-01 1.12589896e+00
4.99673545e-01 1.97455913e-01 -6.46538973e-01 1.32066000e+00
4.34825979e-02 4.91763115e-01 9.55101788e-01 -1.10705304e+00
2.59756565e-01 5.03655791e-01 -1.92972779e-01 -2.12323114e-01
-3.48871559e-01 -7.31136203e-01 -1.17520833e+00 5.37774637e-02
3.48776966e-01 -2.92300373e-01 -1.33165634e+00 1.94545889e+00
5.25606692e-01 4.71030384e-01 4.37117249e-01 1.01211035e+00
1.15313697e+00 6.21470690e-01 4.63164389e-01 -4.44388300e-01
1.42715192e+00 -1.22496188e+00 -6.53628230e-01 -1.01828635e-01
7.13230491e-01 -4.95150357e-01 1.26552022e+00 2.21959680e-01
-1.05349624e+00 -6.49018049e-01 -1.08849764e+00 4.45791334e-02
-6.99260160e-02 -2.97080636e-01 4.31391269e-01 3.31142694e-01
-6.44766271e-01 4.56637263e-01 -7.46495724e-01 -3.15053016e-01
1.01517093e+00 7.72893503e-02 -2.95460671e-01 -3.32506627e-01
-1.25679076e+00 8.07929993e-01 6.14266992e-01 -5.11006057e-01
-1.34084606e+00 -1.09423912e+00 -8.29353869e-01 -3.13121557e-01
3.69232595e-01 -1.18640316e+00 1.28309643e+00 -1.05848920e+00
-1.32369184e+00 1.13900447e+00 -1.85222283e-03 -3.97863507e-01
6.90623343e-01 3.16439569e-01 -5.58869898e-01 3.30159903e-01
5.40789723e-01 1.11216593e+00 9.54088271e-01 -1.52538061e+00
-4.65674639e-01 -3.84766579e-01 -4.53472108e-01 5.50406635e-01
-2.59627789e-01 -6.46573663e-01 -4.76119876e-01 -7.60116637e-01
4.77522388e-02 -9.02170599e-01 -4.92258519e-01 -1.33583575e-01
-5.65575898e-01 -5.27175702e-03 7.59760320e-01 -4.52539772e-01
8.60382736e-01 -2.20336628e+00 2.03670993e-01 -2.87978768e-01
4.91077363e-01 4.44390953e-01 -1.72278956e-01 -3.67500857e-02
-2.21529260e-01 -2.45071813e-01 -8.93964529e-01 -4.24298227e-01
-3.91329765e-01 3.40993583e-01 -2.92517632e-01 3.67923945e-01
8.38619471e-02 1.11659467e+00 -1.39160955e+00 -9.26824749e-01
2.69403726e-01 2.90752798e-01 -4.58210617e-01 3.16700459e-01
-2.26214319e-01 1.11814392e+00 -5.32332361e-01 8.33222151e-01
7.20498085e-01 -6.48385882e-01 8.46423283e-02 -1.94543272e-01
4.09784287e-01 -3.95418197e-01 -5.71404815e-01 2.47536540e+00
-6.44646883e-01 1.25952512e-01 -2.87339121e-01 -1.12544942e+00
7.79480398e-01 6.43875062e-01 8.84708941e-01 -8.30312848e-01
1.00299940e-01 3.90223235e-01 -1.24220550e-01 -7.28757024e-01
2.42413893e-01 -9.98495638e-01 -1.29076764e-01 4.05412048e-01
4.28579688e-01 -2.13989183e-01 -2.16692742e-02 2.45979384e-01
7.51326561e-01 3.81719880e-02 3.49727631e-01 2.47725379e-02
2.13706151e-01 4.26759124e-01 7.74415851e-01 7.27206647e-01
-5.70957303e-01 1.08945668e+00 1.96963519e-01 5.50676174e-02
-1.24854863e+00 -1.39967537e+00 -2.69902140e-01 9.40897882e-01
3.67799908e-01 2.31040090e-01 -7.85767555e-01 -1.00908744e+00
-2.72903860e-01 7.54933953e-01 -6.71752930e-01 -5.81252337e-01
-1.80800602e-01 -9.49069202e-01 6.44926250e-01 5.79011619e-01
6.62292719e-01 -1.35015023e+00 -4.55060750e-01 4.80913222e-01
-4.42520350e-01 -9.87001419e-01 -7.03699529e-01 2.23140672e-01
-1.21165490e+00 -7.46403277e-01 -1.66151702e+00 -1.08034074e+00
7.55535781e-01 2.22536266e-01 9.94727850e-01 -3.50106239e-01
-3.39528322e-01 3.06340456e-01 -3.23242754e-01 -3.39235157e-01
-6.73923016e-01 7.91094601e-02 -2.53792167e-01 -2.03814581e-02
1.71894640e-01 -5.73328435e-01 -6.59872174e-01 2.29103491e-01
-9.88797188e-01 4.25656021e-01 6.23503149e-01 1.31149673e+00
9.26664174e-01 -3.21803063e-01 1.13362777e+00 -1.42159879e+00
6.34018719e-01 -9.29616332e-01 9.48539600e-02 3.74875069e-01
-5.43615937e-01 -1.63053706e-01 5.75645924e-01 -7.58102775e-01
-1.35035980e+00 1.98002741e-01 -8.11387822e-02 -8.27796876e-01
-1.82816550e-01 3.85639608e-01 -2.02500716e-01 3.89420062e-01
9.99969065e-01 5.70397496e-01 2.75565982e-01 -2.26150781e-01
5.87986350e-01 7.07753122e-01 9.09963965e-01 -5.24145305e-01
3.09402138e-01 7.30842471e-01 -2.96254456e-01 -5.74334085e-01
-1.09366453e+00 -7.24681616e-01 -5.57385802e-01 -3.56209977e-03
1.21279764e+00 -1.19098866e+00 -1.77294642e-01 6.37323320e-01
-8.41397941e-01 -4.82431799e-01 -7.62337983e-01 3.89427751e-01
-9.71178114e-01 3.34590495e-01 -4.60163802e-01 -1.78725049e-01
-4.82861727e-01 -1.52338469e+00 1.12658083e+00 3.88162762e-01
-1.27724379e-01 -1.26841378e+00 2.92299807e-01 5.56462049e-01
1.71837598e-01 4.78464514e-01 8.46025169e-01 -6.79198980e-01
-1.70792952e-01 2.53069133e-01 -2.17952326e-01 3.46725643e-01
2.81587064e-01 -8.51991653e-01 -9.98860836e-01 -3.03219318e-01
-2.95786839e-02 -7.29652882e-01 9.57248449e-01 6.28776789e-01
1.14215660e+00 2.36502945e-01 -3.83399516e-01 7.73420334e-01
1.23814332e+00 1.23235554e-01 4.43329692e-01 4.53209952e-02
7.30527222e-01 4.99417782e-01 7.59419978e-01 4.58432883e-01
4.86785263e-01 3.77202421e-01 7.94958547e-02 -6.49956226e-01
-6.28628552e-01 -4.81485665e-01 -4.64711972e-02 8.23724866e-01
4.03165311e-01 -3.98707278e-02 -9.68129277e-01 9.71829653e-01
-1.90132451e+00 -8.15912485e-01 3.13415855e-01 1.83948469e+00
1.31691170e+00 -1.79873645e-01 2.32648909e-01 -3.13854873e-01
8.45037103e-01 3.54480147e-02 -1.05033505e+00 1.00101657e-01
-3.66671868e-02 -1.91798247e-02 2.63733327e-01 1.47448272e-01
-1.01408899e+00 9.99482989e-01 5.74691820e+00 1.02587342e+00
-1.26575696e+00 8.10015738e-01 7.38762617e-01 9.85790566e-02
-4.56530184e-01 -2.74507821e-01 -2.90814459e-01 5.20517528e-01
6.68259799e-01 -1.81892291e-01 -5.49916923e-03 9.37239587e-01
-8.02326668e-03 1.50433511e-01 -9.15676653e-01 1.18298876e+00
9.20698866e-02 -1.53560507e+00 1.35978013e-01 -1.02442689e-01
1.33389843e+00 3.64655219e-02 3.14088196e-01 6.00591540e-01
3.49192530e-01 -9.24018860e-01 3.28375638e-01 4.81202871e-01
1.24715102e+00 -5.48868895e-01 5.74409544e-01 4.71195310e-01
-7.56351590e-01 1.30043700e-01 -3.99663955e-01 7.26594627e-01
4.97443646e-01 5.22808135e-01 -1.16544080e+00 6.61612451e-01
3.36939484e-01 9.73957121e-01 -2.65902907e-01 1.00018251e+00
8.16529244e-02 4.62504923e-01 2.39284322e-01 5.08463740e-01
3.59656006e-01 6.35786960e-03 6.21163547e-01 1.12453341e+00
2.23021254e-01 2.02932209e-01 3.83204699e-01 1.05322957e+00
-3.02473485e-01 2.86079850e-02 -5.27114272e-01 2.98879314e-02
3.11366171e-01 1.18333805e+00 -7.41391480e-01 -7.50327766e-01
-3.22350740e-01 1.18884718e+00 -5.00964895e-02 4.33499336e-01
-8.34031761e-01 -2.16081917e-01 1.68200254e-01 1.05113454e-01
-5.68338148e-02 3.63727808e-01 -6.95892751e-01 -1.30210245e+00
-4.89170700e-01 -1.04032946e+00 7.30791807e-01 -8.91436815e-01
-1.53282976e+00 6.57535493e-01 1.23787232e-01 -1.69832361e+00
-1.23989135e-01 -9.16356817e-02 -5.63409805e-01 8.80987823e-01
-1.62810671e+00 -1.44051588e+00 -3.82350147e-01 9.32422280e-01
9.40036952e-01 -4.52104509e-01 7.74383307e-01 3.64277184e-01
-2.60447085e-01 6.75404370e-01 4.73577753e-02 -3.15121561e-02
1.07791376e+00 -1.31092894e+00 -5.52045181e-03 4.58381951e-01
-8.92225355e-02 5.19338548e-01 5.65931201e-01 -7.94389129e-01
-8.27078760e-01 -1.35274398e+00 1.67780608e-01 -3.34493101e-01
2.93047100e-01 1.38983577e-01 -1.23915410e+00 4.25545424e-01
2.21203640e-01 3.47300768e-01 8.71862829e-01 -6.88119680e-02
-1.82580277e-01 1.71660960e-01 -1.37841725e+00 5.50642788e-01
9.65987265e-01 -5.66900492e-01 -9.15174425e-01 4.69213039e-01
8.11251760e-01 -5.96115172e-01 -1.05342233e+00 4.64148074e-01
1.73433483e-01 -6.39231980e-01 1.01221740e+00 -8.02479804e-01
6.17076159e-01 -3.49263638e-01 6.86465800e-02 -1.62346160e+00
-1.03413515e-01 -4.44667727e-01 4.82756346e-02 1.05726707e+00
4.54231232e-01 -4.41512376e-01 7.85623908e-01 1.93901435e-01
-6.01046085e-01 -7.77606487e-01 -7.71857023e-01 -6.55130804e-01
3.95847738e-01 -3.40751618e-01 4.87166047e-01 1.28062451e+00
1.72384113e-01 3.67540985e-01 -5.89215755e-01 -1.47072980e-02
7.46574521e-01 3.69243175e-01 5.25122762e-01 -8.53438616e-01
-5.23223758e-01 -2.15233371e-01 1.43928349e-01 -8.04633617e-01
2.57832080e-01 -1.47071147e+00 1.64716274e-01 -1.68173909e+00
7.21934021e-01 -5.01480639e-01 -5.18726051e-01 6.31524444e-01
-3.43676716e-01 4.53726798e-01 5.79710715e-02 5.06521583e-01
-8.51150453e-01 6.51221275e-01 2.13740969e+00 -3.10500860e-01
-3.43101710e-01 7.93141127e-02 -9.39956427e-01 6.16379082e-01
5.39823651e-01 -6.08009219e-01 -7.40628123e-01 -2.48909444e-01
-4.16702718e-01 6.80274546e-01 2.77364135e-01 -8.21534991e-01
1.47079706e-01 -4.75518219e-02 3.15352648e-01 -4.61045414e-01
1.84980109e-01 -5.37174404e-01 8.00422579e-02 4.53668982e-01
-4.95278001e-01 -6.51719749e-01 4.24798205e-02 6.67039156e-01
-3.13309610e-01 -1.32140577e-01 1.23480964e+00 -3.82415116e-01
-1.00853968e+00 6.24194503e-01 3.58155183e-02 6.00104868e-01
1.10651815e+00 -3.08607280e-01 -1.08341351e-01 -2.63374537e-01
-1.36529040e+00 4.11077648e-01 4.51281101e-01 4.27444130e-01
6.18337154e-01 -1.42824423e+00 -7.84714282e-01 7.69427493e-02
4.76265758e-01 2.48431593e-01 9.08160031e-01 1.08519745e+00
-6.40523508e-02 -3.22103277e-02 -5.10183394e-01 -9.56174552e-01
-7.15938151e-01 6.63576007e-01 3.29368979e-01 -3.69912207e-01
-8.08527887e-01 6.98299766e-01 5.37331939e-01 -7.27043092e-01
-5.65922484e-02 1.86246596e-02 1.04420170e-01 1.15404375e-01
4.40097123e-01 2.53489345e-01 -1.38418227e-01 -5.24820328e-01
-1.13560095e-01 4.02709007e-01 -1.75605804e-01 -1.02473728e-01
1.13518417e+00 -1.93371683e-01 3.88201982e-01 4.87755328e-01
1.23400247e+00 -5.38582683e-01 -1.58799160e+00 -6.56158745e-01
-3.10998738e-01 -1.58658683e-01 -5.93586825e-02 -1.08689535e+00
-1.05789328e+00 1.01061869e+00 7.54226446e-01 -3.71085823e-01
1.23537481e+00 2.70822555e-01 1.24256134e+00 -6.52164072e-02
2.48421118e-01 -1.07942808e+00 5.72498024e-01 1.06592536e-01
5.51259100e-01 -1.67121363e+00 -2.36907542e-01 -2.78544247e-01
-1.57794178e+00 6.57313704e-01 5.99714339e-01 1.52169749e-01
4.66140389e-01 2.11874694e-02 3.35719854e-01 -1.72552228e-01
-4.39233214e-01 -3.08583170e-01 2.97849029e-01 9.66465950e-01
1.83678567e-01 2.53727883e-01 2.18544202e-03 7.55734384e-01
2.25728706e-01 2.30529383e-01 3.36586028e-01 6.08554125e-01
-3.30742121e-01 -1.06620347e+00 -2.29775429e-01 3.32107812e-01
-2.72149503e-01 -8.82353820e-03 1.91026360e-01 7.64509082e-01
3.03001106e-01 6.05321348e-01 -2.20372915e-01 -7.32292235e-02
2.78667480e-01 1.59462273e-01 3.90247554e-01 -1.00514078e+00
-3.34805965e-01 3.46107364e-01 -3.99604440e-01 -2.61273712e-01
-4.76178974e-01 -6.10386550e-01 -1.62727010e+00 1.76119968e-01
-5.10468334e-02 -3.63111719e-02 2.71058172e-01 1.05216825e+00
3.40595752e-01 8.46344352e-01 5.17559946e-01 -3.99258792e-01
-6.20455086e-01 -8.55692744e-01 -7.93419242e-01 7.14590013e-01
2.73450881e-01 -5.78678668e-01 1.98656663e-01 3.11785251e-01] | [14.61273193359375, -1.967191219329834] |
27a03d97-f619-405f-8d89-bf038cb1a1f3 | a-transformer-based-feature-segmentation-and | 2201.09206 | null | https://arxiv.org/abs/2201.09206v1 | https://arxiv.org/pdf/2201.09206v1.pdf | A Transformer-Based Feature Segmentation and Region Alignment Method For UAV-View Geo-Localization | Cross-view geo-localization is a task of matching the same geographic image from different views, e.g., unmanned aerial vehicle (UAV) and satellite. The most difficult challenges are the position shift and the uncertainty of distance and scale. Existing methods are mainly aimed at digging for more comprehensive fine-grained information. However, it underestimates the importance of extracting robust feature representation and the impact of feature alignment. The CNN-based methods have achieved great success in cross-view geo-localization. However it still has some limitations, e.g., it can only extract part of the information in the neighborhood and some scale reduction operations will make some fine-grained information lost. In particular, we introduce a simple and efficient transformer-based structure called Feature Segmentation and Region Alignment (FSRA) to enhance the model's ability to understand contextual information as well as to understand the distribution of instances. Without using additional supervisory information, FSRA divides regions based on the heat distribution of the transformer's feature map, and then aligns multiple specific regions in different views one on one. Finally, FSRA integrates each region into a set of feature representations. The difference is that FSRA does not divide regions manually, but automatically based on the heat distribution of the feature map. So that specific instances can still be divided and aligned when there are significant shifts and scale changes in the image. In addition, a multiple sampling strategy is proposed to overcome the disparity in the number of satellite images and that of images from other sources. Experiments show that the proposed method has superior performance and achieves the state-of-the-art in both tasks of drone view target localization and drone navigation. Code will be released at https://github.com/Dmmm1997/FSRA | ['Enhui Zheng', 'Jiedong Zhuang', 'Jianhong Hu', 'Ming Dai'] | 2022-01-23 | null | null | null | null | ['drone-navigation', 'drone-view-target-localization'] | ['computer-vision', 'computer-vision'] | [-2.54644781e-01 -4.90570903e-01 6.31593242e-02 -3.92501801e-01
-6.21890783e-01 -9.56234753e-01 3.53923470e-01 -1.28678143e-01
-2.46164963e-01 4.24928576e-01 1.73357464e-02 -2.11100895e-02
-1.65523127e-01 -9.55876052e-01 -4.97708738e-01 -7.28160262e-01
1.74716279e-01 1.61288053e-01 5.43673813e-01 -4.42982793e-01
2.02259481e-01 6.73243940e-01 -1.70887554e+00 -5.45436256e-02
9.42942083e-01 1.08538771e+00 4.05604571e-01 2.00090691e-01
1.04067199e-01 9.40296575e-02 -5.49880922e-01 1.59305379e-01
6.02657259e-01 -2.01312855e-01 -5.54701686e-01 7.59130791e-02
6.03360236e-01 -3.34725201e-01 -2.08933294e-01 1.42235255e+00
3.91159564e-01 1.14958741e-01 3.81217450e-01 -1.27686512e+00
-2.83064753e-01 8.82230327e-02 -8.51737618e-01 1.84785083e-01
2.23087803e-01 -6.07720725e-02 6.98049963e-01 -7.94974148e-01
5.23632944e-01 8.87043178e-01 7.11136699e-01 1.07633188e-01
-7.38921285e-01 -7.66083777e-01 4.55671072e-01 3.47046182e-02
-1.71439993e+00 -9.44226012e-02 6.18561149e-01 -5.21206498e-01
6.04605079e-01 3.84034127e-01 7.61725903e-01 3.88755143e-01
1.38142213e-01 3.44175369e-01 9.66023743e-01 -4.00240980e-02
-2.17563622e-02 -2.75767944e-03 -2.41739392e-01 7.95730352e-01
2.95751274e-01 2.08539680e-01 -3.18175435e-01 -1.29830581e-03
7.97675610e-01 4.08189863e-01 -6.86888337e-01 -7.01841116e-01
-1.44409657e+00 7.31960773e-01 8.65628839e-01 3.80311042e-01
-1.70461982e-01 -1.92252278e-01 4.29936014e-02 -1.39972437e-02
5.19289076e-01 3.71445209e-01 -5.29136837e-01 1.33766174e-01
-1.07595217e+00 1.45153880e-01 3.17424923e-01 1.11763418e+00
1.24725676e+00 -2.21567862e-02 3.34383011e-01 5.63084066e-01
2.17185602e-01 6.89837158e-01 4.52192008e-01 -5.93066990e-01
5.72044313e-01 9.02179360e-01 3.13679278e-01 -1.44269991e+00
-5.15454173e-01 -5.63303053e-01 -6.59290373e-01 3.16528976e-01
3.47253710e-01 -2.80585527e-01 -1.01104760e+00 1.57224178e+00
4.88190413e-01 -3.61223780e-02 -1.75533444e-01 1.17413485e+00
7.26892531e-01 6.13014579e-01 -5.08668005e-01 2.84742743e-01
1.39405024e+00 -8.34848344e-01 -4.35787290e-01 -6.08273923e-01
4.72032458e-01 -7.63970494e-01 6.60005093e-01 -1.19244866e-01
-4.66380328e-01 -5.73181748e-01 -1.19265461e+00 2.02762350e-01
-8.86688709e-01 4.61918622e-01 4.08388168e-01 2.16202825e-01
-1.01200175e+00 3.00885946e-01 -6.97153032e-01 -5.75407505e-01
2.31571142e-02 3.07993978e-01 -6.19602025e-01 -1.43884167e-01
-1.22640705e+00 9.29144382e-01 3.55356514e-01 3.68951619e-01
-5.74021935e-01 -5.58890700e-01 -1.12044013e+00 1.47172213e-02
4.84052986e-01 -6.35367870e-01 7.83280075e-01 -8.64474773e-01
-9.88909662e-01 6.10073507e-01 -1.82359964e-01 -1.96934074e-01
3.82770956e-01 -9.92423743e-02 -5.82049608e-01 -2.40653437e-02
5.73155880e-01 8.05271327e-01 6.29483521e-01 -1.24029756e+00
-1.36625814e+00 -7.76184440e-01 2.15231106e-01 5.16782343e-01
5.51467650e-02 -2.02612758e-01 -6.63169920e-01 -5.48204720e-01
6.55451298e-01 -1.13616729e+00 -1.62208512e-01 -7.56551474e-02
-1.20199740e-01 2.19572559e-01 9.73971307e-01 -5.79928815e-01
1.22492909e+00 -2.19619489e+00 2.00265627e-02 2.35968634e-01
1.70538634e-01 -1.70340808e-03 6.67101890e-02 4.04722124e-01
3.90114449e-02 1.34483874e-01 -1.75510317e-01 2.28666738e-01
-2.47596934e-01 -6.29439810e-03 -2.55409420e-01 6.45631731e-01
-5.66322915e-02 5.54781377e-01 -9.23113465e-01 -1.46863416e-01
3.59001696e-01 4.32602495e-01 -2.38892987e-01 -1.12563774e-01
2.19222441e-01 6.45379245e-01 -5.48097610e-01 9.27713335e-01
1.10076702e+00 2.52116807e-02 -1.69226959e-01 -5.35111070e-01
-6.74691975e-01 2.33752485e-02 -1.36928141e+00 1.67887270e+00
-3.67015302e-01 6.05165124e-01 2.34703813e-02 -6.71595752e-01
1.01037097e+00 -5.52901924e-02 4.40371960e-01 -5.25704265e-01
5.11093028e-02 2.45197818e-01 -1.09257571e-01 -1.34515211e-01
7.77038217e-01 2.12765902e-01 -3.19203079e-01 1.61981866e-01
-1.31377131e-01 -2.69422382e-01 1.04096182e-01 -1.05795257e-01
5.61874747e-01 1.90400600e-01 4.48559642e-01 -2.13250756e-01
4.28634405e-01 4.25845176e-01 8.61341417e-01 4.27254021e-01
-1.83406085e-01 8.34156394e-01 6.62528500e-02 -5.87035954e-01
-7.81691849e-01 -6.70761645e-01 -6.42204210e-02 7.92442322e-01
8.26458812e-01 -4.35874432e-01 -6.88731134e-01 -7.75698543e-01
6.89271390e-02 3.43113035e-01 -6.85957611e-01 -8.24365243e-02
-3.40255916e-01 -6.64813399e-01 2.61419773e-01 4.90303546e-01
9.24455464e-01 -4.75349754e-01 -9.65345979e-01 -1.07393183e-01
-4.80428576e-01 -1.02227116e+00 -6.58466637e-01 -5.33975624e-02
-8.26247156e-01 -1.22511518e+00 -4.91892606e-01 -5.69569707e-01
9.38521266e-01 1.10384250e+00 7.18814075e-01 -1.72322676e-01
4.69849780e-02 2.61584789e-01 -4.00383741e-01 -2.38525435e-01
3.61601949e-01 2.05099091e-01 4.28188145e-02 -1.08286716e-01
3.56936753e-01 -2.07973227e-01 -7.12624609e-01 9.04120743e-01
-7.46563613e-01 -4.47886400e-02 5.60404241e-01 8.16396654e-01
7.67207325e-01 4.28937614e-01 -2.95368787e-02 -4.73832250e-01
2.05419540e-01 -4.58720028e-01 -9.55742896e-01 2.91690797e-01
-4.63324338e-01 -2.38987058e-01 4.96123433e-01 -6.65762648e-02
-7.88755238e-01 2.99306095e-01 2.45339021e-01 -3.83604258e-01
-3.54454368e-01 5.44668853e-01 -2.66333669e-01 -3.67189020e-01
3.83486450e-01 2.96049416e-01 -2.14505214e-02 -2.91843116e-01
2.31209546e-01 6.11153126e-01 3.26554507e-01 -1.75135154e-02
9.83937442e-01 7.21295893e-01 -2.30130032e-01 -4.84365463e-01
-9.30751145e-01 -7.08604217e-01 -8.86025488e-01 -2.46630996e-01
7.11468756e-01 -1.21668458e+00 -2.50962883e-01 3.68736923e-01
-7.87550390e-01 6.42830059e-02 -7.41875246e-02 6.39485836e-01
-2.29217902e-01 2.60534734e-01 8.28174874e-03 -3.68878245e-01
-7.35229626e-02 -1.38446534e+00 1.19588387e+00 6.33401811e-01
1.61663100e-01 -8.46179545e-01 -8.54318663e-02 1.05446026e-01
4.35296029e-01 2.70983636e-01 2.55659640e-01 -3.86964709e-01
-7.77022660e-01 -4.93892103e-01 -2.40018383e-01 5.58663830e-02
4.81450588e-01 -1.09074466e-01 -8.26277673e-01 -5.01636863e-01
-3.34493890e-02 2.14423656e-01 7.81849682e-01 4.35406297e-01
7.33033657e-01 -8.81271809e-02 -5.37920058e-01 9.14421141e-01
1.52230334e+00 4.39891487e-01 4.44843531e-01 6.61298811e-01
6.89076006e-01 6.66541457e-01 1.25353360e+00 3.20567906e-01
6.17300093e-01 8.72417808e-01 8.79370809e-01 -3.85285407e-01
2.09850058e-01 -3.22395325e-01 1.73910409e-01 3.50283831e-01
-9.58535895e-02 -8.30773339e-02 -8.65890980e-01 6.61446214e-01
-1.91943526e+00 -8.56620610e-01 8.26619193e-02 2.39109826e+00
8.01320672e-02 -2.98489153e-01 -7.01062903e-02 -3.06982428e-01
1.07269323e+00 3.84660333e-01 -5.41139424e-01 2.82855958e-01
-1.36117265e-01 -4.53008235e-01 1.06538916e+00 5.11980891e-01
-1.37229705e+00 1.09181118e+00 5.25528765e+00 7.25048840e-01
-1.43126774e+00 -2.21170262e-01 2.46769071e-01 1.79008439e-01
-2.80699451e-02 1.53949067e-01 -1.10151303e+00 5.12851775e-01
1.63044795e-01 7.01792613e-02 3.09173673e-01 9.79010701e-01
6.51634187e-02 -5.26454687e-01 -5.29744506e-01 1.04657769e+00
2.19672248e-01 -1.15336597e+00 -4.84407619e-02 2.39301741e-01
6.67638838e-01 2.48230815e-01 4.45873737e-02 -6.70853481e-02
6.60658181e-02 -6.78861260e-01 8.97808433e-01 3.39375317e-01
7.91196764e-01 -8.03037822e-01 1.12051165e+00 3.32486689e-01
-1.70913815e+00 -1.59847781e-01 -5.19009888e-01 1.11744158e-01
3.35235596e-02 4.28713024e-01 -6.83992088e-01 1.00179911e+00
1.06853175e+00 8.13383698e-01 -7.70465851e-01 1.12230015e+00
-2.51671910e-01 1.63179487e-02 -4.55674946e-01 3.18025410e-01
4.18710470e-01 -4.65492547e-01 4.94888693e-01 7.94095397e-01
8.74318659e-01 -3.72654051e-02 3.97872359e-01 6.09897852e-01
4.81172532e-01 -1.39602616e-01 -9.58396435e-01 4.05637980e-01
6.16672754e-01 1.48071003e+00 -8.57340991e-01 -3.20932537e-01
-4.54288304e-01 7.95120060e-01 1.39999958e-02 3.45821232e-01
-9.26311016e-01 -5.61446548e-01 8.01610172e-01 2.79908836e-01
6.27174735e-01 -3.82401705e-01 -4.45274822e-02 -1.35226345e+00
2.21927688e-02 -6.60543621e-01 4.20592934e-01 -9.34022963e-01
-7.76492059e-01 9.71187115e-01 4.83930483e-02 -1.95433640e+00
-2.39314362e-01 -3.51440012e-01 -3.45920175e-01 9.16618168e-01
-1.56842804e+00 -1.30372453e+00 -8.83959949e-01 5.83732069e-01
5.39902568e-01 -4.10605408e-02 5.22867382e-01 2.84052759e-01
-4.85071063e-01 3.01262558e-01 1.15200363e-01 1.89347893e-01
8.93628240e-01 -1.02819097e+00 2.69985676e-01 1.21653914e+00
6.21165931e-02 5.96451819e-01 5.36924243e-01 -6.76081717e-01
-9.87458706e-01 -1.23947406e+00 5.71498930e-01 -2.82605201e-01
4.75845516e-01 -2.25749105e-01 -5.63941360e-01 7.69715011e-01
-6.40181899e-02 2.14539811e-01 3.99783671e-01 -2.38990203e-01
-2.06936419e-01 -4.18882579e-01 -1.16628301e+00 4.07785922e-01
8.90425205e-01 -4.01365608e-01 -3.86199445e-01 -3.49365585e-02
4.12798762e-01 -7.84944952e-01 -7.92263508e-01 6.84882224e-01
5.53381383e-01 -1.27296209e+00 8.75615060e-01 -1.32685900e-02
-6.57098368e-02 -1.12755716e+00 -3.19387317e-01 -1.74785054e+00
-4.64896500e-01 -5.91119640e-02 5.88090956e-01 1.25163603e+00
3.24145287e-01 -1.04138994e+00 5.21616042e-01 2.66390681e-01
-2.26887017e-01 -6.60533011e-01 -7.33261049e-01 -6.75286889e-01
-4.44030225e-01 -3.13138179e-02 9.74887848e-01 9.94617462e-01
-3.29992205e-01 5.04657030e-02 -3.55914742e-01 7.53199697e-01
2.97829449e-01 6.52908504e-01 8.82940054e-01 -1.16814792e+00
3.42287481e-01 -1.07027590e-01 -5.97600162e-01 -1.07022786e+00
-3.35390925e-01 -4.35210854e-01 3.54930982e-02 -1.57587099e+00
-9.17482153e-02 -5.16020894e-01 -6.14476502e-02 4.75729734e-01
-6.61127567e-02 3.01789701e-01 2.35476404e-01 3.90375167e-01
-5.26319444e-01 4.38170016e-01 1.08516753e+00 -6.31814543e-03
-1.91012636e-01 2.07828194e-01 -6.00253701e-01 8.74580562e-01
1.02979493e+00 -4.55791473e-01 -2.96884239e-01 -6.08928502e-01
1.08881950e-01 -6.16827235e-02 4.60944355e-01 -1.24122500e+00
4.83388335e-01 -1.33078381e-01 6.26391530e-01 -1.03677869e+00
2.17217281e-01 -1.09607494e+00 4.26070660e-01 3.08406204e-01
3.71041685e-01 5.14822960e-01 2.24764913e-01 5.12105405e-01
-7.53852785e-01 -8.07823390e-02 6.57236099e-01 -4.28452194e-01
-1.17811501e+00 4.53850120e-01 -2.09177539e-01 -1.49277195e-01
1.01645279e+00 -5.70016265e-01 -4.14742082e-01 -4.04977143e-01
-4.58875626e-01 3.54679495e-01 1.01907218e+00 4.90692705e-01
5.66048682e-01 -1.29733515e+00 -2.49136880e-01 4.67788994e-01
4.39451396e-01 3.59561622e-01 5.18609941e-01 9.78915334e-01
-6.18977010e-01 5.53569615e-01 -3.94024849e-01 -7.38747120e-01
-1.29512656e+00 3.60862970e-01 5.39232969e-01 -1.14135392e-01
-4.12667304e-01 5.09493649e-01 5.96119583e-01 -6.81037068e-01
-1.47836879e-01 -5.33260524e-01 -3.35941553e-01 3.60962361e-01
5.23500085e-01 2.07522511e-01 6.49688393e-02 -1.05532360e+00
-5.17282784e-01 1.33320689e+00 1.23022899e-01 9.05949995e-02
1.13238144e+00 -5.44094443e-01 -1.10591844e-01 1.28947616e-01
9.12827671e-01 3.09531242e-01 -1.52179503e+00 -1.24216907e-01
-2.72493571e-01 -7.59580851e-01 1.01272099e-01 -6.74289465e-01
-1.31645048e+00 8.24197710e-01 7.94834316e-01 1.33655757e-01
1.23647702e+00 -1.42735153e-01 6.41474843e-01 2.56429799e-02
6.74511075e-01 -1.05829799e+00 -3.96178544e-01 5.99222541e-01
8.49161923e-01 -1.35505331e+00 2.32684761e-01 -5.82824469e-01
-7.74926066e-01 1.21765947e+00 7.89357066e-01 -1.16485627e-02
5.72722793e-01 8.30842108e-02 3.37403953e-01 -3.49156410e-01
-8.33441764e-02 -5.33313632e-01 4.03785825e-01 6.60467029e-01
-3.52797657e-02 5.81898540e-02 1.38335273e-01 3.02130044e-01
-3.16209078e-01 -3.55470419e-01 2.94710189e-01 9.04898345e-01
-5.80699801e-01 -6.52653635e-01 -5.98997295e-01 6.87687173e-02
-1.65171042e-01 -9.06760395e-02 -3.19041222e-01 1.09455156e+00
5.51009119e-01 7.50950873e-01 2.43149742e-01 -7.59617209e-01
4.32556242e-01 -3.50502223e-01 1.11577757e-01 -4.82637227e-01
-2.91685134e-01 1.45930067e-01 -1.28994361e-01 -6.82040274e-01
-5.07664144e-01 -4.81531799e-01 -1.05741203e+00 -1.46913752e-01
-6.09144747e-01 2.53505945e-01 8.21583927e-01 6.82530284e-01
6.87006593e-01 3.96902502e-01 8.15537751e-01 -1.06474400e+00
-1.67835861e-01 -7.82778025e-01 -5.54930866e-01 -5.76981567e-02
3.66480052e-01 -9.94644225e-01 -5.15258312e-01 -2.58858114e-01] | [7.7866621017456055, -1.9335157871246338] |
d3faa150-5bce-444c-a60a-e3caf923becb | intra-and-inter-action-understanding-via | 2005.10229 | null | https://arxiv.org/abs/2005.10229v1 | https://arxiv.org/pdf/2005.10229v1.pdf | Intra- and Inter-Action Understanding via Temporal Action Parsing | Current methods for action recognition primarily rely on deep convolutional networks to derive feature embeddings of visual and motion features. While these methods have demonstrated remarkable performance on standard benchmarks, we are still in need of a better understanding as to how the videos, in particular their internal structures, relate to high-level semantics, which may lead to benefits in multiple aspects, e.g. interpretable predictions and even new methods that can take the recognition performances to a next level. Towards this goal, we construct TAPOS, a new dataset developed on sport videos with manual annotations of sub-actions, and conduct a study on temporal action parsing on top. Our study shows that a sport activity usually consists of multiple sub-actions and that the awareness of such temporal structures is beneficial to action recognition. We also investigate a number of temporal parsing methods, and thereon devise an improved method that is capable of mining sub-actions from training data without knowing the labels of them. On the constructed TAPOS, the proposed method is shown to reveal intra-action information, i.e. how action instances are made of sub-actions, and inter-action information, i.e. one specific sub-action may commonly appear in various actions. | ['Dahua Lin', 'Dian Shao', 'Bo Dai', 'Yue Zhao'] | 2020-05-20 | intra-and-inter-action-understanding-via-1 | http://openaccess.thecvf.com/content_CVPR_2020/html/Shao_Intra-_and_Inter-Action_Understanding_via_Temporal_Action_Parsing_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Shao_Intra-_and_Inter-Action_Understanding_via_Temporal_Action_Parsing_CVPR_2020_paper.pdf | cvpr-2020-6 | ['action-understanding', 'action-parsing'] | ['computer-vision', 'natural-language-processing'] | [ 3.21058124e-01 8.58191699e-02 -5.48087180e-01 -4.83343422e-01
-2.05683812e-01 -4.30062383e-01 6.89444304e-01 9.53114331e-02
-9.77143198e-02 4.75629866e-01 6.42795682e-01 1.13149509e-01
-1.58395842e-01 -7.19234169e-01 -8.95083964e-01 -5.49096823e-01
-4.41362530e-01 2.46231437e-01 6.22516513e-01 -1.17507629e-01
2.84210622e-01 4.08157021e-01 -1.80287385e+00 7.54857838e-01
2.29133114e-01 9.90946829e-01 3.73914503e-02 6.55239820e-01
-2.27333501e-01 1.19126213e+00 -5.80971837e-01 -3.76166850e-01
2.70618796e-02 -6.17641091e-01 -1.08488679e+00 5.24921596e-01
4.94228274e-01 -3.81994784e-01 -2.95676261e-01 7.04374492e-01
-1.42322853e-01 1.17556624e-01 4.14492816e-01 -1.37688851e+00
-2.95837700e-01 6.98236883e-01 -1.87911436e-01 4.03032601e-01
3.26084167e-01 2.59343684e-01 1.19897950e+00 -4.56457853e-01
8.68709683e-01 1.22087753e+00 6.59626722e-01 5.17879188e-01
-9.06674027e-01 -3.72894794e-01 4.16201830e-01 7.37968028e-01
-7.76621521e-01 -3.08166385e-01 7.49342263e-01 -6.83099270e-01
9.08843458e-01 3.99116576e-02 8.19332719e-01 1.39354455e+00
3.57777596e-01 1.09356022e+00 9.75021958e-01 -3.32600951e-01
1.96867928e-01 -1.52906284e-01 3.04166496e-01 5.56676984e-01
1.02706635e-02 -7.90118426e-02 -8.86560261e-01 3.73017341e-01
6.53993487e-01 6.40582591e-02 -3.61813195e-02 -4.82883483e-01
-1.11936712e+00 6.28163159e-01 1.54783890e-01 5.85605681e-01
-4.05380726e-01 3.04617792e-01 6.78331256e-01 -1.06014814e-02
1.51494652e-01 1.28680930e-01 -5.79076946e-01 -6.48264527e-01
-5.93028724e-01 1.56589344e-01 6.67158842e-01 6.93087339e-01
8.12114537e-01 -2.58381039e-01 -1.09726414e-01 5.54425776e-01
-1.46624282e-01 -2.17637107e-01 6.72877848e-01 -1.08824646e+00
6.28391922e-01 9.64754999e-01 8.06330517e-02 -9.96172130e-01
-5.20449877e-01 -4.52181958e-02 -4.07753050e-01 1.38778508e-01
5.11783302e-01 2.22994700e-01 -7.53856361e-01 1.77429867e+00
2.71382540e-01 3.65221053e-01 6.20773621e-02 8.43840718e-01
5.72701454e-01 3.74011517e-01 1.67062744e-01 1.54535342e-02
1.50719595e+00 -9.49451327e-01 -7.31898248e-01 -4.21031028e-01
1.04635477e+00 -3.97666246e-01 9.48664486e-01 3.67055923e-01
-7.68307209e-01 -7.91373849e-01 -9.06155884e-01 -1.41631672e-02
-4.25312042e-01 2.86114186e-01 7.14965820e-01 3.88032436e-01
-5.54252088e-01 8.64923358e-01 -1.17530584e+00 -6.23001933e-01
5.30218720e-01 2.21982017e-01 -6.25886202e-01 5.26899062e-02
-1.07166052e+00 7.75151968e-01 6.87603951e-01 1.56692192e-01
-9.45073545e-01 -3.12674046e-01 -9.31647241e-01 -1.20564595e-01
8.58227491e-01 -1.36896580e-01 1.20240724e+00 -1.41839433e+00
-1.18847156e+00 7.33363032e-01 -2.12826192e-01 -7.33961463e-01
3.70511174e-01 -4.59517390e-01 -6.01270080e-01 3.48602444e-01
1.38612032e-01 5.81928015e-01 6.64702475e-01 -8.86294007e-01
-9.15892541e-01 -4.02996480e-01 6.24151468e-01 4.49446589e-02
-4.18998837e-01 7.07004219e-02 -5.49607337e-01 -5.55104434e-01
5.97687252e-03 -1.08560014e+00 -3.33949253e-02 -2.35822070e-02
-4.86449629e-01 -4.06748414e-01 7.03158200e-01 -6.05962932e-01
1.26961315e+00 -2.12759495e+00 1.12517230e-01 -1.80995062e-01
6.45339265e-02 3.53109241e-01 -4.51020338e-02 6.81907475e-01
-2.23886862e-01 1.39386743e-01 -3.58516723e-02 -2.83537358e-02
2.53270497e-03 6.49534941e-01 -1.00895859e-01 3.18924993e-01
4.58406061e-01 7.43136883e-01 -8.19367349e-01 -5.37119448e-01
3.69028717e-01 1.29220054e-01 -4.80919629e-01 1.74568802e-01
-2.73158193e-01 5.87172568e-01 -6.93518043e-01 5.39708018e-01
1.05926514e-01 -1.60873160e-01 3.75285894e-01 -2.19888017e-01
-1.20138124e-01 4.29054320e-01 -1.19679582e+00 1.52152729e+00
-2.32945696e-01 6.97261751e-01 -3.68510365e-01 -1.46979105e+00
6.84380114e-01 1.95164368e-01 8.25508714e-01 -6.16928160e-01
3.87405045e-02 -6.25812039e-02 8.79212320e-02 -1.04568994e+00
5.27505755e-01 7.86634013e-02 -1.11855567e-01 3.83325279e-01
5.15184738e-03 7.07201600e-01 5.86786509e-01 1.44727165e-02
1.38009322e+00 5.81079185e-01 4.18007106e-01 1.35043740e-01
5.93162715e-01 2.69413352e-01 8.90086114e-01 6.59566462e-01
-3.34499270e-01 4.30969268e-01 9.40808415e-01 -8.75637889e-01
-9.01669085e-01 -5.79580545e-01 1.75595328e-01 1.27648759e+00
2.96778470e-01 -8.04232478e-01 -5.39117634e-01 -8.64438951e-01
-2.32972443e-01 5.02463698e-01 -1.02717984e+00 -2.07710326e-01
-9.27888155e-01 -4.38419253e-01 4.79113340e-01 1.02804279e+00
4.79072213e-01 -1.31043160e+00 -1.12030995e+00 3.07319015e-01
-1.82421356e-01 -1.51790786e+00 -1.77241310e-01 2.39326209e-01
-9.88237679e-01 -1.54893172e+00 -1.69882953e-01 -5.91763973e-01
3.21619481e-01 4.84996736e-02 1.18639791e+00 8.32134113e-02
-2.31288806e-01 5.15144706e-01 -8.81625056e-01 -2.08861679e-01
-4.79781985e-01 -1.40129894e-01 -9.12252143e-02 3.03212613e-01
7.26107717e-01 -5.07223248e-01 -5.78448176e-01 5.11310041e-01
-9.70355809e-01 1.74867585e-01 7.04009473e-01 5.82399487e-01
5.63082576e-01 3.83909605e-02 2.64280867e-02 -8.48770499e-01
8.62638280e-02 -3.72329563e-01 -1.08091615e-01 3.99802804e-01
-2.25280851e-01 2.49240696e-01 5.60888410e-01 -5.53514421e-01
-8.80467594e-01 2.86687762e-01 -3.14493999e-02 -4.97326493e-01
-5.46297550e-01 4.83102143e-01 -1.97869614e-01 4.57688749e-01
5.36266923e-01 3.70998472e-01 -1.67776689e-01 -5.90086758e-01
2.58566201e-01 4.18755442e-01 4.01353866e-01 -5.67939699e-01
4.19366777e-01 5.88730454e-01 4.92674001e-02 -7.25608349e-01
-9.11444068e-01 -5.20116448e-01 -1.11706817e+00 -4.31057155e-01
1.21705544e+00 -6.65072799e-01 -4.78701800e-01 3.21090251e-01
-1.00907052e+00 -3.93635571e-01 -2.31556728e-01 5.58743358e-01
-7.67295122e-01 4.83076841e-01 -5.32378733e-01 -6.32955909e-01
3.65336180e-01 -1.12708235e+00 1.05539119e+00 3.42474543e-02
-4.24247861e-01 -8.63650799e-01 -1.99471065e-03 4.85906243e-01
-2.27243632e-01 2.32645854e-01 8.62593889e-01 -8.78212333e-01
-5.77611625e-01 -7.81779662e-02 9.61053446e-02 4.37270761e-01
2.14875251e-01 -5.86013915e-03 -7.68041313e-01 7.16988966e-02
-2.79045969e-01 -3.81663203e-01 7.31477022e-01 3.50003034e-01
1.37418425e+00 -1.25318319e-01 -3.62017423e-01 2.55605847e-01
1.13769650e+00 3.03104520e-01 9.06192183e-01 5.07649899e-01
5.97728670e-01 8.70223343e-01 9.19985294e-01 4.39926952e-01
2.67668009e-01 1.11908960e+00 6.89905345e-01 3.66419941e-01
-2.04295292e-01 -2.63296515e-01 7.47673333e-01 4.59931314e-01
-3.98620665e-01 -9.90824252e-02 -6.61776423e-01 5.90646088e-01
-2.01966619e+00 -1.11286020e+00 -2.82683402e-01 1.97727633e+00
5.85431993e-01 2.96390623e-01 4.45871860e-01 2.10422814e-01
6.78158820e-01 4.29817080e-01 -6.32075250e-01 -4.68287528e-01
1.82771862e-01 9.02369767e-02 3.82342696e-01 -9.07475129e-02
-1.29865491e+00 9.65262353e-01 5.86800241e+00 5.77637136e-01
-1.01988542e+00 -7.44497851e-02 3.53543162e-01 1.43764943e-01
1.76268592e-01 9.68769118e-02 -9.17400002e-01 4.15271908e-01
8.11465025e-01 1.60451159e-01 4.16594977e-03 9.86520469e-01
3.35555285e-01 -1.32877082e-01 -1.41335881e+00 6.98026955e-01
1.62413120e-01 -1.22924495e+00 2.66841259e-02 6.50508553e-02
2.50316679e-01 -4.74298865e-01 -4.62882012e-01 4.49252576e-01
1.49006499e-02 -8.81546378e-01 8.73446345e-01 5.14407158e-01
2.34325618e-01 -4.18315619e-01 8.34824622e-01 3.94952118e-01
-1.40970266e+00 -2.90272802e-01 -2.25752503e-01 -3.17545772e-01
1.08098708e-01 9.88611057e-02 -5.91832280e-01 6.80633426e-01
8.73731673e-01 1.25787497e+00 -7.52661407e-01 7.88382590e-01
-3.89152318e-01 6.39845073e-01 1.14194699e-01 6.09740056e-02
3.99782389e-01 -2.31943175e-01 3.06375593e-01 1.06588352e+00
3.07648540e-01 1.05132759e-02 2.92456746e-01 5.13742387e-01
2.84832478e-01 1.57812815e-02 -6.79548085e-01 -4.47641134e-01
3.17774229e-02 1.06423533e+00 -8.32667768e-01 -4.68629628e-01
-6.83309078e-01 8.22149873e-01 2.50748605e-01 6.58631027e-02
-1.09699607e+00 8.80096778e-02 9.44892943e-01 2.16950759e-01
7.15176642e-01 -2.16724396e-01 1.04077525e-01 -1.19500065e+00
2.99501359e-01 -7.96464384e-01 5.97301722e-01 -7.21661627e-01
-9.46358383e-01 3.50590587e-01 1.61157712e-01 -1.56267679e+00
-4.45129186e-01 -7.16186583e-01 -5.77576160e-01 8.48633349e-02
-1.18070257e+00 -1.12507570e+00 -2.75748760e-01 4.44457829e-01
9.31685925e-01 4.36807647e-02 5.93223035e-01 1.16292000e-01
-6.77166224e-01 3.87567282e-01 -3.97080094e-01 3.83606583e-01
4.01890218e-01 -9.65222657e-01 4.30215411e-02 9.28392410e-01
6.68987811e-01 2.42277771e-01 7.37648964e-01 -5.92768252e-01
-1.25760639e+00 -9.54052389e-01 7.70865262e-01 -4.83098090e-01
9.04202938e-01 -2.06681103e-01 -9.89265621e-01 9.48633969e-01
1.47392033e-02 5.17834537e-02 6.59984052e-01 2.28666469e-01
-2.18823358e-01 -1.05595998e-01 -5.33797622e-01 5.56207299e-01
1.52494383e+00 -4.55394268e-01 -1.00635254e+00 2.51665473e-01
3.09603900e-01 -1.91174731e-01 -7.72630274e-01 3.97998840e-01
7.46614754e-01 -1.42854226e+00 8.98270667e-01 -1.15323961e+00
9.25578773e-01 -1.90210015e-01 -6.59443140e-02 -9.57016170e-01
-7.96549618e-02 -1.81949958e-01 -2.89801687e-01 1.18484592e+00
-5.08192144e-02 -1.38921753e-01 1.04417133e+00 4.78313416e-01
-4.51950401e-01 -7.34482169e-01 -9.66610551e-01 -1.09579086e+00
-2.59956568e-01 -8.17296982e-01 4.33277428e-01 8.17180097e-01
-1.25023052e-01 3.77988964e-02 -7.22635508e-01 7.34261423e-02
1.61837801e-01 3.16976577e-01 1.01498473e+00 -1.04937363e+00
-3.26917708e-01 -2.78389424e-01 -8.97206187e-01 -1.16035593e+00
2.69959718e-01 -5.09568095e-01 1.12141274e-01 -1.43387616e+00
7.25241080e-02 -3.30515206e-02 -4.23577845e-01 7.57035673e-01
1.13857344e-01 1.59383133e-01 1.44118026e-01 1.17276832e-01
-7.99526811e-01 3.77012789e-01 1.18694842e+00 -1.59829885e-01
-9.93571803e-02 1.11157909e-01 -3.43127191e-01 1.01436257e+00
7.26931512e-01 -5.24194360e-01 -3.87912214e-01 -3.25884879e-01
1.18417040e-01 9.62994769e-02 5.11158645e-01 -1.30623221e+00
-5.56169488e-02 -4.24980462e-01 1.81834653e-01 -4.64882702e-01
4.11090583e-01 -8.94097686e-01 1.91414565e-01 5.51157832e-01
-5.44083893e-01 -5.19836545e-02 -5.96468560e-02 8.42792213e-01
-5.21841764e-01 -4.09933686e-01 3.72014731e-01 -2.78596967e-01
-1.44234562e+00 1.90325141e-01 -5.80980122e-01 -3.22955735e-02
1.24200571e+00 -6.41034782e-01 -6.85665905e-02 -3.25671822e-01
-1.04070044e+00 -1.55576868e-02 3.44612181e-01 7.48057783e-01
3.21743220e-01 -1.28420401e+00 -2.23847643e-01 -2.31234040e-02
4.48331654e-01 -4.53178853e-01 2.54548341e-01 1.05749536e+00
-3.58656973e-01 2.17116043e-01 -4.65431213e-01 -6.47048593e-01
-1.59853554e+00 4.66892838e-01 1.84201181e-01 -3.81023198e-01
-1.03978562e+00 5.09040713e-01 2.54156172e-01 1.09794810e-01
2.94044197e-01 -6.79810286e-01 -6.41070545e-01 2.54257917e-01
5.12805581e-01 1.68375835e-01 -1.44830018e-01 -9.31956470e-01
-4.99389470e-01 6.84057713e-01 5.15103713e-02 3.50152850e-01
1.46037722e+00 2.40337644e-02 1.26263440e-01 7.18716145e-01
1.05670965e+00 -2.35498473e-01 -1.55906188e+00 -3.98146920e-02
4.09241319e-01 -6.16847575e-01 -3.95213544e-01 -5.32067060e-01
-1.11597049e+00 9.67840970e-01 5.77808440e-01 3.51611346e-01
1.18304157e+00 3.63772362e-01 9.08140481e-01 3.31229568e-01
3.96665186e-01 -1.25091994e+00 4.84573007e-01 4.98729497e-01
6.48648620e-01 -1.19532704e+00 2.43020710e-02 -3.96923602e-01
-8.84311497e-01 1.48776138e+00 5.67615271e-01 -1.26274347e-01
3.74789417e-01 -5.98969907e-02 -1.51950926e-01 -4.09592360e-01
-7.71432459e-01 -4.45530891e-01 3.06931108e-01 4.76903319e-01
2.35216141e-01 8.34824964e-02 -4.57208872e-01 6.65844202e-01
6.08334132e-02 2.46160105e-01 5.25908530e-01 1.11126959e+00
-3.43019664e-01 -1.38214707e+00 -2.18448024e-02 4.42784637e-01
-5.31332195e-01 4.08457547e-01 -5.57075918e-01 1.11253572e+00
5.97911537e-01 7.06561148e-01 1.08453162e-01 -7.11713076e-01
3.84295315e-01 2.54798919e-01 4.83013242e-01 -6.95471048e-01
-4.33529824e-01 -1.41888574e-01 4.88888502e-01 -1.07870090e+00
-1.00503910e+00 -9.07637119e-01 -1.16170895e+00 9.17105153e-02
8.23563039e-02 -1.08255342e-01 4.11556661e-01 1.37338305e+00
1.88178673e-01 5.02353013e-01 3.46365094e-01 -7.25868165e-01
-2.77805656e-01 -8.65596235e-01 -6.26352429e-01 8.89055669e-01
-1.77512094e-02 -9.71847177e-01 -2.41534099e-01 3.42931509e-01] | [8.309096336364746, 0.5773602724075317] |
84a12f9b-15d4-44b4-b749-9350c37f00aa | seeing-without-looking-analysis-pipeline-for | 2204.14110 | null | https://arxiv.org/abs/2204.14110v1 | https://arxiv.org/pdf/2204.14110v1.pdf | Seeing without Looking: Analysis Pipeline for Child Sexual Abuse Datasets | The online sharing and viewing of Child Sexual Abuse Material (CSAM) are growing fast, such that human experts can no longer handle the manual inspection. However, the automatic classification of CSAM is a challenging field of research, largely due to the inaccessibility of target data that is - and should forever be - private and in sole possession of law enforcement agencies. To aid researchers in drawing insights from unseen data and safely providing further understanding of CSAM images, we propose an analysis template that goes beyond the statistics of the dataset and respective labels. It focuses on the extraction of automatic signals, provided both by pre-trained machine learning models, e.g., object categories and pornography detection, as well as image metrics such as luminance and sharpness. Only aggregated statistics of sparse signals are provided to guarantee the anonymity of children and adolescents victimized. The pipeline allows filtering the data by applying thresholds to each specified signal and provides the distribution of such signals within the subset, correlations between signals, as well as a bias evaluation. We demonstrated our proposal on the Region-based annotated Child Pornography Dataset (RCPD), one of the few CSAM benchmarks in the literature, composed of over 2000 samples among regular and CSAM images, produced in partnership with Brazil's Federal Police. Although noisy and limited in several senses, we argue that automatic signals can highlight important aspects of the overall distribution of data, which is valuable for databases that can not be disclosed. Our goal is to safely publicize the characteristics of CSAM datasets, encouraging researchers to join the field and perhaps other institutions to provide similar reports on their benchmarks. | ['Jefersson A. dos Santos', 'Sandra Avila', 'João Macedo', 'Camila Laranjeira'] | 2022-04-29 | null | null | null | null | ['pornography-detection'] | ['computer-vision'] | [ 5.27022421e-01 2.86703020e-01 -2.54345745e-01 -7.18795538e-01
-9.24596608e-01 -9.80570138e-01 3.15261573e-01 8.25396419e-01
-3.14541519e-01 3.58373582e-01 2.32275948e-01 -1.52480245e-01
-1.99762970e-01 -7.07237840e-01 -6.38208807e-01 -7.15212107e-01
-4.08170044e-01 2.24672243e-01 7.42817968e-02 1.20768055e-01
3.06757033e-01 7.88743734e-01 -1.41651917e+00 5.35409689e-01
6.00654066e-01 1.08941650e+00 -3.37437600e-01 2.70772994e-01
7.80687183e-02 3.16024125e-01 -6.19171441e-01 -1.06508934e+00
2.36112431e-01 -1.14211276e-01 -5.06824195e-01 -9.85148326e-02
7.85003304e-01 -6.16989315e-01 -1.51252538e-01 1.22076762e+00
3.48906219e-01 -2.15120703e-01 7.00410008e-01 -1.41040516e+00
-5.37638903e-01 6.44352853e-01 -7.73796618e-01 3.76392037e-01
4.04898852e-01 3.25401127e-01 8.49668920e-01 -6.13433659e-01
8.90663624e-01 1.12191391e+00 6.16375268e-01 4.11342353e-01
-1.52469277e+00 -9.47898030e-01 -1.91086486e-01 1.90540224e-01
-1.26828349e+00 -6.23238087e-01 7.04830289e-01 -6.42711163e-01
3.19004834e-01 4.36276436e-01 5.90357363e-01 1.50086939e+00
-3.40576530e-01 6.60578370e-01 1.12254846e+00 -3.31812650e-02
2.43028760e-01 3.20227236e-01 3.64328101e-02 3.71989191e-01
4.95930761e-01 -1.82189733e-01 -7.41774201e-01 -4.81070459e-01
3.68459702e-01 -1.23371184e-01 -2.14167416e-01 -4.35342342e-01
-5.51079333e-01 8.88802290e-01 7.30802938e-02 3.13739955e-01
-3.17892551e-01 -3.84672642e-01 5.22606909e-01 1.23066366e-01
5.41319132e-01 2.47475058e-01 -1.13576196e-01 -3.60892236e-01
-1.30270529e+00 1.62379295e-01 4.41748738e-01 7.33616590e-01
5.61298132e-01 -4.03276294e-01 5.01929373e-02 8.55735242e-01
2.87232678e-02 5.28448999e-01 -6.36960939e-02 -9.13790584e-01
5.93761921e-01 6.53770983e-01 -1.84307739e-01 -1.53800535e+00
-3.44487697e-01 -2.16132954e-01 -5.45660794e-01 1.12468354e-01
6.39945149e-01 6.89911246e-02 -5.51452816e-01 1.62433672e+00
3.74876469e-01 -2.55759414e-02 -3.94231766e-01 8.50642681e-01
9.31772530e-01 2.15163410e-01 -4.61630113e-02 -1.37308195e-01
1.28646529e+00 1.66843161e-02 -6.12276733e-01 1.71138436e-01
3.00755501e-01 -6.16915524e-01 5.96574605e-01 7.04997420e-01
-1.11810946e+00 6.90573007e-02 -8.21239233e-01 1.46648079e-01
-4.42819148e-01 -3.44940610e-02 6.65132761e-01 1.12979901e+00
-7.59311140e-01 6.81518197e-01 -8.59981298e-01 -5.53043067e-01
1.24174058e+00 2.50036716e-01 -8.32518816e-01 -3.13943505e-01
-7.35985756e-01 5.99467635e-01 -1.22368392e-02 1.14295535e-01
-8.63744140e-01 -8.11968684e-01 -9.25822616e-01 5.22988383e-03
2.01516479e-01 3.18007857e-01 5.87617457e-01 -8.46923828e-01
-7.05999076e-01 1.36039722e+00 1.14653014e-01 -3.92927229e-01
5.24996758e-01 1.18469834e-01 -5.19812286e-01 6.73367441e-01
3.48603427e-01 4.65829134e-01 9.93295133e-01 -1.23437440e+00
-6.49316669e-01 -7.15856075e-01 -2.00897187e-01 -4.47856814e-01
-5.86526096e-01 4.36714232e-01 -4.13599223e-01 -5.59481978e-01
2.04599291e-01 -5.72965324e-01 2.09499642e-01 2.69407779e-01
-3.96564305e-01 3.14785480e-01 8.11330974e-01 -1.15769434e+00
1.01089585e+00 -2.63171792e+00 -5.09536982e-01 5.71523488e-01
2.94806242e-01 2.07419977e-01 -2.95077085e-01 4.64027971e-01
-1.70733005e-01 2.56079286e-01 -1.98682532e-01 -3.59280467e-01
-1.83838934e-01 8.17540754e-03 -2.92930573e-01 1.04120004e+00
4.80636150e-01 5.20063341e-01 -9.75170255e-01 -5.09186149e-01
3.18640098e-02 3.93311203e-01 -5.44391274e-01 2.64499281e-02
2.88641363e-01 5.42672336e-01 -2.03147173e-01 9.94669557e-01
1.29213333e+00 2.27802739e-01 8.48551914e-02 -1.77158266e-01
-3.63045372e-02 2.43932098e-01 -1.25317788e+00 1.45635140e+00
-6.30242331e-03 9.18414474e-01 7.64223337e-01 -1.05615973e+00
9.35473502e-01 1.99527070e-01 7.50939906e-01 -6.70488656e-01
1.10078223e-01 4.02156472e-01 2.68219728e-02 -5.99731326e-01
3.58727187e-01 3.53499502e-01 1.58570796e-01 4.02354330e-01
3.34328473e-01 7.03459308e-02 3.03931832e-01 2.91000009e-01
1.21971536e+00 -2.48421133e-01 -1.48163140e-01 -1.59659475e-01
7.09485859e-02 -5.32117598e-02 4.97147709e-01 7.50965655e-01
-2.39887163e-01 9.79628623e-01 8.12530875e-01 -2.30727971e-01
-9.95419979e-01 -1.24192548e+00 -6.50528669e-01 7.40189910e-01
-9.92273539e-02 -2.54088700e-01 -7.15603709e-01 -6.37253463e-01
1.50477126e-01 4.46875513e-01 -8.02005291e-01 7.12146517e-03
-4.77444708e-01 -7.07726181e-01 7.93549478e-01 1.42907500e-01
1.13288268e-01 -8.84920061e-01 -7.27144241e-01 -1.67096630e-02
1.15932617e-03 -1.29740512e+00 -2.99811482e-01 -2.50315666e-02
-5.25806189e-01 -1.37040770e+00 -5.11453211e-01 -1.53029472e-01
7.34176159e-01 1.01249248e-01 7.56963134e-01 9.18116346e-02
-6.65923178e-01 4.16429639e-01 -4.86531258e-01 -3.31796020e-01
-3.86274129e-01 -5.25612496e-02 -2.75406808e-01 5.05360663e-01
3.52600038e-01 -8.83126497e-01 -6.12795651e-01 1.34826764e-01
-9.68151510e-01 -4.83825952e-01 2.61903107e-01 4.23572361e-01
1.17755450e-01 -3.74992602e-02 3.68084878e-01 -1.00033915e+00
3.92873913e-01 -8.95924568e-01 -5.99546432e-01 -3.15177701e-02
-2.69857347e-01 -5.84178507e-01 1.88625932e-01 -4.09587175e-01
-8.25658202e-01 -2.37120479e-01 7.48698935e-02 -3.45681608e-01
-5.73539078e-01 1.76238626e-01 -1.87237784e-01 -3.99072580e-02
5.92575729e-01 -9.12216231e-02 -2.19200365e-02 -6.85976207e-01
-1.56734660e-02 7.88773715e-01 7.67086804e-01 -5.39264858e-01
7.13210940e-01 8.35369051e-01 1.46168917e-01 -8.99132609e-01
-4.15573478e-01 -5.58600366e-01 -5.50560474e-01 -2.55728811e-01
7.40651309e-01 -5.87750256e-01 -8.09665918e-01 4.46646988e-01
-1.19403934e+00 1.98529691e-01 -1.16944738e-01 4.91255432e-01
-1.09212600e-01 3.45616490e-01 -5.37001491e-01 -1.05045712e+00
-6.11102320e-02 -1.15678287e+00 9.66702461e-01 1.45990103e-01
-3.79012287e-01 -5.69462538e-01 -2.24530712e-01 7.73215353e-01
3.59798700e-01 5.03771424e-01 9.38339055e-01 -1.04957688e+00
-3.81041020e-01 -4.82232481e-01 -3.06924760e-01 3.67244869e-01
7.51206130e-02 3.83087277e-01 -1.44130361e+00 -2.66974628e-01
-1.65952548e-01 -2.71879554e-01 6.28626823e-01 3.24038506e-01
1.23913991e+00 -1.78534985e-01 -2.33691245e-01 4.90755379e-01
1.15518892e+00 2.20763147e-01 6.46317244e-01 8.70463476e-02
2.79171020e-01 1.31370687e+00 5.62483728e-01 5.03461301e-01
7.57318959e-02 6.34163618e-01 8.06979060e-01 -9.99871716e-02
4.80468832e-02 -4.48474795e-01 3.15497629e-02 1.20083429e-01
1.26336794e-02 1.17910042e-01 -8.53861809e-01 9.74752963e-01
-1.36589932e+00 -9.98144388e-01 -3.21650416e-01 2.46175456e+00
6.97763264e-01 -5.29718702e-04 2.40683824e-01 3.16321462e-01
8.78643930e-01 7.17739463e-02 -2.39897624e-01 -5.69368482e-01
-1.24477640e-01 4.45616752e-01 5.72541237e-01 2.66568698e-02
-1.06851566e+00 3.86481881e-01 5.85966206e+00 1.01857471e+00
-1.20627582e+00 2.13588506e-01 7.69286454e-01 -3.28173876e-01
-3.80279601e-01 -3.87323610e-02 -4.79204923e-01 8.33339810e-01
6.39349103e-01 3.42113912e-01 5.02745509e-01 7.14291513e-01
4.06114697e-01 -3.51132065e-01 -1.14213467e+00 1.00076330e+00
1.75592259e-01 -1.37565351e+00 -4.37617242e-01 2.15926617e-01
4.65425372e-01 -2.40156770e-01 8.15215483e-02 -2.44781420e-01
-1.52751371e-01 -1.25577855e+00 7.68228292e-01 2.72980303e-01
7.45417714e-01 -6.82616293e-01 5.97462296e-01 1.78247631e-01
-6.78827286e-01 -6.16081543e-02 3.81686389e-02 2.39385977e-01
-9.12485719e-02 8.28988016e-01 -8.07428539e-01 2.95994759e-01
1.10302854e+00 4.62250322e-01 -6.41538620e-01 1.16569149e+00
-7.48400018e-02 8.75175774e-01 -4.74821150e-01 2.27412954e-01
-4.26216200e-02 -2.33494088e-01 5.72975874e-01 1.35317576e+00
2.14070469e-01 2.79308632e-02 -2.86793292e-01 1.02531731e+00
-1.01190425e-01 3.29051226e-01 -7.57784069e-01 -9.56072956e-02
6.50152981e-01 1.54597497e+00 -8.46811116e-01 2.17595711e-01
-6.55079901e-01 4.54002798e-01 1.20640695e-01 1.34661064e-01
-4.27560359e-01 -9.83970612e-02 7.24089265e-01 4.92032200e-01
2.46498704e-01 1.67514622e-01 -4.78990942e-01 -9.14555848e-01
2.30314091e-01 -1.04093194e+00 5.83069801e-01 -5.73924124e-01
-1.31894588e+00 1.59082964e-01 2.12862477e-01 -1.08494258e+00
-5.94009720e-02 -3.73900950e-01 -4.95446712e-01 6.44271374e-01
-9.91876364e-01 -1.27271450e+00 -1.93441272e-01 4.56187338e-01
9.24349390e-03 -7.75927082e-02 5.74571252e-01 7.36047685e-01
-4.99499500e-01 7.72003829e-01 -1.29300043e-01 4.31614280e-01
5.79530001e-01 -6.94051445e-01 -3.19840014e-02 7.82007635e-01
2.48590350e-01 6.05277538e-01 7.31555879e-01 -5.24685919e-01
-1.03442693e+00 -6.92052662e-01 6.00082040e-01 -3.20872515e-01
7.13430107e-01 -7.56303012e-01 -9.15631056e-01 5.19668758e-01
-8.47086906e-02 4.96234559e-02 7.39648104e-01 2.30403859e-02
-3.43883008e-01 -2.86826134e-01 -1.66673219e+00 2.78087020e-01
8.01543176e-01 -4.46218342e-01 -1.18449934e-01 7.22637996e-02
6.43532574e-02 -3.33339959e-01 -9.84623909e-01 2.51211107e-01
5.70350468e-01 -1.15272200e+00 6.26368999e-01 -4.63545233e-01
5.09405255e-01 6.14697672e-02 -1.99137866e-01 -9.33475018e-01
2.89653033e-01 -4.60240752e-01 3.79588932e-01 1.96022511e+00
4.51831758e-01 -5.29490173e-01 8.82573426e-01 9.61145401e-01
1.78202957e-01 -6.25036061e-01 -1.33393395e+00 -4.62137431e-01
-1.25322044e-01 -9.12662268e-01 7.20769882e-01 1.16144383e+00
-1.31862417e-01 -4.20415133e-01 -2.67379850e-01 4.23730373e-01
7.33428478e-01 1.40985465e-02 6.49874985e-01 -1.02546620e+00
-1.23107530e-01 -3.90193343e-01 -8.33470881e-01 -1.41777799e-01
-3.99374738e-02 -8.11148286e-01 -4.95378137e-01 -8.92820179e-01
5.02353370e-01 -4.74641234e-01 1.17470533e-01 5.71516871e-01
4.20440942e-01 5.56078553e-01 9.25565287e-02 -8.94690380e-02
-1.17788099e-01 -6.52936697e-02 9.42590892e-01 -2.35127255e-01
1.42119601e-01 1.24312483e-01 -6.83573186e-01 7.74343073e-01
4.44367141e-01 -8.05078149e-01 -1.42308459e-01 -2.00981647e-01
7.24849701e-02 -4.58591878e-02 6.86303496e-01 -7.71860242e-01
-2.63547990e-02 -2.17458427e-01 3.45515788e-01 -7.85312355e-01
3.60982448e-01 -9.81664538e-01 1.58345193e-01 2.10908845e-01
-2.79731691e-01 -4.37050223e-01 1.09427348e-01 -3.94201577e-02
-2.44293183e-01 -6.60975218e-01 7.59126663e-01 5.74927740e-02
-2.86545545e-01 3.09381634e-01 -5.76777980e-02 1.08668707e-01
9.19450343e-01 -3.08201760e-01 -5.50894439e-01 -6.52572989e-01
-6.51174963e-01 8.94626379e-02 5.78336835e-01 2.92515397e-01
5.41842639e-01 -9.97738183e-01 -7.91547358e-01 4.34173822e-01
1.40681133e-01 -2.69626021e-01 3.79205793e-01 9.42540169e-01
-3.28248352e-01 5.04281819e-02 -2.74032563e-01 -6.03543818e-01
-1.37241340e+00 4.99365300e-01 -9.01773497e-02 1.96032256e-01
-4.12471592e-01 4.85945195e-01 2.31703068e-03 -1.67247683e-01
3.05288851e-01 -2.05023959e-01 -3.58420223e-01 5.21105289e-01
5.50133049e-01 4.80112731e-01 2.74163157e-01 -8.77205193e-01
-5.02217233e-01 1.68531001e-01 1.22190453e-01 -3.58237661e-02
1.79216206e+00 6.70979246e-02 -3.40087801e-01 7.35927746e-02
1.40262568e+00 5.15881360e-01 -9.85354304e-01 2.69713491e-01
-1.45515338e-01 -1.03650522e+00 -3.75335187e-01 -6.13352180e-01
-1.43487620e+00 1.01620674e+00 6.59340918e-01 4.40327466e-01
1.11282170e+00 1.76743284e-01 4.28987950e-01 -3.79990131e-01
4.88794684e-01 -1.02155757e+00 -7.64275491e-02 -2.86212772e-01
8.52830648e-01 -1.11939347e+00 1.23087391e-01 -5.99579215e-01
-3.47405821e-01 9.79670703e-01 2.16315672e-01 1.39584482e-01
3.66701245e-01 4.72790927e-01 -2.31593661e-02 -3.44441921e-01
-4.21368331e-01 2.14746192e-01 1.33732423e-01 1.10951483e+00
1.55376732e-01 -4.05685157e-02 -4.44090873e-01 8.99522007e-01
-2.40195081e-01 -3.90941381e-01 5.70421815e-01 6.66119397e-01
9.96249542e-02 -1.08233547e+00 -6.22651279e-01 8.22854817e-01
-8.48185241e-01 1.50129367e-02 -3.42717141e-01 7.12521911e-01
4.98743415e-01 1.05119026e+00 1.54519513e-01 -5.15355775e-03
3.56275171e-01 -2.80318379e-01 3.50920677e-01 -5.57838798e-01
-6.83446407e-01 -8.19863155e-02 3.36691797e-01 -6.52850211e-01
-2.11729854e-01 -1.09405243e+00 -8.88291538e-01 -3.02843183e-01
-6.36905134e-02 -1.07480474e-01 1.00255060e+00 6.68675542e-01
1.74446568e-01 -2.50201017e-01 6.50886893e-01 -7.01724291e-01
-1.83359534e-01 -7.31201172e-01 -8.95676494e-01 9.34272289e-01
3.94904912e-01 -5.90294778e-01 -3.23614895e-01 -4.65536043e-02] | [12.870583534240723, 1.0257683992385864] |
794f3bae-a674-499d-97ca-a9219de9b495 | h-analysis-and-data-parallel-physics-informed | 2302.08835 | null | https://arxiv.org/abs/2302.08835v1 | https://arxiv.org/pdf/2302.08835v1.pdf | $h$-analysis and data-parallel physics-informed neural networks | We explore the data-parallel acceleration of physics-informed machine learning (PIML) schemes, with a focus on physics-informed neural networks (PINNs) for multiple graphics processing units (GPUs) architectures. In order to develop scale-robust PIML models for sophisticated applications (e.g., involving complex and high-dimensional domains, non-linear operators or multi-physics), which may require a large number of training points, we detail a protocol based on the Horovod training framework. This protocol is backed by $h$-analysis, including a new convergence bound for the generalization error. We show that the acceleration is straightforward to implement, does not compromise training, and proves to be highly efficient, paving the way towards generic scale-robust PIML. Extensive numerical experiments with increasing complexity illustrate its robustness and consistency, offering a wide range of possibilities for real-world simulations. | ['Gonzalo A. Ruz', 'Paul Escapil-Inchauspé'] | 2023-02-17 | null | null | null | null | ['physics-informed-machine-learning'] | ['graphs'] | [ 7.39865974e-02 -5.89909740e-02 8.88303816e-02 -3.76952052e-01
-6.92561448e-01 -3.37551683e-01 6.55328333e-01 -5.75415511e-03
-7.12786555e-01 8.34344208e-01 -6.70750737e-01 -4.68734980e-01
-4.29157317e-01 -7.30344355e-01 -1.00178564e+00 -1.13559163e+00
-5.61326563e-01 8.37695658e-01 2.03003556e-01 -2.16951862e-01
4.33289558e-01 1.20269310e+00 -1.72077024e+00 -1.68675154e-01
5.51793993e-01 1.04755032e+00 1.83867127e-01 1.00151432e+00
7.69804344e-02 4.27489907e-01 -4.07009618e-03 -3.74997497e-01
3.72311741e-01 5.16919559e-03 -7.38991916e-01 -3.33929598e-01
5.25437176e-01 -4.73228544e-02 -6.57706559e-02 9.99012470e-01
6.32091641e-01 5.41924417e-01 5.74465930e-01 -1.20732486e+00
-1.46786794e-01 1.35900453e-01 -3.86281371e-01 1.47880495e-01
-4.37604845e-01 1.57948822e-01 6.55979931e-01 -1.04104972e+00
4.93245065e-01 1.16987622e+00 1.25374794e+00 5.85662901e-01
-1.14390862e+00 -5.23786485e-01 -6.05680607e-02 1.19846836e-01
-1.12118340e+00 -1.30281886e-02 7.51109362e-01 -3.35123181e-01
9.21132326e-01 3.49405169e-01 6.81146264e-01 8.33629966e-01
4.14401859e-01 5.42265952e-01 1.13352180e+00 -6.44960284e-01
7.15011418e-01 3.59680921e-01 5.38525701e-01 9.60357189e-01
5.02743840e-01 3.88361454e-01 -5.45603633e-01 -3.67970139e-01
1.06818581e+00 -4.95798051e-01 -1.44613639e-01 -8.01213384e-01
-9.89364266e-01 1.09753513e+00 6.28221035e-01 -6.09975774e-03
-2.14418635e-01 1.99668556e-01 4.43094134e-01 -6.45639747e-02
5.05494714e-01 4.48052227e-01 -5.93797982e-01 -4.97039743e-02
-8.75928402e-01 5.91093719e-01 9.99415040e-01 9.20073211e-01
8.40837657e-01 1.72136068e-01 4.03133333e-01 5.09422123e-01
1.41675368e-01 5.26604176e-01 4.56742710e-03 -1.33734655e+00
9.25684348e-02 1.14956036e-01 1.63747773e-01 -5.08196950e-01
-8.02701354e-01 -6.13702059e-01 -1.38386261e+00 6.94231212e-01
3.77599239e-01 -3.16342384e-01 -6.45748973e-01 1.65976036e+00
6.19376361e-01 2.82507211e-01 -7.35742450e-02 8.78474891e-01
3.63954037e-01 7.57631481e-01 2.91939765e-01 -1.15317874e-01
1.16390443e+00 -8.71214330e-01 1.75450012e-01 -7.68887103e-02
1.12420189e+00 -2.97309846e-01 1.12244511e+00 4.25657660e-01
-1.30064487e+00 -6.46376550e-01 -1.12425840e+00 -2.27155507e-01
-4.06096429e-01 -1.41472071e-01 1.28036165e+00 6.66107118e-01
-1.10913515e+00 1.35529244e+00 -1.14194965e+00 -3.02967757e-01
3.03727806e-01 6.38350725e-01 -9.60720479e-02 2.42234468e-01
-8.48636091e-01 8.00429702e-01 8.65502298e-01 9.36694592e-02
-3.58399421e-01 -1.03807926e+00 -5.19716203e-01 3.23872343e-02
-2.53127784e-01 -9.92457032e-01 1.10402620e+00 -7.19123662e-01
-1.63231790e+00 6.88261271e-01 1.27126560e-01 -6.49334013e-01
5.10512590e-01 -7.27705136e-02 -1.81773361e-02 -2.39276756e-02
-4.66435641e-01 7.16008544e-01 5.63987374e-01 -9.69710290e-01
-4.25246328e-01 -2.61987090e-01 -8.34000558e-02 3.46972853e-01
-3.12364310e-01 -2.38203127e-02 -2.74887204e-01 -3.11779916e-01
4.15673777e-02 -1.17311084e+00 -7.41622686e-01 2.15166062e-01
-3.59180458e-02 -8.84653479e-02 5.97378910e-01 -2.52743751e-01
2.59033620e-01 -1.83847558e+00 1.42298341e-01 2.28077248e-01
1.38331875e-01 3.51369590e-01 2.23193437e-01 2.36184686e-01
-2.06247687e-01 -2.05122575e-01 -6.20401025e-01 -5.92256129e-01
2.21983507e-01 4.05561447e-01 -4.24631685e-01 6.39136553e-01
1.09876662e-01 6.45439267e-01 -6.01714253e-01 -4.76333290e-01
3.39364588e-01 5.25876045e-01 -8.33692074e-01 -3.27444859e-02
-1.09610707e-01 7.23841488e-01 -5.16146660e-01 2.58534461e-01
8.82821679e-01 -3.52607191e-01 -3.36267501e-01 1.71909370e-02
-2.68452495e-01 -5.59064653e-03 -1.27840161e+00 1.67137349e+00
-6.35664999e-01 5.45346439e-01 3.74192774e-01 -1.34120250e+00
5.86137295e-01 -8.02457258e-02 3.97521317e-01 -3.07230085e-01
1.90863162e-01 4.59507465e-01 -3.20243239e-01 -5.75937442e-02
4.63989258e-01 -4.91279811e-01 9.79972854e-02 1.21165343e-01
3.01253498e-01 -1.88725904e-01 6.73639998e-02 -7.23631158e-02
8.33721101e-01 3.13148767e-01 1.17081389e-01 -8.98591757e-01
5.79403043e-01 4.62349713e-01 3.21994960e-01 9.59267616e-01
-1.00954130e-01 3.24683845e-01 3.36539745e-02 -7.07144499e-01
-1.27494931e+00 -9.31434751e-01 -6.79009259e-01 1.12821066e+00
3.43436822e-02 1.37733936e-01 -7.43762314e-01 -1.61147147e-01
-6.87088445e-02 9.20043111e-01 -3.91736299e-01 -1.16746336e-01
-8.56138229e-01 -1.33098173e+00 2.19045326e-01 5.16757309e-01
6.13481522e-01 -9.04221416e-01 -5.72636902e-01 2.66364485e-01
6.97546780e-01 -1.08757997e+00 2.44201124e-01 4.62855130e-01
-1.24026668e+00 -7.94709146e-01 -9.03726280e-01 -7.25028217e-01
3.99642199e-01 -9.26138088e-02 1.19402158e+00 -1.97979420e-01
-3.28993320e-01 3.83043677e-01 2.50215411e-01 -5.54009855e-01
-3.44550043e-01 1.75528690e-01 4.80735958e-01 -5.15989542e-01
8.17747638e-02 -8.01926553e-01 -7.00962782e-01 5.71325282e-03
-3.96486431e-01 3.87628675e-01 6.60161436e-01 8.44804764e-01
7.05303609e-01 8.21646899e-02 8.50696117e-02 -7.03213096e-01
1.46583065e-01 -1.87937662e-01 -1.14052653e+00 1.44400612e-01
-6.03627980e-01 8.59760717e-02 1.01207304e+00 -5.36560118e-01
-1.11917615e+00 2.33333975e-01 -5.42824686e-01 -2.53488123e-01
-2.73310989e-01 1.01339504e-01 2.75159299e-01 -1.01230967e+00
9.80928838e-01 7.81034231e-02 -4.23979223e-01 -6.44080520e-01
4.46261346e-01 1.86606944e-01 7.22218871e-01 -1.09214497e+00
9.50105071e-01 4.73119348e-01 6.78483486e-01 -1.17805421e+00
-7.05876291e-01 -3.24530393e-01 -6.05485737e-01 -8.00400674e-02
7.88274646e-01 -8.07182252e-01 -9.83480692e-01 4.62398589e-01
-1.19345593e+00 -5.77975750e-01 -5.07306755e-01 8.90549421e-01
-9.52797234e-01 2.82085001e-01 -8.75543714e-01 -1.07198870e+00
-5.53229511e-01 -1.13802743e+00 9.14528906e-01 3.23454231e-01
2.28706688e-01 -1.37244511e+00 1.59833491e-01 -9.80723500e-02
2.74536490e-01 1.77473530e-01 1.09335053e+00 -4.86967832e-01
-6.16161942e-01 -1.57779634e-01 -3.62995535e-01 1.01715291e-03
-7.49907315e-01 5.42599428e-03 -1.17705858e+00 -5.14117241e-01
3.99079114e-01 -3.65726411e-01 6.23805583e-01 4.26307052e-01
1.44476247e+00 -1.77646372e-02 -3.39093506e-01 1.06062031e+00
1.57673156e+00 -1.49326146e-01 9.12280530e-02 2.70925134e-01
7.87292302e-01 4.72564042e-01 1.85284719e-01 2.14760974e-01
4.63487096e-02 5.53488791e-01 2.68148035e-01 -2.44367853e-01
7.69911557e-02 2.38408640e-01 7.97096342e-02 1.10050702e+00
-2.79019743e-01 5.91342300e-02 -1.07475960e+00 1.24789573e-01
-1.74564373e+00 -5.44755459e-01 -6.45255864e-01 2.31302619e+00
3.76942068e-01 3.00093830e-01 -4.92289811e-02 4.90469746e-02
4.23038453e-01 -1.23902246e-01 -7.64298916e-01 -7.60535657e-01
6.00383803e-02 4.89045441e-01 8.69744599e-01 4.54969734e-01
-1.01426113e+00 6.93039656e-01 6.72125864e+00 1.03722751e+00
-1.00643945e+00 3.22735339e-01 5.95359266e-01 9.76926982e-02
1.90405309e-01 2.44750418e-02 -1.11082256e+00 1.11301467e-01
1.41875434e+00 -7.96381757e-02 3.17197710e-01 1.33486426e+00
-5.65221645e-02 -1.05621345e-01 -1.08675039e+00 8.91496301e-01
-5.33454299e-01 -1.74509215e+00 -4.40096296e-02 1.46802207e-02
6.52900994e-01 6.68976903e-01 -1.83076471e-01 5.25288641e-01
4.38104898e-01 -6.09845877e-01 4.51171875e-01 3.53118539e-01
5.48591316e-01 -9.09223557e-01 5.03619075e-01 7.32822239e-01
-9.77838933e-01 1.64058238e-01 -6.91147804e-01 -1.34008199e-01
2.92388555e-02 6.64851725e-01 -1.88539043e-01 4.44685996e-01
7.82678723e-01 3.44081521e-01 -2.33923152e-01 1.15256453e+00
3.95187289e-01 5.25907993e-01 -8.36115718e-01 -3.96293521e-01
3.94034088e-01 -4.74274337e-01 4.96076763e-01 1.38869977e+00
2.55171150e-01 3.52185965e-01 -1.45194992e-01 8.09157014e-01
2.23278940e-01 3.29175740e-02 -5.07680953e-01 2.39533767e-01
-7.48081654e-02 1.34101212e+00 -8.65700901e-01 -4.69774306e-01
-7.48189837e-02 7.29517281e-01 5.78743875e-01 1.72473595e-01
-8.92538548e-01 -1.88624814e-01 5.03843904e-01 -2.50445485e-01
7.35014603e-02 -7.34268785e-01 -6.71147704e-01 -1.08532870e+00
-1.77229494e-01 -3.07251573e-01 1.39028400e-01 -6.57987714e-01
-1.27969480e+00 5.04499197e-01 2.05160633e-01 -9.07250345e-01
-1.27967790e-01 -1.24429762e+00 -6.00063205e-01 9.36680913e-01
-1.41833830e+00 -7.66490281e-01 -1.01062968e-01 3.64123851e-01
2.43770823e-01 7.59251863e-02 1.02650845e+00 1.43041238e-01
-5.62740684e-01 4.45311099e-01 4.32405710e-01 -2.87749410e-01
-6.77971020e-02 -1.24960542e+00 8.73659194e-01 6.22914374e-01
8.80585983e-02 5.53172112e-01 1.06859004e+00 -3.34046602e-01
-1.84507823e+00 -8.81646633e-01 5.96078634e-01 -3.87464613e-01
8.60352635e-01 -5.52549303e-01 -1.08642197e+00 5.18044472e-01
-2.57885158e-01 4.71232869e-02 3.84700626e-01 1.66588679e-01
9.03861821e-02 -3.56276613e-03 -1.14127457e+00 5.82183063e-01
1.12502348e+00 -2.41589397e-01 -1.83393776e-01 1.01099980e+00
3.23303103e-01 -5.46378136e-01 -8.82460952e-01 5.50457835e-01
3.39584649e-01 -1.04816806e+00 1.18041980e+00 -4.11654651e-01
-4.01904732e-02 1.98372841e-01 -7.53070489e-02 -9.08594549e-01
-3.35009992e-01 -5.58606744e-01 -1.45225376e-01 7.09131420e-01
1.19199939e-01 -9.03175414e-01 1.12383270e+00 8.60018432e-01
-3.08106750e-01 -7.52452850e-01 -1.19486153e+00 -9.90893245e-01
5.05006492e-01 -8.73800516e-01 2.09310457e-01 7.89443374e-01
-2.51368493e-01 3.34573444e-03 -9.75787193e-02 5.47861218e-01
1.12894225e+00 6.45646825e-02 5.52409649e-01 -1.40144575e+00
-7.06876040e-01 -5.95889807e-01 -3.01593393e-01 -9.72020686e-01
1.71435982e-01 -7.66152382e-01 -6.46866038e-02 -9.49807167e-01
-6.39211982e-02 -9.95770752e-01 -2.23938320e-02 2.13875771e-01
1.70859098e-01 3.60059410e-01 -1.05426954e-02 2.64999807e-01
-2.88494200e-01 7.00897694e-01 9.39539254e-01 4.09269989e-01
-5.43613806e-02 1.92882687e-01 3.81221473e-02 1.10169446e+00
9.59105968e-01 -4.13796991e-01 -3.44111592e-01 -3.06882977e-01
2.80037578e-02 -8.05433095e-03 7.15134740e-01 -1.63085938e+00
3.68665367e-01 8.40494595e-03 4.41083372e-01 -5.28328836e-01
8.53035867e-01 -7.01064348e-01 3.92007865e-02 6.35459840e-01
-7.79249519e-02 4.69971262e-02 4.91261095e-01 3.95027369e-01
3.05714965e-01 -3.45151097e-01 1.15721965e+00 -2.18105093e-01
-5.86692870e-01 4.24276888e-01 -5.78035936e-02 -1.74248725e-01
7.39070237e-01 3.31237316e-02 6.77714217e-03 1.31810308e-01
-6.60514772e-01 -7.50330314e-02 5.20193577e-01 -3.07355940e-01
3.24788690e-02 -1.13627255e+00 -4.66058433e-01 2.24815890e-01
-1.80626705e-01 2.44612798e-01 3.52225721e-01 6.83415711e-01
-9.18533325e-01 4.53420430e-01 -1.61900803e-01 -8.33905101e-01
-8.51447761e-01 4.61369693e-01 6.38065517e-01 -1.89128697e-01
-9.88005579e-01 1.08972931e+00 1.41457349e-01 -6.02114320e-01
1.54279917e-01 -2.96037763e-01 3.30992222e-01 -5.12666106e-01
4.36100841e-01 5.06494939e-01 3.18500280e-01 -4.13582623e-01
-1.98077396e-01 8.06631386e-01 8.09584707e-02 -2.88496371e-02
1.27786517e+00 3.91191542e-01 7.57251307e-02 5.26570678e-01
1.26408541e+00 -4.56386924e-01 -1.65716517e+00 -1.71481818e-01
-1.81007572e-02 2.84415573e-01 3.43752563e-01 -3.02992135e-01
-8.26875865e-01 1.15638816e+00 7.22337306e-01 -2.70780772e-02
8.40633810e-01 -3.38968784e-01 7.29950845e-01 9.39297080e-01
7.94958830e-01 -1.08278501e+00 -7.33611822e-01 5.18922269e-01
6.79802001e-01 -1.19443405e+00 2.76130766e-01 -4.23740327e-01
-7.11763054e-02 1.38183188e+00 4.16271120e-01 -5.30352116e-01
8.74613523e-01 3.21665853e-01 -3.43008548e-01 -1.16708435e-01
-4.87141967e-01 1.46253064e-01 2.93098420e-01 4.29741204e-01
-6.77482337e-02 -2.54278071e-02 -9.92559865e-02 4.19195779e-02
-4.27573800e-01 -1.24586513e-02 1.46622196e-01 8.67514968e-01
-4.64568526e-01 -9.36926186e-01 -1.77194446e-01 2.31940523e-01
4.73453999e-02 -5.41141406e-02 1.53528407e-01 1.22999942e+00
7.00050369e-02 2.29140490e-01 2.58293021e-02 -4.13758866e-02
1.09142372e-02 2.33172014e-01 6.70195699e-01 -1.00684315e-01
-4.70972806e-01 -3.28878194e-01 -8.37389231e-02 -5.49363434e-01
-2.27059588e-01 -5.59553921e-01 -1.40776801e+00 -7.35216737e-01
2.42848191e-02 2.88600355e-01 1.24310446e+00 1.05775201e+00
3.72636199e-01 1.68505564e-01 2.94942319e-01 -1.80804479e+00
-6.52291238e-01 -4.26524073e-01 -7.18836367e-01 -5.29790409e-02
9.08390954e-02 -7.75050342e-01 -5.09395778e-01 -4.50772226e-01] | [6.4012956619262695, 3.628180980682373] |
bbdfad5e-4d0a-4eac-908f-3b34d4ed7a2e | encouraging-divergent-thinking-in-large | 2305.19118 | null | https://arxiv.org/abs/2305.19118v1 | https://arxiv.org/pdf/2305.19118v1.pdf | Encouraging Divergent Thinking in Large Language Models through Multi-Agent Debate | Modern large language models (LLMs) like ChatGPT have shown remarkable performance on general language tasks but still struggle on complex reasoning tasks, which drives the research on cognitive behaviors of LLMs to explore human-like problem-solving strategies. Along this direction, one representative strategy is self-reflection, which asks an LLM to refine the solution with the feedback generated by itself iteratively. However, our study shows that such reflection-style methods suffer from the Degeneration-of-Thought (DoT) problem: once the LLM has established confidence in its solutions, it is unable to generate novel thoughts later through reflection even if its initial stance is incorrect. To address the DoT problem, we propose a Multi-Agent Debate (MAD) framework, in which multiple agents express their arguments in the state of "tit for tat" and a judge manages the debate process to obtain a final solution. Clearly, our MAD framework encourages divergent thinking in LLMs which would be helpful for tasks that require deep levels of contemplation. Experiment results on two challenging datasets, commonsense machine translation and counter-intuitive arithmetic reasoning, demonstrate the effectiveness of our MAD framework. Extensive analyses suggest that the adaptive break of debate and the modest level of "tit for tat" state are required for MAD to obtain good performance. Moreover, we find that LLMs might not be a fair judge if different LLMs are used for agents. Codes: https://github.com/Skytliang/Multi-Agents-Debate | ['Shuming Shi', 'Zhaopeng Tu', 'Yujiu Yang', 'Rui Wang', 'Yan Wang', 'Xing Wang', 'Wenxiang Jiao', 'Zhiwei He', 'Tian Liang'] | 2023-05-30 | null | null | null | null | ['arithmetic-reasoning'] | ['reasoning'] | [-1.22838646e-01 5.89993954e-01 -2.67022476e-02 -2.88975567e-01
-6.25799954e-01 -6.64435446e-01 9.28760111e-01 2.34348655e-01
-4.01216596e-01 7.34083474e-01 3.53684336e-01 -6.96910381e-01
1.13050915e-01 -8.26591969e-01 -5.29363453e-01 -5.19665658e-01
3.01726252e-01 7.79245794e-01 3.98002239e-03 -6.90382421e-01
5.39491713e-01 -1.71713531e-01 -9.37569261e-01 4.27915633e-01
1.30125666e+00 1.47465810e-01 1.57092869e-01 4.08639669e-01
-3.53570223e-01 1.44478810e+00 -7.24235535e-01 -7.18769729e-01
3.10899973e-01 -8.00811529e-01 -1.24340534e+00 3.02437004e-02
-8.77253264e-02 -2.23173618e-01 1.82235822e-01 1.48518527e+00
1.89375013e-01 1.59667015e-01 4.44356918e-01 -1.26372015e+00
-8.51856709e-01 1.29466963e+00 -5.63405275e-01 1.71382964e-01
4.18876261e-01 5.32769382e-01 9.59675610e-01 -7.14658678e-01
5.35453975e-01 1.56771564e+00 5.25499761e-01 5.56100309e-01
-9.45284009e-01 -6.16951168e-01 3.58195394e-01 2.66579866e-01
-9.14063931e-01 -1.84820935e-01 6.42521381e-01 -3.35940719e-01
7.84333766e-01 4.87360060e-01 8.00249875e-01 9.34199393e-01
4.50760394e-01 6.18295789e-01 1.79156613e+00 -3.79303217e-01
5.52845180e-01 9.12878364e-02 2.39942446e-01 7.91905820e-01
4.50115144e-01 -2.79951483e-01 -6.67315781e-01 -3.73581499e-01
5.38609385e-01 -4.59236950e-01 -1.34844601e-01 9.97052789e-02
-1.27373326e+00 1.07961714e+00 3.88460547e-01 5.45416415e-01
-5.58631718e-01 3.99215519e-02 1.92386851e-01 7.39054620e-01
4.51762408e-01 8.74374211e-01 -7.65731856e-02 -3.01133245e-01
-7.54092157e-01 4.61099625e-01 1.00976562e+00 4.19263572e-01
5.61934888e-01 -1.66481107e-01 -7.99883008e-02 6.92493558e-01
4.31950122e-01 2.26759613e-01 5.81482530e-01 -1.55324078e+00
3.83167833e-01 9.06293690e-01 2.34925970e-01 -1.03476632e+00
-3.24277967e-01 -4.66318697e-01 -6.58785760e-01 5.04600942e-01
6.81129932e-01 -2.65844584e-01 -3.63853693e-01 1.91815364e+00
3.94827098e-01 -1.19955666e-01 3.06930661e-01 1.19014418e+00
4.19198483e-01 6.93026900e-01 6.57920465e-02 -3.40178698e-01
1.35194743e+00 -1.20531368e+00 -5.43786526e-01 -5.60561478e-01
8.08142185e-01 -7.04718471e-01 1.30958080e+00 4.54999328e-01
-1.50380433e+00 -4.31719311e-02 -7.84422696e-01 -1.31514624e-01
1.75185248e-01 -3.20508569e-01 6.26486123e-01 3.09545040e-01
-1.08736503e+00 3.50896835e-01 -6.59278989e-01 -3.07887316e-01
2.92452365e-01 -1.35927007e-01 1.98036470e-02 6.35215640e-02
-1.18269277e+00 1.18038273e+00 1.15687996e-01 4.09582585e-01
-7.50022233e-01 -3.28600943e-01 -4.05697733e-01 -9.13031027e-03
7.71210909e-01 -8.51670742e-01 1.57277501e+00 -1.14973867e+00
-1.72537529e+00 8.70398879e-01 -5.00203855e-02 -4.23895091e-01
8.85029018e-01 -1.97614640e-01 1.33470744e-01 1.33685051e-02
3.30802619e-01 4.60045755e-01 5.60694754e-01 -1.21719217e+00
-2.42495939e-01 -2.18542144e-01 5.93887389e-01 3.86541158e-01
8.67076293e-02 2.12917000e-01 1.87955454e-01 -4.94803995e-01
1.26788139e-01 -1.09226251e+00 -4.23235893e-01 -1.85562417e-01
-3.11911941e-01 -6.40415013e-01 1.45433724e-01 -4.15515393e-01
1.12197065e+00 -1.70325994e+00 2.97715127e-01 2.47038044e-02
5.43942988e-01 2.04684690e-01 -2.86585510e-01 6.53636038e-01
2.55644500e-01 3.75714242e-01 -7.24431202e-02 -1.50476873e-01
2.63207197e-01 2.06792235e-01 -6.56491876e-01 1.21002175e-01
8.29844326e-02 1.16973197e+00 -1.00589633e+00 -3.71678561e-01
-3.54872912e-01 -8.31553563e-02 -8.06726098e-01 1.46483645e-01
-6.98008657e-01 4.96889561e-01 -6.17435157e-01 3.32614452e-01
5.31581581e-01 -7.67484844e-01 5.46747029e-01 2.43131384e-01
-2.19381526e-01 5.82879126e-01 -9.44364250e-01 1.36382580e+00
-1.00219488e-01 4.65729415e-01 2.85119802e-01 -7.65437484e-01
6.93516850e-01 7.52854720e-02 -2.29209945e-01 -8.18927586e-01
2.76987433e-01 2.46118844e-01 6.94514871e-01 -4.78986353e-01
4.23925698e-01 -3.01354200e-01 -4.91454452e-02 1.29184771e+00
-6.46473467e-01 -4.05191511e-01 3.05909663e-01 6.28082871e-01
9.74178433e-01 -1.01707853e-01 3.30456406e-01 -5.31505525e-01
5.62768102e-01 4.02029872e-01 7.36172855e-01 1.00737941e+00
-2.78809607e-01 -1.57460362e-01 8.40960562e-01 -5.10473371e-01
-9.56513882e-01 -7.27955878e-01 4.79162306e-01 1.25991201e+00
1.57313153e-01 -5.40680707e-01 -9.75634933e-01 -3.58884871e-01
-3.44711155e-01 1.21145821e+00 -3.50108206e-01 -1.91774908e-02
-8.02870691e-01 -6.40364826e-01 6.88775480e-01 1.62130490e-01
8.48292708e-01 -1.19256914e+00 -1.11572027e+00 2.34808087e-01
-6.30929947e-01 -7.55577445e-01 -1.88476831e-01 -1.69198662e-01
-6.63678527e-01 -9.71406460e-01 -3.40983003e-01 -4.50728059e-01
7.86855638e-01 2.83441842e-01 1.04708982e+00 6.31188214e-01
3.19949090e-01 2.03380153e-01 -2.97735840e-01 -3.66410702e-01
-9.60207880e-01 -2.21807480e-01 -1.97358001e-02 -4.78559881e-01
2.72056818e-01 -6.70743644e-01 -4.67702061e-01 3.58052015e-01
-7.23262489e-01 6.38746858e-01 3.89736265e-01 7.18832195e-01
-1.40669063e-01 -8.42339322e-02 4.90813315e-01 -6.90733552e-01
1.38732553e+00 -4.49010760e-01 -3.71576309e-01 3.55395228e-01
-5.97240150e-01 2.36629263e-01 4.90642339e-01 -5.14762163e-01
-1.21964145e+00 -9.97447073e-01 1.54783443e-01 2.29258582e-01
3.16977501e-01 7.76431084e-01 4.98135775e-01 2.39393517e-01
8.11695218e-01 3.39754283e-01 3.81283820e-01 -1.47423550e-01
3.16145420e-01 4.40369993e-01 2.24877164e-01 -1.25390470e+00
8.25796545e-01 3.53984386e-01 -5.86911559e-01 -3.81934971e-01
-1.06765711e+00 3.36899191e-01 -7.72515610e-02 -2.64627665e-01
5.87418139e-01 -9.42304015e-01 -1.04236245e+00 5.14386714e-01
-1.42432642e+00 -9.00489032e-01 1.75902192e-02 2.80130148e-01
-5.70454240e-01 4.67268080e-01 -9.53503847e-01 -8.94875884e-01
-4.08446610e-01 -1.12848544e+00 3.61349821e-01 5.60703039e-01
-7.97376156e-01 -8.86843443e-01 5.32671288e-02 8.62353086e-01
7.72059977e-01 -2.59178430e-01 1.30330718e+00 -6.18066907e-01
-6.72108948e-01 4.48027849e-01 -2.31511313e-02 -5.73709458e-02
-7.07897171e-02 -1.45631969e-01 -4.38509792e-01 -1.45340636e-01
4.67511624e-01 -8.17127824e-01 4.17024136e-01 -5.58253080e-02
7.07618713e-01 -6.51419997e-01 1.52265981e-01 -2.60255113e-02
8.31530571e-01 1.09371684e-01 4.47615415e-01 7.09223092e-01
7.43579492e-02 5.69573879e-01 6.29589558e-01 3.94191146e-01
9.89564419e-01 2.46888816e-01 9.86398607e-02 1.62231117e-01
1.63109019e-01 -1.62358582e-01 6.47096157e-01 9.71187174e-01
-2.73896784e-01 -2.20882222e-01 -1.18949068e+00 1.45008326e-01
-2.19432712e+00 -9.81860101e-01 -1.47184819e-01 1.58155000e+00
1.31724000e+00 2.38936827e-01 -9.66129154e-02 -2.14686245e-01
5.53106785e-01 3.21778744e-01 -5.11542797e-01 -7.61985362e-01
-5.27396947e-02 -8.63472447e-02 -3.23273271e-01 8.22502613e-01
-1.40433833e-01 1.21985233e+00 5.56679153e+00 5.57657659e-01
-1.04589164e+00 3.13268572e-01 6.91461384e-01 4.83272150e-02
-6.48717284e-01 3.58141094e-01 -3.10629934e-01 3.21700811e-01
7.86807895e-01 -6.69795692e-01 6.53464615e-01 4.34549898e-01
4.86492723e-01 -6.24396145e-01 -9.17645693e-01 5.35453618e-01
2.75138374e-02 -1.47314417e+00 1.03343084e-01 -1.82561148e-02
6.14172518e-01 1.48223266e-02 -1.84799612e-01 4.43523616e-01
8.99062812e-01 -9.09125507e-01 1.00986695e+00 3.03643256e-01
-2.07745172e-02 -3.92810136e-01 4.28054065e-01 1.15201569e+00
-5.60521066e-01 -1.21646181e-01 -3.43863964e-01 -7.74817288e-01
3.65067810e-01 2.13373303e-01 -7.31957614e-01 1.96269155e-01
2.45130762e-01 1.00508429e-01 -4.90021139e-01 3.95456582e-01
-6.58140779e-01 5.99975288e-01 -9.79945436e-02 -2.43942454e-01
5.20913064e-01 -4.30295736e-01 6.18671775e-01 7.34493136e-01
1.15260750e-01 5.48785150e-01 3.56659472e-01 1.37268615e+00
5.75563982e-02 -8.93850848e-02 -1.70653015e-01 -3.87350321e-01
4.95252550e-01 1.15306020e+00 -8.25051785e-01 -5.27379513e-01
-6.55642077e-02 7.88688660e-01 6.94391251e-01 2.16702089e-01
-8.41341615e-01 3.34545940e-01 4.23916459e-01 5.58104813e-02
-3.50900263e-01 -5.04790127e-01 -5.68880081e-01 -1.40053582e+00
2.40929388e-02 -1.55071115e+00 2.02247009e-01 -1.09741616e+00
-1.21632183e+00 5.75394750e-01 -8.67044181e-02 -6.08173966e-01
-3.08872283e-01 -2.41917759e-01 -1.01590240e+00 5.54242373e-01
-1.21997726e+00 -9.49514210e-01 -1.98179260e-01 3.54489803e-01
8.45073104e-01 3.07333712e-02 5.41341662e-01 -2.41634771e-01
-4.90656763e-01 9.98331830e-02 -5.51530480e-01 8.62062946e-02
5.74458957e-01 -8.79070580e-01 4.01609331e-01 7.68571079e-01
1.41414523e-01 1.16716516e+00 1.00691044e+00 -8.96644413e-01
-1.45399272e+00 -4.90655392e-01 1.03170884e+00 -5.19825280e-01
9.92189705e-01 -1.93858877e-01 -1.22997260e+00 7.29357779e-01
6.02241158e-01 -7.07582474e-01 4.74007547e-01 5.99974301e-03
-3.88954729e-01 4.67398882e-01 -8.59220684e-01 1.11336863e+00
1.01567054e+00 -4.07206774e-01 -1.29449952e+00 4.92691040e-01
6.30417585e-01 -3.37916315e-01 -3.62376511e-01 2.05395203e-02
2.98409373e-01 -1.10895252e+00 5.09609699e-01 -7.68257678e-01
7.98530102e-01 -4.44374651e-01 2.48082243e-02 -1.46692097e+00
-4.07248586e-01 -8.83672535e-01 1.90230548e-01 1.03617203e+00
3.74685735e-01 -9.70558167e-01 3.17347646e-01 1.02300286e+00
5.89193590e-02 -6.03319764e-01 -8.43452454e-01 -4.72571075e-01
4.78716791e-01 -4.97686088e-01 4.19304341e-01 1.16222501e+00
5.16151428e-01 5.42656362e-01 -1.63446605e-01 3.30487117e-02
5.55691957e-01 3.80189836e-01 9.27402198e-01 -1.03903639e+00
-3.80502999e-01 -6.30796254e-01 4.72720712e-01 -9.56160724e-01
4.42677885e-01 -1.00940859e+00 -4.21677046e-02 -1.69093561e+00
5.33851624e-01 -4.32200670e-01 1.39232114e-01 6.28766358e-01
-2.66916186e-01 -1.25135228e-01 6.36626065e-01 5.42099595e-01
-8.65992308e-01 4.46947008e-01 1.31557548e+00 -2.16013826e-02
-2.17069745e-01 -3.20902586e-01 -1.25353336e+00 1.17836857e+00
8.58636975e-01 -4.83831614e-01 -3.23137939e-01 -5.93981564e-01
8.75038922e-01 2.10235953e-01 6.05129480e-01 -4.84720677e-01
6.35959148e-01 -7.42886186e-01 -1.60253927e-01 -1.51846454e-01
7.63902143e-02 -1.61835745e-01 2.34116659e-01 9.18219388e-01
-6.78408444e-01 4.32588279e-01 -1.16969205e-01 2.07549945e-01
4.51387651e-02 -1.33874089e-01 6.26854539e-01 -5.33921897e-01
-4.29336637e-01 -3.96674871e-01 -7.44671345e-01 3.53139132e-01
9.07054722e-01 -5.46173304e-02 -9.20601070e-01 -6.45882308e-01
-4.88549590e-01 6.34587884e-01 5.79908550e-01 3.53686124e-01
3.32161844e-01 -9.92044926e-01 -1.05656660e+00 -1.94811702e-01
-2.13406920e-01 -1.73870362e-02 7.67418742e-02 1.17818451e+00
-4.19696301e-01 2.85794824e-01 -2.06490144e-01 -3.92623633e-01
-9.61800814e-01 2.27911800e-01 3.39562565e-01 -3.30964595e-01
-7.01059163e-01 8.03387940e-01 2.49576271e-01 -4.83940840e-01
-2.79027551e-01 -4.73429114e-01 7.32121989e-03 5.65933920e-02
6.61593020e-01 2.44390994e-01 -3.92446429e-01 -2.80787796e-01
-3.39781851e-01 2.24872082e-01 -2.65773803e-01 -3.98131609e-01
1.36637068e+00 -4.31846231e-01 -7.04132855e-01 5.08708179e-01
1.75285459e-01 3.24151576e-01 -1.00228059e+00 -3.07828337e-01
1.23369180e-01 -3.54843795e-01 -3.90805095e-01 -1.15795612e+00
-3.84776056e-01 5.47576189e-01 -3.30435753e-01 3.76843154e-01
5.67204833e-01 2.08740413e-01 5.06776631e-01 6.47697926e-01
6.65115178e-01 -1.12710857e+00 4.80131090e-01 8.31527352e-01
1.22014034e+00 -1.13877428e+00 -5.12608662e-02 -1.87973946e-01
-1.10492587e+00 1.05638409e+00 8.87740791e-01 -3.16700041e-02
9.10894424e-02 2.64188051e-01 3.19022954e-01 -4.37923521e-01
-1.44159222e+00 1.76038742e-01 -2.50378668e-01 6.60559833e-02
3.88676852e-01 2.15337723e-01 -8.16475928e-01 6.04421616e-01
-6.37813449e-01 2.24191900e-02 9.61365521e-01 9.07610059e-01
-6.07481182e-01 -9.58776355e-01 -5.68789661e-01 5.16462065e-02
-2.25476444e-01 -1.00960948e-01 -7.33167648e-01 5.82591832e-01
-1.49219558e-01 1.14193881e+00 -8.97154063e-02 6.78816438e-02
-8.85819793e-02 1.76381528e-01 3.71356606e-01 -6.01495683e-01
-9.70613062e-01 -4.58220951e-02 2.72946626e-01 -6.29046679e-01
-3.39580595e-01 -4.04747158e-01 -1.62682462e+00 -8.69637549e-01
-5.38509339e-02 4.34961021e-01 2.72529840e-01 1.28856850e+00
3.45017135e-01 1.68147519e-01 2.20333219e-01 -4.27314967e-01
-1.03781831e+00 -8.85893047e-01 -2.70861108e-02 3.23186725e-01
1.55427735e-02 -4.08858061e-01 -4.32563275e-01 -2.70833880e-01] | [9.603625297546387, 7.381516933441162] |
4c659cdd-bac0-4255-9808-22756fb038a1 | decoupled-spatial-temporal-transformer-for | 2104.06637 | null | https://arxiv.org/abs/2104.06637v1 | https://arxiv.org/pdf/2104.06637v1.pdf | Decoupled Spatial-Temporal Transformer for Video Inpainting | Video inpainting aims to fill the given spatiotemporal holes with realistic appearance but is still a challenging task even with prosperous deep learning approaches. Recent works introduce the promising Transformer architecture into deep video inpainting and achieve better performance. However, it still suffers from synthesizing blurry texture as well as huge computational cost. Towards this end, we propose a novel Decoupled Spatial-Temporal Transformer (DSTT) for improving video inpainting with exceptional efficiency. Our proposed DSTT disentangles the task of learning spatial-temporal attention into 2 sub-tasks: one is for attending temporal object movements on different frames at same spatial locations, which is achieved by temporally-decoupled Transformer block, and the other is for attending similar background textures on same frame of all spatial positions, which is achieved by spatially-decoupled Transformer block. The interweaving stack of such two blocks makes our proposed model attend background textures and moving objects more precisely, and thus the attended plausible and temporally-coherent appearance can be propagated to fill the holes. In addition, a hierarchical encoder is adopted before the stack of Transformer blocks, for learning robust and hierarchical features that maintain multi-level local spatial structure, resulting in the more representative token vectors. Seamless combination of these two novel designs forms a better spatial-temporal attention scheme and our proposed model achieves better performance than state-of-the-art video inpainting approaches with significant boosted efficiency. | ['Hongsheng Li', 'Jifeng Dai', 'Xiaogang Wang', 'Wenxiu Sun', 'Lewei Lu', 'Xiaoyu Shi', 'Yangyi Huang', 'Hanming Deng', 'Rui Liu'] | 2021-04-14 | null | null | null | null | ['video-inpainting'] | ['computer-vision'] | [ 2.80243009e-02 -3.08504313e-01 -2.10877955e-01 2.56709987e-03
-6.19205475e-01 -8.78005698e-02 4.42438602e-01 -2.28808314e-01
-1.83787182e-01 7.47715652e-01 3.74842376e-01 1.72486439e-01
-2.16624495e-02 -7.08588481e-01 -1.07846987e+00 -8.45445931e-01
2.88584650e-01 -1.01102181e-01 6.54726803e-01 -1.75437942e-01
9.43542793e-02 3.88188124e-01 -1.37926114e+00 5.30034781e-01
8.49678040e-01 1.08541954e+00 5.85024178e-01 4.60844874e-01
-2.31056929e-01 1.29457486e+00 -3.36949706e-01 -3.73914927e-01
2.62956977e-01 -4.94574785e-01 -5.20481944e-01 2.96757281e-01
6.27274811e-01 -7.42684901e-01 -8.80188227e-01 9.12503004e-01
2.07068637e-01 2.63339013e-01 2.13669255e-01 -1.02580273e+00
-1.00310266e+00 3.95492405e-01 -1.07825196e+00 3.14681470e-01
1.46900907e-01 2.96083927e-01 8.43325973e-01 -7.90793598e-01
3.80893022e-01 1.40697706e+00 4.92553711e-01 4.90556657e-01
-1.10306227e+00 -6.69541001e-01 5.53687751e-01 5.52728355e-01
-1.31886292e+00 -3.94120783e-01 9.84031141e-01 -3.18108767e-01
7.12032914e-01 3.53722334e-01 7.79026389e-01 9.86775696e-01
2.15722814e-01 1.07668531e+00 9.65381920e-01 -1.34823024e-01
4.41038609e-02 -3.26445252e-01 -3.27720702e-01 6.45063519e-01
-1.18657194e-01 1.11150071e-01 -5.35943627e-01 2.55483687e-01
1.37534237e+00 6.06115997e-01 -5.78571618e-01 -3.08071762e-01
-1.21590745e+00 5.55290699e-01 6.63491309e-01 5.38894415e-01
-6.10177696e-01 3.76794368e-01 3.19315910e-01 1.30156428e-01
4.84186679e-01 6.94593340e-02 -4.18117017e-01 -6.78742304e-02
-1.28539109e+00 2.77851969e-01 2.37854738e-02 9.42766964e-01
7.87468374e-01 2.63014883e-01 -5.26953101e-01 8.57274592e-01
1.95345819e-01 1.79996192e-01 3.93170506e-01 -9.92394626e-01
5.87817132e-01 4.64024931e-01 3.39582115e-01 -1.29996514e+00
1.50732994e-02 -5.34571886e-01 -1.06702876e+00 1.81700602e-01
1.68497831e-01 1.40432328e-01 -8.41512024e-01 1.86829531e+00
3.69868934e-01 6.61741793e-01 -2.32241914e-01 1.04895318e+00
6.52131796e-01 1.14437056e+00 9.71568301e-02 -4.46727514e-01
1.43526375e+00 -1.55506408e+00 -9.56662059e-01 -2.64259219e-01
4.25802125e-03 -8.19496214e-01 9.16442037e-01 2.91491956e-01
-1.34386396e+00 -8.18828285e-01 -1.00457823e+00 -5.68912208e-01
2.27916971e-01 9.56506133e-02 5.87701321e-01 3.02999914e-02
-9.64462519e-01 5.20532608e-01 -8.76483381e-01 -1.16584850e-02
5.63263237e-01 1.16453826e-01 -3.89656097e-01 -3.61611307e-01
-1.04879498e+00 6.38043582e-01 1.92609444e-01 2.81093806e-01
-1.02930868e+00 -7.50913203e-01 -8.28034163e-01 2.41556257e-01
4.06584680e-01 -7.48006344e-01 1.02927983e+00 -1.16399932e+00
-1.44558704e+00 5.88632166e-01 -4.29206729e-01 -3.02689224e-01
7.82374918e-01 -4.78743136e-01 -2.62088209e-01 2.13023275e-01
1.98221162e-01 6.96930945e-01 1.28455698e+00 -1.33696580e+00
-8.55916440e-01 -2.20137879e-01 9.86948013e-02 3.35052669e-01
-5.46224892e-01 -7.56007358e-02 -1.03350997e+00 -1.12614739e+00
-1.77577732e-03 -3.90940994e-01 -1.79271162e-01 3.36782604e-01
-2.13155523e-03 -7.63375685e-02 1.27383661e+00 -1.04954100e+00
1.51943171e+00 -2.21422052e+00 6.76373005e-01 -5.49283862e-01
4.58021849e-01 4.34414595e-01 -1.96096256e-01 2.82600433e-01
-3.09376474e-02 -2.67198265e-01 -2.24288836e-01 -5.33780754e-01
-3.41693878e-01 3.11228096e-01 -4.34437186e-01 1.94929406e-01
3.45720291e-01 8.67864072e-01 -1.07339871e+00 -3.91731501e-01
5.75360656e-01 6.28114402e-01 -7.08259463e-01 4.20684993e-01
-2.65400052e-01 5.83862841e-01 -4.91397649e-01 4.64855015e-01
1.01052225e+00 -5.45411855e-02 -1.67613000e-01 -4.98024762e-01
-2.65297592e-01 -4.86675389e-02 -9.25272882e-01 1.90238416e+00
-6.01503789e-01 5.81358254e-01 2.70742804e-01 -9.31090117e-01
7.82959640e-01 1.16957672e-01 4.12195385e-01 -9.74870503e-01
-2.67251283e-02 7.67484456e-02 -3.83247554e-01 -7.61143386e-01
4.92272556e-01 -1.56758308e-01 3.04931849e-01 1.07760429e-01
-6.85128942e-02 2.23234177e-01 -1.41412439e-02 1.74444243e-02
9.27769482e-01 5.48309326e-01 1.23005033e-01 -1.08340450e-01
6.08454645e-01 -4.75851089e-01 8.06232929e-01 4.93781865e-01
-2.17872009e-01 8.08322728e-01 3.55904877e-01 -8.33967388e-01
-1.20454121e+00 -8.93199563e-01 2.83820391e-01 9.46044505e-01
5.95874727e-01 -3.02698106e-01 -6.84742391e-01 -5.20990252e-01
-1.73466593e-01 4.48924303e-01 -7.40217447e-01 -1.21111743e-01
-9.69048560e-01 -3.92856330e-01 1.31674975e-01 5.67745149e-01
8.87756348e-01 -1.18098354e+00 -5.87189734e-01 3.00321847e-01
-5.56075692e-01 -1.08888841e+00 -9.64010656e-01 -2.12803602e-01
-8.40267777e-01 -8.23855937e-01 -1.18531406e+00 -1.02567565e+00
4.04401928e-01 6.65873468e-01 8.92438412e-01 1.75273731e-01
-1.66769654e-01 -1.37812898e-01 -4.77455109e-01 2.26508960e-01
-6.36750385e-02 -1.96722284e-01 -2.98828930e-01 5.98549962e-01
3.17479186e-02 -7.90512025e-01 -1.03637052e+00 2.59676993e-01
-1.32712126e+00 5.51455379e-01 6.58152819e-01 1.04143584e+00
4.57832694e-01 6.91153407e-02 2.65320361e-01 -3.04401577e-01
1.85658365e-01 -4.45057809e-01 -4.46860164e-01 2.50285298e-01
6.29202724e-02 -2.84276698e-02 8.73774409e-01 -3.83491814e-01
-1.18597424e+00 -1.90732107e-01 -3.13267410e-01 -1.06338418e+00
5.27147995e-03 1.94444321e-02 -3.41813236e-01 -2.67227236e-02
1.18498906e-01 6.30852163e-01 -4.70577367e-02 -5.89605570e-01
3.30072165e-01 3.64684582e-01 5.67181945e-01 -6.39267564e-01
7.59519994e-01 5.70939362e-01 -2.23901004e-01 -6.40835881e-01
-7.59035051e-01 -1.65675074e-01 -3.45201582e-01 -8.53945985e-02
9.75042760e-01 -1.00228405e+00 -5.38304210e-01 7.97342241e-01
-1.30274796e+00 -3.94472718e-01 -2.48173013e-01 2.59293467e-01
-5.31904876e-01 6.78417861e-01 -9.52580571e-01 -5.50848663e-01
-2.44258389e-01 -1.35179710e+00 1.30592942e+00 2.43567750e-01
1.76447213e-01 -7.85898328e-01 -1.62077844e-01 3.53553593e-01
4.76021588e-01 1.72894657e-01 8.84928286e-01 2.63196677e-01
-1.10856152e+00 2.05329731e-01 -6.31708443e-01 3.59156817e-01
3.44027996e-01 -1.10501051e-01 -7.94303417e-01 -4.54921752e-01
3.13561589e-01 -1.35553941e-01 1.02729356e+00 4.14243966e-01
1.36253107e+00 -4.56280589e-01 -2.30328441e-01 8.35927546e-01
1.45287549e+00 3.69280934e-01 9.47019696e-01 4.40480471e-01
1.07030463e+00 5.17703056e-01 7.14952648e-01 3.86453956e-01
2.34120920e-01 1.01447535e+00 5.69250524e-01 -3.91246498e-01
-3.71271551e-01 -3.89772236e-01 2.63915807e-01 7.93846309e-01
-1.32462367e-01 -1.98820159e-01 -3.08776021e-01 6.34293079e-01
-2.17483926e+00 -1.23381543e+00 1.34537756e-01 2.10205936e+00
8.54895830e-01 -8.07578713e-02 -2.27798261e-02 1.14932962e-01
7.93809593e-01 6.41988933e-01 -4.57356215e-01 -5.17778769e-02
-8.65780041e-02 1.47474661e-01 1.46414220e-01 6.28525198e-01
-1.04928899e+00 9.45718467e-01 4.81791973e+00 1.45566630e+00
-1.17409480e+00 3.62457603e-01 7.86566973e-01 -1.46669105e-01
-3.84779036e-01 2.90050842e-02 -4.76414651e-01 7.59988308e-01
1.10648349e-01 2.23370939e-01 5.35867810e-01 4.80448008e-01
4.70583498e-01 7.72870630e-02 -9.29296196e-01 1.22585797e+00
1.12719759e-01 -1.64857554e+00 4.56001461e-01 -2.03270778e-01
7.72050321e-01 -4.57641393e-01 1.77485898e-01 1.86025411e-01
-1.24390073e-01 -8.09205413e-01 1.12145853e+00 4.53491747e-01
7.52636969e-01 -7.79046535e-01 5.02000868e-01 1.91035613e-01
-1.50131583e+00 -3.70961308e-01 -3.28243047e-01 -1.19891644e-01
3.51473600e-01 5.28934240e-01 2.48521969e-01 8.10220301e-01
8.13419342e-01 1.17703688e+00 -2.89255261e-01 1.08045959e+00
-5.69393933e-02 2.30515063e-01 -3.86053929e-03 3.92039597e-01
5.72771668e-01 -3.39862317e-01 5.12747705e-01 1.03807032e+00
3.79148841e-01 2.29398772e-01 1.40579775e-01 8.61980259e-01
-5.35861515e-02 -1.51273742e-01 -3.37064475e-01 3.67857337e-01
3.62537801e-01 1.15699196e+00 -4.44779634e-01 -4.84658480e-01
-6.44141853e-01 1.38420010e+00 5.16241550e-01 4.56807256e-01
-1.21491885e+00 -2.32814029e-01 7.49544740e-01 2.25857407e-01
6.44073367e-01 -1.31653756e-01 2.06573829e-02 -1.41309893e+00
2.85744637e-01 -9.45111930e-01 1.35413617e-01 -9.24567461e-01
-1.20940840e+00 7.24215865e-01 -2.06341118e-01 -1.50686169e+00
3.41834456e-01 -3.21135253e-01 -5.55270314e-01 8.41189921e-01
-1.64155281e+00 -1.42488444e+00 -5.96373498e-01 8.45123470e-01
1.05167103e+00 1.81058884e-01 3.51566166e-01 7.83021033e-01
-7.03857064e-01 5.83135366e-01 1.27638072e-01 -1.19321629e-01
7.47254133e-01 -8.93312335e-01 2.83604890e-01 1.08032966e+00
-1.09164990e-01 3.75326544e-01 5.32541335e-01 -5.51202714e-01
-1.24680781e+00 -1.39362299e+00 7.26809204e-01 -1.00875326e-01
4.22596484e-01 -7.50455037e-02 -1.03053272e+00 6.01378798e-01
4.19235229e-01 1.01404525e-01 -1.15372010e-01 -5.28504491e-01
-4.13553178e-01 -4.23953921e-01 -9.40529644e-01 8.44426632e-01
1.09304023e+00 -4.01074976e-01 -5.63439012e-01 2.65581459e-01
9.76247966e-01 -4.42073464e-01 -5.03165066e-01 3.55424196e-01
4.61816818e-01 -1.30095005e+00 1.03991091e+00 -2.62001485e-01
9.56054807e-01 -6.87356234e-01 -7.64039084e-02 -9.07794356e-01
-6.43143296e-01 -9.08576548e-01 -2.78388888e-01 1.27970099e+00
-3.00761223e-01 -3.39402497e-01 6.66302085e-01 2.38312513e-01
-3.76067132e-01 -9.98496354e-01 -8.46940756e-01 -4.94531304e-01
-9.72038731e-02 -1.52132213e-01 5.00373781e-01 1.00390732e+00
-4.45011318e-01 1.64346218e-01 -1.07980192e+00 7.01777413e-02
5.70313454e-01 1.64063781e-01 6.20876133e-01 -6.07960582e-01
-5.47435999e-01 -5.38905740e-01 -3.51757705e-01 -1.70577145e+00
-1.71167210e-01 -3.52033198e-01 1.76430102e-02 -1.41322207e+00
4.33330685e-01 -3.68090957e-01 -2.78393984e-01 3.77040625e-01
-4.70906854e-01 4.43733990e-01 3.09561342e-01 3.45918208e-01
-5.29449999e-01 9.34293628e-01 1.60650778e+00 -2.36521065e-01
-1.50354370e-01 -3.12241226e-01 -5.88495195e-01 6.77807808e-01
3.28570634e-01 -2.89260715e-01 -5.67934752e-01 -8.53817821e-01
-2.89965689e-01 5.38618386e-01 5.83200514e-01 -9.52100694e-01
1.66631684e-01 -2.83375800e-01 5.19609630e-01 -5.00256181e-01
4.57942426e-01 -8.70660365e-01 2.88555026e-01 3.51462334e-01
-1.34743586e-01 2.13563189e-01 3.00254941e-01 7.59312451e-01
-5.61471760e-01 1.00527763e-01 9.36360955e-01 -1.24217734e-01
-8.15969408e-01 6.51432991e-01 -1.00875303e-01 -7.17618316e-02
1.22213769e+00 -3.20606351e-01 -5.67713603e-02 -2.57330447e-01
-5.78523397e-01 1.30202532e-01 6.05089188e-01 6.37650311e-01
7.99988389e-01 -1.49156618e+00 -5.86680889e-01 2.99967229e-01
-2.30696708e-01 1.02422714e-01 8.88281941e-01 8.33068192e-01
-8.63105893e-01 1.79963827e-01 -5.98351061e-01 -6.57068670e-01
-9.81882572e-01 7.97783017e-01 2.64541149e-01 -4.04520303e-01
-8.34576309e-01 1.07356906e+00 9.14917946e-01 2.73710251e-01
2.72263378e-01 -2.08724678e-01 -6.48131296e-02 -1.09663695e-01
7.44314075e-01 1.95974678e-01 -1.97519854e-01 -6.43600881e-01
-7.96291903e-02 9.20372427e-01 -3.74943286e-01 1.88306779e-01
1.21861768e+00 -4.23078656e-01 -3.16849828e-01 1.71996146e-01
1.14414191e+00 3.88065092e-02 -1.92797446e+00 -2.56178230e-01
-4.82341379e-01 -9.67179477e-01 -7.63718113e-02 -2.82835811e-01
-1.45561218e+00 1.05530250e+00 3.65714580e-01 -6.03896603e-02
1.46057045e+00 -3.42095613e-01 1.30015051e+00 -4.07330066e-01
3.91903609e-01 -5.24676979e-01 4.60189551e-01 2.80183107e-01
1.00148213e+00 -9.92056906e-01 -9.03433189e-02 -5.03313363e-01
-5.18865705e-01 9.36987579e-01 8.90486777e-01 -3.55700642e-01
1.96003959e-01 7.87153840e-02 -2.99534440e-01 1.11825287e-01
-7.13103592e-01 4.74033877e-02 2.67675370e-01 5.02326548e-01
2.73873508e-01 -1.82326391e-01 -3.33812207e-01 5.01351178e-01
4.61739421e-01 3.84854004e-02 1.99043885e-01 7.54714727e-01
-2.88782746e-01 -1.00058806e+00 -2.66568214e-01 1.10991955e-01
-3.90637130e-01 -2.18076244e-01 1.60109252e-01 6.55245662e-01
4.44297969e-01 6.72546983e-01 4.68833335e-02 -2.68221617e-01
3.07824045e-01 -4.04431641e-01 5.50844014e-01 -2.85951138e-01
-3.81381065e-01 4.13177341e-01 -2.95328498e-01 -8.87439132e-01
-4.60209757e-01 -3.99454057e-01 -7.56590605e-01 -3.49741638e-01
-7.08934572e-03 7.53705353e-02 -2.00991541e-01 8.50792348e-01
4.02428508e-01 7.16549635e-01 6.15727305e-01 -1.30053866e+00
-2.27462634e-01 -7.53932595e-01 -4.50423926e-01 5.59680581e-01
6.53896809e-01 -7.29892313e-01 -1.27808824e-01 1.62948504e-01] | [10.869129180908203, -1.3348575830459595] |
66d978a7-20bc-48d4-a18c-f8f9181852b0 | neuro-symbolic-approaches-for-context-aware | 2306.05058 | null | https://arxiv.org/abs/2306.05058v1 | https://arxiv.org/pdf/2306.05058v1.pdf | Neuro-Symbolic Approaches for Context-Aware Human Activity Recognition | Deep Learning models are a standard solution for sensor-based Human Activity Recognition (HAR), but their deployment is often limited by labeled data scarcity and models' opacity. Neuro-Symbolic AI (NeSy) provides an interesting research direction to mitigate these issues by infusing knowledge about context information into HAR deep learning classifiers. However, existing NeSy methods for context-aware HAR require computationally expensive symbolic reasoners during classification, making them less suitable for deployment on resource-constrained devices (e.g., mobile devices). Additionally, NeSy approaches for context-aware HAR have never been evaluated on in-the-wild datasets, and their generalization capabilities in real-world scenarios are questionable. In this work, we propose a novel approach based on a semantic loss function that infuses knowledge constraints in the HAR model during the training phase, avoiding symbolic reasoning during classification. Our results on scripted and in-the-wild datasets show the impact of different semantic loss functions in outperforming a purely data-driven model. We also compare our solution with existing NeSy methods and analyze each approach's strengths and weaknesses. Our semantic loss remains the only NeSy solution that can be deployed as a single DNN without the need for symbolic reasoning modules, reaching recognition rates close (and better in some cases) to existing approaches. | ['Claudio Bettini', 'Gabriele Civitarese', 'Luca Arrotta'] | 2023-06-08 | null | null | null | null | ['human-activity-recognition', 'human-activity-recognition'] | ['computer-vision', 'time-series'] | [ 0.24119183 0.16610295 -0.5668742 -0.51644737 -0.25298545 -0.3161402
0.50821394 0.02179015 -0.6594799 0.9752991 -0.13416228 -0.19056088
-0.4767167 -0.7748696 -0.77746755 -0.40373668 0.17179866 0.5984843
0.18930385 -0.09326394 -0.2359384 0.5294418 -1.7524023 0.22156788
0.88845116 1.3251237 -0.01001026 0.30164558 -0.13810469 1.1999191
-0.84402454 -0.30773386 0.01989499 -0.4689398 -0.7005932 -0.27020052
0.3548022 -0.02826271 -0.09457415 0.8256148 0.37463588 0.13134676
0.18882008 -1.6849698 -0.33594227 0.82477504 0.2073175 0.03561771
0.35478202 0.10810463 0.8233284 -0.43348876 0.56383544 0.99366105
1.1225471 0.7030353 -1.2355158 -0.5569366 0.3194575 0.35289982
-1.46318 -0.3822689 0.92040414 -0.14888807 1.2978661 0.39507255
0.8013535 1.7529609 -0.27236336 1.0258168 1.1147472 -0.40475973
0.8558967 0.05534011 0.08985608 0.59589654 0.40538293 -0.08154938
-0.79998463 -0.09035292 0.5681336 0.25737706 0.01549945 -0.5375772
-1.0747852 0.5757289 0.54705626 0.72494876 -0.5091058 0.44239783
0.43150163 0.09033134 0.20138924 0.69318753 -0.5516422 -0.56656414
-1.2913765 0.43911847 0.8809182 0.90756917 0.46531242 0.22608124
-0.11356568 0.7755561 -0.00609329 0.44016466 0.6420487 -0.982814
0.31720772 0.8139599 0.09744798 -0.85707235 -0.7481254 -0.507236
-0.70001495 -0.08720171 0.50644493 -0.15597795 -0.86904794 1.8648067
-0.03097223 0.74512345 0.15493812 0.96850115 0.56086236 0.06623743
0.51030177 0.09470607 1.1797894 -0.9833756 -0.7363621 -0.4942465
0.7739154 0.08150349 1.3187985 0.64564955 -0.8024834 -0.30789414
-1.2889208 -0.03802845 -0.78961736 0.07133518 1.0771519 0.78142136
-0.8147298 0.71029437 -1.0396641 -0.6757313 0.671956 0.36834356
-0.0698422 0.23680675 -1.331688 1.0213808 0.49329415 0.13648531
-0.8801918 -0.4266618 -0.817996 0.13578959 0.6617361 -0.6120269
1.3273257 -1.1184546 -1.4399643 0.6026554 0.10144507 -0.9885854
0.5801935 -0.28003 -0.6393909 -0.10760231 -0.13868923 0.39460054
0.51889795 -0.94896424 -0.35282326 -0.24483341 0.39940438 -0.12171508
-0.4233532 -0.26849976 -0.12437364 -0.69679713 -0.13796759 -0.9817815
-0.11590393 -0.01442998 -0.22718818 -0.24441047 0.8938485 -0.45749438
1.2510403 -2.1027546 0.08819795 0.1400648 0.02767529 0.7120442
0.05427806 0.1377408 0.22466986 0.02646504 -0.39565638 -0.32680374
0.20387955 0.7723019 -0.22393507 0.05737973 0.0391118 0.94550747
-0.9804365 -0.36231112 0.2181522 0.48750404 -0.44755033 0.02890679
-0.65500855 0.16477002 -0.2813999 0.8381614 0.15837206 -0.19693707
0.54063886 0.07438881 0.26614758 0.34256405 -1.0462502 1.9155078
-0.868256 0.46637404 -0.24057451 -1.2414443 0.91484296 0.29266763
0.39304253 -0.8443916 0.32319027 0.39261064 -0.17936508 -0.509616
0.19957677 -0.00981932 -0.35765338 0.01088514 0.19399078 0.28957802
0.03157261 -0.30602196 1.1735331 0.4468517 0.28434333 0.03156561
0.37946597 0.0191279 0.8517506 0.9772488 -0.3520076 0.30459705
0.30129272 -0.5956372 -0.66912127 -0.82798153 -0.00897334 1.2218821
0.08666127 -0.483968 -1.0153861 -0.94590527 0.06175386 1.0120155
-0.52065444 -0.27704182 -0.58140326 -0.61276746 1.2401843 0.9214565
0.8828887 -1.2244931 -1.1440909 0.42480308 -0.24383563 -1.4319817
0.38186488 0.5336655 -1.0288324 -1.0230324 -0.38677427 -0.34260374
0.14431275 -0.28655687 1.008897 0.11045536 -0.10631828 0.24606816
-0.44147184 -0.64251006 -0.02076312 0.36579105 0.2812055 -0.01999077
0.87818563 -0.898962 -0.3014587 0.2947696 -0.59041685 -0.16767089
0.49248305 0.7499435 0.4261144 -0.06302773 0.67584664 -0.88767666
0.42999306 -0.5225921 -0.46850005 0.40700164 -0.89048856 0.09972923
0.7276712 -0.6180772 -0.93755424 0.02111972 -0.02561738 -0.6140154
-0.37985265 0.5864038 -0.25094777 0.01741743 0.90382594 0.07998445
-0.28224903 -0.7937452 0.12580049 0.649044 0.69745755 -0.77493787
0.37102807 0.28179055 0.07621595 -0.6978214 -1.0162748 -0.13668554
-0.38757238 0.10564512 0.84097165 -0.87249446 -0.7447317 0.45384356
-0.87804604 -0.55987054 -0.4784495 0.3988105 -0.73697114 0.0155919
-0.3533531 -0.99414647 -0.23755763 -0.85176665 1.0196702 0.20548591
-0.53542805 -0.95902985 -0.2947626 0.6006791 0.66011566 0.45745862
0.6354983 -0.98142445 -0.26901537 -0.22468361 -0.02946717 0.38682383
-0.3218126 -0.5622205 -1.3121531 0.12136097 -0.16397321 -0.44985384
0.541704 -0.01528865 1.4879277 -0.405761 -0.2528317 0.70047075
1.2716395 0.2886391 0.6557096 0.58940464 0.77895963 0.24752448
0.57144564 0.3894499 0.4209671 0.93954915 0.30276126 0.0296719
-0.29146203 -0.46905965 0.29056555 0.20896241 -0.3562298 -0.2247519
-1.1318928 0.4798115 -2.2430792 -0.94906425 0.14354505 1.8956175
0.7551147 0.19204043 0.34050432 0.51327854 0.2559157 0.05512174
-0.6711047 -0.62296 -0.20986767 0.34009323 0.3238379 0.16427879
-1.2101798 1.0037707 5.9332337 0.61567366 -1.2434164 0.37416288
0.1128794 -0.3421395 0.08347369 -0.0596756 -0.7302067 0.64841396
1.3056827 0.32836506 0.5877606 1.4265306 0.03582419 -0.15939474
-1.373662 1.4085786 0.1191007 -1.3141196 -0.27634853 -0.1218775
0.36294442 0.08248202 -0.3579582 0.61615103 0.16367976 -1.0273216
0.950644 0.6785737 0.62454355 -0.6141562 0.8781308 0.48864865
-0.7346938 -0.36935183 0.02518852 -0.45484796 -0.08919729 0.66631776
-0.78373814 0.2499965 0.8389599 0.5982238 -0.7407074 0.6573686
-0.397174 0.76510656 -0.48944533 -0.2687279 0.17561354 0.2512857
0.32430044 1.2613925 0.22485314 -0.2517344 0.29023924 0.88602394
-0.05622675 -0.29454592 -0.6670797 -0.14645323 0.6283733 0.6908743
-0.58465636 -0.31809983 -0.18417777 0.97716415 0.30836722 0.3569669
-1.0381014 -0.4774746 0.62470514 0.25679705 0.16361888 -0.24670428
-0.6870891 -1.1765366 0.29987726 -0.8941888 0.47025794 -0.6655673
-0.9318381 0.39928138 0.2459535 -0.9671549 -0.48414668 -0.5671285
-0.22617857 0.43455163 -1.4658219 -1.2676404 -0.46247348 0.6257829
0.20586918 -0.1978992 1.0494645 0.4659642 -0.60298043 0.9123301
-0.3037285 0.150459 0.15596305 -1.0746783 0.11295862 0.7355553
0.2426132 0.70437074 0.67152125 -0.39305818 -1.3076296 -0.953684
1.0945601 -0.61238223 0.42094696 -0.7752829 -0.84795743 0.86038107
-0.3710906 0.2115671 0.7628244 0.44524726 -0.5343494 -0.21947646
-1.4756562 0.5343031 1.5913299 -0.61786675 -0.70824045 0.12984535
0.5748636 -0.24038753 -0.8048877 0.40848827 0.5819674 -0.8600113
0.8845577 -0.60756195 0.02611224 -0.32932138 -0.36580968 -0.94929415
0.14585966 -0.38807476 -0.6488925 1.0351617 0.24164969 -0.6669028
0.9449156 0.7433287 -0.08978414 -0.6631339 -1.0901418 -1.3359085
-0.42705467 -0.87623686 0.94298905 1.1314621 0.04470625 0.21032082
-0.40160733 -0.09592555 0.39874583 -0.15304069 0.58642566 -1.4726641
-0.36611307 -0.3235075 -0.76508784 -0.41696748 0.37702626 -0.91782916
-0.02927727 -1.3308232 -0.2248008 -0.7021689 -0.53641903 1.251466
0.47940418 0.15701874 0.3591982 0.07848577 -0.70581245 0.4248241
0.5546295 -0.11450157 -0.2444649 -0.24598238 -0.49639118 1.006827
0.9551655 -0.4177405 -0.690715 -0.34590048 0.2696612 -0.13222377
0.7563986 -1.7205602 0.20819958 -0.34636387 0.291871 -0.02965336
0.53431946 -1.1062055 0.2051415 0.34551108 -0.28768253 -0.34299764
0.12242327 0.396175 -0.14899588 -0.15479213 0.4957841 0.02607374
-0.97610235 -0.14838673 -0.20134127 0.06811707 0.9718922 -0.4042438
-0.1933881 -0.03020026 -0.8444823 -0.01112555 0.33823746 0.6373127
0.35646963 -1.1876296 -0.01743405 0.17799994 0.2675877 0.07689651
-0.00799564 0.7922222 -0.37350798 0.5782653 -0.21123493 -0.4611396
-0.9397986 0.57605374 0.58398575 -0.23641504 -0.69173366 0.6296991
-0.55641556 -0.5855131 0.75517577 -0.6510513 0.01182042 -0.01602506
0.42131004 0.5553412 0.35649562 -0.23683241 -0.726851 0.07933128
0.30845544 0.0122769 1.2847134 0.375368 0.28435263 0.570511
0.85867023 -0.5584558 -1.1211371 -0.22161989 0.5431692 -0.20134872
0.05128942 -1.2973042 -0.8722125 0.8587527 0.7646543 -0.05792458
1.2036042 -0.12110168 0.92947495 0.9183057 0.79811704 -1.4441826
-0.12758267 0.47256398 0.61614084 -1.0414716 -0.17465067 -0.1233317
-0.62566245 0.700946 0.660241 0.00797902 0.4889025 0.20923917
-0.06224624 -0.24334857 -0.5098766 -0.2701456 -0.04384389 0.9083304
0.25408208 0.30389026 -0.34583673 1.1684074 -0.22732614 0.714008
0.19346327 1.2512925 -0.12272786 -1.0894381 -0.2487306 0.26742905
-0.33499065 0.16408317 -0.41384187 0.77482754 0.6104704 0.9073587
-0.22625023 -0.595468 0.6398374 0.61392546 0.38220146 -0.5186594
-0.66640645 -0.457631 0.4561066 -0.93430233 -0.5052716 -0.65573716
-1.4444561 -0.22201338 0.17548203 -0.1696399 0.8692136 1.1147598
0.6506804 0.53849006 0.03935297 -0.6919742 -0.52690697 -0.6929406
-0.5198783 0.48404172 0.10954316 -1.0637413 -0.15937185 -0.20223977] | [7.808737277984619, 0.9772263169288635] |
eb1253c6-922b-477f-ad01-d2264263f1be | a-structured-span-selector | 2205.03977 | null | https://arxiv.org/abs/2205.03977v2 | https://arxiv.org/pdf/2205.03977v2.pdf | A Structured Span Selector | Many natural language processing tasks, e.g., coreference resolution and semantic role labeling, require selecting text spans and making decisions about them. A typical approach to such tasks is to score all possible spans and greedily select spans for task-specific downstream processing. This approach, however, does not incorporate any inductive bias about what sort of spans ought to be selected, e.g., that selected spans tend to be syntactic constituents. In this paper, we propose a novel grammar-based structured span selection model which learns to make use of the partial span-level annotation provided for such problems. Compared to previous approaches, our approach gets rid of the heuristic greedy span selection scheme, allowing us to model the downstream task on an optimal set of spans. We evaluate our model on two popular span prediction tasks: coreference resolution and semantic role labeling. We show empirical improvements on both. | ['Mrinmaya Sachan', 'Ryan Cotterell', 'Yuchen Eleanor Jiang', 'Tianyu Liu'] | 2022-05-08 | null | https://aclanthology.org/2022.naacl-main.189 | https://aclanthology.org/2022.naacl-main.189.pdf | naacl-2022-7 | ['semantic-role-labeling'] | ['natural-language-processing'] | [ 5.5699050e-01 6.2005937e-01 -6.4461058e-01 -6.1335194e-01
-1.1220137e+00 -8.5002315e-01 3.9281285e-01 6.0116422e-01
-5.3844756e-01 9.7011989e-01 8.7378234e-01 -4.1056314e-01
-2.1336296e-01 -7.4760723e-01 -4.2218128e-01 -2.6778615e-01
1.0162680e-01 8.9137125e-01 5.2195334e-01 -2.7521592e-01
5.7996869e-01 3.5062772e-01 -1.4079061e+00 5.1495284e-01
8.3619028e-01 5.2978265e-01 3.5611093e-01 4.7766283e-01
-3.4808543e-01 8.9174789e-01 -2.6082882e-01 -2.8243986e-01
-1.1548639e-01 -5.9513134e-01 -1.5598683e+00 -1.3049698e-01
3.8543975e-01 3.1429514e-02 -5.1713455e-03 9.3102926e-01
3.2178968e-01 6.3388610e-01 4.8637301e-01 -9.3643916e-01
6.4553939e-02 1.4085987e+00 -3.7369475e-01 4.2498466e-01
5.3834850e-01 -2.5426784e-01 1.7872163e+00 -4.9656093e-01
9.7582150e-01 1.4560357e+00 5.3158087e-01 6.6374862e-01
-1.5211453e+00 -3.0296376e-01 5.0103062e-01 9.2135713e-02
-1.0319920e+00 -6.3672632e-01 9.4452280e-01 -2.0798483e-01
9.3397146e-01 2.8076476e-01 2.3749335e-01 7.9335618e-01
-2.2312821e-01 8.0631280e-01 9.8337841e-01 -7.1316344e-01
1.3377593e-01 -3.0427736e-01 5.5075121e-01 6.2204146e-01
2.3037402e-01 -1.4385919e-01 -8.0836976e-01 -6.1731172e-01
3.0199930e-01 -5.9669948e-01 -3.3764514e-01 -2.2280914e-01
-1.2130961e+00 8.4997624e-01 -6.0277976e-02 4.2132255e-01
-2.9843119e-01 3.4493897e-02 6.0879219e-01 2.5197002e-01
4.2405632e-01 8.7730771e-01 -5.9391922e-01 2.8197050e-02
-1.0664636e+00 5.5207658e-01 9.9666578e-01 1.0336620e+00
7.8015846e-01 -7.1208549e-01 -4.1580412e-01 9.8918980e-01
1.3538381e-01 -1.6967386e-01 2.0011394e-01 -1.3629494e+00
6.4108354e-01 4.7951716e-01 3.7031391e-01 -6.2280166e-01
-7.0507020e-01 1.6927122e-01 4.8645675e-02 -2.2589266e-01
9.2319864e-01 -1.3691936e-01 -6.2450373e-01 2.2157390e+00
3.3782136e-01 -4.5016389e-02 -3.1619266e-02 8.5030091e-01
2.7721879e-01 3.6798567e-01 6.3275051e-01 -5.9520346e-01
1.6285633e+00 -5.6933975e-01 -4.5734435e-01 -2.6986396e-01
8.9291245e-01 -6.6934043e-01 1.1145474e+00 3.4043720e-01
-1.1288437e+00 -9.7669773e-02 -6.8735129e-01 -3.8592067e-01
2.1194394e-01 -3.1349838e-01 7.0180190e-01 3.1828474e-02
-9.3728912e-01 9.6450585e-01 -5.2703512e-01 -5.8106428e-01
5.1145818e-02 2.6464090e-01 -2.0578779e-01 1.4379592e-02
-1.4193369e+00 8.8044220e-01 8.6809999e-01 -5.0329781e-01
-4.3961361e-01 -5.6090003e-01 -9.4605762e-01 2.6324129e-01
8.2567596e-01 -7.2855717e-01 1.6487615e+00 -9.6964502e-01
-9.7406691e-01 1.2277418e+00 -6.3176841e-01 -6.6048700e-01
1.0523957e-01 -3.1209210e-01 -9.9486925e-02 3.7431234e-01
3.6908916e-01 7.7377254e-01 4.6018097e-01 -1.1060457e+00
-8.1830674e-01 -3.7179092e-01 3.6495158e-01 6.0304916e-01
-5.0736990e-02 5.2391189e-01 -1.8471541e-01 -4.4894961e-01
3.6034361e-01 -9.9578011e-01 -3.6418161e-01 -2.2513483e-01
-5.6320900e-01 -9.0776646e-01 2.4132478e-01 -4.0032390e-01
1.4941413e+00 -1.7558299e+00 1.3110508e-01 1.5516856e-01
-2.7547333e-02 -2.1277906e-01 -5.6714684e-02 6.1550105e-01
-8.3630398e-02 2.5678983e-01 -2.6873249e-01 -1.8671612e-01
3.3286806e-02 2.5512898e-01 -4.7702968e-01 2.6302972e-01
5.2995842e-02 4.0176451e-01 -1.1469356e+00 -9.7031510e-01
-2.8367692e-01 -2.8086007e-01 -6.3522887e-01 1.1039255e-01
-6.8280888e-01 2.9505220e-01 -6.9839352e-01 3.8513672e-01
1.6008435e-01 -9.7956046e-02 7.0279843e-01 -6.6672288e-02
-2.2227333e-01 1.2309802e+00 -9.3424404e-01 1.8862617e+00
-2.8082916e-01 2.7823400e-01 1.4351866e-01 -9.7716957e-01
6.5587336e-01 3.5925484e-01 5.5647057e-01 -3.5846165e-01
-8.7012298e-02 2.2882618e-01 7.2362185e-02 -4.1609380e-01
5.7100403e-01 -5.1800323e-01 -5.8654708e-01 7.7943909e-01
-1.2207907e-01 1.4952973e-01 4.8416302e-01 3.7372631e-01
1.1693890e+00 3.9885533e-01 7.9973859e-01 -6.3906640e-01
5.4735237e-01 5.7281429e-01 1.1345989e+00 7.6748151e-01
-2.3024665e-01 3.7289321e-01 8.8861692e-01 -3.0998203e-01
-1.0006196e+00 -6.4657128e-01 3.6872551e-02 1.6416907e+00
1.9749887e-01 -5.1361650e-01 -8.8250548e-01 -1.0664452e+00
-2.3790509e-01 1.1169506e+00 -2.6389363e-01 -1.4573442e-02
-1.1671131e+00 -2.9624960e-01 6.2720680e-01 4.8956162e-01
-6.9160396e-03 -1.3837247e+00 -7.4331665e-01 5.2805668e-01
-6.2887168e-01 -1.0109848e+00 -7.6109058e-01 2.0705995e-01
-9.5496994e-01 -1.4152045e+00 2.2581983e-01 -7.8282893e-01
4.0173188e-01 1.2510231e-01 1.4279836e+00 1.8456064e-01
1.5296189e-01 1.7574987e-01 -4.0093070e-01 -2.4878442e-01
-3.8370436e-01 3.9039734e-01 -2.3951367e-02 -3.3634073e-01
7.2258061e-01 -6.3225132e-01 -5.4556674e-01 1.3478623e-01
-4.8752296e-01 4.8094876e-02 3.0499530e-01 7.3829818e-01
6.4032340e-01 -3.3809051e-01 7.2390693e-01 -1.5288293e+00
7.8350711e-01 -4.7946790e-01 -2.4173571e-01 4.2813271e-01
-7.1888614e-01 4.6208569e-01 6.6246158e-01 -1.9988576e-01
-1.3317850e+00 8.8102743e-02 -1.4577729e-01 -3.8477369e-02
-3.9154145e-01 4.8803097e-01 -3.9283180e-01 5.4015809e-01
7.6382530e-01 -2.0577000e-01 -2.7024096e-01 -6.3774937e-01
3.7170026e-01 3.2314813e-01 4.1342333e-01 -1.2133653e+00
3.6321849e-01 2.2740920e-01 -1.3497868e-01 -4.4372651e-01
-1.4506049e+00 -7.1882331e-01 -5.0429088e-01 1.7042878e-01
7.6588237e-01 -6.5747613e-01 -5.9568942e-01 -2.8931266e-01
-1.3723321e+00 -4.1719636e-01 -1.9691810e-01 3.2910711e-01
-9.1104603e-01 4.7854853e-01 -6.8203729e-01 -8.0227321e-01
-4.2753297e-01 -4.8997158e-01 9.9304616e-01 8.5015506e-02
-9.9749619e-01 -8.6372542e-01 1.9895218e-02 2.7920970e-01
-8.8881172e-02 2.6049581e-01 1.2423292e+00 -1.1259778e+00
-2.0591530e-01 3.1028047e-01 3.8131978e-02 -3.1009772e-01
-1.4652745e-01 -4.9637431e-01 -7.4699658e-01 -1.0187153e-02
-9.6690699e-02 -4.2904514e-01 9.5138806e-01 2.3415188e-01
1.2021420e+00 -5.1548046e-01 -6.2421691e-01 3.0851948e-01
1.2937694e+00 1.7131133e-01 4.0340763e-01 3.3575708e-01
4.3817383e-01 1.0366280e+00 1.1584314e+00 1.2453454e-01
3.8474849e-01 6.6495341e-01 1.1591434e-01 5.0328720e-01
2.2968287e-02 -5.5231816e-01 1.1017466e-01 6.2979542e-02
-4.2583518e-02 -3.4877822e-01 -9.7164756e-01 7.6004797e-01
-1.9531833e+00 -1.1593153e+00 -3.2510094e-02 2.0337689e+00
1.1596773e+00 4.4046393e-01 1.6491371e-01 4.5512699e-02
9.1575843e-01 3.3354706e-01 -5.1965249e-01 -5.4983234e-01
2.1380480e-01 2.2751838e-01 3.1625536e-01 6.8850571e-01
-1.1304908e+00 1.2742820e+00 6.1724787e+00 5.0908917e-01
-6.2036270e-01 8.4909819e-02 3.8884225e-01 -1.6136119e-01
-5.4255241e-01 4.4880888e-01 -1.0714004e+00 2.4389979e-01
8.8525295e-01 -2.7085274e-01 2.8041330e-01 7.5606251e-01
3.2147610e-01 -1.9938062e-01 -1.3341730e+00 4.0375647e-01
-2.1661345e-02 -1.1519946e+00 7.0124529e-02 -2.0804517e-01
1.9316342e-01 -2.7131817e-01 -7.3015869e-01 2.5288835e-01
7.5100625e-01 -7.9799408e-01 8.1086594e-01 3.1322059e-01
6.7102855e-01 -7.8236634e-01 2.8333902e-01 5.1659644e-01
-1.0309635e+00 -3.1689018e-01 -3.0541012e-01 -8.3307751e-02
3.7707859e-01 5.2197808e-01 -7.1370882e-01 3.3193335e-01
3.4128314e-01 3.0214801e-01 7.3649525e-03 7.4801350e-01
-6.7740673e-01 9.9060076e-01 -2.0043573e-01 8.8183582e-02
1.5106417e-01 -4.8755392e-02 8.5927808e-01 1.3180163e+00
2.0241848e-01 4.9454880e-01 5.4205269e-01 6.0529453e-01
-1.7416300e-01 1.8419872e-01 -5.5448997e-01 8.8161588e-02
1.0904208e+00 1.0654562e+00 -7.6930237e-01 -3.6063525e-01
-3.0081195e-01 5.2107406e-01 7.4343026e-01 1.4642200e-01
-5.2012199e-01 -2.8771809e-01 7.1481007e-01 3.7566423e-01
1.4249726e-01 -6.6881850e-02 -5.5224484e-01 -9.4157356e-01
-2.0279452e-01 -7.1084774e-01 1.1728041e+00 -4.9828497e-01
-1.0192086e+00 3.1826308e-01 3.2224688e-01 -7.8586441e-01
-4.0003395e-01 -3.1393501e-01 -8.1851840e-01 7.3384714e-01
-1.4043552e+00 -9.1842854e-01 3.4469470e-01 3.2020453e-01
7.7670753e-01 3.8642469e-01 7.1686798e-01 -2.2874837e-01
-1.2305561e-01 2.2116731e-01 -6.6045868e-01 1.3784964e-01
9.6694964e-01 -1.3974674e+00 8.3771452e-02 8.5343635e-01
6.1969366e-02 7.5896734e-01 1.1512048e+00 -6.0665798e-01
-8.8744789e-01 -9.4343781e-01 1.6486872e+00 -3.8570914e-01
5.9280568e-01 -2.0730160e-01 -1.0494785e+00 9.6013319e-01
2.3957348e-01 -3.1948239e-01 8.0554098e-01 7.4697810e-01
-5.1562256e-01 5.9088144e-02 -1.0236168e+00 4.9147394e-01
1.7329054e+00 -3.2234547e-01 -1.3116928e+00 1.0368701e-01
8.8796657e-01 -3.0417877e-01 -7.0018548e-01 2.1431889e-01
2.3411159e-01 -7.4650186e-01 6.9496876e-01 -9.5753664e-01
4.5580524e-01 -2.2074363e-01 1.9294951e-02 -1.2863355e+00
-4.9757639e-01 -6.9029009e-01 -1.3086937e-01 1.2244351e+00
6.0798180e-01 -3.3416107e-01 6.2970889e-01 9.4518822e-01
-4.8522779e-01 -5.9598404e-01 -8.3505332e-01 -4.2949182e-01
1.7757792e-02 -2.4247894e-01 6.6702497e-01 9.3207413e-01
5.0501823e-01 8.2709116e-01 -1.8132411e-01 2.0226486e-02
6.7091650e-01 7.9778087e-01 1.7716597e-01 -1.5647050e+00
-2.9706222e-01 -4.4318920e-01 3.3523402e-01 -9.2279798e-01
6.5756381e-01 -9.9869341e-01 3.0139786e-01 -1.4922947e+00
1.8307923e-01 -7.3465884e-01 -4.2629454e-01 7.9567903e-01
-4.2103848e-01 -1.7720535e-01 1.3506266e-01 2.1501596e-01
-9.2755693e-01 -3.2311395e-02 9.8233849e-01 1.2321425e-01
-3.9429280e-01 -8.5589126e-02 -1.3229411e+00 9.6374547e-01
1.0185202e+00 -6.3434565e-01 -5.8715212e-01 -2.9162550e-01
4.2059571e-01 4.7534680e-01 -7.4551329e-02 -4.5472178e-01
2.5659582e-01 -7.1553415e-01 -3.8467523e-02 -3.0759174e-01
-1.8951509e-02 -3.8332394e-01 -1.6001278e-01 1.5777358e-01
-9.1631937e-01 -2.0876376e-01 -9.2910022e-02 5.3095210e-01
-9.2138164e-02 -5.6564695e-01 6.5841091e-01 -5.7560086e-01
-9.4755489e-01 1.2123818e-01 -3.3682710e-01 4.4132236e-01
8.2355785e-01 -1.7219868e-02 -1.9651774e-01 -5.2798443e-02
-1.0383382e+00 5.0930166e-01 5.4393989e-01 2.9087356e-01
2.9637367e-01 -1.0101191e+00 -7.1698862e-01 -4.6828341e-01
2.2766671e-01 2.4460629e-01 -1.4207202e-01 7.0550430e-01
1.7555714e-02 5.0388169e-01 1.4931396e-01 -1.9179417e-02
-1.4856061e+00 6.4955527e-01 1.3351546e-01 -7.3729169e-01
-7.2843719e-01 6.9302905e-01 2.3347053e-01 -1.2285444e-01
-1.0284738e-02 -4.2573456e-02 -5.7666665e-01 2.8671885e-01
6.4274508e-01 1.5260351e-01 -1.9755740e-01 -4.8065525e-01
-4.0472811e-01 9.1208078e-02 -1.6208866e-01 -2.6960081e-01
1.1607665e+00 -3.2982361e-01 -4.4019145e-01 4.3011066e-01
6.5102541e-01 1.7597964e-01 -1.1187820e+00 -4.7505438e-01
9.8368889e-01 -2.9133755e-01 -4.0872768e-01 -6.0778111e-01
-3.0510277e-01 4.1517580e-01 -4.0708372e-01 2.3617494e-01
1.0754673e+00 3.7254435e-01 7.6012200e-01 4.0518722e-01
6.7842948e-01 -1.3224300e+00 -2.8242272e-01 6.2611133e-01
7.1030849e-01 -7.9457688e-01 -1.1644766e-01 -7.7467531e-01
-6.2498510e-01 1.1412445e+00 8.4951717e-01 -8.5312665e-02
1.7384762e-01 1.3172214e-01 -2.1234237e-01 -2.6819634e-01
-1.2564412e+00 -4.2985958e-01 3.9736517e-02 3.4387764e-01
8.2849878e-01 1.9197650e-01 -1.1162393e+00 4.2170036e-01
-2.7871692e-01 -2.2529407e-01 3.6155891e-01 9.4563431e-01
-9.3188578e-01 -1.3775885e+00 -4.9962261e-01 7.0882893e-01
-7.4089217e-01 -1.5908636e-01 -5.5546886e-01 4.4542101e-01
1.4144877e-01 1.0025120e+00 8.5978828e-02 1.0685466e-01
3.0412883e-01 7.3375702e-01 5.4496980e-01 -1.1153899e+00
-4.8986948e-01 2.9423418e-02 9.4672793e-01 -5.9001189e-01
-5.5302739e-01 -9.7425425e-01 -1.6784704e+00 -3.8060542e-02
-1.8612918e-01 6.0494339e-01 -2.2308032e-03 1.1009603e+00
1.9974397e-01 1.2606469e-01 2.6082140e-01 -4.1818005e-01
-7.6854652e-01 -8.8400269e-01 -4.5706901e-01 8.1106406e-01
1.0694211e-02 -6.6991162e-01 -7.1662597e-02 2.7500698e-02] | [10.1627197265625, 9.27225399017334] |
e0595421-3fc7-4ab7-9d0a-2e1e3cd16c9c | principled-and-efficient-motif-finding-for | 2302.04599 | null | https://arxiv.org/abs/2302.04599v3 | https://arxiv.org/pdf/2302.04599v3.pdf | Principled and Efficient Motif Finding for Structure Learning of Lifted Graphical Models | Structure learning is a core problem in AI central to the fields of neuro-symbolic AI and statistical relational learning. It consists in automatically learning a logical theory from data. The basis for structure learning is mining repeating patterns in the data, known as structural motifs. Finding these patterns reduces the exponential search space and therefore guides the learning of formulas. Despite the importance of motif learning, it is still not well understood. We present the first principled approach for mining structural motifs in lifted graphical models, languages that blend first-order logic with probabilistic models, which uses a stochastic process to measure the similarity of entities in the data. Our first contribution is an algorithm, which depends on two intuitive hyperparameters: one controlling the uncertainty in the entity similarity measure, and one controlling the softness of the resulting rules. Our second contribution is a preprocessing step where we perform hierarchical clustering on the data to reduce the search space to the most relevant data. Our third contribution is to introduce an O(n ln n) (in the size of the entities in the data) algorithm for clustering structurally-related data. We evaluate our approach using standard benchmarks and show that we outperform state-of-the-art structure learning approaches by up to 6% in terms of accuracy and up to 80% in terms of runtime. | ['Efthymia Tsamoura', 'Dominic Phillips', 'Jonathan Feldstein'] | 2023-02-09 | null | null | null | null | ['relational-reasoning'] | ['natural-language-processing'] | [ 3.59875560e-01 2.65509486e-01 -4.05831695e-01 -3.12862754e-01
-5.92062056e-01 -4.57436383e-01 4.44943905e-01 6.54566526e-01
-2.49392390e-01 4.96919900e-01 4.09690179e-02 -5.75799108e-01
-7.14669526e-01 -9.17170763e-01 -1.07775104e+00 -6.11570358e-01
-5.15002608e-01 9.97304738e-01 5.67677319e-01 9.17823464e-02
4.78333950e-01 2.90169030e-01 -1.72024727e+00 5.94649196e-01
6.31846130e-01 9.71570253e-01 -1.21722326e-01 4.82088059e-01
-3.23244691e-01 9.51815009e-01 1.00300349e-02 -1.12827659e-01
-2.36682724e-02 -3.92736733e-01 -1.06942546e+00 -9.69732106e-02
2.08320856e-01 5.02926052e-01 3.98803540e-02 1.09777093e+00
-2.05628231e-01 -3.59335877e-02 8.06928933e-01 -1.05321562e+00
-8.16256776e-02 1.20187855e+00 -4.58623350e-01 3.56120504e-02
3.57205808e-01 -5.26311636e-01 1.61081803e+00 -7.51031578e-01
7.14817464e-01 1.27890229e+00 4.55913901e-01 2.30054170e-01
-1.74197161e+00 -3.63287508e-01 1.64969936e-01 4.70952839e-01
-1.46886611e+00 -4.15971488e-01 7.66654849e-01 -5.69813311e-01
8.19292724e-01 2.41727322e-01 4.58381385e-01 4.75607246e-01
-9.95613560e-02 7.57238984e-01 8.99111807e-01 -8.49551380e-01
6.69712305e-01 -3.58445160e-02 4.30232555e-01 1.09115303e+00
3.92106563e-01 -1.54173290e-02 -6.77543998e-01 -4.87122506e-01
2.32032076e-01 -1.62683323e-01 2.85123050e-01 -8.67603660e-01
-8.06790352e-01 9.69547093e-01 1.49894297e-01 4.64776129e-01
1.77236423e-02 2.47591212e-01 2.28305832e-01 3.47703367e-01
-6.63199276e-02 4.23120141e-01 -4.76115316e-01 7.72528350e-02
-8.81631076e-01 3.11927646e-01 1.06800759e+00 7.14530528e-01
8.34437609e-01 -5.12593865e-01 3.52400720e-01 5.26015818e-01
5.01869977e-01 2.05900341e-01 6.58714697e-02 -7.11292148e-01
2.85531044e-01 9.28977966e-01 -2.14532584e-01 -1.00210607e+00
-3.21682096e-01 -1.07172050e-01 -5.42450905e-01 -8.25295746e-02
4.47982401e-01 1.41241342e-01 -6.49340510e-01 1.71565664e+00
1.68158397e-01 2.25758161e-02 -2.56233484e-01 2.25756332e-01
2.57111937e-01 4.93202955e-01 -1.48377746e-01 -3.75220805e-01
1.05584097e+00 -3.47642809e-01 -3.02119881e-01 -4.69319299e-02
8.73802304e-01 -2.96071440e-01 8.03330839e-01 6.06728375e-01
-9.28833008e-01 -1.01499416e-01 -1.02564251e+00 1.75336033e-01
-1.73739582e-01 -1.88244313e-01 8.32843482e-01 4.14689273e-01
-6.44126236e-01 8.07065427e-01 -8.41181397e-01 -1.25988439e-01
3.81782860e-01 6.50204360e-01 -3.14687788e-01 6.06755614e-02
-9.86535788e-01 6.64228857e-01 6.59922361e-01 -3.58866960e-01
-4.22272384e-01 -6.01861119e-01 -9.43584204e-01 5.73418476e-02
9.20472383e-01 -3.84886473e-01 9.44337547e-01 -6.80211961e-01
-1.03266585e+00 8.44393373e-01 -4.98953819e-01 -8.28012884e-01
1.37191921e-01 1.26325935e-01 -2.21429035e-01 -1.72543209e-02
-1.31037936e-01 2.63591617e-01 6.52351797e-01 -1.23903608e+00
-7.09713161e-01 -4.81378555e-01 -3.07943016e-01 -2.78538734e-01
1.34318575e-01 -1.41246781e-01 -3.22724849e-01 -3.06490093e-01
5.05404472e-01 -9.28044200e-01 -2.82544702e-01 -3.37255627e-01
-4.22491848e-01 -6.26028657e-01 3.46960783e-01 -3.35567445e-02
1.41171396e+00 -2.10414290e+00 4.26168710e-01 1.06493306e+00
2.68109500e-01 -3.72593373e-01 3.68904501e-01 3.90795499e-01
-1.04914628e-01 1.97222486e-01 -3.75614792e-01 -3.09803039e-02
2.18123227e-01 3.49795461e-01 -5.30409992e-01 3.11874628e-01
1.48621425e-01 6.93861544e-01 -7.26714432e-01 -5.59704244e-01
-7.26669207e-02 -2.17568815e-01 -7.15571582e-01 -1.38981774e-01
-7.20160127e-01 -2.25781545e-01 -3.05095524e-01 3.44547302e-01
7.63117373e-02 -4.33869332e-01 7.14353859e-01 -1.50891556e-03
-6.56274259e-02 6.83094263e-01 -1.66067755e+00 1.25615239e+00
1.00441873e-01 2.21856892e-01 -3.45322430e-01 -1.29633665e+00
7.95068979e-01 -3.92760411e-02 5.83232284e-01 -4.08681184e-01
8.42553675e-02 3.05329353e-01 2.03439206e-01 -3.58107924e-01
3.40308212e-02 -1.91528022e-01 -2.54973292e-01 6.17933393e-01
-4.16962020e-02 2.57148176e-01 6.11943841e-01 2.09484473e-01
1.27719367e+00 -1.16455518e-01 3.46405447e-01 -3.56262773e-01
5.07614791e-01 -8.18322599e-02 5.52747011e-01 6.81512237e-01
4.81100708e-01 1.85229238e-02 1.08360481e+00 -6.48035526e-01
-7.69284844e-01 -9.36634719e-01 -6.52093887e-02 1.16056132e+00
-1.81349114e-01 -8.37415874e-01 -5.13957977e-01 -5.57363749e-01
1.56663522e-01 6.13878012e-01 -7.70667195e-01 -1.95618898e-01
-4.08225983e-01 -5.35764039e-01 3.26141417e-01 4.64421481e-01
-2.06590176e-01 -9.50477421e-01 -5.61679304e-01 4.29217964e-02
1.64620295e-01 -9.08620059e-01 -3.72035988e-02 6.67684019e-01
-1.07406974e+00 -1.19458759e+00 3.42023194e-01 -7.29963064e-01
6.43459320e-01 -3.89159828e-01 1.09948468e+00 -3.95042449e-02
-2.50500798e-01 2.06783437e-03 -4.70588058e-02 -4.34882939e-01
-5.01630545e-01 1.52057901e-01 1.03257023e-01 -4.01928201e-02
5.54651618e-01 -7.38280654e-01 1.18069261e-01 1.88664868e-02
-8.90425861e-01 -8.67355540e-02 7.83136368e-01 5.87301254e-01
9.77473378e-01 4.41447645e-01 -3.54473218e-02 -1.24249995e+00
2.51746625e-01 -3.69183481e-01 -9.49281096e-01 6.34410918e-01
-8.20484400e-01 9.27745342e-01 4.08471644e-01 -3.00482899e-01
-2.14320302e-01 4.75174665e-01 3.84314299e-01 -2.38069728e-01
-1.81052625e-01 9.51357007e-01 -2.08006442e-01 2.17395246e-01
5.94651163e-01 1.39685825e-01 -1.01042986e-01 -3.11691046e-01
3.43062192e-01 1.85243085e-01 4.64600861e-01 -9.30988610e-01
6.84277177e-01 3.77137929e-01 6.12537384e-01 -7.98845828e-01
-8.45342815e-01 -4.37855482e-01 -7.09484279e-01 2.68190622e-01
4.71540570e-01 -4.36713487e-01 -1.04341674e+00 -1.22601263e-01
-6.85356736e-01 -1.71025679e-01 -2.65928298e-01 1.97934732e-01
-6.79688096e-01 3.76650244e-01 -2.53809303e-01 -9.38191235e-01
-3.38837430e-02 -7.35853970e-01 5.35320520e-01 -2.44028151e-01
-5.92211485e-01 -8.08424413e-01 2.41070539e-01 1.29519984e-01
-3.38801473e-01 1.24471821e-01 1.72496927e+00 -9.27036226e-01
-6.37148201e-01 -6.26376942e-02 1.37200892e-01 -4.08074558e-02
-7.33908564e-02 6.84284642e-02 -5.35421073e-01 1.43789649e-01
-1.93664357e-01 -2.74752557e-01 9.63835716e-01 2.01800987e-01
1.18916953e+00 -3.91061217e-01 -5.34035325e-01 1.04265846e-01
1.29441774e+00 7.39811435e-02 4.67475384e-01 2.20113754e-01
5.27927935e-01 7.46780813e-01 3.41244757e-01 2.98702896e-01
2.73136526e-01 5.82615674e-01 1.65911704e-01 4.58770990e-01
4.42049086e-01 -4.63793635e-01 2.72135466e-01 5.40407896e-01
4.02993374e-02 3.36271971e-01 -1.18951702e+00 4.86041605e-01
-2.11888671e+00 -1.21298826e+00 -1.12414308e-01 2.38876271e+00
1.19965184e+00 6.59941196e-01 5.51639915e-01 7.00959444e-01
3.89928818e-01 -4.49266106e-01 -3.26410621e-01 -3.05765331e-01
5.77504225e-02 5.09103358e-01 4.25282568e-01 8.24985743e-01
-1.17206311e+00 7.66605437e-01 5.84954691e+00 6.37390137e-01
-7.33097613e-01 -5.56907952e-01 3.94651264e-01 -8.84980545e-06
-2.35354200e-01 2.30624899e-01 -8.54076982e-01 3.35761249e-01
1.09474218e+00 -3.10971197e-02 5.11972129e-01 9.42932010e-01
-6.33261055e-02 -3.75238955e-01 -1.70733237e+00 7.91357517e-01
8.67805257e-03 -1.47331107e+00 1.17339425e-01 2.82405645e-01
5.88502824e-01 -1.46005392e-01 -1.18553087e-01 -5.07815331e-02
3.25947791e-01 -1.21426034e+00 7.60882258e-01 7.16116905e-01
3.36093485e-01 -1.13814366e+00 5.04756153e-01 5.64047158e-01
-1.01474202e+00 -2.93076068e-01 -1.64512098e-01 -4.16776724e-02
-2.81321675e-01 7.60867655e-01 -8.02219808e-01 2.91850477e-01
5.25193036e-01 5.17403543e-01 -4.92652208e-01 8.86668384e-01
-1.49460733e-01 6.60525858e-01 -6.56935513e-01 -2.68244088e-01
3.68912891e-02 -1.00201249e-01 3.47489327e-01 1.02651668e+00
-1.11269102e-01 4.84090708e-02 3.06176662e-01 8.81026149e-01
-9.59242061e-02 -9.16630998e-02 -5.21483541e-01 -5.93873933e-02
5.31104326e-01 7.67200589e-01 -7.79508173e-01 -6.04944229e-02
-1.01729140e-01 2.37508222e-01 4.96478260e-01 -1.96244657e-01
-2.96018690e-01 -3.01782250e-01 1.03569269e-01 4.01729286e-01
6.61329985e-01 -2.14635596e-01 -4.70141232e-01 -8.21931779e-01
2.60138333e-01 -1.05604315e+00 6.95604146e-01 1.71138532e-02
-1.22547269e+00 1.85154662e-01 2.34775320e-01 -6.17604554e-01
-5.23829818e-01 -8.77800643e-01 -3.05794328e-01 4.14313316e-01
-9.30778503e-01 -7.74913669e-01 2.86353022e-01 4.00281012e-01
4.43003587e-02 -8.71060714e-02 9.51693475e-01 -1.67220339e-01
-2.99701691e-01 5.32169878e-01 4.88829836e-02 2.12848932e-01
3.24064344e-01 -1.44831097e+00 1.03501320e-01 5.66715777e-01
7.44466960e-01 9.35601354e-01 7.70165324e-01 -4.88048106e-01
-1.52928233e+00 -6.53908014e-01 1.37749875e+00 -5.75312257e-01
1.00135815e+00 -4.65561569e-01 -8.89978409e-01 5.65055132e-01
-3.35029960e-01 -1.82135571e-02 1.02515352e+00 6.36025608e-01
-6.92868590e-01 -2.51157224e-01 -6.92419589e-01 5.47494590e-01
9.90259171e-01 -5.17275751e-01 -9.80721176e-01 -2.43105367e-02
2.46780574e-01 -3.86629589e-02 -6.28053188e-01 3.88732165e-01
5.51917970e-01 -7.63228118e-01 6.19201899e-01 -9.46686268e-01
3.93253714e-01 -5.38355172e-01 -3.95495594e-01 -6.21295512e-01
-3.88155192e-01 -7.67673790e-01 -6.89406812e-01 7.88258314e-01
8.44757318e-01 -1.29061803e-01 9.93533313e-01 5.15877604e-01
4.46655869e-01 -1.07878613e+00 -7.77891219e-01 -6.96526051e-01
-6.91878051e-02 -6.63200855e-01 1.88894987e-01 6.29136801e-01
4.38072771e-01 8.75086129e-01 -3.36041823e-02 1.11788489e-01
7.83551157e-01 5.77123702e-01 5.49113452e-01 -1.74019229e+00
-5.55762172e-01 -5.90001464e-01 -6.05843484e-01 -7.53163278e-01
2.40425602e-01 -1.08739293e+00 -7.78661575e-03 -1.15245485e+00
3.58458191e-01 -4.08161640e-01 -3.31390321e-01 4.95421827e-01
2.71170318e-01 -3.68438870e-01 -1.31754383e-01 9.58527774e-02
-1.02066743e+00 1.12872057e-01 3.35540473e-01 -4.64469530e-02
-3.24307889e-01 2.20555440e-01 -6.12655103e-01 9.01530921e-01
5.09926677e-01 -6.75996721e-01 -3.99529934e-01 1.28338533e-03
8.07872534e-01 -3.03247452e-01 3.21580380e-01 -8.53743494e-01
5.44434369e-01 -1.75278261e-01 2.31007099e-01 -8.00179482e-01
-1.94537655e-01 -8.33133161e-01 5.28230481e-02 5.92605889e-01
-6.94829106e-01 -4.72846217e-02 -3.68101858e-02 6.59061253e-01
-3.08656115e-02 -4.53567982e-01 6.55613720e-01 -3.30182090e-02
-4.53363210e-01 2.84296144e-02 -2.99605459e-01 1.28982633e-01
7.79197276e-01 7.42326453e-02 2.18141943e-01 -7.65863582e-02
-8.91628683e-01 2.17031747e-01 1.94115534e-01 8.98324400e-02
5.21208048e-01 -1.04307914e+00 -3.57258230e-01 3.92790101e-02
1.96598634e-01 1.34969249e-01 -3.36105347e-01 7.69367993e-01
-1.21015012e-01 6.28970802e-01 2.57465035e-01 -5.59272826e-01
-1.49009597e+00 6.78296208e-01 1.13174677e-01 -5.05761743e-01
-2.89208889e-01 6.61353588e-01 -3.01134080e-01 -1.19846389e-01
6.35151565e-01 -5.48820078e-01 -5.33813015e-02 5.68788238e-02
5.67678154e-01 3.41767699e-01 1.39764786e-01 -1.24636203e-01
-5.59231579e-01 4.37923402e-01 -3.31603467e-01 -1.41378969e-01
1.54618704e+00 3.57795686e-01 -5.79616904e-01 1.00730085e+00
7.28828788e-01 -2.01284457e-02 -6.29780173e-01 -5.83933532e-01
1.15245247e+00 8.65673423e-02 -2.40048945e-01 -4.78725225e-01
-3.06936681e-01 5.67504346e-01 2.00674608e-01 4.15289044e-01
8.27203512e-01 5.94165325e-01 2.72797108e-01 1.06686592e+00
4.34336692e-01 -1.11016774e+00 1.88366568e-03 8.64782572e-01
4.71226692e-01 -9.06092882e-01 2.49779657e-01 -3.94928515e-01
-3.28259945e-01 1.12298334e+00 1.23911567e-01 -2.21098214e-01
7.89939344e-01 4.43185091e-01 -5.54797173e-01 -3.93264383e-01
-1.09267044e+00 -2.27209777e-01 6.15416944e-01 1.73411205e-01
3.72092634e-01 2.84968410e-02 -2.33051166e-01 8.61557424e-01
-3.03716153e-01 -1.76744714e-01 -6.48566484e-02 1.00870907e+00
-8.04314971e-01 -1.43842816e+00 -1.20181330e-01 4.72118378e-01
-5.06702542e-01 -3.97427827e-02 -8.31724346e-01 5.47128856e-01
1.85425743e-01 6.57324493e-01 -1.12505265e-01 -3.48264933e-01
2.57516801e-01 5.02025783e-01 7.25332677e-01 -7.51557350e-01
2.94058025e-02 -7.24762976e-02 -4.21746559e-02 -4.55195695e-01
-3.77480268e-01 -9.10224259e-01 -1.52936912e+00 -2.54675776e-01
-1.12652615e-01 3.95403296e-01 4.38730180e-01 1.34357202e+00
2.79701203e-01 1.55814029e-02 5.78017652e-01 -1.33545026e-01
-6.36814535e-01 -5.55514872e-01 -4.13079202e-01 4.97181565e-02
9.73978192e-02 -6.51631534e-01 -2.14673996e-01 1.92777246e-01] | [8.777297973632812, 6.781511306762695] |
0f04ad35-759c-447a-a05d-bde8284dcac1 | generalizability-of-machine-learning-models | 2202.01337 | null | https://arxiv.org/abs/2202.01337v2 | https://arxiv.org/pdf/2202.01337v2.pdf | Generalizability of Machine Learning Models: Quantitative Evaluation of Three Methodological Pitfalls | Purpose: Despite the potential of machine learning models, the lack of generalizability has hindered their widespread adoption in clinical practice. We investigate three methodological pitfalls: (1) violation of independence assumption, (2) model evaluation with an inappropriate performance indicator or baseline for comparison, and (3) batch effect. Materials and Methods: Using several retrospective datasets, we implement machine learning models with and without the pitfalls to quantitatively illustrate these pitfalls' effect on model generalizability. Results: Violation of independence assumption, more specifically, applying oversampling, feature selection, and data augmentation before splitting data into train, validation, and test sets, respectively, led to misleading and superficial gains in F1 scores of 71.2% in predicting local recurrence and 5.0% in predicting 3-year overall survival in head and neck cancer as well as 46.0% in distinguishing histopathological patterns in lung cancer. Further, randomly distributing data points for a subject across training, validation, and test sets led to a 21.8% superficial increase in F1 score. Also, we showed the importance of the choice of performance measures and baseline for comparison. In the presence of batch effect, a model built for pneumonia detection led to F1 score of 98.7%. However, when the same model was applied to a new dataset of normal patients, it only correctly classified 3.86% of the samples. Conclusions: These methodological pitfalls cannot be captured using internal model evaluation, and the inaccurate predictions made by such models may lead to wrong conclusions and interpretations. Therefore, understanding and avoiding these pitfalls is necessary for developing generalizable models. | ['Reza Forghani', 'Alan Spatz', 'Caroline Reinhold', 'Rajiv Gupta', 'Katie Ovens', 'Farhad Maleki'] | 2022-02-01 | null | null | null | null | ['pneumonia-detection'] | ['medical'] | [ 4.48928088e-01 6.93102777e-02 -4.25181329e-01 -2.92391181e-01
-9.77444530e-01 -3.03291380e-01 4.26880240e-01 5.34367025e-01
-7.18672574e-01 9.33421135e-01 9.27492604e-02 -8.84444058e-01
-3.95927012e-01 -5.42666376e-01 -4.08050895e-01 -9.69974399e-01
3.77309732e-02 4.63593006e-01 1.04352117e-01 5.42412102e-01
8.60216245e-02 7.51862288e-01 -1.25926685e+00 2.58593380e-01
7.32874155e-01 6.03226900e-01 7.54121318e-02 5.86750329e-01
3.48709859e-02 8.75304639e-01 -6.58726454e-01 -2.20330227e-02
-2.20709685e-02 -4.37247008e-01 -5.75083315e-01 1.76858649e-01
7.75798932e-02 -5.27227461e-01 1.02082752e-01 3.86863440e-01
5.69575071e-01 -2.12140635e-01 1.04938734e+00 -1.14097226e+00
-2.66226768e-01 3.98051560e-01 -5.08859038e-01 1.54179391e-02
-1.68774515e-01 5.24046183e-01 6.78124428e-01 -5.14408827e-01
3.50538969e-01 4.84026700e-01 9.15611148e-01 5.06468356e-01
-1.39504671e+00 -6.91751659e-01 -1.37060151e-01 -1.51410982e-01
-1.31307590e+00 -4.65135902e-01 4.56770211e-02 -6.41106248e-01
6.80856705e-01 6.99112356e-01 7.29100049e-01 9.03658032e-01
4.40603167e-01 2.84089059e-01 1.18883753e+00 -3.48297864e-01
2.45764852e-01 5.05009174e-01 2.16746509e-01 4.14189726e-01
7.66600490e-01 2.68027365e-01 1.27103671e-01 -5.97861052e-01
6.45350754e-01 7.98669904e-02 -2.43698671e-01 -2.53572315e-02
-9.44951713e-01 7.40250349e-01 5.27467206e-02 4.71287996e-01
-3.69429141e-01 -2.80235887e-01 3.62718642e-01 -5.43291233e-02
1.26980379e-01 3.19880664e-01 -3.76661509e-01 2.14604869e-01
-9.65218425e-01 -2.01423734e-01 6.69143558e-01 3.64959866e-01
1.23408228e-01 -2.42050618e-01 -3.62124443e-02 8.21763754e-01
2.07517862e-01 4.49430317e-01 8.67840528e-01 -7.02021599e-01
-2.32010148e-02 6.00996494e-01 6.26396239e-02 -6.30242527e-01
-9.26382184e-01 -7.51327634e-01 -8.24867249e-01 1.97266452e-02
6.72075808e-01 -4.18405943e-02 -9.40840364e-01 1.64468241e+00
8.76683220e-02 -3.45097929e-01 9.76904631e-02 6.37896538e-01
4.20158476e-01 2.95803770e-02 5.37587762e-01 -5.14063120e-01
1.37114859e+00 -4.30156797e-01 -3.96687210e-01 -1.13451608e-01
1.28279221e+00 -5.99727988e-01 1.06165075e+00 2.77002752e-01
-9.60531116e-01 -3.15628290e-01 -7.22605169e-01 4.49122816e-01
4.17163707e-02 1.44484192e-01 4.80234087e-01 9.41526651e-01
-9.13714349e-01 4.53899622e-01 -1.03732073e+00 -7.45888531e-01
3.12804252e-01 6.39923632e-01 -4.50668067e-01 -4.74200137e-02
-8.46089125e-01 9.34722245e-01 1.54586285e-01 2.74686553e-02
-4.33852792e-01 -6.98242724e-01 -3.43918204e-01 9.43514332e-02
-2.00478099e-02 -9.80476022e-01 9.80177522e-01 -9.22817349e-01
-7.94133186e-01 8.34148288e-01 -2.26088196e-01 -3.38875741e-01
5.21344423e-01 1.56504706e-01 -4.16021854e-01 -1.38362765e-01
1.21778980e-01 2.03840911e-01 2.09643438e-01 -1.05825508e+00
-7.01745927e-01 -6.23642266e-01 -7.09714711e-01 1.12462930e-01
-3.38509858e-01 -8.16299859e-03 -1.47203624e-01 -3.75188440e-01
1.40818939e-01 -7.81298876e-01 -5.97716987e-01 -1.16857596e-01
-1.03594176e-01 6.02506883e-02 2.47050360e-01 -7.25535154e-01
1.32554567e+00 -2.19471550e+00 -6.41739547e-01 3.27941835e-01
2.90753275e-01 1.19207814e-01 1.64926483e-03 2.55874902e-01
-2.94588894e-01 5.29851377e-01 -1.79202035e-01 5.53844217e-03
-4.69857275e-01 6.72668591e-02 2.01909736e-01 6.83844864e-01
3.19065750e-01 6.45583391e-01 -7.53875554e-01 -7.31293201e-01
3.38730514e-01 5.09113193e-01 -6.20818555e-01 -1.05937026e-01
4.64115053e-01 3.28594863e-01 -2.63191968e-01 5.20566285e-01
4.39613342e-01 -4.45449471e-01 5.27506471e-01 -1.22769378e-01
-1.06346734e-01 3.38187605e-01 -7.89484918e-01 7.71880805e-01
-3.60668272e-01 4.26264822e-01 -2.96455979e-01 -6.94972873e-01
8.67675126e-01 4.26164389e-01 6.08687162e-01 -6.63850069e-01
2.17490405e-01 6.12760663e-01 5.19156575e-01 -4.58862871e-01
-1.66747980e-02 -7.28649795e-01 2.46441066e-01 3.48903298e-01
-2.93963552e-01 -4.34940448e-03 -1.33031234e-01 -1.43303335e-01
1.18089855e+00 -4.51443821e-01 4.52268422e-01 -2.90307403e-01
3.77158642e-01 2.06295714e-01 5.87251723e-01 8.33991647e-01
-2.91582942e-01 8.06834519e-01 5.22590995e-01 -2.63302654e-01
-7.36911118e-01 -9.74004805e-01 -7.82755196e-01 4.14738804e-01
-5.29655516e-01 -3.35689858e-02 -1.64444610e-01 -6.72398984e-01
2.20053554e-01 1.01467824e+00 -7.98359632e-01 -4.57447290e-01
-2.97340751e-01 -1.40736115e+00 5.89838982e-01 5.91181695e-01
-7.42801814e-04 -5.85466623e-01 -7.01089144e-01 2.93348491e-01
-1.17702365e-01 -6.24705434e-01 -6.22834526e-02 4.19755280e-01
-1.09191108e+00 -1.48077118e+00 -7.79404223e-01 -4.54878360e-01
7.89928973e-01 2.58184910e-01 8.88404489e-01 4.73199695e-01
-2.99365968e-01 3.18958610e-01 -5.01718968e-02 -5.17300785e-01
-7.02402115e-01 -1.49312511e-01 -2.17614025e-02 -2.93127567e-01
5.41898251e-01 -2.26614922e-01 -6.79703712e-01 4.65701252e-01
-9.22972262e-01 -1.89504266e-01 9.28128958e-01 9.65896070e-01
5.40636539e-01 -1.05656162e-01 7.62261569e-01 -1.08719695e+00
1.95025891e-01 -6.22268796e-01 -2.23140955e-01 2.16246977e-01
-1.02341652e+00 -1.40695646e-01 4.69219029e-01 -3.82724941e-01
-8.18170309e-01 -1.94419086e-01 -1.11636159e-03 3.07861660e-02
-1.14102378e-01 5.60345888e-01 2.62052566e-01 2.57663727e-01
8.78803730e-01 -8.30467883e-03 6.06210470e-01 -2.35813692e-01
-5.95465600e-01 8.19836617e-01 1.05838947e-01 -2.78558403e-01
5.31501949e-01 5.17964244e-01 2.10774958e-01 -8.15272093e-01
-5.45151353e-01 -7.04881608e-01 -4.30271626e-01 1.21389143e-01
5.57501614e-01 -8.59463751e-01 -3.95124495e-01 3.22659552e-01
-4.60227698e-01 -5.15972853e-01 -4.20069426e-01 1.10677564e+00
-3.78505588e-01 3.74438971e-01 -3.03209305e-01 -7.76611745e-01
-4.16319728e-01 -1.26618469e+00 5.97428918e-01 1.02294669e-01
-8.29059958e-01 -1.11220193e+00 1.72644645e-01 4.01865810e-01
5.08710980e-01 2.60550529e-01 1.12888193e+00 -1.22173142e+00
1.08734302e-01 -5.44880629e-01 -2.28552565e-01 2.59548992e-01
7.93828666e-01 4.36821550e-01 -9.59276140e-01 -3.61402392e-01
1.49797231e-01 -2.00722247e-01 5.51929176e-01 5.10576010e-01
1.09374475e+00 -1.84697673e-01 -6.72719836e-01 2.90495038e-01
1.35423398e+00 5.72931528e-01 5.28833270e-01 3.67608041e-01
2.22517520e-01 6.77750587e-01 2.72178233e-01 1.58495113e-01
1.51302397e-01 4.76050436e-01 9.05088410e-02 -2.92827487e-01
-4.88136634e-02 -1.05107293e-01 5.05121686e-02 3.23003501e-01
-2.49407347e-02 -6.08166121e-02 -1.19208288e+00 6.84415281e-01
-1.12055635e+00 -7.53652513e-01 -4.29921031e-01 2.71350455e+00
6.38347924e-01 4.19765204e-01 2.09512889e-01 1.90160379e-01
6.21455967e-01 -2.84968436e-01 -3.65270555e-01 -3.51838797e-01
1.51250418e-03 -6.50323778e-02 6.87667847e-01 2.86932975e-01
-5.89069068e-01 8.59763995e-02 6.32587481e+00 4.60717410e-01
-1.30988717e+00 -2.09254056e-01 1.01190829e+00 -1.56254306e-01
-2.49245554e-01 1.29185885e-01 -7.04138279e-01 3.31455976e-01
1.22451639e+00 -1.11519404e-01 -2.73079157e-01 3.49206775e-01
5.19270718e-01 -4.16981101e-01 -9.38043714e-01 3.32101077e-01
-6.70957007e-03 -8.96535158e-01 -1.61594763e-01 2.98207641e-01
3.70571136e-01 -2.35053729e-02 -9.25459117e-02 3.42163742e-01
6.74173683e-02 -1.07133341e+00 3.34236085e-01 3.39337617e-01
1.00190794e+00 -3.90230358e-01 1.17901874e+00 3.51804227e-01
-4.73923296e-01 8.81584641e-03 -8.38905498e-02 7.68285170e-02
-2.33373791e-02 6.87899053e-01 -1.61295509e+00 4.63528037e-01
4.76788580e-01 1.21504433e-01 -5.45300722e-01 1.00767100e+00
2.16855451e-01 1.08537865e+00 -6.37732089e-01 -1.58726513e-01
1.05478518e-01 2.47368932e-01 3.74579839e-02 1.02484632e+00
2.09440112e-01 1.18010662e-01 -4.98915553e-01 4.16783214e-01
4.74704653e-01 2.86494553e-01 -3.38345140e-01 -1.08774185e-01
4.76713330e-01 1.10698700e+00 -9.10799444e-01 -1.58859387e-01
-6.98345661e-01 3.59309137e-01 9.74854827e-02 1.99605227e-01
-8.07039142e-01 -1.66339636e-01 1.33143485e-01 6.52165174e-01
-1.79784119e-01 4.15712208e-01 -6.62542582e-01 -7.24529564e-01
-1.42391488e-01 -8.30879092e-01 7.31718659e-01 -3.78400683e-01
-1.04198635e+00 3.07083607e-01 -6.33391440e-02 -1.10923958e+00
-1.86283976e-01 -5.03999293e-01 -5.95704973e-01 1.12871480e+00
-1.18529785e+00 -7.40880311e-01 -3.10290486e-01 2.06531249e-02
5.54916002e-02 2.54938722e-01 1.05383766e+00 1.05208173e-01
-6.02427959e-01 8.50532174e-01 3.91969264e-01 1.99556910e-02
1.00333011e+00 -9.94400740e-01 -1.49554685e-01 2.45845929e-01
-4.53284115e-01 6.53247952e-01 4.18369949e-01 -5.95190644e-01
-6.37725651e-01 -9.94527102e-01 1.17486572e+00 -5.48020065e-01
5.93147278e-01 2.66988724e-01 -1.01345801e+00 5.93744934e-01
-5.25903404e-01 -4.73166913e-01 1.30069089e+00 1.73024476e-01
-1.32044882e-01 -6.80560991e-02 -1.32910192e+00 7.95424283e-01
4.07100677e-01 -3.64390425e-02 -1.77655101e-01 2.63233453e-01
1.16510272e-01 3.50103676e-02 -1.21673632e+00 7.55001783e-01
9.45168078e-01 -1.04137731e+00 8.39685202e-01 -7.71263421e-01
2.29982287e-01 -1.25355981e-02 -1.04759678e-01 -6.99774384e-01
-6.02359653e-01 7.23626763e-02 7.07295656e-01 1.04228771e+00
8.94477367e-01 -9.73434925e-01 8.35712254e-01 1.12012482e+00
-2.74761654e-02 -1.14152169e+00 -7.44981945e-01 -5.41849673e-01
4.53264087e-01 -3.30786616e-01 3.41089815e-01 8.28777671e-01
-1.64895728e-02 -4.76487838e-02 2.20493242e-01 8.40878040e-02
3.00334007e-01 -2.32498825e-01 6.08143091e-01 -1.07979941e+00
-3.70998412e-01 -4.56880242e-01 -3.20160419e-01 -1.86105549e-01
-3.65888536e-01 -7.76686668e-01 -3.39790136e-01 -1.66009295e+00
6.78641498e-01 -9.67692852e-01 -4.69259918e-01 5.38041413e-01
-4.71634328e-01 1.28440186e-02 3.86378951e-02 5.42484283e-01
2.14371517e-01 -4.47001755e-02 8.95402193e-01 1.28632784e-01
-3.73426884e-01 4.78570282e-01 -9.79719937e-01 5.50375104e-01
1.10739732e+00 -5.66670597e-01 -4.40118939e-01 -5.65455258e-02
-1.54070556e-01 2.56899774e-01 6.18820727e-01 -8.88550758e-01
3.23796868e-02 -4.24985290e-01 8.49063993e-01 -4.24111992e-01
6.04396313e-02 -9.27158475e-01 6.78013921e-01 1.24093914e+00
-7.18989223e-02 4.64876089e-03 4.87847775e-01 3.31207931e-01
1.81510806e-01 -3.93019736e-01 7.30152667e-01 -3.80492322e-02
-1.04433604e-01 -2.37631425e-01 -6.49656892e-01 -2.13612586e-01
1.02678299e+00 -6.12193227e-01 -3.44053447e-01 -1.28799319e-01
-7.08262265e-01 6.98565990e-02 6.68816388e-01 -1.33952415e-02
2.01270506e-01 -8.87365878e-01 -8.79130304e-01 2.99281895e-01
1.77691281e-01 -8.54343027e-02 4.11049545e-01 1.50323462e+00
-5.05245686e-01 5.73484600e-01 1.30101278e-01 -6.51137531e-01
-1.29341519e+00 3.82964849e-01 5.41599989e-01 -5.46338141e-01
-2.39002138e-01 4.80610788e-01 3.94470900e-01 -2.15423450e-01
2.70091258e-02 -2.31853992e-01 6.40275627e-02 -2.88345926e-02
2.95481950e-01 4.59321320e-01 3.11069608e-01 -3.37765217e-01
-5.79423428e-01 2.80424118e-01 -4.15300369e-01 -1.91545486e-03
8.68032694e-01 1.76418006e-01 1.81371182e-01 5.55704594e-01
1.17570221e+00 1.01733997e-01 -7.72699296e-01 1.34870812e-01
-1.13441296e-01 -1.19506218e-01 -1.60436332e-01 -9.36250567e-01
-6.60667598e-01 5.41418850e-01 6.48033142e-01 -5.38231991e-03
1.19388127e+00 -1.08668119e-01 1.18415773e-01 9.63963941e-03
1.70493364e-01 -5.99819720e-01 -4.78753805e-01 -1.54610932e-01
5.44744194e-01 -1.15307617e+00 1.93550006e-01 -1.44172966e-01
-5.40601432e-01 8.98668230e-01 4.43149209e-01 3.55589777e-01
5.54083824e-01 1.31776944e-01 2.20702380e-01 3.14825587e-02
-9.99131739e-01 2.07896486e-01 -4.63794805e-02 5.29592633e-01
6.94534421e-01 2.00960293e-01 -7.70519972e-01 7.45537579e-01
-4.17342521e-02 3.14418226e-01 5.12810946e-01 8.87836337e-01
-3.05763483e-01 -9.66583610e-01 -4.52634573e-01 1.06523335e+00
-6.42513096e-01 9.86250564e-02 -2.29596376e-01 1.41576505e+00
-2.16700628e-01 8.13461304e-01 2.70729154e-01 -2.18042031e-01
4.27286714e-01 2.28717893e-01 7.76848122e-02 -4.64209408e-01
-7.23310828e-01 7.02389702e-02 1.50387108e-01 -2.58472171e-02
-3.70271206e-01 -9.06104088e-01 -1.19615042e+00 -2.05252275e-01
-7.31302857e-01 4.41430449e-01 5.10382116e-01 7.62283444e-01
2.88654655e-01 5.22495568e-01 6.34806275e-01 1.62570644e-02
-9.92243350e-01 -1.02640426e+00 -4.99065965e-01 1.72551095e-01
3.00475091e-01 -4.28792536e-01 -1.06905878e+00 -9.67779607e-02] | [15.168002128601074, -2.8172976970672607] |
f415d219-37f7-4808-807d-93b5059b9c4f | gpr-net-geometric-prototypical-network-for | 2304.06007 | null | https://arxiv.org/abs/2304.06007v1 | https://arxiv.org/pdf/2304.06007v1.pdf | GPr-Net: Geometric Prototypical Network for Point Cloud Few-Shot Learning | In the realm of 3D-computer vision applications, point cloud few-shot learning plays a critical role. However, it poses an arduous challenge due to the sparsity, irregularity, and unordered nature of the data. Current methods rely on complex local geometric extraction techniques such as convolution, graph, and attention mechanisms, along with extensive data-driven pre-training tasks. These approaches contradict the fundamental goal of few-shot learning, which is to facilitate efficient learning. To address this issue, we propose GPr-Net (Geometric Prototypical Network), a lightweight and computationally efficient geometric prototypical network that captures the intrinsic topology of point clouds and achieves superior performance. Our proposed method, IGI++ (Intrinsic Geometry Interpreter++) employs vector-based hand-crafted intrinsic geometry interpreters and Laplace vectors to extract and evaluate point cloud morphology, resulting in improved representations for FSL (Few-Shot Learning). Additionally, Laplace vectors enable the extraction of valuable features from point clouds with fewer points. To tackle the distribution drift challenge in few-shot metric learning, we leverage hyperbolic space and demonstrate that our approach handles intra and inter-class variance better than existing point cloud few-shot learning methods. Experimental results on the ModelNet40 dataset show that GPr-Net outperforms state-of-the-art methods in few-shot learning on point clouds, achieving utmost computational efficiency that is $170\times$ better than all existing works. The code is publicly available at https://github.com/TejasAnvekar/GPr-Net. | ['Dena Bazazian', 'Tejas Anvekar'] | 2023-04-12 | null | null | null | null | ['metric-learning', 'few-shot-3d-point-cloud-classification', 'gpr', 'metric-learning', 'gpr'] | ['computer-vision', 'computer-vision', 'computer-vision', 'methodology', 'miscellaneous'] | [-2.71195590e-01 -3.84658456e-01 1.00028470e-01 -2.29406103e-01
-9.08118069e-01 -2.74004668e-01 5.27014077e-01 1.04972132e-01
-1.65884048e-01 2.36563981e-01 -2.92201370e-01 -1.00764900e-01
-4.33593184e-01 -1.03833747e+00 -7.84990191e-01 -6.25813425e-01
-1.72436520e-01 5.96563339e-01 3.91492933e-01 -2.54858226e-01
3.60886127e-01 8.88666809e-01 -1.86457109e+00 -2.66474158e-01
7.78417587e-01 1.21947408e+00 3.54272246e-01 5.15524507e-01
-4.55141038e-01 5.36031425e-01 -1.57909602e-01 -6.17199652e-02
3.97659481e-01 2.61375815e-01 -2.71766275e-01 -1.13861978e-01
6.35803223e-01 -3.20230871e-01 -4.54234749e-01 1.03210032e+00
7.09974885e-01 4.61638838e-01 7.58196414e-01 -1.31525528e+00
-5.37084401e-01 6.71889186e-02 -7.05855370e-01 3.71634424e-01
-7.63490945e-02 4.16280389e-01 1.05913103e+00 -1.46003139e+00
4.82147545e-01 9.99772429e-01 8.88672233e-01 3.21243942e-01
-8.70689929e-01 -8.12968314e-01 -1.26235962e-01 3.07590634e-01
-1.73380387e+00 -2.46129215e-01 1.17505670e+00 -5.02206564e-01
1.12282002e+00 3.84140946e-02 7.03474700e-01 7.59464324e-01
-1.91672429e-01 5.54086387e-01 5.08258700e-01 -2.41636522e-02
5.85275471e-01 -2.07173452e-01 2.03988105e-01 8.22471738e-01
2.91888475e-01 2.58960038e-01 -4.27002609e-01 -3.08322042e-01
7.80313373e-01 5.67418575e-01 -1.91269591e-01 -8.10430348e-01
-7.48489499e-01 9.18940485e-01 5.82607627e-01 6.59037754e-02
-4.41944271e-01 2.39388734e-01 3.20501268e-01 2.19966084e-01
6.51679218e-01 1.31479740e-01 -3.28543574e-01 -3.52686703e-01
-9.32975888e-01 2.76122838e-01 6.33499026e-01 1.31214345e+00
1.20101905e+00 1.73673317e-01 -4.93318355e-03 8.98495972e-01
2.39867881e-01 6.05713725e-01 5.22094890e-02 -7.30219483e-01
3.95863980e-01 7.07406759e-01 -1.21058621e-01 -1.13304245e+00
-2.47067720e-01 -3.84126872e-01 -9.69529212e-01 3.54056507e-01
7.84913450e-03 1.32321179e-01 -1.06773448e+00 1.15970099e+00
4.55081224e-01 7.88987398e-01 -2.96056211e-01 8.43172073e-01
1.06898606e+00 6.07049584e-01 -1.63801879e-01 -2.95727998e-02
9.92576957e-01 -5.26065946e-01 -4.73713055e-02 4.54375967e-02
5.54250062e-01 -4.52470183e-01 1.24023151e+00 2.15475447e-02
-8.57124984e-01 -5.07271111e-01 -1.08568287e+00 4.66950238e-02
-4.14582431e-01 -4.39150453e-01 7.16630876e-01 4.00844812e-01
-8.19554746e-01 9.18996453e-01 -9.39972460e-01 -2.63672054e-01
1.09423685e+00 2.94370383e-01 -6.40947223e-02 -4.44671243e-01
-7.28506565e-01 4.17751729e-01 1.75267696e-01 -2.04455629e-01
-7.35059559e-01 -1.36440420e+00 -1.15463686e+00 2.59644955e-01
5.04791081e-01 -6.39171541e-01 1.07278216e+00 8.80966410e-02
-1.20032847e+00 6.37004375e-01 2.04123016e-02 -2.80895382e-01
4.25351739e-01 -8.51014555e-02 5.61337406e-03 1.63775027e-01
3.60178873e-02 3.83625537e-01 8.99792433e-01 -1.17373455e+00
-5.65069973e-01 -5.33291638e-01 -1.47557110e-01 5.73232919e-02
-4.00689960e-01 -3.71136755e-01 -6.66306078e-01 -4.72222924e-01
1.35866031e-01 -7.71501184e-01 -1.98665008e-01 3.20988357e-01
-4.41034883e-02 -5.36312222e-01 1.09938967e+00 -4.00455631e-02
8.43739092e-01 -2.43842912e+00 -2.57754236e-01 2.02115878e-01
6.03415430e-01 3.53932351e-01 -1.13289610e-01 2.43460298e-01
1.41370595e-02 -7.59940296e-02 -3.69352162e-01 -5.10306954e-01
8.54891166e-02 2.35740349e-01 -2.98728019e-01 7.68791676e-01
3.14342141e-01 1.02740502e+00 -1.05943096e+00 -5.40986419e-01
9.36392307e-01 7.21705139e-01 -3.89514148e-01 3.51382941e-02
-2.76151448e-01 1.35410316e-02 -4.13623959e-01 1.16351449e+00
1.02618706e+00 -4.89314795e-01 -6.12072408e-01 -1.60094902e-01
-1.37062922e-01 -2.60498613e-01 -1.20220530e+00 2.05088758e+00
-4.41287190e-01 3.26523095e-01 -3.12740743e-01 -8.33669484e-01
1.20075512e+00 -5.57013135e-03 8.81156147e-01 -5.43590605e-01
3.82369488e-01 1.62255108e-01 -3.53634596e-01 -3.31782460e-01
2.98034728e-01 -1.74057603e-01 1.21052578e-01 2.48272672e-01
2.13687286e-01 -5.92739224e-01 -3.96099426e-02 2.68802881e-01
1.28795373e+00 5.08858822e-02 2.35887617e-01 -1.86473466e-02
1.64563775e-01 4.20291722e-03 5.81900597e-01 6.73011124e-01
-2.91339755e-01 9.72129524e-01 -2.18650978e-02 -7.16407299e-01
-1.11985803e+00 -1.09287906e+00 -1.12995937e-01 7.31361866e-01
3.99947941e-01 -4.58441287e-01 -1.64412647e-01 -3.70124727e-01
4.12615180e-01 8.15297604e-01 -5.05984545e-01 -1.76201969e-01
-4.62147236e-01 -4.50889915e-01 1.11145467e-01 6.49015725e-01
1.58988208e-01 -7.90813804e-01 -6.93026483e-01 1.53148040e-01
3.59919310e-01 -1.08402765e+00 -2.06364378e-01 7.46325403e-02
-1.05309236e+00 -1.13768458e+00 -6.58604264e-01 -5.20431757e-01
5.07349014e-01 8.17237198e-01 1.05445111e+00 1.07813202e-01
-6.00690663e-01 4.07097369e-01 -3.49621624e-01 -6.42859638e-01
3.48131001e-01 1.14344418e-01 -5.01725189e-02 -2.07345143e-01
7.60257602e-01 -1.15323746e+00 -5.61382353e-01 7.37591162e-02
-6.78625345e-01 -2.89282143e-01 5.25293410e-01 7.29945600e-01
8.28943908e-01 8.39412808e-02 1.78393677e-01 -6.86529577e-01
3.00856739e-01 -6.10631704e-01 -6.55460715e-01 -1.79085031e-03
-5.27155578e-01 -2.63741046e-01 5.48126101e-01 -2.24536851e-01
-6.22162640e-01 1.69366393e-02 -1.86857476e-03 -1.50631976e+00
-1.71471328e-01 2.53735989e-01 3.38247232e-02 -6.20891929e-01
6.75655246e-01 3.12243462e-01 4.83005121e-02 -4.72540110e-01
3.69002640e-01 4.50993866e-01 3.61287802e-01 -5.45804083e-01
1.24018753e+00 7.43234396e-01 1.73856616e-01 -1.29614258e+00
-6.39505386e-01 -9.98744190e-01 -6.64028883e-01 -2.67101079e-01
3.97611350e-01 -9.62388515e-01 -6.58288479e-01 3.02047342e-01
-1.15695930e+00 -4.60367836e-02 -6.58851504e-01 4.02393579e-01
-7.15660930e-01 2.66799808e-01 -3.59383613e-01 -8.31427872e-01
-6.07667446e-01 -7.82293260e-01 1.26867199e+00 1.90615743e-01
1.37596235e-01 -8.15264642e-01 1.08169898e-01 5.04010096e-02
2.76958734e-01 4.71970588e-01 7.74584711e-01 -4.92808789e-01
-9.39105749e-01 -3.79353881e-01 -5.07479370e-01 1.27068833e-01
1.00846469e-01 1.41449831e-02 -9.99465406e-01 -3.98888707e-01
9.42185521e-02 -2.99701482e-01 8.33837628e-01 4.37179238e-01
1.45663095e+00 1.24493703e-01 -2.67643809e-01 1.18219602e+00
1.88022327e+00 -1.03663683e-01 5.36118627e-01 1.78060010e-02
1.10184240e+00 2.04943016e-01 5.81971705e-01 8.97902489e-01
4.84682351e-01 2.97214866e-01 7.69216478e-01 1.63549587e-01
-1.77846208e-01 -1.01815321e-01 -3.03763956e-01 1.00281107e+00
-2.82682955e-01 1.28011465e-01 -1.25843620e+00 7.46884227e-01
-1.89012575e+00 -1.01901805e+00 4.34559248e-02 2.06090093e+00
3.44035476e-01 2.40852118e-01 -8.95892456e-02 -2.86925230e-02
5.60916364e-01 4.27661836e-01 -8.25141132e-01 1.58606619e-01
1.00681737e-01 4.37096864e-01 5.21044314e-01 9.35142636e-02
-1.00612962e+00 1.02853870e+00 4.61853504e+00 9.88958478e-01
-1.12026238e+00 2.48483300e-01 1.86653465e-01 -4.01018977e-01
-8.70148465e-02 -2.28615656e-01 -7.62739122e-01 4.41926003e-01
5.95031500e-01 -3.43745947e-01 3.40141028e-01 1.21309340e+00
3.40378694e-02 2.75992125e-01 -9.57100272e-01 1.57342327e+00
2.52149999e-01 -1.73643494e+00 -7.83238336e-02 7.97125399e-02
5.92672408e-01 7.21571386e-01 -7.40907565e-02 4.74814177e-01
1.97895780e-01 -8.00062656e-01 4.46256250e-01 5.84823310e-01
1.07393146e+00 -1.01801813e+00 6.56925321e-01 3.81408066e-01
-1.58743358e+00 -3.08488086e-02 -8.72205377e-01 -1.76755264e-01
6.94931298e-02 7.79422879e-01 -7.90428698e-01 6.04448557e-01
9.43636417e-01 1.11817014e+00 -3.18066359e-01 1.36034489e+00
1.37165710e-01 3.39116573e-01 -5.92938364e-01 -1.32067367e-01
3.14341217e-01 -1.60121262e-01 7.91523516e-01 8.52733493e-01
4.69886869e-01 2.67184585e-01 2.40019709e-01 1.06079853e+00
-1.18233435e-01 2.17324831e-02 -1.05502009e+00 5.62547743e-02
7.65912473e-01 1.32988000e+00 -7.49702334e-01 -2.96867132e-01
-6.60258234e-01 4.23157007e-01 4.81478363e-01 1.37584776e-01
-6.40953839e-01 -6.64045453e-01 9.68984067e-01 1.02149755e-01
7.22146451e-01 -4.62496638e-01 -5.46624482e-01 -9.90792632e-01
9.50704291e-02 -3.06169897e-01 1.36599839e-01 -6.29995108e-01
-1.57915807e+00 3.91150832e-01 -6.05146959e-02 -1.62476039e+00
5.33617288e-02 -4.29660082e-01 -8.70142043e-01 6.27354681e-01
-1.78990197e+00 -1.25601912e+00 -8.30574512e-01 7.33204722e-01
8.15547049e-01 -2.10698992e-01 5.22737682e-01 3.54592562e-01
-3.26837897e-01 3.88537258e-01 1.14422943e-02 2.47562435e-02
2.69005239e-01 -1.08484471e+00 7.68419981e-01 5.20043194e-01
1.75044671e-01 4.98832881e-01 4.44100827e-01 -7.34533966e-01
-1.78586102e+00 -1.47746336e+00 4.11962330e-01 -3.96329522e-01
6.66155040e-01 -4.35784101e-01 -1.10650027e+00 2.61822939e-01
-4.98292804e-01 5.84473193e-01 6.03540242e-01 -2.99583357e-02
-2.80425876e-01 -1.03759706e-01 -1.03431249e+00 3.62926096e-01
1.37509525e+00 -5.29158771e-01 -5.90855002e-01 3.96059424e-01
9.42599714e-01 -4.40571159e-01 -7.83221781e-01 6.12281144e-01
1.84387073e-01 -8.69358301e-01 1.25836742e+00 -3.55508894e-01
2.17543066e-01 -2.21480101e-01 -3.72811556e-01 -1.09657121e+00
-4.74782825e-01 -4.36773181e-01 -5.29550433e-01 9.33952034e-01
-9.88107994e-02 -3.45599234e-01 1.05305374e+00 4.19453859e-01
-4.18223500e-01 -1.01384127e+00 -9.44872737e-01 -1.11798394e+00
-5.87936230e-02 -8.49402606e-01 8.74306738e-01 1.00906420e+00
-5.19322157e-01 5.64734861e-02 -8.99918601e-02 3.14764738e-01
1.16086817e+00 2.41730392e-01 9.51335549e-01 -1.76030815e+00
1.19727179e-01 -4.83410507e-01 -8.68336618e-01 -6.60626054e-01
8.35135803e-02 -9.48340714e-01 2.58916691e-02 -1.54084885e+00
-5.93143627e-02 -6.38641417e-01 -2.92391092e-01 4.12977338e-01
3.89542803e-02 1.78903088e-01 2.67421156e-01 3.73152614e-01
-6.58883572e-01 1.02803934e+00 9.06252742e-01 -3.39765996e-01
-2.63598472e-01 -2.33653560e-02 -4.35619861e-01 7.84187436e-01
7.18088210e-01 -4.49735671e-01 -4.78099257e-01 -3.78074825e-01
5.08758798e-02 -2.72732288e-01 5.93084812e-01 -1.37257349e+00
6.33612275e-01 -1.08070701e-01 1.89780638e-01 -1.04886556e+00
5.71025670e-01 -8.28478575e-01 -1.10854864e-01 1.60907716e-01
4.27572459e-01 -1.22290470e-01 1.06532909e-01 7.94991970e-01
-5.90819716e-02 -4.62679518e-03 8.05810273e-01 -3.05145890e-01
-1.11397505e+00 1.18336976e+00 5.59878290e-01 1.54735684e-01
1.16858530e+00 -5.22592247e-01 -1.86316460e-01 2.86286529e-02
-3.41235042e-01 2.95393258e-01 6.49708867e-01 4.39604193e-01
1.20626843e+00 -1.37406492e+00 -5.06481528e-01 3.07318807e-01
4.99864459e-01 6.66471481e-01 6.00152731e-01 7.02219546e-01
-5.80513179e-01 3.89934294e-02 -1.29557967e-01 -1.04549694e+00
-9.25755978e-01 5.66981912e-01 6.99462891e-02 3.46550256e-01
-1.29638624e+00 1.06790006e+00 -4.45599742e-02 -4.79803920e-01
2.47696742e-01 -1.39799058e-01 8.83385316e-02 3.75084393e-02
5.67685187e-01 5.94485343e-01 2.27660164e-01 -2.94976026e-01
-4.84846950e-01 9.00673389e-01 -1.17415063e-01 4.66873109e-01
1.89523208e+00 1.50209546e-01 1.30228207e-01 6.28701746e-01
1.33183455e+00 -5.22394359e-01 -1.50774884e+00 -5.85440814e-01
-1.26618356e-01 -8.15669894e-01 3.53384286e-01 -1.58472866e-01
-1.11660957e+00 1.07774103e+00 6.27209663e-01 -7.02371672e-02
6.24848425e-01 1.81512311e-01 9.38642442e-01 4.86099422e-01
7.39862025e-01 -9.90246236e-01 7.99252391e-02 6.18514538e-01
6.76284909e-01 -1.49449050e+00 6.90608323e-02 -3.63101304e-01
-2.21441060e-01 1.02385271e+00 7.24934876e-01 -4.87024426e-01
1.09424949e+00 2.47728482e-01 -2.79289961e-01 -7.39810228e-01
-6.90612078e-01 -3.14221650e-01 2.69436866e-01 7.06589282e-01
-1.03806071e-01 -1.24484792e-01 4.36398029e-01 3.68924409e-01
-2.50548750e-01 1.73731357e-01 1.37234986e-01 1.17538452e+00
-6.20444477e-01 -4.53568071e-01 -2.00082213e-01 7.54529476e-01
2.55384624e-01 -7.33746663e-02 1.81109808e-05 7.15804160e-01
8.68165269e-02 5.44707537e-01 3.61728102e-01 -5.67961514e-01
4.69844252e-01 -2.38920450e-01 3.14012438e-01 -8.54139626e-01
-5.31107746e-02 -1.86720788e-01 -5.80701590e-01 -8.32491934e-01
-1.34182662e-01 -5.87280929e-01 -1.20800674e+00 -5.40053546e-01
-2.90852785e-01 -1.53931364e-01 7.76773930e-01 8.51145923e-01
6.38975918e-01 4.93610322e-01 7.51919687e-01 -1.30823016e+00
-6.74252152e-01 -7.78939426e-01 -7.39761174e-01 2.94979692e-01
3.90120059e-01 -1.04573965e+00 -5.97472310e-01 -4.74044323e-01] | [8.01235580444336, -3.395920753479004] |
0ed61d08-5921-40b5-9915-8d80be4e69ab | a-large-corpus-of-product-reviews-in | null | null | https://aclanthology.org/L14-1354 | https://aclanthology.org/L14-1354.pdf | A Large Corpus of Product Reviews in Portuguese: Tackling Out-Of-Vocabulary Words | Web 2.0 has allowed a never imagined communication boom. With the widespread use of computational and mobile devices, anyone, in practically any language, may post comments in the web. As such, formal language is not necessarily used. In fact, in these communicative situations, language is marked by the absence of more complex syntactic structures and the presence of internet slang, with missing diacritics, repetitions of vowels, and the use of chat-speak style abbreviations, emoticons and colloquial expressions. Such language use poses severe new challenges for Natural Language Processing (NLP) tools and applications, which, so far, have focused on well-written texts. In this work, we report the construction of a large web corpus of product reviews in Brazilian Portuguese and the analysis of its lexical phenomena, which support the development of a lexical normalization tool for, in future work, subsidizing the use of standard NLP products for web opinion mining and summarization purposes. | ['ra', "S Alu{\\'\\i}sio", 'Maria das Gra{\\c{c}}as Volpe Nunes', 'Lucas Avan{\\c{c}}o', 'Thiago Pardo', 'Nathan Hartmann', 'Magali Duran', 'Pedro Balage'] | 2014-05-01 | null | null | null | lrec-2014-5 | ['lexical-normalization'] | ['natural-language-processing'] | [ 1.21482136e-02 6.94106519e-02 6.75350875e-02 -1.31368294e-01
-2.18156248e-01 -8.28554392e-01 8.06202650e-01 8.61478209e-01
-5.03462195e-01 7.54951060e-01 3.00720602e-01 -5.99214733e-01
1.25263527e-01 -6.09233975e-01 1.43791452e-01 -4.41128105e-01
4.03755575e-01 2.80412316e-01 1.55167446e-01 -6.54254079e-01
5.03779829e-01 4.31680411e-01 -1.61030614e+00 3.45255852e-01
8.26755762e-01 4.81994390e-01 2.01079652e-01 4.10351932e-01
-8.05933952e-01 6.86962187e-01 -8.26948464e-01 -1.00357151e+00
-2.29554310e-01 -4.54416484e-01 -6.16729736e-01 2.41815656e-01
5.61956465e-02 2.42756248e-01 3.55298370e-01 1.06182861e+00
2.99838156e-01 1.22426175e-01 4.93839234e-01 -6.30794644e-01
-4.07859981e-01 6.85965002e-01 -4.54249769e-01 1.10892102e-01
7.03472674e-01 -1.93731681e-01 1.13251865e+00 -6.93027914e-01
9.99944508e-01 1.00519168e+00 3.35726172e-01 2.14542001e-01
-9.61049318e-01 -1.15989819e-01 -3.48727703e-02 -1.75196439e-01
-1.08367288e+00 -2.69579411e-01 7.31384754e-01 -5.09368539e-01
1.15876329e+00 3.78473252e-01 7.03037798e-01 1.07533884e+00
6.53585494e-02 3.88800204e-01 9.49566126e-01 -9.93605435e-01
2.87962079e-01 6.63418055e-01 3.49782437e-01 3.56953561e-01
4.15072888e-01 -7.92178988e-01 -3.90348345e-01 -2.28170410e-01
-1.00820147e-01 -2.41098031e-01 -2.68511713e-01 1.65780261e-01
-8.97630036e-01 9.09441233e-01 -3.75576705e-01 1.19083333e+00
-4.95159477e-01 -5.60219824e-01 8.50143433e-01 3.43992978e-01
7.59009063e-01 4.47695732e-01 -6.24808669e-01 -6.10677660e-01
-6.51818573e-01 1.47776157e-02 1.25741327e+00 6.42894149e-01
5.96860647e-01 -3.17313582e-01 4.98383254e-01 1.09888947e+00
1.57312170e-01 5.03523171e-01 8.75070155e-01 -6.37524843e-01
4.37465698e-01 1.07982981e+00 1.30970731e-01 -1.25120211e+00
-3.46305996e-01 1.79448053e-02 -3.82409096e-01 -2.03319103e-01
3.77099007e-01 -3.07558447e-01 -7.40354136e-02 1.29509139e+00
3.64168048e-01 -7.32696295e-01 2.28186801e-01 3.69471639e-01
6.11348033e-01 6.64273500e-01 5.16151660e-04 -6.79296613e-01
1.43239558e+00 -4.11268532e-01 -1.15209472e+00 -4.24007811e-02
7.98108876e-01 -1.40673983e+00 1.41596353e+00 5.21500528e-01
-9.57628727e-01 -2.99301475e-01 -8.90526950e-01 -5.71355876e-03
-9.19142962e-01 -6.21485673e-02 5.21787465e-01 1.05859947e+00
-5.01340330e-01 4.98181671e-01 -5.35564184e-01 -8.26229513e-01
-5.97393364e-02 -2.88978182e-02 -4.78735924e-01 1.42355993e-01
-9.83868003e-01 9.89269018e-01 1.99292615e-01 1.63062066e-01
5.95894098e-01 -1.99727967e-01 -8.16921830e-01 -2.15227649e-01
4.61403996e-01 6.87221438e-02 1.13671911e+00 -1.41088831e+00
-1.63053799e+00 1.27948868e+00 -1.86450556e-01 -2.61681437e-01
2.81384021e-01 -2.82872081e-01 -7.44108140e-01 -4.21239585e-02
7.69716501e-02 -1.41626924e-01 5.85689366e-01 -8.30212295e-01
-6.20763659e-01 -3.23537707e-01 1.10860847e-01 1.83467269e-02
-6.93980634e-01 5.62505066e-01 -1.91711113e-01 -4.64615107e-01
-2.18792096e-01 -7.10597992e-01 -3.22161689e-02 -4.79840815e-01
-2.42396861e-01 -5.21616876e-01 4.58659142e-01 -8.38073492e-01
1.58328247e+00 -2.60611224e+00 3.85631155e-03 4.33521211e-01
1.11810021e-01 3.53477716e-01 2.08689585e-01 9.09941375e-01
1.49667442e-01 4.40242469e-01 -2.13046283e-01 -9.28660631e-02
2.06867933e-01 4.08405781e-01 -6.62937714e-03 3.03979665e-01
-1.65989816e-01 5.61734021e-01 -8.34464729e-01 -5.44907153e-01
3.22985560e-01 3.84560019e-01 -2.48908862e-01 -2.34082118e-01
-1.52639791e-01 -1.75961152e-01 -4.47543800e-01 4.32465047e-01
2.33413383e-01 -2.33317036e-02 5.98268807e-01 2.58307695e-01
-7.68645942e-01 5.29031217e-01 -8.67266297e-01 1.27749038e+00
-8.00374866e-01 7.94717014e-01 5.97212054e-02 -6.30164146e-01
8.78489256e-01 2.46229544e-01 5.45046926e-01 -7.84244955e-01
4.84848201e-01 5.95482230e-01 1.18839152e-01 -7.39339411e-01
6.48096263e-01 7.59742036e-02 -9.41293612e-02 3.28301251e-01
-2.50992417e-01 -2.97903270e-01 7.96612620e-01 2.07834214e-01
7.74988055e-01 -5.44497035e-02 8.07837844e-01 -3.77849519e-01
8.74463558e-01 8.63175318e-02 1.33778557e-01 1.31076425e-01
1.38306795e-02 4.28045541e-01 8.09327424e-01 -2.65801966e-01
-8.92158151e-01 -5.36346674e-01 -1.77004904e-01 9.56879675e-01
-3.94133359e-01 -7.97524154e-01 -9.36970592e-01 -5.36060214e-01
-2.78066218e-01 9.58383739e-01 -7.95576572e-02 2.85975903e-01
-6.37369215e-01 -5.64011157e-01 4.89783734e-02 -3.83488327e-01
-2.90942844e-02 -1.45014250e+00 -5.58155894e-01 6.33858681e-01
-3.18095684e-01 -1.36908662e+00 -8.12398344e-02 2.08493754e-01
-5.84673285e-01 -1.11746323e+00 -5.32003582e-01 -7.10968494e-01
3.19268078e-01 -1.76370349e-02 1.00485218e+00 1.69543505e-01
-5.61266430e-02 1.23883873e-01 -1.00662804e+00 -5.57726860e-01
-7.85628140e-01 2.96343863e-01 -1.41224638e-01 -1.06021389e-02
7.68042624e-01 -6.07939899e-01 9.10826307e-03 -1.32484421e-01
-1.00650525e+00 -2.81073034e-01 2.32602045e-01 4.68651175e-01
-2.16795951e-02 -1.38635442e-01 6.58482969e-01 -1.25088358e+00
1.08953118e+00 -3.52394372e-01 -4.42815304e-01 3.79370749e-02
-3.79025728e-01 -1.70287281e-01 7.81143665e-01 -2.65362471e-01
-1.15007436e+00 -2.35150114e-01 -5.96394360e-01 6.78866446e-01
-2.46971071e-01 9.30675089e-01 -1.85648605e-01 2.19268471e-01
8.51284146e-01 -5.31997532e-02 1.49481058e-01 -5.62021315e-01
3.26689035e-01 1.11507487e+00 -3.06669958e-02 -1.60130709e-01
5.44290602e-01 2.40249917e-01 -4.60733354e-01 -1.64043188e+00
-6.44734442e-01 -6.98718071e-01 -5.86578488e-01 -2.55604953e-01
7.56514490e-01 -3.51312757e-01 -4.95406032e-01 2.96726674e-01
-1.25097609e+00 8.55949447e-02 -5.60365319e-01 2.53081828e-01
2.55508944e-02 9.22683537e-01 -4.67177510e-01 -8.27828288e-01
-2.47670174e-01 -7.89858580e-01 4.97949421e-01 1.90810427e-01
-8.92830193e-01 -9.37948108e-01 1.34034202e-01 5.11457682e-01
2.57659644e-01 2.05892652e-01 1.13861907e+00 -9.63021457e-01
2.53448844e-01 -4.87316489e-01 2.09930807e-01 7.07713246e-01
3.79267246e-01 4.80286211e-01 -5.72166383e-01 2.79256105e-01
9.74986181e-02 -4.69894186e-02 1.62534416e-01 -1.08327240e-01
4.47085381e-01 -4.34178382e-01 1.95388332e-01 -3.37160319e-01
1.24215865e+00 2.60518044e-01 6.37656033e-01 4.78162557e-01
2.20532387e-01 1.02454007e+00 4.47268426e-01 6.44351959e-01
2.16498077e-01 4.83585447e-01 1.40397549e-01 3.06553781e-01
1.74261048e-01 1.08463116e-01 5.55914819e-01 1.39699316e+00
-3.93431447e-02 -2.87205577e-01 -7.52017140e-01 5.91874480e-01
-1.57091308e+00 -5.81923187e-01 -6.67408824e-01 2.13858914e+00
8.37360144e-01 2.98103958e-01 -1.33999959e-01 3.72133732e-01
6.33207977e-01 2.12865263e-01 2.66315192e-01 -9.42262650e-01
-2.51904637e-01 4.29990381e-01 1.72579393e-01 6.05364323e-01
-6.66558981e-01 9.11940515e-01 4.88730097e+00 8.13383460e-01
-1.11485386e+00 1.22199848e-01 1.78776771e-01 1.95582524e-01
-3.33304942e-01 -1.62405804e-01 -5.18575668e-01 6.47189200e-01
8.74953091e-01 -1.17809609e-01 4.09436047e-01 7.98267663e-01
5.19852519e-01 -3.99624556e-01 -6.69850707e-01 9.15539324e-01
2.68005729e-01 -1.05735373e+00 5.97812701e-03 -5.57359047e-02
5.72394907e-01 -3.07048019e-03 -2.35715896e-01 7.55039304e-02
-2.21880168e-01 -4.99318719e-01 6.04010165e-01 -2.99921855e-02
1.92028984e-01 -7.31347024e-01 9.93169665e-01 2.58676946e-01
-6.17105365e-01 1.93349794e-01 -2.37782389e-01 -2.13386327e-01
2.76884824e-01 7.98510194e-01 -3.96221995e-01 5.83128810e-01
3.99666458e-01 5.43999612e-01 -5.08831680e-01 6.30505264e-01
-3.97320300e-01 5.38552999e-01 -5.13182640e-01 -7.44625509e-01
2.23772943e-01 -4.96697426e-01 6.07612908e-01 1.37009859e+00
2.43204474e-01 -1.35989025e-01 -9.74385068e-02 -2.78887060e-02
6.31907908e-03 1.11661649e+00 -5.29365718e-01 -4.73260999e-01
2.84809202e-01 1.47903025e+00 -1.09122396e+00 -1.39333040e-01
-8.17616224e-01 8.93273413e-01 -5.21023870e-02 1.42248854e-01
-2.92802513e-01 -4.55868989e-01 5.33983827e-01 4.33112979e-01
5.45342751e-02 -3.79090011e-01 -2.83764392e-01 -1.19398320e+00
3.92420441e-01 -1.11008596e+00 8.88920277e-02 -4.41268265e-01
-1.43879569e+00 6.79686904e-01 -2.97212362e-01 -8.59933317e-01
-4.15356517e-01 -8.41765583e-01 -4.59009141e-01 5.51948905e-01
-1.31388891e+00 -7.38474786e-01 2.60133058e-01 3.84724110e-01
5.22542238e-01 -1.08877726e-01 9.73151147e-01 5.28259933e-01
-3.02011281e-01 7.29412287e-02 3.06408286e-01 -1.36826620e-01
6.12868965e-01 -1.07521260e+00 1.38816416e-01 5.10962367e-01
4.26935852e-01 6.35449767e-01 1.05773771e+00 -4.07653987e-01
-9.94294822e-01 -3.62495244e-01 1.85407841e+00 -1.94656953e-01
1.21783304e+00 -3.84997904e-01 -7.57271171e-01 2.37808779e-01
6.36367440e-01 -7.78234661e-01 9.40106094e-01 3.70716691e-01
1.30560268e-02 5.77664301e-02 -8.96667659e-01 8.02034378e-01
4.93040472e-01 -5.85448623e-01 -8.71231019e-01 6.25359774e-01
4.34793323e-01 1.70187011e-01 -6.46041095e-01 -1.68586135e-01
4.54928607e-01 -8.65208149e-01 3.41972381e-01 -2.91092306e-01
2.79535085e-01 -9.55428854e-02 6.73788860e-02 -1.13651621e+00
2.94339865e-01 -9.87814307e-01 4.49743658e-01 1.46874905e+00
9.64267135e-01 -6.41624331e-01 4.78150129e-01 5.57948589e-01
-8.81549418e-02 -2.30057076e-01 -9.61333394e-01 -2.56015480e-01
-8.30668285e-02 -6.83336437e-01 6.13097325e-02 1.01716030e+00
7.46230364e-01 6.18402600e-01 -2.33300943e-02 -4.43291098e-01
-7.62208402e-02 -5.47371693e-02 4.60118890e-01 -1.41673398e+00
-3.43825668e-02 -5.20048499e-01 -3.15719068e-01 -5.08415163e-01
1.16831206e-01 -5.88717461e-01 -2.96000749e-01 -1.29682350e+00
-3.21204782e-01 7.41546750e-02 2.36391842e-01 1.96841702e-01
2.55094260e-01 1.37037024e-01 2.69696951e-01 -1.01015501e-01
-4.74926978e-01 2.08748057e-01 9.10249650e-01 1.53887495e-01
-4.66552973e-01 -3.16452682e-02 -7.10973084e-01 1.09172821e+00
9.01960075e-01 -5.20615399e-01 -1.97220787e-01 1.67454436e-01
1.15635002e+00 -3.61069500e-01 -1.98569402e-01 -6.33957386e-01
-8.22496042e-02 -7.67145082e-02 -2.49582514e-01 -3.26622903e-01
3.29902843e-02 -1.19104099e+00 -6.74193650e-02 3.80416423e-01
-1.57428626e-02 2.61118144e-01 -3.31782587e-02 -6.09608181e-02
-3.80830616e-01 -8.27889442e-01 3.85441512e-01 -2.93531299e-01
-3.51449549e-01 -4.05214399e-01 -1.17943788e+00 4.71327417e-02
9.95055258e-01 -3.11138004e-01 9.42818820e-03 -2.93770701e-01
-5.74767530e-01 -1.93452388e-01 6.27328396e-01 4.72050697e-01
2.51553684e-01 -6.29305959e-01 -4.96598333e-01 -6.11467138e-02
1.79892957e-01 -2.38788649e-01 1.10478342e-01 5.57382941e-01
-9.00743544e-01 1.56434685e-01 -1.21003374e-01 1.37326449e-01
-1.44153154e+00 4.41317290e-01 -9.51014459e-02 -2.77735084e-01
-4.53405678e-01 1.94622219e-01 -3.63276958e-01 -3.47061455e-01
-6.83367625e-02 -3.71342331e-01 -6.73484325e-01 8.38411093e-01
5.04572093e-01 2.30476156e-01 3.35783392e-01 -8.39490592e-01
-1.56266004e-01 1.76133960e-01 4.97664735e-02 -3.13007176e-01
1.28726208e+00 -4.56320226e-01 -7.11665154e-01 7.67397344e-01
1.02141893e+00 7.22227216e-01 -3.39467257e-01 2.35981330e-01
5.38247764e-01 -1.50462374e-01 -3.87709796e-01 -7.40978301e-01
-5.72409451e-01 4.95401978e-01 4.42768931e-02 9.62343931e-01
8.00912499e-01 -1.45619780e-01 5.68163514e-01 6.45852387e-01
1.17635064e-01 -1.57603037e+00 -3.73363405e-01 7.52192795e-01
6.69827402e-01 -9.73330557e-01 -1.05511211e-01 -5.83658636e-01
-6.95899546e-01 1.30374503e+00 8.79577175e-02 1.42783031e-01
8.61392558e-01 2.88610071e-01 4.00091439e-01 -2.39255518e-01
-4.68271673e-01 -3.70319605e-01 -1.36683494e-01 4.04409468e-01
8.89052868e-01 1.44677954e-02 -1.38664305e+00 5.16426265e-01
-4.86879498e-01 -1.39038771e-01 7.77565837e-01 8.95349681e-01
-3.17342281e-01 -1.57838202e+00 -2.01545179e-01 2.76152045e-01
-1.11615241e+00 -4.05621082e-01 -6.97988451e-01 9.99377608e-01
1.25593796e-01 1.23772490e+00 4.31272797e-02 1.12565547e-01
4.36791778e-01 3.33815843e-01 2.16883674e-01 -6.63427472e-01
-1.10901964e+00 5.80920801e-02 6.10679209e-01 -1.18081778e-01
-7.00234294e-01 -9.24668670e-01 -1.09986711e+00 -4.95391488e-01
-3.28554690e-01 4.90439296e-01 1.11605060e+00 1.11187589e+00
2.49926895e-02 2.44795799e-01 3.15885991e-01 -4.28646445e-01
-2.30783954e-01 -1.14292490e+00 -7.55006254e-01 4.98077780e-01
-7.60136768e-02 -1.13064595e-01 -3.87041181e-01 1.55187786e-01] | [11.024356842041016, 7.100668907165527] |
3f16e847-c9b7-4861-a31f-26443ae88990 | at-bert-adversarial-training-bert-for-acronym | 2101.03700 | null | https://arxiv.org/abs/2101.03700v2 | https://arxiv.org/pdf/2101.03700v2.pdf | AT-BERT: Adversarial Training BERT for Acronym Identification Winning Solution for SDU@AAAI-21 | Acronym identification focuses on finding the acronyms and the phrases that have been abbreviated, which is crucial for scientific document understanding tasks. However, the limited size of manually annotated datasets hinders further improvement for the problem. Recent breakthroughs of language models pre-trained on large corpora clearly show that unsupervised pre-training can vastly improve the performance of downstream tasks. In this paper, we present an Adversarial Training BERT method named AT-BERT, our winning solution to acronym identification task for Scientific Document Understanding (SDU) Challenge of AAAI 2021. Specifically, the pre-trained BERT is adopted to capture better semantic representation. Then we incorporate the FGM adversarial training strategy into the fine-tuning of BERT, which makes the model more robust and generalized. Furthermore, an ensemble mechanism is devised to involve the representations learned from multiple BERT variants. Assembling all these components together, the experimental results on the SciAI dataset show that our proposed approach outperforms all other competitive state-of-the-art methods. | ['Jiayu Tang', 'Weilin Wu', 'Guanxiong Zeng', 'Qiwei Zhong', 'Yang Zhang', 'Wangli Lin', 'Danqing Zhu'] | 2021-01-11 | null | null | null | null | ['unsupervised-pre-training'] | ['methodology'] | [ 3.90898198e-01 1.60760930e-04 -1.67715251e-01 -3.77468377e-01
-7.50945091e-01 -7.76808023e-01 7.89428473e-01 -7.04657612e-03
-3.61954331e-01 9.45116580e-01 2.30582014e-01 -4.31124955e-01
-1.74097985e-01 -7.38213718e-01 -8.26480150e-01 -6.18805468e-01
5.38960159e-01 5.46905100e-01 -3.33130389e-01 -1.55606031e-01
5.05167067e-01 4.77115661e-01 -8.55661273e-01 2.82800198e-01
1.08050609e+00 8.42495382e-01 2.74814844e-01 3.29752415e-01
-5.88333547e-01 5.79314053e-01 -8.52049291e-01 -4.73359883e-01
-3.08664162e-02 -1.68263420e-01 -9.48534667e-01 -4.94791478e-01
1.69743150e-01 -1.30131165e-03 -2.57776320e-01 9.41751957e-01
3.61845076e-01 1.01079322e-01 1.05003214e+00 -1.05982149e+00
-8.84521604e-01 1.09967148e+00 -5.83277106e-01 2.61887044e-01
2.78315365e-01 -2.62967765e-01 1.17814720e+00 -8.43299747e-01
6.67180777e-01 1.45919466e+00 5.39515972e-01 7.28742838e-01
-6.92868054e-01 -1.06391025e+00 3.38568658e-01 1.44671097e-01
-1.37286711e+00 -1.07949041e-01 6.75357521e-01 -3.05766106e-01
7.73326635e-01 7.51328617e-02 3.12554464e-02 1.45716572e+00
-1.11444838e-01 8.85163248e-01 7.06818521e-01 -3.76809120e-01
-2.44694818e-02 1.05004527e-01 3.68713379e-01 4.24901515e-01
4.30096537e-01 -2.24073991e-01 -3.93010944e-01 -5.94897605e-02
6.82791233e-01 -9.14320499e-02 -3.96307021e-01 -2.67130762e-01
-1.28792846e+00 8.24267089e-01 5.00480056e-01 4.33678657e-01
-2.27915183e-01 1.08071886e-01 5.33438504e-01 -1.43181365e-02
2.60402918e-01 1.12415349e+00 -6.01564527e-01 -8.00794885e-02
-8.66510630e-01 1.32387325e-01 5.38005531e-01 1.23127317e+00
3.54201376e-01 -1.28120720e-01 -1.15454026e-01 9.44527447e-01
2.71352977e-01 3.85804504e-01 5.72291493e-01 -6.61005259e-01
7.92244673e-01 8.21619868e-01 9.56501067e-03 -6.59623563e-01
-4.46703911e-01 -8.73407781e-01 -9.25948143e-01 -3.71949285e-01
3.32898259e-01 -9.30316001e-02 -1.03444374e+00 1.76365030e+00
-1.52064368e-01 4.61847574e-01 4.17392492e-01 6.65178061e-01
1.24689960e+00 8.66525352e-01 5.92249393e-01 1.79298788e-01
1.49558818e+00 -1.18542409e+00 -6.86243892e-01 -2.32360691e-01
7.12806582e-01 -7.14838803e-01 7.76349664e-01 4.65796441e-01
-6.03324711e-01 -7.62914956e-01 -1.35286772e+00 -2.72167236e-01
-9.16831195e-01 3.51292789e-01 8.49981606e-01 5.91162324e-01
-3.35393667e-01 6.75596297e-01 -3.40689719e-01 -1.53378010e-01
5.90373576e-01 2.97720939e-01 -2.18146235e-01 -2.69656599e-01
-1.52613485e+00 7.06475079e-01 6.40862405e-01 1.50293067e-01
-7.76129067e-01 -1.11281276e+00 -6.89263701e-01 2.54714340e-01
2.80769199e-01 -9.77457404e-01 8.86301517e-01 -6.63001776e-01
-1.26543891e+00 9.27103519e-01 2.85283327e-02 -6.02010250e-01
4.79375511e-01 -5.14838517e-01 -4.34299290e-01 1.54184327e-01
1.24643400e-01 7.96977460e-01 7.35709965e-01 -1.33456552e+00
-3.24380815e-01 -2.94684678e-01 9.14931446e-02 1.97873428e-01
-4.26724374e-01 2.98003275e-02 -6.64639831e-01 -9.12865520e-01
-2.00273499e-01 -7.52843380e-01 -8.88729542e-02 -3.07565868e-01
-3.65735769e-01 -6.10681415e-01 5.82358539e-01 -5.96385956e-01
1.43150675e+00 -2.17392302e+00 4.90062445e-01 2.06006408e-01
-2.60435250e-02 5.46244740e-01 -1.86945304e-01 3.87051910e-01
-3.35740536e-01 4.69859481e-01 -3.53753954e-01 -3.60605031e-01
-1.50312304e-01 2.34684274e-01 -5.88606477e-01 -6.39273226e-02
2.48769850e-01 9.47397888e-01 -9.81059074e-01 -3.07667971e-01
1.16339386e-01 1.98634252e-01 -2.33086780e-01 2.78217763e-01
-3.01236808e-01 7.21563160e-01 -1.02325165e+00 5.56832373e-01
7.82782614e-01 -4.51688021e-01 6.09546788e-02 -2.06025943e-01
1.81518033e-01 2.62353927e-01 -9.22166824e-01 2.35984540e+00
-6.88310385e-01 1.67146623e-01 -3.18053097e-01 -1.38849640e+00
1.12073374e+00 1.92254350e-01 2.51630932e-01 -3.31357032e-01
4.38500270e-02 4.68179047e-01 -9.81229916e-02 -2.10081711e-01
3.05638075e-01 1.44834936e-01 -3.79647881e-01 -2.63333437e-03
1.07167177e-01 -3.61647099e-01 3.18938732e-01 5.36039293e-01
8.34466755e-01 4.79833424e-01 4.04927343e-01 -2.40628660e-01
1.16920042e+00 -1.47817239e-01 5.81719816e-01 9.47351038e-01
9.69544947e-02 6.25901937e-01 3.41547281e-01 -3.62036675e-01
-8.57276499e-01 -7.95478642e-01 -3.24232757e-01 7.93288052e-01
2.07801715e-01 -4.50228363e-01 -6.80879772e-01 -9.64703023e-01
9.63856801e-02 9.39272285e-01 -6.44411922e-01 -2.11664245e-01
-7.55349576e-01 -7.88466454e-01 8.06509435e-01 6.89691246e-01
6.95647359e-01 -1.11959541e+00 1.08252391e-01 1.76706210e-01
-1.26169875e-01 -1.14366865e+00 -3.07118595e-01 1.00207359e-01
-6.73480988e-01 -9.66550291e-01 -1.07602060e+00 -7.54296660e-01
5.90095282e-01 7.35119507e-02 1.05845392e+00 6.80213571e-02
-2.53624201e-01 5.09234853e-02 -4.94755626e-01 -6.41111374e-01
-4.67764974e-01 3.67649734e-01 -2.13629693e-01 -2.53383130e-01
5.08944213e-01 -2.98953235e-01 -5.28311253e-01 2.80624032e-02
-7.60105908e-01 9.65906382e-02 7.08245218e-01 9.38669980e-01
3.29861641e-01 -2.19415084e-01 1.02782404e+00 -1.35751665e+00
3.70131224e-01 -4.17132318e-01 -2.87333369e-01 5.04554987e-01
-5.99368870e-01 3.02999049e-01 9.38748956e-01 -2.57585466e-01
-1.21786916e+00 -2.09374234e-01 -3.25812459e-01 -1.61051318e-01
-9.23707560e-02 5.81816137e-01 -6.12493992e-01 8.09517056e-02
3.97651851e-01 1.69547632e-01 -4.41775411e-01 -7.79599369e-01
6.20780110e-01 9.80739474e-01 6.07448399e-01 -7.77876675e-01
8.40276182e-01 1.53389275e-01 -5.04448563e-02 -5.26691020e-01
-1.29837763e+00 -6.96963370e-01 -5.29397249e-01 3.28111917e-01
8.39759290e-01 -1.10551095e+00 -4.55834240e-01 3.98269206e-01
-1.34233260e+00 3.14689338e-01 6.25668466e-02 3.08102608e-01
-4.35315758e-01 6.13833129e-01 -3.99787635e-01 -4.30613101e-01
-6.71742141e-01 -1.17918324e+00 1.08161974e+00 5.70594311e-01
-3.23036015e-01 -8.57045770e-01 -2.17319176e-01 8.36279988e-01
3.03357333e-01 -6.43931106e-02 1.29888129e+00 -1.10833728e+00
-3.77751142e-01 -1.39581710e-02 -6.50185287e-01 5.35974383e-01
1.41226411e-01 -2.44122893e-01 -9.51593459e-01 -1.06393538e-01
-3.23034912e-01 -2.50599235e-01 1.19987583e+00 1.40840828e-01
1.82412326e+00 1.85818911e-01 -6.15783572e-01 8.08905125e-01
1.21920705e+00 3.43610168e-01 6.83023870e-01 7.20932841e-01
9.21337008e-01 3.21953952e-01 7.13604927e-01 2.71101773e-01
8.48650839e-03 4.55420226e-01 2.93034762e-01 1.18838213e-02
-1.72519669e-01 -3.50073487e-01 -1.42321959e-01 6.89497948e-01
-2.97946744e-02 -5.39956510e-01 -8.72113228e-01 3.95645946e-01
-1.66984701e+00 -6.13871336e-01 1.01464994e-01 1.93281913e+00
9.50335801e-01 2.26057842e-01 -4.04607594e-01 -3.16590853e-02
6.65067673e-01 1.60292476e-01 -5.44257998e-01 -5.07690668e-01
-2.23850012e-01 3.63117158e-01 5.55209219e-01 1.60781309e-01
-1.28671336e+00 1.32619059e+00 5.66386318e+00 1.27021706e+00
-7.68534184e-01 -1.74903199e-01 5.28380334e-01 5.86881161e-01
-5.66188157e-01 -1.06320478e-01 -1.03605831e+00 5.94981611e-01
9.32945073e-01 -2.51845539e-01 3.14784288e-01 8.06341648e-01
-4.10552584e-02 4.82019991e-01 -1.21309173e+00 9.76801455e-01
1.53420925e-01 -1.46576011e+00 6.90060675e-01 -2.27749795e-01
7.39413202e-01 -3.57002825e-01 -1.26623083e-02 6.20564461e-01
2.10239947e-01 -1.22884226e+00 2.36312017e-01 5.14450848e-01
6.60518467e-01 -8.43023479e-01 1.11308205e+00 1.85171783e-01
-8.43500674e-01 -4.86632772e-02 -4.84885603e-01 3.57185632e-01
1.38749294e-02 4.61070418e-01 -6.95868015e-01 1.23830450e+00
2.87807345e-01 1.03187203e+00 -6.14463329e-01 8.70067060e-01
-5.70220351e-01 6.50622547e-01 -3.93468551e-02 -7.98793584e-02
4.85587448e-01 -3.83385628e-01 4.14347023e-01 1.47127223e+00
2.52280384e-01 -5.63323684e-03 -1.03598535e-01 9.47892904e-01
-6.79574430e-01 1.20294914e-01 -5.88800669e-01 -6.36518538e-01
6.14441097e-01 1.17092848e+00 -3.77424777e-01 -5.02697051e-01
-4.84426826e-01 1.00513864e+00 2.08850697e-01 3.91504914e-01
-9.66905892e-01 -5.60932100e-01 5.17506778e-01 -6.60694301e-01
3.31617236e-01 -1.63391218e-01 -5.03924787e-01 -1.13585281e+00
-3.56001258e-01 -9.97336984e-01 5.29516757e-01 -8.21384132e-01
-1.33499265e+00 4.15516555e-01 -2.34711647e-01 -1.14892042e+00
4.07232009e-02 -8.44717741e-01 -6.82903647e-01 9.13982391e-01
-1.69107890e+00 -1.37803733e+00 -1.77124485e-01 1.28204793e-01
8.19551170e-01 -4.26474839e-01 1.04322612e+00 3.74238819e-01
-8.34725499e-01 8.30134451e-01 4.48511034e-01 2.02365041e-01
1.00499356e+00 -1.28189099e+00 3.45918417e-01 7.92399406e-01
2.70236850e-01 1.11203277e+00 4.96664226e-01 -6.10725105e-01
-1.17786205e+00 -9.92329836e-01 7.78454900e-01 -3.78777832e-01
8.77761602e-01 -2.43245631e-01 -1.00374162e+00 4.59634602e-01
8.13964680e-02 -5.01150429e-01 7.83741653e-01 1.47909433e-01
-5.28564095e-01 -1.12923859e-02 -7.71930218e-01 3.62750351e-01
1.04308033e+00 -2.67317355e-01 -1.03592718e+00 4.98755127e-01
9.57616270e-01 -2.76925862e-01 -9.55307841e-01 5.22535920e-01
3.47085238e-01 -5.21153547e-02 1.41275883e+00 -1.00462151e+00
8.36751580e-01 -2.39509493e-01 1.32547125e-01 -1.20623589e+00
-1.21091932e-01 -5.31204462e-01 1.06412746e-01 1.54957676e+00
5.26083648e-01 -3.31498176e-01 6.79202080e-01 3.48310769e-01
-3.23244393e-01 -5.66314459e-01 -7.77603507e-01 -7.15644062e-01
4.82176840e-01 -1.24137208e-01 6.72745883e-01 9.26490366e-01
-3.18326712e-01 6.17236972e-01 -3.64469975e-01 1.12856567e-01
4.86309499e-01 3.17600846e-01 6.40391767e-01 -1.22412431e+00
-1.21167295e-01 -4.41101909e-01 -2.42414340e-01 -1.41230690e+00
4.60669875e-01 -1.07347131e+00 -2.49300480e-01 -1.50809884e+00
3.06109488e-01 -5.20806193e-01 -7.66247094e-01 3.74248326e-01
-6.89367473e-01 -1.18596435e-01 1.62316803e-02 1.89774632e-01
-4.49881583e-01 6.52963102e-01 1.30135298e+00 -6.29307389e-01
1.33489043e-01 -1.02107801e-01 -9.82513666e-01 7.00285375e-01
6.81158900e-01 -3.54251355e-01 -3.26295704e-01 -5.40052056e-01
6.16689660e-02 -4.07428652e-01 2.19587423e-02 -7.48281300e-01
8.08843449e-02 4.77866121e-02 4.33053434e-01 -6.36748850e-01
2.16518059e-01 -6.07033968e-01 -3.78899038e-01 2.42360651e-01
-6.11196339e-01 -3.01523149e-01 2.56205767e-01 7.67107606e-01
-2.25662857e-01 -8.35930705e-01 6.66812599e-01 -3.19326788e-01
-8.94753635e-01 1.69435903e-01 -1.07633680e-01 2.14825794e-01
7.19149828e-01 1.49271771e-01 -4.02667493e-01 3.08158845e-02
-5.74175119e-01 4.04327661e-01 3.52876894e-02 6.34313881e-01
3.03020746e-01 -8.84345472e-01 -6.96617961e-01 -1.37421489e-01
2.51620829e-01 1.56240270e-01 3.15208495e-01 2.78026670e-01
-5.93583405e-01 7.80985892e-01 -7.90924653e-02 -1.66671962e-01
-1.06265092e+00 7.58956492e-01 -7.71910399e-02 -6.80791497e-01
-6.03519797e-01 1.18266189e+00 4.71382827e-01 -4.78636950e-01
4.57271636e-01 -1.51754647e-01 -6.79197609e-01 -2.45056897e-02
4.79353756e-01 1.27189264e-01 -5.72180189e-02 -1.62872136e-01
-4.99706477e-01 7.32883096e-01 -5.68566561e-01 2.72323281e-01
1.42127967e+00 -6.68960214e-02 3.05327233e-02 1.51419058e-01
1.27173114e+00 -7.84663484e-02 -5.66037893e-01 -3.27412337e-01
1.02925405e-01 -1.36491945e-02 9.48480591e-02 -1.13547981e+00
-8.53062868e-01 1.07136297e+00 3.08792919e-01 -3.27995867e-01
8.92743826e-01 -1.98104545e-01 8.95017564e-01 6.09150887e-01
1.37599722e-01 -9.87408638e-01 -6.01401627e-02 4.65777338e-01
9.86574054e-01 -1.23368990e+00 1.81918457e-01 -7.57668197e-01
-6.89171851e-01 1.31222367e+00 5.47998011e-01 4.14082669e-02
1.80031359e-01 -1.29472777e-01 -1.32049387e-02 9.49660316e-03
-2.29540035e-01 1.03415407e-01 3.98346752e-01 3.55783761e-01
4.06648308e-01 -6.18668608e-02 -6.50497019e-01 1.07619607e+00
-1.70716524e-01 -6.33594170e-02 4.19162214e-01 5.68795383e-01
-1.24384925e-01 -1.47752726e+00 -1.44064084e-01 2.63266116e-01
-4.63787347e-01 -4.31069940e-01 -5.41183054e-01 7.44602561e-01
1.45727038e-01 7.74260163e-01 -3.71182144e-01 -2.13158280e-02
1.40678614e-01 3.64663631e-01 2.36245602e-01 -6.84768438e-01
-5.78422248e-01 -2.41969779e-01 2.33331650e-01 -2.03169212e-01
-4.34144616e-01 -3.60181421e-01 -1.36720514e+00 1.55109614e-01
-2.43596315e-01 5.55488169e-01 7.58910239e-01 9.96783316e-01
4.64226842e-01 9.56407845e-01 5.33309937e-01 -3.03074300e-01
-7.47192383e-01 -1.17882133e+00 -2.35403091e-01 6.16576254e-01
-1.58103228e-01 -9.03449476e-01 -4.13918912e-01 -1.68952402e-02] | [10.011470794677734, 8.566057205200195] |
0af80438-25b2-43d3-a127-6d3b2e689fa0 | sparse-in-space-and-time-audio-visual | 2210.07055 | null | https://arxiv.org/abs/2210.07055v1 | https://arxiv.org/pdf/2210.07055v1.pdf | Sparse in Space and Time: Audio-visual Synchronisation with Trainable Selectors | The objective of this paper is audio-visual synchronisation of general videos 'in the wild'. For such videos, the events that may be harnessed for synchronisation cues may be spatially small and may occur only infrequently during a many seconds-long video clip, i.e. the synchronisation signal is 'sparse in space and time'. This contrasts with the case of synchronising videos of talking heads, where audio-visual correspondence is dense in both time and space. We make four contributions: (i) in order to handle longer temporal sequences required for sparse synchronisation signals, we design a multi-modal transformer model that employs 'selectors' to distil the long audio and visual streams into small sequences that are then used to predict the temporal offset between streams. (ii) We identify artefacts that can arise from the compression codecs used for audio and video and can be used by audio-visual models in training to artificially solve the synchronisation task. (iii) We curate a dataset with only sparse in time and space synchronisation signals; and (iv) the effectiveness of the proposed model is shown on both dense and sparse datasets quantitatively and qualitatively. Project page: v-iashin.github.io/SparseSync | ['Andrew Zisserman', 'Esa Rahtu', 'Weidi Xie', 'Vladimir Iashin'] | 2022-10-13 | null | null | null | null | ['audio-visual-synchronization', 'audio-visual-synchronization'] | ['audio', 'computer-vision'] | [ 2.73642033e-01 1.30494162e-01 7.84547776e-02 1.51125550e-01
-8.60391319e-01 -3.98828000e-01 7.46922433e-01 -1.80217057e-01
5.23524582e-02 4.42859292e-01 6.19848251e-01 1.35002837e-01
-1.65472195e-01 -7.35242292e-02 -9.15439010e-01 -6.99299574e-01
-4.75107908e-01 4.63117920e-02 2.93299943e-01 -2.42248271e-02
8.03318769e-02 1.34006262e-01 -2.01616549e+00 6.52703881e-01
2.83418834e-01 9.06019688e-01 5.92083991e-01 1.11509061e+00
1.33032382e-01 1.19081545e+00 -6.22658014e-01 1.63159937e-01
1.68533593e-01 -9.66221392e-01 -6.34526551e-01 1.49743900e-01
3.72461826e-01 1.72896296e-01 -6.53575599e-01 7.36162066e-01
6.74157560e-01 1.07283227e-01 3.99971008e-01 -1.37868381e+00
-9.43310708e-02 5.55255532e-01 -4.82886970e-01 4.94558632e-01
1.08077908e+00 1.40659600e-01 9.04012442e-01 -7.43254662e-01
9.59003210e-01 1.10476375e+00 7.93123841e-01 3.39254767e-01
-1.11535430e+00 -6.67283893e-01 -4.36318442e-02 5.65144181e-01
-1.54166293e+00 -1.02752972e+00 1.01831043e+00 -5.13311625e-01
9.74452496e-01 5.03679395e-01 1.00296688e+00 1.51163971e+00
1.33281827e-01 7.36915588e-01 7.28124499e-01 -5.28438270e-01
9.68708843e-02 -6.37663901e-02 -7.86338806e-01 6.76013827e-02
-6.96352482e-01 3.90455067e-01 -1.18043411e+00 1.43581182e-01
7.24072933e-01 -2.87698686e-01 -5.17320871e-01 -1.17981136e-01
-1.49318552e+00 6.84454441e-01 1.59529820e-01 6.80859566e-01
-3.41584682e-01 1.82161838e-01 6.40282393e-01 6.29142106e-01
3.10265273e-01 3.02827060e-01 9.31285173e-02 -5.11702180e-01
-1.44402146e+00 1.52047008e-01 6.81931019e-01 1.12055516e+00
3.63628656e-01 3.61182243e-01 2.92965975e-02 9.08900976e-01
2.11649567e-01 4.04351354e-01 7.61683583e-01 -9.86851037e-01
5.71070552e-01 -3.16272408e-01 -1.06300041e-01 -1.21857977e+00
-1.83006257e-01 -7.15639889e-02 -8.90408278e-01 -2.28190467e-01
2.92310655e-01 1.14236370e-01 -5.63358545e-01 1.57366538e+00
1.22875400e-01 9.19861317e-01 -9.91589576e-02 8.65753531e-01
9.33387756e-01 1.00781536e+00 -2.16291875e-01 -6.96866810e-01
1.20046175e+00 -5.67471385e-01 -1.04528499e+00 -1.64189935e-01
3.06117922e-01 -1.14748454e+00 8.13779175e-01 3.58766109e-01
-1.41437054e+00 -6.82890773e-01 -8.84879768e-01 1.89499289e-01
1.96860190e-02 -3.20223540e-01 2.50149310e-01 2.98021622e-02
-1.11484170e+00 5.20909488e-01 -5.90408802e-01 -3.91488373e-01
1.17137514e-01 9.18276086e-02 -3.50593030e-01 8.05245936e-02
-1.39069927e+00 7.58719325e-01 3.04284573e-01 1.23226978e-01
-1.16984129e+00 -7.54978776e-01 -9.96585965e-01 -2.18640834e-01
1.83082640e-01 -2.74241418e-01 1.20619535e+00 -1.37408948e+00
-1.38893628e+00 7.87979126e-01 -1.91742271e-01 -4.29248154e-01
4.53518182e-01 1.34256855e-03 -7.63864756e-01 6.90224648e-01
9.06450078e-02 6.76150382e-01 1.17913914e+00 -1.22215819e+00
-6.61781192e-01 3.21762621e-01 -2.42496982e-01 3.95676613e-01
1.62319876e-02 2.98389137e-01 -6.08114362e-01 -9.74327564e-01
4.71235551e-02 -7.94504881e-01 1.79120004e-01 -2.03449801e-01
-2.48764336e-01 3.83869827e-01 1.10061884e+00 -7.55896211e-01
1.42903161e+00 -2.47976041e+00 4.97993737e-01 2.06561208e-01
5.13871238e-02 1.29475966e-02 -6.33896440e-02 6.80544138e-01
-4.06376153e-01 -3.07855815e-01 1.15549713e-01 -3.62561405e-01
-1.96977973e-01 2.26156086e-01 -4.08331394e-01 7.26025462e-01
-6.06917962e-03 5.87986946e-01 -8.27930450e-01 -7.45841563e-01
4.55859780e-01 6.89768434e-01 -4.44899321e-01 5.39831519e-01
1.49429232e-01 7.81360209e-01 1.51205674e-01 5.45358539e-01
2.09782243e-01 -9.03980434e-02 8.09703320e-02 -3.24290663e-01
-2.30393037e-01 2.69178212e-01 -1.37140250e+00 1.74577832e+00
-6.30770683e-01 1.15858424e+00 9.53426212e-02 -1.08922195e+00
5.81887603e-01 9.60248709e-01 7.87113726e-01 -1.00464034e+00
8.38617906e-02 1.01126343e-01 -1.98565140e-01 -7.33548999e-01
1.93617746e-01 -2.39690468e-01 1.13482960e-02 1.91342384e-01
3.83667469e-01 -2.85797954e-01 2.24211812e-01 2.93039739e-01
1.01824212e+00 -1.05311805e-02 3.27489197e-01 1.27418235e-01
4.23285663e-01 -5.07068217e-01 3.56901348e-01 3.63044053e-01
4.30503190e-02 1.01858950e+00 5.01704335e-01 -1.57075778e-01
-1.30509293e+00 -9.12170410e-01 -6.23475686e-02 9.89081740e-01
1.57775715e-01 -7.37780452e-01 -4.92825091e-01 7.97215402e-02
-3.66471559e-01 3.54457080e-01 -5.14186442e-01 -1.33465827e-01
-5.62959671e-01 -7.46154934e-02 5.42313337e-01 2.66843081e-01
-3.28048766e-02 -1.21599889e+00 -7.50746608e-01 3.08202565e-01
-6.34146690e-01 -1.27295983e+00 -5.93068600e-01 2.84129739e-01
-6.00908041e-01 -9.40321445e-01 -7.44451940e-01 -7.27516294e-01
2.70197153e-01 3.05715472e-01 1.13315964e+00 -1.35691404e-01
-2.39251435e-01 6.57078803e-01 -6.13835752e-01 -1.63565814e-01
-3.49977493e-01 -4.38718259e-01 6.18963465e-02 1.75774336e-01
-2.11195573e-02 -1.01032162e+00 -5.44579983e-01 3.93800318e-01
-9.49388981e-01 2.00717464e-01 3.08292300e-01 7.81600237e-01
4.23431158e-01 3.97487581e-02 4.09676880e-01 -4.72003520e-01
2.39141017e-01 -7.61291742e-01 -3.19781959e-01 -2.33406007e-01
1.62920326e-01 -4.00008619e-01 7.15517581e-01 -9.45275247e-01
-6.40538096e-01 6.81700185e-02 -9.39736366e-02 -9.91747379e-01
-5.71660465e-03 3.46908927e-01 -7.77355582e-02 3.56135182e-02
6.84389651e-01 1.65042788e-01 8.99599027e-03 -3.21975946e-01
1.51777148e-01 5.61861038e-01 7.71340668e-01 -2.37519830e-01
6.01777375e-01 4.86420512e-01 -2.01655880e-01 -1.25433791e+00
-4.30873185e-01 -6.96517050e-01 -4.90745366e-01 -7.77131796e-01
6.62393093e-01 -1.24517608e+00 -4.95891273e-01 3.30957115e-01
-1.01374722e+00 -3.96889508e-01 -5.57039261e-01 6.83801889e-01
-9.50833201e-01 3.15505147e-01 -5.14382005e-01 -6.88540816e-01
2.17546612e-01 -1.04519629e+00 1.05801690e+00 1.26856238e-01
-7.58962870e-01 -1.18251514e+00 2.63833225e-01 -1.22296497e-01
1.43939435e-01 3.82649750e-01 2.48675466e-01 -2.44934514e-01
-4.49690759e-01 -1.28439218e-01 1.82884380e-01 1.15462273e-01
-3.59563380e-02 1.14405036e-01 -1.31752181e+00 -3.69979739e-01
1.64532289e-01 -3.72637182e-01 3.96963805e-01 7.65162587e-01
9.29452956e-01 -4.43519056e-01 -2.75528342e-01 8.17998648e-01
1.20812249e+00 2.32488468e-01 8.79887104e-01 2.30391085e-01
6.38536692e-01 6.96489453e-01 6.57201588e-01 6.83957040e-01
2.62604021e-02 1.07835555e+00 2.43244335e-01 -1.23746708e-01
-3.90678644e-01 -4.71207172e-01 6.43729687e-01 1.38076115e+00
-1.79872274e-01 -1.06044665e-01 -6.88753486e-01 9.00780380e-01
-1.82389534e+00 -1.41355348e+00 -8.13851878e-03 2.13368082e+00
9.54379261e-01 1.73762683e-02 3.09142947e-01 5.34713030e-01
8.71299028e-01 4.14686710e-01 1.15238994e-01 -3.84565085e-01
-2.93840885e-01 2.09032431e-01 1.90476716e-01 5.66440463e-01
-8.65831316e-01 4.10798490e-01 6.56078053e+00 9.31176305e-01
-1.23577452e+00 1.60768390e-01 1.20847344e-01 -5.96155465e-01
-3.02067131e-01 1.11789070e-01 -3.11295360e-01 9.49298739e-01
1.49622774e+00 -4.87736315e-02 4.95260000e-01 3.28939825e-01
6.43414021e-01 -1.63566038e-01 -1.31323159e+00 1.35869277e+00
1.56268686e-01 -1.38072848e+00 -3.32879305e-01 -1.02973998e-01
6.13036633e-01 -2.80660182e-01 1.21420756e-01 -1.53192773e-01
-9.29733887e-02 -1.13877046e+00 1.21533918e+00 5.39714038e-01
1.15928578e+00 -6.94399834e-01 4.79650080e-01 1.58686474e-01
-1.54159296e+00 -4.89306413e-02 -1.05969325e-01 1.22693956e-01
4.83923197e-01 4.48309898e-01 -5.59350610e-01 5.22278786e-01
8.18275809e-01 1.20012891e+00 -2.73576021e-01 1.22129250e+00
-2.94232927e-02 6.65975630e-01 -4.54725236e-01 5.82376957e-01
1.76620007e-01 7.36645609e-02 7.46681273e-01 1.46960092e+00
6.33161902e-01 -1.35324031e-01 -8.84649009e-02 3.85566890e-01
3.44997525e-01 -1.95729241e-01 -1.04287601e+00 9.97316837e-02
6.54451132e-01 7.67042756e-01 -5.98963439e-01 -9.87397358e-02
-4.33805734e-01 9.03549194e-01 -3.13898087e-01 4.45033371e-01
-1.01325846e+00 -2.76467055e-01 2.73062706e-01 2.84871697e-01
5.28574586e-01 5.45530841e-02 1.98701292e-01 -1.08059120e+00
5.72720170e-02 -1.09370136e+00 3.66269231e-01 -1.12805486e+00
-8.68920326e-01 6.00272834e-01 1.19560510e-01 -1.89752889e+00
-8.05283546e-01 1.03041887e-01 -5.55361271e-01 7.22240746e-01
-1.08270776e+00 -1.05766320e+00 -1.38733402e-01 9.54588771e-01
6.97627008e-01 -3.06959391e-01 7.03026235e-01 6.32201791e-01
-8.92931148e-02 3.86634946e-01 8.78127217e-02 -3.26716959e-01
7.99253404e-01 -1.02448845e+00 -3.15705612e-02 7.88390636e-01
4.67894614e-01 1.80863753e-01 1.26494670e+00 -3.88563871e-01
-1.31937242e+00 -8.00809860e-01 1.03619266e+00 -2.72457451e-01
8.36100936e-01 -5.46539903e-01 -9.09906507e-01 5.76175094e-01
4.86667335e-01 1.97226137e-01 6.70222521e-01 -2.68945873e-01
-3.06159854e-01 -1.13681689e-01 -8.11755061e-01 3.30865085e-01
8.66821527e-01 -1.07209599e+00 -6.41443729e-01 2.82044113e-01
4.41220373e-01 -6.16441607e-01 -8.07901978e-01 -1.66770622e-01
7.11535990e-01 -1.13014650e+00 1.11941230e+00 -1.05904050e-01
6.10342741e-01 -3.57215226e-01 -2.04729989e-01 -1.27488828e+00
-8.82169828e-02 -1.39452755e+00 -3.79585713e-01 1.31449783e+00
-3.63588100e-03 -6.99948967e-02 2.74839640e-01 -2.31911123e-01
2.22384445e-02 -2.83397347e-01 -1.31941628e+00 -7.42005467e-01
-5.40403843e-01 -8.25233817e-01 1.34020835e-01 1.10080433e+00
2.39788443e-01 3.14616024e-01 -9.61735368e-01 -9.21901874e-03
3.33374053e-01 -1.77012473e-01 5.43869615e-01 -8.09678912e-01
-3.95430863e-01 -9.54042673e-02 -6.12951875e-01 -9.27145481e-01
5.28351124e-03 -5.77879548e-01 2.98853904e-01 -9.43760157e-01
-1.18331529e-01 -1.53735965e-01 -2.16132123e-02 9.29999799e-02
3.78227592e-01 4.24381435e-01 3.49915922e-01 4.38081294e-01
-6.65574551e-01 5.21727026e-01 1.10300350e+00 9.11095589e-02
-1.92971036e-01 -7.73685873e-02 -3.33901227e-01 5.06041646e-01
1.42607987e-01 -4.36052442e-01 -6.08465254e-01 -2.28031278e-02
3.04611087e-01 7.17852831e-01 5.87873340e-01 -1.12262082e+00
2.48504505e-01 6.31600246e-02 1.78558230e-01 -4.64357913e-01
5.85568607e-01 -1.07636905e+00 8.03595066e-01 1.34299383e-01
-4.38110679e-01 -5.49239339e-03 2.70714998e-01 6.18911386e-01
-7.62968421e-01 -3.39392461e-02 9.52428401e-01 -5.03809489e-02
-5.48986554e-01 -1.01684734e-01 -6.61535740e-01 1.94325581e-01
1.01929808e+00 -5.84583521e-01 1.87949911e-01 -8.72670770e-01
-8.26164067e-01 -1.44596398e-01 3.83039951e-01 3.63751918e-01
7.03118622e-01 -1.71776044e+00 -5.30771613e-01 2.15213358e-01
3.74220423e-02 -6.06090687e-02 5.49709201e-01 1.12720549e+00
-5.18339634e-01 3.51287663e-01 -3.81753027e-01 -1.02673197e+00
-1.37841201e+00 5.93170047e-01 7.75064230e-02 1.86748967e-01
-1.10511804e+00 8.58190060e-01 2.45618686e-01 4.58869278e-01
5.22022247e-01 -1.83132112e-01 -2.87292421e-01 4.21353251e-01
6.16938651e-01 1.17778748e-01 -6.44437447e-02 -1.18118489e+00
-2.84668475e-01 6.31712914e-01 2.91197062e-01 -3.21347952e-01
1.51854122e+00 -4.65925932e-01 2.08547592e-01 9.70236421e-01
1.47455239e+00 8.94621164e-02 -1.32789445e+00 -5.27969822e-02
-1.76520348e-01 -7.50674427e-01 -1.17247753e-01 -1.72776893e-01
-9.35238719e-01 8.61430168e-01 5.52757323e-01 5.38980961e-01
1.42354703e+00 2.26317853e-01 5.88838100e-01 -3.57055902e-01
3.56063899e-03 -1.19640541e+00 3.27600241e-01 3.50475609e-01
1.16862845e+00 -6.95764899e-01 -2.57068854e-02 -1.85816318e-01
-8.55421901e-01 1.02915657e+00 -9.85868182e-03 8.14432651e-03
4.51980948e-01 4.88160819e-01 -4.92934771e-02 -2.93007046e-01
-9.80650663e-01 -4.69483063e-02 3.73757660e-01 8.80857468e-01
3.69134367e-01 -3.46859604e-01 2.24988133e-01 1.46611303e-01
-4.26154643e-01 -1.24172188e-01 5.67692041e-01 8.52214456e-01
-1.49571225e-01 -6.83282375e-01 -6.55404091e-01 3.73447165e-02
-3.61965120e-01 6.60239384e-02 -1.89100876e-01 6.60337210e-01
1.38009280e-01 8.88289750e-01 1.36160940e-01 -4.19978529e-01
2.12213650e-01 -6.43658787e-02 4.36050862e-01 -4.23411727e-01
-4.60481703e-01 6.98863029e-01 1.70124024e-01 -9.13911819e-01
-8.25042129e-01 -8.51240098e-01 -8.28485131e-01 -3.15625757e-01
-1.10588253e-01 2.15575546e-01 2.62192965e-01 8.78904700e-01
5.15761301e-02 9.03935313e-01 8.14023614e-01 -1.37195349e+00
9.21973139e-02 -9.13814485e-01 -7.53578663e-01 6.20477915e-01
6.72703385e-01 -7.15246975e-01 -7.25658178e-01 7.46488333e-01] | [14.98043155670166, 5.116672992706299] |
e9737010-034c-4595-a767-9ec73c5e2ae5 | vip-cnn-visual-phrase-guided-convolutional | 1702.07191 | null | http://arxiv.org/abs/1702.07191v2 | http://arxiv.org/pdf/1702.07191v2.pdf | ViP-CNN: Visual Phrase Guided Convolutional Neural Network | As the intermediate level task connecting image captioning and object
detection, visual relationship detection started to catch researchers'
attention because of its descriptive power and clear structure. It detects the
objects and captures their pair-wise interactions with a
subject-predicate-object triplet, e.g. person-ride-horse. In this paper, each
visual relationship is considered as a phrase with three components. We
formulate the visual relationship detection as three inter-connected
recognition problems and propose a Visual Phrase guided Convolutional Neural
Network (ViP-CNN) to address them simultaneously. In ViP-CNN, we present a
Phrase-guided Message Passing Structure (PMPS) to establish the connection
among relationship components and help the model consider the three problems
jointly. Corresponding non-maximum suppression method and model training
strategy are also proposed. Experimental results show that our ViP-CNN
outperforms the state-of-art method both in speed and accuracy. We further
pretrain ViP-CNN on our cleansed Visual Genome Relationship dataset, which is
found to perform better than the pretraining on the ImageNet for this task. | ["Xiao'ou Tang", 'Yikang Li', 'Wanli Ouyang', 'Xiaogang Wang'] | 2017-02-23 | vip-cnn-visual-phrase-guided-convolutional-1 | http://openaccess.thecvf.com/content_cvpr_2017/html/Li_ViP-CNN_Visual_Phrase_CVPR_2017_paper.html | http://openaccess.thecvf.com/content_cvpr_2017/papers/Li_ViP-CNN_Visual_Phrase_CVPR_2017_paper.pdf | cvpr-2017-7 | ['visual-relationship-detection'] | ['computer-vision'] | [ 3.54354471e-01 1.09267242e-01 -3.26061368e-01 -3.82660091e-01
-4.09846395e-01 -1.85051173e-01 7.04760492e-01 1.39309049e-01
-3.69894266e-01 4.29654181e-01 1.29680768e-01 -1.98584080e-01
2.29601227e-02 -6.15998983e-01 -1.05461192e+00 -4.13285762e-01
-9.19195078e-03 3.48112255e-01 2.75612295e-01 -1.60064101e-01
8.20193961e-02 1.55043975e-01 -1.54666948e+00 6.40863121e-01
4.96127069e-01 8.14643085e-01 5.02504110e-01 5.91439664e-01
-1.81208253e-01 9.94078398e-01 -5.05848050e-01 -5.85818529e-01
-2.56105155e-01 -3.42409581e-01 -8.57453942e-01 3.58279377e-01
6.53356969e-01 -1.35298550e-01 -2.97208667e-01 1.18839765e+00
1.36092678e-01 -3.85939586e-03 5.82883775e-01 -1.81927299e+00
-1.28785527e+00 6.38081491e-01 -1.03038573e+00 3.11192095e-01
4.20454651e-01 -3.09487358e-02 1.16336453e+00 -1.13809454e+00
6.19332850e-01 1.53691173e+00 3.66602659e-01 3.83021533e-01
-1.12655902e+00 -7.13078260e-01 3.95570397e-01 5.91153145e-01
-1.48231959e+00 -4.45422903e-02 7.24629283e-01 -5.56871891e-01
1.19298697e+00 2.32037604e-01 7.99091637e-01 1.04057610e+00
-6.64377660e-02 9.55235362e-01 5.27144969e-01 -3.31986606e-01
-4.12838817e-01 3.10461372e-01 5.19717932e-01 7.96488762e-01
2.55880028e-01 1.21028777e-02 -6.59626126e-01 2.20909595e-01
8.46469820e-01 -1.22286424e-01 -2.80618787e-01 -4.09472823e-01
-1.47795486e+00 5.89988351e-01 8.69567752e-01 2.27208927e-01
-1.68311045e-01 4.95187819e-01 1.96824104e-01 1.35004846e-02
1.48104355e-01 1.82585165e-01 -1.23975262e-01 3.17359239e-01
-4.88442600e-01 2.68015027e-01 4.57251281e-01 1.48106647e+00
7.33659208e-01 -3.40118736e-01 -5.90627551e-01 8.43786895e-01
6.29167914e-01 3.04494888e-01 -3.88480201e-02 -4.81420666e-01
8.11927319e-01 8.08363199e-01 -2.16945130e-02 -1.58429134e+00
-4.05861944e-01 -5.16830623e-01 -1.05389977e+00 -3.96056384e-01
-1.83204487e-02 2.31611550e-01 -9.29730177e-01 1.64903307e+00
1.35920838e-01 3.17798376e-01 8.99000317e-02 1.21291006e+00
1.80276871e+00 1.01336098e+00 4.36601490e-01 -2.19429687e-01
2.03940821e+00 -1.43278205e+00 -9.56354260e-01 -4.97704327e-01
1.94658086e-01 -5.29770434e-01 8.30277264e-01 -1.20116383e-01
-9.92952108e-01 -1.00686109e+00 -1.02095902e+00 -4.33043897e-01
-5.16658485e-01 1.74313083e-01 7.17911839e-01 -7.96004832e-02
-8.95677209e-01 -7.54382983e-02 -4.10628885e-01 -6.50419474e-01
6.39304221e-01 3.49043459e-01 -5.66060901e-01 1.20153897e-01
-1.23463893e+00 7.80431092e-01 6.01013541e-01 3.65891486e-01
-8.29771698e-01 -3.92158151e-01 -1.01687551e+00 2.57162213e-01
3.51320237e-01 -9.49258268e-01 8.99077833e-01 -8.39381635e-01
-6.94306672e-01 1.33376265e+00 -2.84088880e-01 -3.88931215e-01
4.05925699e-03 -8.62040073e-02 -4.43633497e-01 2.60204583e-01
2.27459446e-01 1.32099223e+00 9.19706702e-01 -1.79246390e+00
-8.06021869e-01 5.65293059e-02 1.87644497e-01 4.00155723e-01
-3.26608568e-01 5.67618430e-01 -1.24959207e+00 -6.13772452e-01
1.53688371e-01 -6.74931884e-01 4.02555317e-02 1.58687770e-01
-6.75647259e-01 -5.23068488e-01 9.10273612e-01 -4.61934000e-01
9.78826344e-01 -2.09335065e+00 2.29780301e-01 -1.56657308e-01
5.30249834e-01 3.33760142e-01 -3.31756026e-01 2.87142962e-01
-4.54228073e-01 2.11812660e-01 -2.47287825e-01 -5.86707294e-01
-2.08893735e-02 3.84556502e-01 -2.61598140e-01 3.28304738e-01
6.38920903e-01 1.44302619e+00 -9.79509532e-01 -1.01339126e+00
4.02433090e-02 5.99801242e-01 -1.60778284e-01 5.18156290e-01
-2.43944094e-01 3.31681281e-01 -2.97359139e-01 6.87833965e-01
7.65032411e-01 -5.94882071e-01 -1.90712251e-02 -6.93239033e-01
1.02937765e-01 -7.06545562e-02 -7.31214285e-01 1.50312185e+00
-7.55903348e-02 9.47149634e-01 -1.09877042e-01 -1.08033133e+00
1.11045468e+00 5.64178750e-02 4.72897142e-02 -8.86692882e-01
1.07600674e-01 -4.12946582e-01 -2.13510562e-02 -1.07803476e+00
5.67142010e-01 1.82798475e-01 7.62432516e-02 -1.05106123e-01
6.32079095e-02 2.32098132e-01 1.26914099e-01 3.57689947e-01
5.67206979e-01 2.40588859e-01 3.78501117e-01 -1.33658767e-01
7.62906671e-01 1.12238042e-01 4.96799022e-01 6.39983296e-01
-2.38020375e-01 7.06234276e-01 6.49764717e-01 -4.13887292e-01
-6.80224895e-01 -8.68951082e-01 1.40307145e-03 9.61985469e-01
8.97784352e-01 -5.25398970e-01 -2.37872168e-01 -7.07180202e-01
-2.67644711e-02 4.76247817e-01 -6.61963224e-01 -7.19038099e-02
-7.86228120e-01 -6.06602967e-01 1.43349588e-01 6.51561916e-01
8.08205605e-01 -1.43959785e+00 -3.14616680e-01 -5.58300875e-02
-3.32851291e-01 -1.47053230e+00 -5.10983527e-01 -3.20795039e-03
-2.27278560e-01 -1.11073053e+00 -5.22468150e-01 -1.63111079e+00
9.22545254e-01 4.93269205e-01 1.34889245e+00 3.65449607e-01
-1.44770458e-01 2.58828580e-01 -3.87371093e-01 -4.66011852e-01
-1.19251966e-01 -1.82098597e-01 -4.12094802e-01 2.30616122e-01
6.18350863e-01 -3.19760233e-01 -4.86937314e-01 3.22302729e-01
-7.74083972e-01 5.85890114e-01 7.53899276e-01 8.69059086e-01
6.53625786e-01 -4.18998785e-02 1.17505960e-01 -5.12214303e-01
5.84490776e-01 -4.18299586e-01 -3.60000789e-01 5.37975669e-01
-6.99064657e-02 -1.24616884e-01 1.80647984e-01 -4.70564604e-01
-7.81274915e-01 1.84321344e-01 1.75143495e-01 -5.37942290e-01
-2.99525261e-01 4.98450547e-01 -3.19582283e-01 1.85696352e-02
1.84210330e-01 4.27014977e-01 -1.98567063e-01 -3.25844169e-01
6.02311492e-01 5.64209521e-01 8.73498261e-01 -2.54408717e-01
9.46466982e-01 2.76993871e-01 4.65492047e-02 -5.99355161e-01
-1.02342069e+00 -7.31266558e-01 -6.30047321e-01 -3.98471683e-01
1.30842412e+00 -1.05391347e+00 -1.08537149e+00 1.62243277e-01
-1.85656440e+00 2.48557866e-01 2.31500685e-01 1.76946342e-01
-3.11173528e-01 3.25900912e-01 -3.69953543e-01 -5.99025488e-01
-3.80567878e-01 -1.15825427e+00 1.25797236e+00 2.34296918e-01
1.57057792e-01 -8.02247167e-01 -2.67633587e-01 5.92161953e-01
4.52518044e-03 2.42827103e-01 8.14614832e-01 -5.29702961e-01
-7.19849467e-01 4.37210537e-02 -1.07764769e+00 1.52329803e-01
-1.18779195e-02 -5.49833067e-02 -7.38979578e-01 -1.61746457e-01
-3.02039683e-01 -2.55755544e-01 1.12822175e+00 5.94467893e-02
1.21326911e+00 -2.55780101e-01 -7.66173482e-01 6.65058792e-01
1.43373764e+00 1.14328228e-01 6.36199415e-01 4.69733179e-01
1.34846735e+00 7.92860389e-01 6.96221709e-01 -8.15778226e-02
7.44632542e-01 9.10476327e-01 9.31222737e-01 -5.32990873e-01
-4.98805255e-01 -3.64554018e-01 9.26536620e-02 8.36220503e-01
1.02651156e-01 -4.28899288e-01 -9.04859960e-01 7.32393503e-01
-2.37563920e+00 -1.02160573e+00 -7.90483236e-01 1.53158677e+00
7.23984301e-01 9.35809165e-02 7.85389021e-02 -3.52093071e-01
1.07158709e+00 1.58472687e-01 -9.90512446e-02 -2.72248656e-01
-1.69144243e-01 -3.60061467e-01 2.71431446e-01 3.16411734e-01
-1.46076250e+00 1.04245389e+00 5.96732950e+00 6.42169237e-01
-7.61918485e-01 3.09535731e-02 5.73962092e-01 4.34511155e-01
3.11468318e-02 -1.20189853e-01 -1.01143265e+00 2.94015795e-01
3.34307760e-01 8.59500095e-02 1.12118106e-02 7.45946884e-01
-5.81797808e-02 6.24831393e-02 -1.30820584e+00 1.51751018e+00
4.63648558e-01 -1.21469283e+00 5.42157352e-01 -2.02461630e-01
3.79426986e-01 -5.26406109e-01 1.41917020e-01 2.56122679e-01
-1.28869668e-01 -1.28638327e+00 8.43495429e-01 5.87853014e-01
4.97680932e-01 -6.00854993e-01 8.08387995e-01 1.86903983e-01
-1.62135637e+00 8.68954957e-02 -4.13974583e-01 -2.98381560e-02
2.17152476e-01 2.16310933e-01 -7.10443437e-01 6.52565479e-01
7.94705570e-01 1.13056636e+00 -6.36028469e-01 1.12265706e+00
-4.58671540e-01 2.84643859e-01 -3.14675551e-03 -1.55691132e-01
3.23140353e-01 -2.15891958e-03 6.24728799e-01 1.44331229e+00
-1.09670974e-01 1.10566005e-01 2.06607819e-01 1.33436716e+00
-1.19119376e-01 -5.12246117e-02 -5.54466724e-01 1.33808643e-01
3.56566191e-01 1.44097042e+00 -6.83990359e-01 -3.82361233e-01
-6.97153211e-01 9.17009950e-01 5.87500274e-01 3.33064288e-01
-1.15275300e+00 -1.86847225e-01 5.43165684e-01 -1.96106166e-01
5.53311348e-01 -1.81329012e-01 -1.07663486e-03 -1.01269543e+00
1.82180777e-01 -6.94849968e-01 3.40340286e-01 -1.15830851e+00
-1.41243613e+00 7.39242017e-01 2.88012922e-01 -1.34949076e+00
2.94729888e-01 -6.50073409e-01 -4.32046860e-01 7.41378486e-01
-1.83877194e+00 -1.72050464e+00 -6.04480386e-01 6.73752189e-01
5.94773889e-01 1.83360167e-02 5.05186081e-01 3.12795460e-01
-8.83287609e-01 6.71554625e-01 -8.03381383e-01 4.65489179e-01
4.00391012e-01 -9.99769092e-01 5.16884923e-01 1.02665114e+00
3.45126033e-01 6.58365667e-01 7.08888888e-01 -6.25925601e-01
-1.41336763e+00 -1.40731537e+00 1.13118696e+00 -4.53833550e-01
6.36488199e-01 -5.88295102e-01 -1.02559400e+00 7.92857528e-01
6.81990445e-01 1.37084290e-01 2.64492899e-01 -1.85976565e-01
-5.76367080e-01 -1.83767244e-01 -6.14397824e-01 6.19032085e-01
1.34068668e+00 -4.70677376e-01 -9.37813282e-01 5.28886020e-01
1.25432336e+00 -4.05368418e-01 -6.06385589e-01 4.55610335e-01
1.94222108e-01 -6.40186787e-01 1.30477202e+00 -5.95992684e-01
6.37739003e-01 -6.60123050e-01 -1.47175528e-02 -7.62873709e-01
-6.74363017e-01 -3.97388011e-01 -2.13156641e-01 1.53824317e+00
5.02970815e-01 -1.09298304e-01 5.74625850e-01 1.71357200e-01
1.14366021e-02 -5.92552662e-01 -5.90348661e-01 -7.00027764e-01
-5.09442687e-01 -3.90232116e-01 4.25101608e-01 9.90254700e-01
1.68387406e-02 8.84965658e-01 -6.89682543e-01 5.83656788e-01
6.81417882e-01 2.81256050e-01 7.60054290e-01 -8.45649183e-01
-2.98142672e-01 -3.92099798e-01 -5.66086292e-01 -1.38202453e+00
9.23583880e-02 -9.27699089e-01 2.84562230e-01 -1.91384661e+00
7.81570554e-01 -4.95012403e-02 -3.14982772e-01 8.51246834e-01
-5.78534842e-01 4.26957786e-01 4.17341739e-01 2.63865978e-01
-1.16287315e+00 5.27457952e-01 1.46540046e+00 -7.12136626e-01
-8.09797794e-02 -3.50835830e-01 -6.93742514e-01 4.62322980e-01
4.13624048e-01 -4.86841857e-01 -5.32026350e-01 -7.25461602e-01
3.92558604e-01 6.18192963e-02 8.23789001e-01 -8.72426152e-01
5.86910486e-01 -1.37818428e-02 1.75169289e-01 -1.04963911e+00
5.05099893e-01 -8.98658514e-01 7.95979947e-02 3.16991746e-01
-4.95141536e-01 1.86221838e-01 2.90751100e-01 6.12905383e-01
-4.65449005e-01 3.55802849e-02 4.47575152e-01 -1.72273777e-02
-1.14357662e+00 3.17685604e-01 -1.53480461e-02 -3.18188995e-01
1.20992899e+00 -2.83196092e-01 -6.32112443e-01 -2.51937449e-01
-7.23092020e-01 6.53018713e-01 -1.07699297e-01 7.33482599e-01
1.05440819e+00 -1.44862962e+00 -8.38732004e-01 2.30028760e-02
6.58214331e-01 2.89725550e-02 1.30496323e-01 6.95599496e-01
-4.13112402e-01 5.14564574e-01 -2.15628669e-01 -9.43284333e-01
-1.65247095e+00 1.09065795e+00 2.26501644e-01 3.17005776e-02
-6.61530256e-01 1.36483467e+00 5.84405839e-01 2.67473166e-03
6.90145433e-01 -3.78240675e-01 -8.96296501e-01 -1.79024576e-03
5.83034813e-01 -3.08377177e-01 -4.28484142e-01 -1.07206154e+00
-6.46506369e-01 9.25616503e-01 -1.70926675e-01 4.95613545e-01
1.22993255e+00 -1.64894328e-01 -6.12239540e-01 1.75795317e-01
1.45411897e+00 -6.34721816e-01 -9.85786915e-01 -3.24218601e-01
-9.51099321e-02 -3.30470026e-01 -8.91781375e-02 -5.54011166e-01
-1.19524479e+00 7.03964710e-01 5.08719027e-01 4.12338287e-01
1.06261098e+00 3.94639879e-01 6.36757195e-01 1.19168781e-01
-2.47289538e-02 -4.54315811e-01 3.76440108e-01 2.22639620e-01
1.32554924e+00 -1.40692472e+00 1.57043248e-01 -1.09142089e+00
-6.94844961e-01 9.39978182e-01 1.00668824e+00 4.74971123e-02
5.78889966e-01 -2.56908871e-02 -1.38420850e-01 -6.02807164e-01
-9.58598256e-01 -5.76659620e-01 7.51415908e-01 9.08025622e-01
1.20324843e-01 -1.14363238e-01 -2.86766201e-01 5.96571743e-01
1.43036202e-01 -2.24473938e-01 7.55824596e-02 6.16423845e-01
-3.31652582e-01 -8.87368023e-01 -3.83200198e-01 1.03003472e-01
-9.55239534e-02 -4.15716738e-01 -4.93815184e-01 9.53358591e-01
4.65745658e-01 1.11304438e+00 4.54827309e-01 -4.96755540e-01
5.31482458e-01 -5.07246196e-01 2.38720357e-01 -6.07926786e-01
-5.93733847e-01 -1.04055949e-01 2.54236698e-01 -5.83646297e-01
-1.01361752e+00 -2.57984728e-01 -1.15157330e+00 -1.38417911e-02
-3.08867991e-01 -1.68509111e-01 3.94907475e-01 9.04288888e-01
2.84941256e-01 7.97041714e-01 3.54335457e-01 -6.75319433e-01
1.97400987e-01 -7.83109307e-01 -1.98037177e-01 7.23601997e-01
4.24565166e-01 -6.99792862e-01 4.37274054e-02 3.04002613e-01] | [10.373311996459961, 1.5671747922897339] |
23963f3c-4368-4fb4-ad3a-22a076e57091 | online-illumination-invariant-moving-object | 1808.01066 | null | http://arxiv.org/abs/1808.01066v1 | http://arxiv.org/pdf/1808.01066v1.pdf | Online Illumination Invariant Moving Object Detection by Generative Neural Network | Moving object detection (MOD) is a significant problem in computer vision
that has many real world applications. Different categories of methods have
been proposed to solve MOD. One of the challenges is to separate moving objects
from illumination changes and shadows that are present in most real world
videos. State-of-the-art methods that can handle illumination changes and
shadows work in a batch mode; thus, these methods are not suitable for long
video sequences or real-time applications. In this paper, we propose an
extension of a state-of-the-art batch MOD method (ILISD) to an
online/incremental MOD using unsupervised and generative neural networks, which
use illumination invariant image representations. For each image in a sequence,
we use a low-dimensional representation of a background image by a neural
network and then based on the illumination invariant representation, decompose
the foreground image into: illumination change and moving objects. Optimization
is performed by stochastic gradient descent in an end-to-end and unsupervised
fashion. Our algorithm can work in both batch and online modes. In the batch
mode, like other batch methods, optimizer uses all the images. In online mode,
images can be incrementally fed into the optimizer. Based on our experimental
evaluation on benchmark image sequences, both the online and the batch modes of
our algorithm achieve state-of-the-art accuracy on most data sets. | ['Nilanjan Ray', 'Moein Shakeri', 'Fateme Bahri'] | 2018-08-03 | null | null | null | null | ['moving-object-detection'] | ['computer-vision'] | [ 4.00808424e-01 -7.63469160e-01 5.90007231e-02 -4.62689400e-01
-3.59169513e-01 -3.36746424e-01 4.24574494e-01 -5.48554301e-01
-5.01969635e-01 3.18520993e-01 -3.15627754e-01 -1.18543148e-01
2.56271124e-01 -4.66854602e-01 -7.10867822e-01 -1.03541887e+00
1.08361848e-01 2.51666009e-01 8.62095058e-01 1.40837029e-01
1.11398824e-01 4.40033913e-01 -1.59469450e+00 3.67060781e-01
6.83176816e-01 1.05676758e+00 5.67414999e-01 9.57296252e-01
-3.23455244e-01 1.01824868e+00 -5.37704110e-01 -8.04969575e-03
5.84652781e-01 -6.23704314e-01 -4.17640209e-01 6.41737700e-01
6.61454678e-01 -4.75045800e-01 -4.40037578e-01 1.06696045e+00
6.66925371e-01 5.07371724e-01 3.70841801e-01 -1.22331738e+00
-4.35847163e-01 1.32557645e-01 -7.06351519e-01 5.85361063e-01
-4.03520390e-02 2.45673478e-01 4.43867922e-01 -1.06476426e+00
4.60181326e-01 1.36039793e+00 4.52316284e-01 4.66254592e-01
-1.13329828e+00 -4.00306940e-01 7.14479625e-01 6.18560016e-01
-1.09339213e+00 -4.47071761e-01 8.38302791e-01 -4.99631047e-01
7.86985457e-01 5.79281338e-02 6.31308436e-01 6.14157259e-01
2.46110708e-01 1.18117213e+00 1.00862515e+00 -3.41699898e-01
3.59018326e-01 -1.45991623e-01 1.20382458e-01 8.41004431e-01
5.02609946e-02 -6.90202788e-02 -4.53397095e-01 1.91184387e-01
5.71946919e-01 5.11474252e-01 -3.10452372e-01 -3.77334535e-01
-1.07191193e+00 5.28873444e-01 3.33288193e-01 4.33828086e-02
-4.65312272e-01 1.68243706e-01 2.46075287e-01 1.18286610e-01
5.19929528e-01 -3.23912084e-01 -4.39469635e-01 -1.41819939e-02
-1.16388047e+00 7.60993659e-02 7.42568791e-01 5.94943702e-01
7.41317034e-01 1.69565842e-01 -1.40440017e-01 8.50148201e-01
3.65588337e-01 7.01551557e-01 6.82312906e-01 -7.30897307e-01
2.34954283e-01 5.10925949e-01 4.64308187e-02 -1.01939452e+00
-2.86648005e-01 -2.76891649e-01 -8.79024804e-01 3.99885118e-01
2.75498599e-01 -2.70821843e-02 -1.26246154e+00 1.36340487e+00
7.35301614e-01 5.49607992e-01 5.76344579e-02 1.02745283e+00
8.85864615e-01 1.00615549e+00 -2.41009831e-01 -5.75170398e-01
1.19042432e+00 -1.57996833e+00 -9.05172050e-01 -5.51120102e-01
2.65744049e-02 -9.46711600e-01 7.41016865e-01 6.10245705e-01
-8.98903251e-01 -9.29493666e-01 -9.65335131e-01 1.37704119e-01
-3.06700170e-01 2.49483481e-01 2.95433611e-01 7.17112243e-01
-7.88962007e-01 5.69681168e-01 -1.14891481e+00 -2.51407504e-01
2.69483268e-01 2.43213296e-01 3.08546685e-02 -2.57447839e-01
-7.49662936e-01 5.72997510e-01 4.72844183e-01 5.13069332e-01
-1.21843994e+00 -2.29545265e-01 -8.27587187e-01 -2.73248881e-01
7.39589214e-01 -3.88845205e-01 1.21220541e+00 -1.47153544e+00
-1.75136733e+00 7.09737957e-01 -5.72536230e-01 -4.44211066e-01
4.65654761e-01 -5.90120077e-01 -3.52513909e-01 9.22780707e-02
-1.47945493e-01 2.28115022e-01 1.33981192e+00 -1.23416710e+00
-8.13533187e-01 -1.61121756e-01 -1.05794653e-01 1.62568733e-01
-2.69981861e-01 2.81960279e-01 -1.06000507e+00 -6.18599415e-01
1.52518600e-01 -1.10136139e+00 -1.38170332e-01 1.62265673e-01
-8.25847387e-02 5.33391088e-02 1.45623124e+00 -7.56877661e-01
1.23014176e+00 -2.11527777e+00 2.74358839e-01 -2.12744877e-01
-1.35483548e-01 7.11839437e-01 -5.41469902e-02 3.18387002e-02
1.26182944e-01 -5.28161466e-01 -3.98032039e-01 -6.45229638e-01
-2.13637203e-01 4.98838753e-01 -1.22644357e-01 7.45341241e-01
6.44535795e-02 6.04047477e-01 -8.23278964e-01 -5.87140799e-01
5.14397025e-01 3.18235457e-01 -3.53189826e-01 5.58650732e-01
-2.35160768e-01 2.82020032e-01 -1.01558492e-01 6.95904315e-01
9.57902670e-01 -2.56457418e-01 1.52777717e-01 -2.12284893e-01
-2.10156590e-01 -2.54987836e-01 -1.66137278e+00 1.36072981e+00
-2.87556589e-01 8.06198597e-01 2.39566058e-01 -1.04677510e+00
5.35550117e-01 3.34308594e-02 3.19374144e-01 -3.93152505e-01
5.04862741e-02 1.15885697e-01 -8.73689950e-02 -8.04301798e-01
3.82015288e-01 8.07306170e-02 5.48946261e-01 3.07510436e-01
-1.01370336e-02 2.16983274e-01 5.45790076e-01 -4.48748581e-02
8.46127033e-01 4.29284990e-01 1.64995730e-01 1.07192777e-01
8.40720177e-01 -2.20246106e-01 9.64667439e-01 8.15873265e-01
-2.48567462e-01 7.99161315e-01 -5.77756055e-02 -7.35773623e-01
-5.08005679e-01 -8.72612774e-01 2.16541544e-01 1.32316053e+00
4.42995518e-01 -1.76535949e-01 -7.85927117e-01 -7.09625483e-01
-2.44425490e-01 3.86967331e-01 -3.78010541e-01 2.29037199e-02
-8.46884131e-01 -1.10866630e+00 5.47678359e-02 5.41270137e-01
7.75353730e-01 -1.11431503e+00 -9.95234966e-01 4.32008654e-01
-2.65608221e-01 -1.35911667e+00 -6.54257536e-01 1.26741678e-01
-7.99620688e-01 -1.04285967e+00 -7.41414070e-01 -8.06289554e-01
8.49300325e-01 8.41518283e-01 8.76845300e-01 -6.23993352e-02
-5.73553741e-01 3.03015888e-01 -4.37558502e-01 -4.98842865e-01
-1.43497974e-01 -3.86445314e-01 6.64736778e-02 7.34681904e-01
1.44924849e-01 -1.94990650e-01 -8.98983061e-01 5.79172134e-01
-1.23008108e+00 6.02415912e-02 5.90833485e-01 7.45569468e-01
7.92204261e-01 2.68963009e-01 -1.04846440e-01 -7.78824449e-01
-8.85314792e-02 -2.87515670e-01 -1.06589699e+00 1.87422156e-01
-4.83914375e-01 -5.38689718e-02 5.75129151e-01 -7.10614741e-01
-1.35857236e+00 4.62968349e-01 -1.30273681e-02 -6.60958827e-01
-2.55500320e-02 2.25267261e-01 -2.82243192e-01 -1.22395813e-01
3.04161668e-01 5.40243208e-01 -1.53046682e-01 -6.35553479e-01
2.15869278e-01 6.47612989e-01 6.33791804e-01 -2.08639920e-01
8.76906812e-01 8.63098502e-01 -1.04919031e-01 -9.66443837e-01
-9.88648772e-01 -9.08273399e-01 -6.38728380e-01 -4.97063190e-01
1.08834088e+00 -7.38883257e-01 -2.65678912e-01 1.14518726e+00
-1.02694893e+00 -5.59835553e-01 8.82690698e-02 4.46029842e-01
-3.31162721e-01 5.34860611e-01 -4.27704096e-01 -1.08009994e+00
-3.50923657e-01 -1.27868032e+00 1.02199984e+00 6.35309041e-01
5.08022785e-01 -9.13091540e-01 3.78470495e-02 3.59943539e-01
4.96890843e-01 2.09888890e-01 3.45330864e-01 -3.74857336e-01
-8.15494478e-01 -1.48852259e-01 -1.19931422e-01 6.29969954e-01
4.45909679e-01 1.24024563e-01 -9.46604967e-01 -4.08143282e-01
2.50392884e-01 6.86230734e-02 1.13133931e+00 5.83286583e-01
1.16194952e+00 -2.21144721e-01 -2.43997782e-01 9.31678474e-01
1.59234130e+00 5.38313270e-01 5.82746863e-01 4.61519152e-01
8.48198116e-01 1.40621215e-01 7.93125987e-01 4.80470061e-01
2.11775228e-01 6.58826232e-01 4.42063600e-01 -2.29805976e-01
-1.87255517e-01 2.21270695e-01 7.29879022e-01 6.43119097e-01
-1.34517699e-01 -3.34384739e-01 -7.83635199e-01 4.51063901e-01
-2.27441001e+00 -1.10964060e+00 -1.64841801e-01 2.34710550e+00
5.98190963e-01 1.63542345e-01 4.21623923e-02 -9.76663381e-02
8.46712291e-01 5.00765800e-01 -6.11432016e-01 2.83341967e-02
-1.26241997e-01 -1.70993388e-01 4.60123837e-01 2.09410444e-01
-1.45690942e+00 8.98147821e-01 5.69865561e+00 5.65957010e-01
-1.50553823e+00 2.23382652e-01 6.34300947e-01 -3.07078034e-01
5.94211400e-01 5.19708404e-03 -9.84073102e-01 6.13947928e-01
6.79894209e-01 3.07101190e-01 6.69066906e-01 8.93684030e-01
3.88677806e-01 -3.27241004e-01 -8.87321174e-01 1.19144726e+00
4.84689176e-01 -1.06956840e+00 -2.92280823e-01 -2.55983800e-01
1.01780939e+00 2.18149140e-01 -2.08369583e-01 2.35635281e-01
-4.97604012e-02 -5.92190862e-01 8.05845320e-01 4.95130152e-01
3.36460203e-01 -4.34989899e-01 8.66461396e-01 4.74928170e-01
-1.34202158e+00 -2.06005424e-01 -4.42481488e-01 9.03125703e-02
4.56446826e-01 7.96098411e-01 -3.66634995e-01 4.03545886e-01
9.80423868e-01 7.02460170e-01 -5.16836941e-01 1.16257119e+00
-2.30540842e-01 6.90574348e-01 -3.95049989e-01 1.55544609e-01
2.75665194e-01 -3.98145616e-01 6.47915184e-01 1.51987910e+00
1.91123579e-02 1.45857304e-01 6.54737532e-01 3.66485059e-01
3.97017635e-02 -5.03985994e-02 -1.69628307e-01 -1.02276662e-02
7.32105831e-03 1.37131429e+00 -1.01135767e+00 -5.66897392e-01
-6.08188212e-01 1.37455189e+00 -5.70040867e-02 4.73567486e-01
-8.81731331e-01 -3.51525486e-01 3.93663406e-01 -1.63824305e-01
9.31309760e-01 -4.45112437e-01 2.85929412e-01 -1.28229809e+00
3.12711418e-01 -9.45805192e-01 4.39933777e-01 -5.80885291e-01
-1.07594526e+00 3.34529579e-01 6.92596519e-03 -1.15587664e+00
-1.30970463e-01 -8.37566793e-01 -8.29807222e-01 4.27023023e-01
-1.65379679e+00 -9.68501270e-01 -7.64000237e-01 6.06273293e-01
1.18413150e+00 -4.56924029e-02 3.38063627e-01 3.61547858e-01
-9.66191113e-01 2.47319058e-01 4.58978444e-01 2.22898662e-01
7.89673448e-01 -1.24282324e+00 2.37313882e-01 1.35141063e+00
2.33192578e-01 3.60352546e-01 6.08109176e-01 -5.90015411e-01
-1.76981735e+00 -1.36112118e+00 3.59423876e-01 -3.10186371e-02
3.14502954e-01 -2.79487848e-01 -9.41844463e-01 5.03101289e-01
2.20016152e-01 5.11225104e-01 4.06949311e-01 -3.17041963e-01
-2.79080961e-02 -4.71743912e-01 -8.84423018e-01 4.03845102e-01
9.69630420e-01 -7.61893913e-02 -4.54600781e-01 5.64511180e-01
4.00944650e-01 -7.62918472e-01 -2.39622653e-01 1.68397501e-01
4.32461917e-01 -1.02700448e+00 8.76511037e-01 -3.46334398e-01
1.87753692e-01 -8.59650850e-01 -1.58838943e-01 -1.11307108e+00
-3.14806819e-01 -8.37351203e-01 -5.46777725e-01 1.10962307e+00
-1.08347282e-01 -6.57907307e-01 4.85264719e-01 3.49743396e-01
1.91154499e-02 -7.47992992e-01 -7.33367205e-01 -8.53472769e-01
-5.94071031e-01 -2.65972018e-01 1.86760888e-01 3.93208593e-01
-9.34166133e-01 9.44905430e-02 -5.83572745e-01 3.97529423e-01
8.94449890e-01 5.19211054e-01 9.89432573e-01 -8.45451415e-01
-6.36327088e-01 -2.15667799e-01 -4.12533313e-01 -1.20016944e+00
1.73082389e-02 -4.02523398e-01 5.04584849e-01 -1.62707078e+00
3.62058878e-01 3.65000740e-02 -4.21820432e-01 3.53501678e-01
-5.11845648e-01 3.63535881e-01 4.13625926e-01 3.04267287e-01
-9.36204076e-01 5.26443720e-01 6.77032232e-01 -3.03509951e-01
-3.50176781e-01 4.24495377e-02 -5.52226007e-02 1.09786534e+00
5.32478392e-01 -4.93829370e-01 -2.88879544e-01 -6.44114554e-01
-2.36725241e-01 -3.70935798e-01 3.02752912e-01 -1.01613200e+00
2.73597062e-01 -5.04864812e-01 5.47048986e-01 -8.35157871e-01
3.33322644e-01 -7.78468311e-01 4.43403199e-02 5.19152880e-01
1.57454863e-01 -3.18446122e-02 1.93172157e-01 6.83359265e-01
-2.90179372e-01 -3.00971776e-01 1.11918569e+00 -1.97331458e-01
-1.09873819e+00 4.04958993e-01 -5.12413979e-01 -4.19676583e-03
1.09142733e+00 -3.22804391e-01 -3.67169492e-02 -2.67107964e-01
-4.57826555e-01 2.55564898e-01 3.33742708e-01 5.09326994e-01
7.01033115e-01 -1.04646242e+00 -7.03657925e-01 1.18274555e-01
-1.48775712e-01 1.78820133e-01 3.30004752e-01 1.01049709e+00
-6.01758122e-01 4.28858027e-02 6.21238053e-02 -8.97456706e-01
-1.65693283e+00 6.94786251e-01 4.35524702e-01 -1.44221306e-01
-6.34172201e-01 8.52786720e-01 5.26699364e-01 -2.59248465e-02
3.67908269e-01 -9.21030343e-02 -1.49597740e-02 -7.28699937e-02
1.04321778e+00 4.52020824e-01 8.86602327e-02 -6.88614249e-01
-4.48015392e-01 7.12066770e-01 -9.48892608e-02 -1.17944693e-02
1.46853435e+00 -2.15477556e-01 -2.00730652e-01 5.82733393e-01
1.18861043e+00 -1.41992196e-01 -1.52551305e+00 -4.82423842e-01
-3.96718830e-01 -8.34389806e-01 2.08330587e-01 -4.55976129e-01
-1.38660395e+00 8.12087417e-01 1.09730470e+00 1.17144007e-02
1.53005016e+00 -3.13243926e-01 8.17316294e-01 4.92027551e-01
1.01356685e-01 -1.30393791e+00 1.57515004e-01 6.15921736e-01
6.63881898e-01 -1.41646242e+00 2.20548511e-01 -2.26055920e-01
-5.27338862e-01 1.24920857e+00 5.41460693e-01 -1.39775695e-02
5.45731783e-01 2.77022302e-01 3.07514101e-01 2.21151099e-01
-6.11082613e-01 -3.83083612e-01 3.59623492e-01 2.95705676e-01
2.05331236e-01 -2.23565444e-01 -7.32661113e-02 1.95569173e-01
5.15090883e-01 1.57199539e-02 2.54340857e-01 1.33059144e+00
-6.39908850e-01 -9.26356554e-01 -7.84430802e-01 2.86753535e-01
-5.52172244e-01 9.91135463e-02 -2.23547414e-01 2.51206487e-01
3.70787978e-01 1.00795102e+00 -1.75189972e-02 -9.66219753e-02
2.14873493e-01 -1.95146631e-03 3.68062913e-01 -3.85956705e-01
-4.23842877e-01 3.39097649e-01 -2.83152312e-01 -7.40456402e-01
-6.50652409e-01 -8.54798913e-01 -1.29657662e+00 1.47214293e-01
-5.17884612e-01 -3.44411790e-01 8.20783257e-01 1.01861441e+00
2.70223498e-01 6.36298120e-01 7.06015289e-01 -1.38308012e+00
-4.09965426e-01 -7.83993602e-01 -2.62085080e-01 5.35119772e-01
5.10992169e-01 -6.63100839e-01 -3.46005619e-01 6.28357291e-01] | [8.996256828308105, -0.6542775630950928] |
df7a856c-ff33-4fcd-907b-3fd3538b263c | gravitational-dimensionality-reduction-using | 2211.01369 | null | https://arxiv.org/abs/2211.01369v1 | https://arxiv.org/pdf/2211.01369v1.pdf | Gravitational Dimensionality Reduction Using Newtonian Gravity and Einstein's General Relativity | Due to the effectiveness of using machine learning in physics, it has been widely received increased attention in the literature. However, the notion of applying physics in machine learning has not been given much awareness to. This work is a hybrid of physics and machine learning where concepts of physics are used in machine learning. We propose the supervised Gravitational Dimensionality Reduction (GDR) algorithm where the data points of every class are moved to each other for reduction of intra-class variances and better separation of classes. For every data point, the other points are considered to be gravitational particles, such as stars, where the point is attracted to the points of its class by gravity. The data points are first projected onto a spacetime manifold using principal component analysis. We propose two variants of GDR -- one with the Newtonian gravity and one with the Einstein's general relativity. The former uses Newtonian gravity in a straight line between points but the latter moves data points along the geodesics of spacetime manifold. For GDR with relativity gravitation, we use both Schwarzschild and Minkowski metric tensors to cover both general relativity and special relativity. Our simulations show the effectiveness of GDR in discrimination of classes. | ['Smriti Sharma', 'Benyamin Ghojogh'] | 2022-10-30 | null | null | null | null | ['supervised-dimensionality-reduction'] | ['computer-vision'] | [-3.21471542e-01 -2.48290692e-02 2.01237991e-01 -3.08757931e-01
2.47305512e-01 -4.78455812e-01 7.65045404e-01 -3.30222607e-01
-3.29335213e-01 4.28765267e-01 3.60407010e-02 -3.15630943e-01
-5.87944090e-01 -1.10089362e+00 -2.48571157e-01 -1.15939069e+00
-2.48186648e-01 4.73685056e-01 3.93794924e-01 -4.30486798e-01
5.93388259e-01 6.33446336e-01 -1.47670352e+00 -2.56392717e-01
7.75056601e-01 4.82218355e-01 2.34982312e-01 8.00407529e-01
-8.24544057e-02 1.56656981e-01 -2.76892751e-01 -2.25438789e-01
7.24634349e-01 -5.15827358e-01 -8.25126350e-01 -1.51690260e-01
2.20392575e-03 2.69827813e-01 -7.06864476e-01 1.33551478e+00
3.55603248e-01 6.36824489e-01 7.88668215e-01 -1.08885634e+00
-1.13092053e+00 2.65285075e-01 -8.60399961e-01 4.82822895e-01
2.11607322e-01 -2.46690422e-01 9.23061252e-01 -5.81121504e-01
1.75344735e-01 1.60182166e+00 9.89996567e-02 7.78304040e-01
-7.14436591e-01 -2.98290610e-01 -3.81837070e-01 4.30015981e-01
-1.16655338e+00 2.27005258e-01 9.86142933e-01 -6.06456101e-01
5.59407890e-01 4.33121204e-01 4.90500391e-01 3.43167722e-01
2.86354870e-01 3.17847162e-01 1.04923344e+00 -6.96901023e-01
5.91405630e-01 1.09156027e-01 6.15430176e-01 4.35886413e-01
3.97969514e-01 3.47625941e-01 -1.71436951e-01 -6.17963597e-02
5.33569157e-01 6.75632153e-04 -1.47497609e-01 -3.62029284e-01
-1.23087120e+00 1.08359981e+00 5.44145644e-01 5.80926239e-01
-2.40569375e-02 -1.62637338e-01 -1.41251326e-01 -7.77653307e-02
2.92549551e-01 5.96231461e-01 -2.62096792e-01 1.29174277e-01
-3.63414109e-01 3.97802234e-01 6.18548512e-01 5.68461239e-01
7.72394121e-01 -1.56190917e-01 4.89505291e-01 5.93442380e-01
4.93766069e-01 7.39911377e-01 9.25452113e-01 -6.68180346e-01
2.79539645e-01 9.36866522e-01 5.74797764e-03 -1.39199638e+00
-6.79464340e-01 -1.91485524e-01 -6.52723014e-01 4.12437797e-01
3.65508288e-01 -1.20162286e-01 -4.46349651e-01 1.46245646e+00
6.29846394e-01 5.38836956e-01 5.09510458e-01 1.34486938e+00
8.37051451e-01 6.89600348e-01 -2.97162354e-01 -2.99405549e-02
1.52041543e+00 -5.21340013e-01 -2.59208649e-01 7.49712810e-02
7.90830910e-01 -6.10912204e-01 8.55378389e-01 3.44459325e-01
-8.15270483e-01 -5.09748876e-01 -9.78337228e-01 1.65696852e-02
-6.25670493e-01 -2.40725446e-02 7.30597615e-01 8.46732140e-01
-8.23916614e-01 8.03344309e-01 -8.54575217e-01 -5.71187258e-01
-1.53953984e-01 5.26829123e-01 -3.02809596e-01 5.55308104e-01
-1.16388237e+00 8.80229175e-01 5.99841535e-01 -1.87530518e-01
-1.85361177e-01 -3.00641179e-01 -3.76274675e-01 -2.26107091e-01
-3.13606799e-01 -4.85609591e-01 6.44301653e-01 -3.40252787e-01
-1.46945357e+00 8.96374702e-01 1.46217182e-01 -3.08437467e-01
1.69803966e-02 -5.93056306e-02 -8.40323567e-01 -2.23121233e-02
-1.60090461e-01 2.16110885e-01 5.38765073e-01 -8.73656452e-01
-7.21582115e-01 -7.98569739e-01 2.74419665e-01 3.05421680e-01
-3.85662735e-01 1.60249658e-02 2.75593009e-02 -3.63108516e-01
9.81820762e-01 -1.35629690e+00 -3.77979991e-03 -6.02890372e-01
-2.82633096e-01 -7.21127212e-01 1.24366713e+00 -4.25000548e-01
9.30887341e-01 -2.28898096e+00 5.40021360e-01 1.07376963e-01
3.67395222e-01 3.05606782e-01 2.34736562e-01 -5.41237816e-02
-3.64971697e-01 -6.71748668e-02 -4.02882174e-02 2.01896548e-01
2.55822856e-02 2.68340975e-01 -2.22672075e-01 8.53127301e-01
-2.53264576e-01 5.09474933e-01 -1.10011446e+00 -2.45502710e-01
4.60176319e-01 1.54109389e-01 -5.27410328e-01 3.92642580e-02
8.74140859e-02 8.08961868e-01 -7.68262684e-01 7.20796138e-02
1.05454612e+00 1.01909608e-01 -4.07993346e-01 -2.59486824e-01
-3.13743353e-01 7.72091821e-02 -1.34725094e+00 1.44262016e+00
-5.08958846e-02 3.00848186e-01 -4.16656315e-01 -1.20921361e+00
1.31277406e+00 -8.58276337e-02 4.15804327e-01 -5.26588798e-01
2.80634969e-01 1.92042008e-01 4.81603175e-01 -7.39694953e-01
5.16112685e-01 -3.06673348e-01 2.80299783e-01 4.31220651e-01
-1.29656509e-01 -3.58327448e-01 2.15776548e-01 1.16152585e-01
7.96432734e-01 2.55539976e-02 1.07742541e-01 -6.88106358e-01
9.27473426e-01 -3.09638768e-01 4.53264773e-01 2.50306696e-01
-9.56655368e-02 4.99337286e-01 3.73711996e-03 -7.34317958e-01
-9.35699463e-01 -1.12232864e+00 -4.71771926e-01 9.56236959e-01
6.19797170e-01 -7.05284402e-02 -7.16704071e-01 -5.47960579e-01
7.43478388e-02 9.92567718e-01 -4.23611104e-01 -5.19897580e-01
-4.87618417e-01 -1.40363848e+00 -4.23445255e-02 -6.84596077e-02
7.52285004e-01 -8.91077042e-01 -6.31930470e-01 -3.19060147e-01
1.43987373e-01 -6.39120221e-01 2.10803971e-01 -2.21115232e-01
-1.00341642e+00 -1.09530020e+00 -5.46470344e-01 -3.52959037e-01
5.68936169e-01 4.62842822e-01 7.36417770e-01 -3.65899242e-02
-3.39982152e-01 3.19995493e-01 -6.84058189e-01 -4.94509459e-01
-3.77725363e-01 -1.39084190e-01 7.85602808e-01 1.68170288e-01
1.09950483e+00 -6.19833887e-01 -7.19147563e-01 5.78932047e-01
-6.57748520e-01 -2.43707493e-01 3.13932747e-01 -5.10405144e-03
-1.27721503e-01 4.97930080e-01 1.41321421e-01 -2.07220614e-01
2.39516780e-01 -6.69519007e-01 -5.74645877e-01 -2.32643005e-03
-2.64830858e-01 2.74327457e-01 3.11382830e-01 -1.73798114e-01
-8.53381097e-01 -4.08869296e-01 3.58622611e-01 -8.83470252e-02
-1.44340828e-01 2.20419951e-02 -9.31428224e-02 -4.30531114e-01
8.95112455e-01 6.47975877e-02 -2.61135250e-01 -7.39797533e-01
4.34715420e-01 9.34445083e-01 6.02960289e-01 -7.41119087e-01
1.09853959e+00 5.69321096e-01 6.00809455e-01 -1.04478633e+00
-7.34854817e-01 -5.22768676e-01 -9.77613032e-01 -2.30973631e-01
1.20886755e+00 -3.68058234e-01 -8.05759490e-01 2.21163318e-01
-9.05175507e-01 5.37308156e-01 -1.76495999e-01 1.35581768e+00
-3.41194987e-01 7.28430271e-01 -1.61437452e-01 -9.99013901e-01
-3.26892510e-02 -9.65294659e-01 7.40948737e-01 7.14504659e-01
2.07311541e-01 -1.31727207e+00 4.42834854e-01 1.95575893e-01
3.03741284e-02 8.38034377e-02 9.41306949e-01 -8.89315188e-01
-3.20734441e-01 -1.93843171e-01 -3.67721431e-02 1.75508872e-01
4.74405676e-01 -1.13205034e-02 -8.70019495e-01 -4.43585187e-01
7.61549413e-01 2.89103448e-01 5.43352544e-01 2.86930919e-01
8.54088664e-01 -6.32482693e-02 -4.32560056e-01 5.93822956e-01
1.25538266e+00 6.13026738e-01 6.33725882e-01 5.35567403e-01
8.28310132e-01 6.78413749e-01 4.02287066e-01 2.26998821e-01
1.22458518e-01 6.46239758e-01 4.95479643e-01 1.02097861e-01
1.28018886e-01 4.15119439e-01 1.26459539e-01 1.20570052e+00
-5.68340182e-01 1.48228526e-01 -9.53031838e-01 2.11888641e-01
-1.71747506e+00 -8.60464871e-01 -5.71960211e-01 2.57064700e+00
3.20933998e-01 -3.71442400e-02 5.82951009e-02 1.03378989e-01
1.01609981e+00 -1.48553178e-01 -3.10097337e-01 -3.69213611e-01
-2.11689204e-01 -1.85315728e-01 4.90844756e-01 4.55269694e-01
-1.08521831e+00 6.00276709e-01 5.76003790e+00 5.99597812e-01
-1.13496828e+00 -5.68504035e-02 2.06440151e-01 -1.72216278e-02
2.20046304e-02 3.79360050e-01 -8.15586805e-01 7.72804797e-01
8.52912307e-01 -2.79999942e-01 5.19314587e-01 8.38430643e-01
2.08140969e-01 -1.55332759e-01 -1.04399061e+00 1.20041740e+00
6.68966323e-02 -7.88694382e-01 1.25284076e-01 2.33814150e-01
7.88083613e-01 -1.56692695e-02 -2.57189497e-02 1.66592017e-01
2.60647058e-01 -6.72160685e-01 2.62118608e-01 6.30966604e-01
-1.28275948e-02 -5.17102480e-01 5.53074062e-01 3.94164294e-01
-1.01577187e+00 -1.59964427e-01 -1.15877509e+00 -3.89230311e-01
-5.00400066e-02 8.17479253e-01 -4.74454850e-01 9.05180275e-01
6.19768322e-01 5.66550493e-01 -7.51134276e-01 9.84596312e-01
6.44026846e-02 4.03507203e-01 -4.23125416e-01 -1.50025919e-01
3.11550379e-01 -1.06535232e+00 8.77286255e-01 7.56795883e-01
4.74007785e-01 4.91624862e-01 -1.29541028e-02 8.00632536e-01
3.90229017e-01 2.12530479e-01 -5.46425700e-01 2.65764564e-01
2.15772465e-01 1.37785399e+00 -8.27087045e-01 -2.65458316e-01
-5.64654887e-01 3.54004592e-01 -1.04429901e-01 7.97431991e-02
-7.44840980e-01 -4.95761126e-01 5.07110775e-01 -1.76116964e-03
-2.19435051e-01 -3.98904502e-01 -3.29517275e-01 -1.16189110e+00
-3.07210088e-01 -1.60633281e-01 4.72164959e-01 -7.68836379e-01
-1.26085174e+00 3.61626148e-01 1.65587127e-01 -1.27616823e+00
1.67846277e-01 -1.09253764e+00 -8.37327182e-01 1.13279116e+00
-8.56082141e-01 -6.02435946e-01 -3.36030126e-01 6.60120726e-01
5.97402453e-02 -7.05552578e-01 5.31858802e-01 4.69570085e-02
-4.60803926e-01 -1.50306039e-02 5.15646279e-01 -1.52305616e-02
3.93054098e-01 -1.51082802e+00 5.06609142e-01 7.46834874e-01
1.12580128e-01 7.94076920e-01 1.10774839e+00 -6.63904667e-01
-1.47874570e+00 -8.44844639e-01 6.25871956e-01 -5.57429075e-01
8.10354590e-01 -2.67856181e-01 -9.58958268e-01 4.35810119e-01
-1.89040273e-01 -1.89242419e-02 4.95810539e-01 9.04026181e-02
-2.30608881e-01 -7.81455934e-02 -1.33484590e+00 5.93738794e-01
8.00684750e-01 -2.26878226e-01 -1.36255014e+00 8.49263132e-01
5.10827541e-01 -7.32029229e-02 -5.78122258e-01 3.98600757e-01
3.80540900e-02 -9.76613343e-01 1.07788646e+00 -7.24692583e-01
6.92175254e-02 -6.97163403e-01 -2.73753077e-01 -1.27649534e+00
-7.01975524e-01 -4.90196377e-01 1.67638585e-01 9.05122161e-01
-1.73105910e-01 -8.00128222e-01 7.75255263e-01 6.81228459e-01
3.57147753e-02 -1.27608135e-01 -1.03568792e+00 -1.00632882e+00
5.05067289e-01 -9.83044785e-03 5.72980225e-01 1.05694258e+00
3.43910962e-01 3.19041640e-01 -6.01135790e-02 5.30547976e-01
8.50587070e-01 2.31416315e-01 6.29199922e-01 -1.75978100e+00
-2.72580445e-01 -4.46557939e-01 -1.05066204e+00 -8.19224477e-01
4.69545759e-02 -1.17202568e+00 -4.12510931e-01 -1.37729526e+00
9.28678662e-02 -4.29589391e-01 -3.23942572e-01 -1.37229010e-01
-5.11370637e-02 4.06144746e-02 5.14754653e-03 5.42906165e-01
-2.96743482e-01 4.80498135e-01 1.15422797e+00 1.71579763e-01
-3.78315210e-01 4.65392955e-02 -3.37169528e-01 1.03069913e+00
8.88870955e-01 -2.60424495e-01 -3.75428945e-01 -3.38842720e-01
6.89538643e-02 -3.77825707e-01 3.39215696e-01 -1.40185273e+00
1.44704252e-01 -2.57638425e-01 2.07706213e-01 -4.83837664e-01
1.34552702e-01 -5.27867436e-01 4.68581952e-02 4.23733145e-01
-9.96693224e-03 -3.64473313e-02 -2.82689452e-01 4.72649932e-01
3.00155789e-01 -6.21539712e-01 1.21889055e+00 -3.22519869e-01
-6.17776513e-01 2.36389652e-01 1.01326734e-01 -1.17354140e-01
1.30107307e+00 -1.11047268e-01 -4.38601047e-01 1.41999543e-01
-8.04920435e-01 -1.50795609e-01 3.36187422e-01 5.64253330e-01
1.48885965e-01 -1.36099350e+00 -7.06010580e-01 1.88127846e-01
-1.18960701e-02 -2.72009850e-01 1.80719167e-01 6.99132204e-01
-6.31883025e-01 5.61209500e-01 -2.51906663e-01 -7.37191558e-01
-1.01761878e+00 1.07702267e+00 4.30217713e-01 4.91230130e-01
-8.31889451e-01 7.74016619e-01 8.33346665e-01 -6.58063650e-01
-3.07245076e-01 -1.54162437e-01 -6.63625479e-01 -3.95577461e-01
6.79790735e-01 8.25230718e-01 -7.31638912e-03 -1.00592828e+00
-3.56379539e-01 1.00355506e+00 8.68609995e-02 -1.88230928e-02
1.01412773e+00 -1.71844140e-01 -6.57953620e-01 5.28061986e-01
1.16119540e+00 4.97260876e-02 -5.60976326e-01 -1.81930065e-01
2.58644223e-01 -6.56507492e-01 5.73160015e-02 -3.89626771e-01
-9.20326591e-01 9.35196161e-01 8.03525805e-01 7.57849753e-01
7.98474371e-01 1.15513712e-01 4.93616164e-01 4.66648489e-01
5.81124663e-01 -1.02448535e+00 -1.23961702e-01 5.75115621e-01
6.21496379e-01 -9.39765215e-01 -4.44677137e-02 -2.45389521e-01
-3.20794761e-01 1.36131239e+00 4.33774471e-01 -5.91173410e-01
9.35673237e-01 -3.30670387e-01 -2.08881244e-01 -2.93893129e-01
-5.31847365e-02 8.21438357e-02 4.27917659e-01 5.23554683e-01
1.44665629e-01 2.70865977e-01 -7.34409034e-01 3.82140905e-01
-7.81675637e-01 -4.18923825e-01 6.15619779e-01 6.69928908e-01
-9.14163888e-01 -9.19257164e-01 -8.40017557e-01 2.23599821e-01
-1.82215750e-01 1.86976716e-01 -1.78234115e-01 4.67349946e-01
5.04728913e-01 1.03060448e+00 2.91745722e-01 -4.30460095e-01
1.31998703e-01 1.51654836e-02 3.81915241e-01 -4.62165475e-01
1.55038714e-01 -2.77185291e-01 -6.09982729e-01 -1.90821499e-01
-5.45555711e-01 -7.18546212e-01 -1.50143623e+00 -5.57544410e-01
-3.32185656e-01 8.10349584e-01 1.05007398e+00 1.12111938e+00
1.00901999e-01 1.32091045e-01 9.31037784e-01 -7.63803244e-01
-5.02884984e-01 -9.72106099e-01 -1.02240717e+00 6.61429405e-01
9.67428833e-02 -1.13599682e+00 -1.05886173e+00 -6.39311895e-02] | [7.849867343902588, 4.152456283569336] |
b36e71e3-9929-4f79-a2f8-c54f241f51df | an-industry-4-0-example-real-time-quality | 2206.05818 | null | https://arxiv.org/abs/2206.05818v1 | https://arxiv.org/pdf/2206.05818v1.pdf | An Industry 4.0 example: real-time quality control for steel-based mass production using Machine Learning on non-invasive sensor data | Insufficient steel quality in mass production can cause extremely costly damage to tooling, production downtimes and low quality products. Automatic, fast and cheap strategies to estimate essential material properties for quality control, risk mitigation and the prediction of faults are highly desirable. In this work we analyse a high throughput production line of steel-based products. Currently, the material quality is checked using manual destructive testing, which is slow, wasteful and covers only a tiny fraction of the material. To achieve complete testing coverage our industrial collaborator developed a contactless, non-invasive, electromagnetic sensor to measure all material during production in real-time. Our contribution is three-fold: 1) We show in a controlled experiment that the sensor can distinguish steel with deliberately altered properties. 2) 48 steel coils were fully measured non-invasively and additional destructive tests were conducted on samples to serve as ground truth. A linear model is fitted to predict from the non-invasive measurements two key material properties (yield strength and tensile strength) that normally are obtained by destructive tests. The performance is evaluated in leave-one-coil-out cross-validation. 3) The resulting model is used to analyse the material properties and the relationship with logged product faults on real production data of ~108 km of processed material measured with the non-invasive sensor. The model achieves an excellent performance (F3-score of 0.95) predicting material running out of specifications for the tensile strength. The combination of model predictions and logged product faults shows that if a significant percentage of estimated yield stress values is out of specification, the risk of product faults is high. Our analysis demonstrates promising directions for real-time quality control, risk monitoring and fault detection. | ['Kerstin Bunte', 'Nick Goet', 'Kevin Koster', 'Michiel Straat'] | 2022-06-12 | null | null | null | null | ['fault-detection'] | ['miscellaneous'] | [ 2.65296936e-01 2.93856651e-01 2.49338999e-01 -1.80737510e-01
-5.56219339e-01 -3.47468168e-01 1.20040737e-01 3.97968113e-01
2.34548658e-01 5.67382216e-01 -6.29828811e-01 -1.69385806e-01
-6.02633834e-01 -8.61506343e-01 -7.46517241e-01 -6.70630097e-01
-6.50850460e-02 6.62840843e-01 6.38421714e-01 -9.19146836e-02
1.90983713e-01 8.59217286e-01 -1.48812091e+00 2.74627805e-01
6.31658614e-01 1.40336704e+00 2.94941306e-01 4.27221090e-01
5.92521906e-01 8.80689919e-01 -7.75499940e-01 1.51331052e-01
1.59417436e-01 -2.99629956e-01 -8.72059643e-01 8.90066549e-02
-2.72709996e-01 -2.94197351e-01 3.94646972e-01 7.86791801e-01
1.86632484e-01 -2.82213748e-01 7.22608507e-01 -1.30813766e+00
1.35735437e-01 4.51708853e-01 -1.41399384e-01 -4.34053630e-01
4.95398551e-01 2.02594891e-01 4.43409055e-01 -4.86538053e-01
4.88359749e-01 5.73212922e-01 7.73467124e-01 4.39749053e-03
-1.15365887e+00 -4.24145103e-01 -6.06635809e-01 -1.20912328e-01
-1.13267493e+00 -6.90941289e-02 9.54087973e-01 -8.37550223e-01
1.13277698e+00 5.02383947e-01 5.71738780e-01 7.54313886e-01
1.15348446e+00 2.07728818e-02 1.30103076e+00 -3.80834401e-01
5.72609603e-01 3.92890960e-01 -2.74335980e-01 6.10331535e-01
3.93618703e-01 5.25419354e-01 -2.54726440e-01 -2.69588362e-02
6.22733653e-01 -1.19863734e-01 1.08144052e-01 -3.06821555e-01
-8.07574511e-01 3.30491632e-01 3.58446874e-02 7.25676239e-01
-4.25462902e-01 1.11822970e-01 5.32444954e-01 6.50452435e-01
3.01074952e-01 8.38013947e-01 -9.68554318e-01 -4.15128261e-01
-8.47874284e-01 1.32477164e-01 9.06330466e-01 6.97632134e-01
4.54767525e-01 2.26000518e-01 8.83948952e-02 6.63016021e-01
3.59216601e-01 9.84889030e-01 5.19505367e-02 -7.04523742e-01
-4.91540581e-02 4.52279121e-01 2.33250335e-01 -9.23041224e-01
-7.21881628e-01 -2.39627436e-01 -5.02174616e-01 4.10819829e-01
1.16838910e-01 -2.03028321e-01 -6.59003854e-01 8.72375846e-01
1.62648186e-01 -5.00430703e-01 -3.21549475e-01 7.30681241e-01
2.37567753e-01 3.48191559e-01 -1.16385691e-01 -3.88588667e-01
1.05938947e+00 -2.17740327e-01 -6.80121839e-01 5.69355711e-02
6.41793430e-01 -9.10176992e-01 6.96308076e-01 1.05831122e+00
-9.12712812e-01 -5.78961492e-01 -1.44548869e+00 8.13029468e-01
-2.60019958e-01 1.66038647e-01 5.25435865e-01 7.90358543e-01
-3.79339039e-01 1.27609134e+00 -1.03178561e+00 -2.67154397e-03
4.07765917e-02 3.40362251e-01 -5.05895495e-01 4.82598729e-02
-1.24617696e+00 1.08142006e+00 7.05417478e-04 3.22267234e-01
-1.02162921e+00 -7.90351331e-01 -5.86551487e-01 -3.61202925e-01
2.35424981e-01 8.98294989e-03 1.20300031e+00 -1.58530384e-01
-1.87402892e+00 3.93872291e-01 6.85182154e-01 -1.61241874e-01
5.50339401e-01 -3.45423847e-01 -6.40653729e-01 3.41142833e-01
8.99292082e-02 -4.39879626e-01 7.34937072e-01 -1.20955324e+00
-3.36699545e-01 -1.36091083e-01 -4.18128312e-01 -9.11186755e-01
2.36929223e-01 1.30362958e-01 4.19850051e-01 -3.19484144e-01
3.87318224e-01 -8.79415393e-01 -1.26846418e-01 -5.70492923e-01
-5.43424189e-01 1.17841721e-01 6.04163945e-01 -8.34823012e-01
7.05185235e-01 -1.78670216e+00 -3.41017962e-01 7.57446647e-01
-1.14152268e-01 -2.70835310e-01 5.37199497e-01 7.90896595e-01
-2.46880069e-01 -2.64751583e-01 -3.34766775e-01 5.38861573e-01
-2.72277948e-02 -2.66566157e-01 8.48425701e-02 7.82085836e-01
5.37024379e-01 6.34045124e-01 -8.83459747e-01 3.57309245e-02
5.95296562e-01 7.11635780e-03 -1.05767801e-01 3.33343387e-01
-6.90005198e-02 2.96883106e-01 -4.02898043e-01 9.34612691e-01
4.96648490e-01 4.31300849e-01 -4.56856415e-02 -4.88306522e-01
1.01345487e-01 -5.67196123e-02 -1.12997091e+00 1.25423276e+00
-7.07160711e-01 1.67754367e-01 2.10019842e-01 -1.15521896e+00
1.57230055e+00 5.47318399e-01 9.24106359e-01 -9.71098423e-01
4.48358417e-01 6.60243809e-01 -5.27582876e-02 -8.37867975e-01
1.31335586e-01 -5.04337728e-01 -4.28587973e-01 -3.16006541e-02
1.02026574e-01 -9.98228550e-01 -6.47865310e-02 -2.31394663e-01
1.61917794e+00 2.50376493e-01 -3.02615017e-01 -6.59467041e-01
4.20706064e-01 -9.70037747e-03 4.79345024e-01 3.02160233e-01
1.45098582e-01 4.19787705e-01 7.40707874e-01 -5.37904203e-02
-1.02715039e+00 -1.10739565e+00 -2.42976785e-01 2.59249806e-01
4.84362468e-02 -1.02556951e-01 -6.03557289e-01 -4.73911285e-01
2.75807768e-01 7.55334735e-01 -2.65134424e-01 -7.79067814e-01
-3.24264914e-01 -3.99315238e-01 2.60272026e-01 6.62067235e-01
3.24091762e-02 -9.84282017e-01 -8.38847160e-01 5.50952137e-01
4.81851399e-01 -9.62820709e-01 5.77551186e-01 6.84713602e-01
-8.93261313e-01 -1.29283226e+00 -7.88593069e-02 -2.52154052e-01
5.81033289e-01 -5.40361166e-01 1.06728220e+00 9.03710574e-02
-6.69671357e-01 1.78883746e-01 -4.26489979e-01 -5.32313347e-01
-8.94411504e-01 -3.23676616e-01 2.72853315e-01 -2.85483569e-01
-7.63841048e-02 -5.12887001e-01 -1.19321071e-01 7.77048349e-01
-6.38895512e-01 -4.84760374e-01 7.57313192e-01 7.05096781e-01
5.63243091e-01 8.19652736e-01 8.96814108e-01 -1.09457827e+00
3.53484452e-01 -4.05556083e-01 -8.50389004e-01 4.34462503e-02
-9.01054442e-01 2.12756693e-02 7.82512605e-01 -2.53371686e-01
-9.34152126e-01 1.02312975e-01 -2.53765941e-01 -2.45743036e-01
-3.59155059e-01 5.46339095e-01 -3.83319587e-01 9.78195295e-02
5.63705683e-01 -4.56710100e-01 1.40309617e-01 -6.10707283e-01
-2.83809125e-01 7.41189659e-01 5.39796531e-01 -7.93687165e-01
8.79151881e-01 -9.03858766e-02 5.69986761e-01 -8.32102418e-01
-2.75779426e-01 -2.56394148e-01 -5.78870952e-01 -6.64625943e-01
4.07234669e-01 -5.36807597e-01 -9.34046268e-01 7.51411498e-01
-6.54556096e-01 -4.62183565e-01 -5.15372813e-01 6.80225134e-01
-7.55209446e-01 1.10803895e-01 -7.37639070e-01 -1.26092851e+00
-3.11772227e-01 -1.13336432e+00 1.16726291e+00 -3.96824956e-01
-4.57710803e-01 -6.05550587e-01 -2.85434276e-01 2.83710837e-01
4.51549172e-01 7.92920828e-01 7.63313651e-01 -3.52775574e-01
-1.22102380e-01 -9.39787030e-01 4.14068788e-01 7.58025408e-01
3.10903013e-01 9.91953686e-02 -8.64140868e-01 -2.67882884e-01
5.80704272e-01 -3.03843707e-01 3.28176379e-01 1.77083120e-01
8.30385149e-01 2.87772328e-01 -3.03084970e-01 -9.41053405e-02
1.47770464e+00 4.54041809e-01 6.91886783e-01 9.40637514e-02
4.35501009e-01 8.59272718e-01 1.32920909e+00 3.40608716e-01
-8.74332428e-01 5.97883284e-01 4.77430075e-01 5.91410026e-02
3.54154110e-01 -4.23388779e-02 4.91423458e-01 7.99870849e-01
-5.82676157e-02 8.72534811e-02 -8.14623296e-01 4.20742780e-01
-1.25460446e+00 -7.31246531e-01 -6.91040814e-01 2.41264057e+00
6.49499297e-01 8.65762115e-01 1.32182753e-02 9.47453320e-01
3.92979026e-01 -7.57468581e-01 -1.67890757e-01 -6.94112480e-01
3.57859999e-01 3.99257362e-01 8.89353752e-01 2.98168838e-01
-6.75270200e-01 3.82583439e-02 6.60933161e+00 6.03118420e-01
-1.03888869e+00 -4.42482866e-02 4.08909917e-01 -5.43362424e-02
-2.15310827e-02 4.90841866e-02 -4.10182327e-01 6.24036312e-01
1.40141487e+00 3.42657030e-01 -9.62746665e-02 8.42423201e-01
2.18202591e-01 -7.33536839e-01 -1.20962596e+00 4.33850795e-01
-3.93260688e-01 -9.66594517e-01 -1.01948428e+00 2.83241775e-02
3.94777089e-01 -3.45732212e-01 -5.39248347e-01 -9.53251272e-02
-8.52692779e-03 -6.99397624e-01 1.07816803e+00 9.27366018e-01
9.03838515e-01 -9.47787285e-01 1.28895092e+00 2.12039009e-01
-7.53879607e-01 -2.55997628e-01 -4.67195176e-02 -1.39328465e-02
3.93678129e-01 1.54443109e+00 -8.34861040e-01 1.01402581e+00
8.26380014e-01 2.04449475e-01 -4.27510828e-01 5.59386611e-01
-5.34145124e-02 8.77119303e-01 -5.82395613e-01 -7.08545446e-02
-3.48628074e-01 -1.21818013e-01 2.48234197e-01 7.26243973e-01
2.68864751e-01 -5.14845908e-01 -1.09785527e-01 1.04498518e+00
7.12277532e-01 -3.21452975e-01 -5.86241186e-01 -1.14140799e-02
4.03540224e-01 1.02780616e+00 -9.81717944e-01 2.20007449e-01
-1.26123503e-01 5.58964074e-01 -4.98935252e-01 -1.60921007e-01
-7.96337366e-01 -7.85187483e-01 9.50723886e-02 7.07137108e-01
8.85818377e-02 -3.72976921e-02 -3.50528240e-01 -7.47314841e-02
4.29507583e-01 -4.37403470e-01 -3.29306990e-01 -7.69459188e-01
-1.42919731e+00 2.28536561e-01 4.20525163e-01 -1.28496623e+00
-4.70390707e-01 -9.16073024e-01 -4.62774932e-01 7.65195608e-01
-5.36008894e-01 -8.34496975e-01 1.10277329e-02 -1.25535339e-01
1.64298803e-01 -1.32879749e-01 6.98530257e-01 2.80843824e-01
-5.03987312e-01 3.73088211e-01 1.33978650e-01 -2.23558620e-01
3.63868386e-01 -1.11320460e+00 -2.95079891e-02 5.14984727e-01
-6.30084813e-01 2.33989224e-01 1.19276643e+00 -1.22011149e+00
-1.71117330e+00 -8.79103482e-01 4.99417782e-01 -4.10176933e-01
9.55524266e-01 -5.41346371e-01 -6.89464331e-01 3.70196812e-02
-2.87083775e-01 3.56623456e-02 2.96126664e-01 -1.30298585e-01
3.58364284e-01 -4.19699103e-01 -1.59431744e+00 -3.36561829e-01
3.93701762e-01 -5.11249363e-01 -4.90461141e-01 5.14711022e-01
4.10390377e-01 -1.59536824e-01 -1.68218780e+00 7.09703207e-01
5.31774759e-01 -9.44177151e-01 6.32121146e-01 3.87392640e-02
4.95153069e-01 -7.83553272e-02 5.38575649e-02 -9.60725129e-01
-1.87991038e-01 -4.92915004e-01 2.27531902e-02 1.56467187e+00
4.63568270e-01 -6.01338744e-01 6.04133725e-01 6.69371188e-01
-4.39549059e-01 -1.00321913e+00 -9.68293488e-01 -1.23284423e+00
-6.45403937e-02 -7.17518866e-01 4.92583930e-01 4.68207777e-01
3.10199857e-01 5.96990362e-02 -3.84081975e-02 1.23144291e-01
2.80385643e-01 -2.78413296e-01 2.78513998e-01 -1.52820981e+00
-4.67895508e-01 1.93918452e-01 -8.46198976e-01 9.44463164e-02
-2.05544278e-01 -3.23678493e-01 5.77608943e-01 -1.33261645e+00
-3.92976463e-01 -4.83084172e-01 -2.15746984e-01 3.24949056e-01
4.47666258e-01 5.12295999e-02 -3.09846878e-01 -1.70736492e-01
1.79991145e-02 2.11598590e-01 9.26549494e-01 -2.81643122e-01
2.70473678e-02 1.90971151e-01 8.91729891e-02 4.63219970e-01
7.88574398e-01 -5.65791428e-01 -2.60948539e-01 3.61237884e-01
6.14643216e-01 4.58223641e-01 4.79008347e-01 -1.44989789e+00
-3.20558727e-01 -1.42180398e-01 5.16299903e-01 -7.54312575e-01
-5.84084615e-02 -1.49098492e+00 8.03806484e-01 8.17005873e-01
1.05899155e-01 -1.44201607e-01 1.12706929e-01 2.92816788e-01
-6.13781363e-02 -7.19058514e-01 5.71105421e-01 5.20181991e-02
-1.24843195e-01 -3.79641235e-01 -5.66977739e-01 -6.51799560e-01
1.07735968e+00 -3.32581669e-01 7.81660229e-02 2.70559877e-01
-7.32548475e-01 -1.16343796e-01 4.32945430e-01 2.46184796e-01
3.05482805e-01 -1.22288322e+00 -4.81762409e-01 4.88284528e-01
9.25581828e-02 -5.23794070e-02 3.72116178e-01 1.06421566e+00
-7.23184824e-01 2.79131770e-01 -2.08774656e-01 -7.28319585e-01
-7.66432643e-01 7.13880479e-01 4.39237475e-01 -2.42888734e-01
-4.69996035e-01 6.86717093e-01 -6.10934734e-01 -1.92668065e-01
-3.83515120e-01 -6.28312111e-01 2.25248784e-01 -9.10429731e-02
6.76462725e-02 7.03244746e-01 1.00412297e+00 -2.95924276e-01
-4.68081236e-01 4.84943569e-01 3.40990633e-01 4.22021784e-02
1.50787818e+00 2.71085858e-01 -2.22458199e-01 1.00768542e+00
1.07976902e+00 1.52207389e-01 -1.15245974e+00 6.08285666e-01
2.57045120e-01 -2.14820355e-01 8.10995176e-02 -1.11714530e+00
-1.27092946e+00 4.11435604e-01 7.67628133e-01 6.28814220e-01
1.05944490e+00 1.74958240e-02 5.84211767e-01 2.12073281e-01
8.55577350e-01 -1.64404106e+00 -1.46125495e-01 7.05157593e-02
1.09179568e+00 -7.15422511e-01 3.13157648e-01 -7.20562458e-01
-2.26466596e-01 9.87552881e-01 4.99877036e-01 -1.93670139e-01
8.75934422e-01 9.43503261e-01 -1.63048849e-01 -5.50014496e-01
-5.04799545e-01 5.92554927e-01 -1.01367719e-01 7.67117083e-01
4.17601645e-01 2.08123013e-01 -1.27767414e-01 8.86520386e-01
-2.81143546e-01 4.00346853e-02 2.32717291e-01 1.44688773e+00
-4.81166661e-01 -9.17787611e-01 -7.43711829e-01 9.09808934e-01
-4.80856031e-01 6.24357760e-01 -2.97231883e-01 7.10554063e-01
6.28399327e-02 1.39235127e+00 -6.53514788e-02 -7.69590259e-01
9.62907493e-01 -1.08433301e-02 5.70140243e-01 -3.86883348e-01
-8.04071844e-01 1.59243178e-02 4.49225843e-01 -9.26676214e-01
-9.73242298e-02 -6.02607667e-01 -1.34441829e+00 -1.15461715e-01
-9.23707128e-01 2.07344100e-01 1.04092741e+00 9.91111577e-01
-3.63882072e-02 1.05085146e+00 1.03987622e+00 -7.15874314e-01
-6.80809915e-01 -1.35207546e+00 -1.42527092e+00 3.09931576e-01
-9.30085853e-02 -1.10079479e+00 -5.63645005e-01 -3.67483683e-02] | [6.793285369873047, 2.3606481552124023] |
4b811d15-fb00-4e6d-adea-12fad4822e5b | matcha-enhancing-visual-language-pretraining | 2212.09662 | null | https://arxiv.org/abs/2212.09662v2 | https://arxiv.org/pdf/2212.09662v2.pdf | MatCha: Enhancing Visual Language Pretraining with Math Reasoning and Chart Derendering | Visual language data such as plots, charts, and infographics are ubiquitous in the human world. However, state-of-the-art vision-language models do not perform well on these data. We propose MatCha (Math reasoning and Chart derendering pretraining) to enhance visual language models' capabilities in jointly modeling charts/plots and language data. Specifically, we propose several pretraining tasks that cover plot deconstruction and numerical reasoning which are the key capabilities in visual language modeling. We perform the MatCha pretraining starting from Pix2Struct, a recently proposed image-to-text visual language model. On standard benchmarks such as PlotQA and ChartQA, the MatCha model outperforms state-of-the-art methods by as much as nearly 20%. We also examine how well MatCha pretraining transfers to domains such as screenshots, textbook diagrams, and document figures and observe overall improvement, verifying the usefulness of MatCha pretraining on broader visual language tasks. | ['Julian Martin Eisenschlos', 'Nigel Collier', 'Yasemin Altun', 'Mandar Joshi', 'Kenton Lee', 'Chenxi Pang', 'Syrine Krichene', 'Francesco Piccinno', 'Fangyu Liu'] | 2022-12-19 | null | null | null | null | ['chart-question-answering', 'chart-question-answering', 'data-summarization'] | ['computer-code', 'computer-vision', 'miscellaneous'] | [-1.74780875e-01 1.07720951e-02 -4.00320701e-02 -2.80364215e-01
-5.93711793e-01 -8.09794664e-01 1.12209189e+00 3.22100222e-01
-1.18681416e-01 1.17620319e-01 3.02649021e-01 -1.08978033e+00
5.15125036e-01 -5.85204303e-01 -1.05550182e+00 1.10383913e-01
1.79323718e-01 3.05833489e-01 5.68108261e-02 -7.59039894e-02
1.02691971e-01 5.54064393e-01 -1.38088655e+00 1.10036802e+00
8.09374452e-01 7.42810905e-01 3.01263463e-02 9.20642197e-01
-7.49081790e-01 1.35858881e+00 -5.70491135e-01 -6.88206017e-01
-6.07566983e-02 -2.91060716e-01 -7.06384361e-01 1.42905682e-01
1.05298567e+00 -5.99543571e-01 -5.26247621e-01 8.91008198e-01
8.28330219e-02 -1.06328346e-01 9.35574651e-01 -1.47585952e+00
-1.46601629e+00 7.09161282e-01 -9.27938282e-01 4.59378213e-02
6.25040770e-01 5.95506430e-01 1.14066279e+00 -1.40416479e+00
7.56634712e-01 1.80064511e+00 6.31560087e-01 9.89104882e-02
-1.33329368e+00 -4.26101983e-01 4.50685292e-01 6.12461828e-02
-1.26056004e+00 -3.19842339e-01 6.29566193e-01 -7.04054177e-01
1.32131004e+00 2.13788539e-01 8.25924695e-01 8.88803303e-01
1.29001215e-01 1.33290362e+00 1.11322236e+00 -5.45397878e-01
-5.86486468e-03 8.75645503e-02 1.61283821e-01 1.15258968e+00
2.87457351e-02 -1.59267411e-01 -6.43759191e-01 2.02125549e-01
1.08840775e+00 3.91263925e-02 9.78753939e-02 -5.14325798e-01
-1.31808639e+00 6.66864812e-01 7.00858891e-01 -8.96687359e-02
-1.81704015e-01 4.57694620e-01 4.60400343e-01 2.99995184e-01
1.61343291e-01 3.19200397e-01 1.96444452e-01 -2.53725480e-02
-1.11078870e+00 3.48693848e-01 5.61109126e-01 1.42789602e+00
6.42074049e-01 2.53156215e-01 -4.39741939e-01 5.94166338e-01
5.55625260e-01 7.93694615e-01 -1.27228513e-01 -8.18131506e-01
9.65743959e-01 9.05213892e-01 -1.47910208e-01 -8.38858008e-01
-3.86526227e-01 -2.40559027e-01 -6.89663649e-01 6.00039184e-01
5.47082305e-01 9.35519934e-02 -1.58153355e+00 1.17195034e+00
-2.12294966e-01 -3.94366205e-01 -3.77883390e-02 7.42569029e-01
1.52213895e+00 9.56915259e-01 4.38378811e-01 2.49520749e-01
1.46910071e+00 -1.27050626e+00 -6.50464058e-01 -4.35188323e-01
8.15153360e-01 -9.29704070e-01 1.93877792e+00 3.92127573e-01
-1.47056162e+00 -7.88089275e-01 -1.07509649e+00 -8.46951783e-01
-7.24177241e-01 5.61931849e-01 7.22979486e-01 3.19407672e-01
-1.29677939e+00 8.34309384e-02 -7.57402539e-01 -5.88470995e-01
6.81421697e-01 -3.85251254e-01 -2.94588745e-01 -2.76800781e-01
-5.52490294e-01 8.74740243e-01 2.29574144e-01 -2.04791605e-01
-9.55306649e-01 -1.13777256e+00 -1.16799867e+00 3.10672879e-01
3.27462494e-01 -7.95344055e-01 1.49576521e+00 -5.82564592e-01
-1.18428123e+00 1.19091856e+00 -1.52301997e-01 -5.41782081e-01
8.79708350e-01 -3.50117803e-01 -2.29972795e-01 1.36602908e-01
-1.76041439e-01 1.16898537e+00 6.99968338e-01 -1.50856364e+00
-2.46497750e-01 -2.90982593e-02 4.20977235e-01 9.62578058e-02
-1.27968341e-01 3.15482207e-02 -8.74999225e-01 -6.75597668e-01
-3.49614263e-01 -4.25592542e-01 1.08073488e-01 7.79952466e-01
-5.65764487e-01 -1.54345453e-01 6.87786043e-01 -8.80541146e-01
1.28828442e+00 -2.14777684e+00 -4.67157699e-02 1.50546491e-01
5.05981266e-01 1.23020932e-01 -4.74011540e-01 7.70636261e-01
-2.08075449e-01 3.01652431e-01 5.67445979e-02 -5.03914297e-01
2.68846363e-01 -4.06943634e-03 -6.05884671e-01 7.99422264e-02
2.24964157e-01 1.39412749e+00 -5.39857090e-01 -7.42976248e-01
5.16720355e-01 4.39824104e-01 -4.48403835e-01 2.99784690e-01
-6.57392502e-01 -2.92989612e-01 -2.25915830e-03 7.92660475e-01
5.50352216e-01 -8.27063560e-01 1.06791019e-01 -3.72103482e-01
-1.29863515e-01 -1.98871233e-02 -7.77386129e-01 1.93827081e+00
-4.29606050e-01 1.21834326e+00 -2.83576906e-01 -4.32575583e-01
7.29207218e-01 -1.44879177e-01 -2.24600792e-01 -1.16983688e+00
-8.16784203e-02 -3.92694652e-01 -1.27119943e-01 -5.17246902e-01
7.87757218e-01 1.71081185e-01 2.16889828e-01 3.22727919e-01
6.65974468e-02 -4.19401139e-01 4.12128419e-01 7.60496557e-01
6.97841883e-01 2.55545914e-01 5.37799954e-01 -2.99650785e-02
3.36095899e-01 3.07762444e-01 -5.57774007e-01 8.86438191e-01
1.42427906e-01 6.03325844e-01 8.00707936e-01 -3.65885049e-01
-1.35444570e+00 -1.40450156e+00 3.58651072e-01 1.31097174e+00
-1.79232359e-01 -9.85105753e-01 -4.82533842e-01 -5.80305278e-01
2.50709355e-01 1.10134721e+00 -6.64641142e-01 3.10938954e-01
-2.73155868e-01 9.76867303e-02 4.92921889e-01 1.12771678e+00
3.88260633e-01 -1.03057921e+00 -4.61107969e-01 -3.53302002e-01
4.69052494e-01 -1.24718535e+00 -3.33991736e-01 -4.62558419e-02
-6.67942405e-01 -9.58898723e-01 -9.42882955e-01 -8.53051186e-01
6.94370389e-01 3.52470100e-01 1.51620376e+00 2.61616290e-01
-2.75618494e-01 6.58777177e-01 3.79076530e-03 -5.41702151e-01
-6.85409904e-01 -2.45449468e-01 -6.96901321e-01 -5.51674902e-01
-3.73828709e-02 -2.40262046e-01 -1.95991710e-01 -2.75156021e-01
-9.64888692e-01 9.08333898e-01 6.65060103e-01 4.49409366e-01
3.86726707e-01 -4.07489568e-01 -1.41867369e-01 -9.23936129e-01
7.54997849e-01 -8.37353691e-02 -6.25026822e-01 7.12009251e-01
-4.39149320e-01 1.78507522e-01 6.75754547e-01 -3.71383518e-01
-8.73287678e-01 -1.18854977e-01 1.36264220e-01 -7.28592992e-01
-2.45892763e-01 8.76752615e-01 -1.15551986e-02 2.31471673e-01
5.71577609e-01 5.64377196e-02 -1.08666942e-02 -4.46402669e-01
1.18187630e+00 1.09367371e-01 7.53799677e-01 -5.02553284e-01
1.01209235e+00 4.82023120e-01 -6.63804561e-02 -9.47667539e-01
-5.22828758e-01 -3.49999964e-01 -4.79996115e-01 -3.32370132e-01
9.89024520e-01 -1.19071150e+00 -9.22455907e-01 1.67561278e-01
-1.49580896e+00 -6.40773773e-01 -1.49918020e-01 8.02315772e-02
-5.83109856e-01 3.65367085e-01 -6.19140387e-01 -7.65899658e-01
-2.23550111e-01 -1.08826804e+00 1.07188308e+00 2.80024588e-01
-2.06945658e-01 -1.09628117e+00 -3.91516417e-01 2.14219466e-01
8.87361914e-02 3.83815020e-01 1.49855590e+00 -2.95807511e-01
-6.83826327e-01 2.42345795e-01 -1.00316453e+00 5.11099584e-02
-1.90656751e-01 4.79891658e-01 -7.54985988e-01 -1.00566074e-01
-9.32883978e-01 -7.51667380e-01 1.05432343e+00 1.81386709e-01
1.05897880e+00 -1.48575053e-01 -2.91664720e-01 6.69771135e-01
1.32932079e+00 3.05539101e-01 6.25454128e-01 2.00336620e-01
1.20200682e+00 2.34810963e-01 2.70903885e-01 2.99274653e-01
8.43694508e-01 1.82564139e-01 5.11720955e-01 -6.90838635e-01
-5.36604464e-01 -7.73326039e-01 2.42012948e-01 6.96797550e-01
1.37565866e-01 -4.53741312e-01 -1.35339510e+00 5.97589195e-01
-1.87325823e+00 -7.27904081e-01 -2.44408324e-01 1.61326730e+00
7.18674600e-01 3.00510883e-01 2.24229515e-01 -1.93735987e-01
7.64403716e-02 5.21319091e-01 -5.23331285e-01 -6.37264907e-01
-1.89537659e-01 5.94766475e-02 2.21206442e-01 3.57838333e-01
-9.86521125e-01 1.19998336e+00 6.83461094e+00 6.55742764e-01
-1.08788323e+00 -4.28609371e-01 5.38998187e-01 8.88761654e-02
-5.99894881e-01 -2.45848410e-02 -3.33437830e-01 -1.36240408e-01
6.95979655e-01 -1.90398887e-01 4.77388114e-01 1.05350876e+00
1.37524322e-01 1.69482194e-02 -1.46771240e+00 1.42227411e+00
3.29942495e-01 -1.85490489e+00 8.88966262e-01 -1.78456143e-01
4.92957234e-01 -3.16515237e-01 3.27671230e-01 6.55985653e-01
6.28856957e-01 -1.55709410e+00 1.24406385e+00 6.41880274e-01
1.04311860e+00 -5.34805894e-01 7.10356906e-02 -1.37764309e-02
-1.22876811e+00 2.83247530e-01 -2.57352501e-01 -1.30695343e-01
1.59902513e-01 7.64510185e-02 -9.92042959e-01 5.08596301e-01
5.71299016e-01 1.08922827e+00 -1.44601452e+00 9.57063973e-01
-3.97953421e-01 6.07551932e-01 8.77945349e-02 -1.21952690e-01
5.59155285e-01 6.90709800e-02 5.78290410e-02 1.48509634e+00
4.72145453e-02 -2.09072366e-01 2.64835268e-01 1.20756435e+00
-3.27455133e-01 1.96517527e-01 -7.14491665e-01 -7.17493296e-01
1.09300189e-01 8.72812986e-01 -7.90115893e-01 -6.66310310e-01
-9.38506603e-01 8.37795913e-01 5.64465642e-01 8.43600452e-01
-1.04200065e+00 -2.48061642e-01 4.05605376e-01 2.70727247e-01
4.22466815e-01 -5.77619016e-01 -5.99523604e-01 -1.06244957e+00
-5.33850566e-02 -1.20047796e+00 4.29688424e-01 -1.80349600e+00
-1.10752749e+00 4.31903660e-01 1.47556633e-01 -1.04825902e+00
-2.75919557e-01 -1.05023122e+00 -5.00120401e-01 7.43171632e-01
-1.42055655e+00 -1.70487666e+00 -5.87021947e-01 6.05165243e-01
7.88070738e-01 -1.78332880e-01 5.17311811e-01 -1.29010484e-01
-5.23569658e-02 3.67185533e-01 -1.50230840e-01 5.38086474e-01
5.84444702e-01 -1.41015923e+00 1.11668694e+00 9.46484685e-01
6.16596878e-01 7.29077160e-01 6.19049430e-01 -6.15226150e-01
-1.70099926e+00 -7.90108800e-01 4.31440294e-01 -4.82430071e-01
9.84558046e-01 -7.64119864e-01 -9.99511898e-01 1.00692189e+00
8.17308128e-01 -4.47998010e-02 2.12495953e-01 1.05608471e-01
-1.04138684e+00 3.18274856e-01 -4.70436186e-01 1.04647219e+00
1.03918755e+00 -1.04928756e+00 -7.62409329e-01 3.10979068e-01
9.01031435e-01 -8.09831142e-01 -4.82647747e-01 -1.42125919e-01
6.65955484e-01 -7.99471974e-01 1.29799902e+00 -1.08528852e+00
1.00045788e+00 -2.59991974e-01 -7.16310292e-02 -1.06534863e+00
-3.05185378e-01 -4.59273010e-01 -2.40958109e-01 1.19698966e+00
4.53608543e-01 -7.39642456e-02 4.59452569e-01 3.33059847e-01
-2.00751461e-02 -4.12147194e-01 -2.90680647e-01 -7.29169548e-01
2.19101176e-01 -6.71542108e-01 4.36097860e-01 9.50437963e-01
-7.80403018e-02 4.29940104e-01 -1.54052764e-01 -8.00536424e-02
4.81794596e-01 1.39429256e-01 1.15762007e+00 -8.93447816e-01
-1.18873157e-01 -7.68618524e-01 -2.44064242e-01 -1.16904831e+00
1.60602555e-01 -1.05593359e+00 -3.08689177e-01 -2.34344244e+00
2.28330970e-01 2.60604680e-01 1.64405107e-01 6.64071918e-01
-7.32946470e-02 3.66475917e-02 9.82247174e-01 -5.03085274e-03
-6.94066584e-01 2.48292908e-01 1.57173336e+00 -6.08678818e-01
-1.06592469e-01 -6.25601172e-01 -5.78604758e-01 8.30267906e-01
2.30182350e-01 1.54541239e-01 -7.08789706e-01 -8.23837221e-01
4.32755798e-01 2.62087345e-01 7.31821179e-01 -9.20393467e-01
1.93882272e-01 -3.43356550e-01 7.70817697e-01 -1.15832591e+00
2.10292116e-01 -5.70641279e-01 -2.42217466e-01 4.60102484e-02
-6.76539719e-01 5.58339953e-01 7.45544493e-01 4.02239591e-01
-1.91743121e-01 3.03246737e-01 6.06586516e-01 -1.12368852e-01
-8.65425229e-01 -7.20185935e-02 -4.02784228e-01 2.73035318e-01
6.31310165e-01 -2.73791313e-01 -9.40982044e-01 -7.02071965e-01
-7.08247721e-01 3.12511474e-01 6.53431773e-01 6.08732104e-01
8.74188602e-01 -1.31280577e+00 -5.13459682e-01 4.07298170e-02
5.01824856e-01 -7.29992017e-02 1.76999375e-01 4.76283222e-01
-1.07757807e+00 5.91202319e-01 -3.69827569e-01 -7.61533201e-01
-1.50376368e+00 1.09782732e+00 -4.58998121e-02 -1.89781666e-01
-8.72505128e-01 5.26714027e-01 7.21380889e-01 -4.54255678e-02
5.22068739e-01 -1.35004365e+00 8.11020881e-02 -7.93312639e-02
6.26987219e-01 1.48280963e-01 -2.07206994e-01 -2.38245532e-01
-3.74908358e-01 4.63185102e-01 -2.15603799e-01 -2.29995728e-01
1.08430016e+00 8.25785175e-02 2.54744053e-01 5.67657173e-01
9.81470346e-01 1.10106990e-02 -1.25633287e+00 -1.13672078e-01
5.28304391e-02 -1.81473345e-01 -1.14937477e-01 -1.05975139e+00
-7.47804642e-01 1.48266029e+00 1.60342157e-01 1.16387554e-01
8.00481558e-01 9.42996219e-02 2.10785210e-01 5.20289838e-01
-1.25651434e-01 -6.14007115e-01 3.78518999e-01 4.85676944e-01
1.44350088e+00 -1.20723164e+00 2.73285955e-01 -2.84178525e-01
-9.82781768e-01 1.36190355e+00 7.33378410e-01 -4.18225043e-02
3.04248720e-01 4.80806142e-01 2.51655340e-01 -3.79869640e-01
-9.04342175e-01 -3.98347855e-01 8.33808661e-01 6.70090079e-01
7.01597691e-01 -6.14667777e-03 3.07557285e-01 2.55177826e-01
-2.78431058e-01 -7.00897872e-02 5.62292099e-01 9.41679418e-01
-8.53750482e-02 -3.74341220e-01 -4.29548115e-01 2.06684962e-01
1.15906537e-01 -5.52756190e-01 -7.33470559e-01 1.44796371e+00
-5.61869442e-01 5.66460133e-01 4.15762752e-01 3.60793434e-02
4.15864170e-01 1.37664750e-01 6.27309978e-01 -4.45966959e-01
-6.18800759e-01 1.24812730e-01 1.92928568e-01 -6.38446271e-01
-1.62910819e-01 -2.84106493e-01 -1.57628179e+00 -3.89227539e-01
4.86169755e-01 -4.99914527e-01 5.59291780e-01 7.34444082e-01
2.09840611e-01 7.82019138e-01 -4.68222231e-01 -4.96184140e-01
-1.01274848e-01 -7.40679741e-01 -2.56151646e-01 5.69173932e-01
4.64352131e-01 -4.13587093e-01 9.67260748e-02 4.99458641e-01] | [11.109640121459961, 2.0356361865997314] |
ac48f33e-0040-4aeb-8603-c8196f75f355 | readability-based-sentence-ranking-for | 1603.06009 | null | http://arxiv.org/abs/1603.06009v1 | http://arxiv.org/pdf/1603.06009v1.pdf | Readability-based Sentence Ranking for Evaluating Text Simplification | We propose a new method for evaluating the readability of simplified
sentences through pair-wise ranking. The validity of the method is established
through in-corpus and cross-corpus evaluation experiments. The approach
correctly identifies the ranking of simplified and unsimplified sentences in
terms of their reading level with an accuracy of over 80%, significantly
outperforming previous results. To gain qualitative insights into the nature of
simplification at the sentence level, we studied the impact of specific
linguistic features. We empirically confirm that both word-level and syntactic
features play a role in comparing the degree of simplification of authentic
data. To carry out this research, we created a new sentence-aligned corpus from
professionally simplified news articles. The new corpus resource enriches the
empirical basis of sentence-level simplification research, which so far relied
on a single resource. Most importantly, it facilitates cross-corpus evaluation
for simplification, a key step towards generalizable results. | ['Sowmya Vajjala', 'Detmar Meurers'] | 2016-03-18 | null | null | null | null | ['cross-corpus'] | ['computer-vision'] | [ 2.48266563e-01 2.42655948e-01 -1.25461206e-01 -4.35588151e-01
-1.13750827e+00 -7.22920418e-01 7.05968082e-01 8.20496917e-01
-6.66488290e-01 5.87489724e-01 8.67689371e-01 -2.62516916e-01
-2.83213884e-01 -5.27773559e-01 -4.01574582e-01 -2.11773198e-02
4.72431362e-01 2.07679868e-01 -1.07308790e-01 -5.49998283e-01
7.60832012e-01 2.64848471e-01 -1.53449225e+00 4.32876766e-01
1.37656546e+00 2.20909655e-01 3.68567735e-01 5.05123317e-01
-1.21957131e-01 6.60111964e-01 -9.74107802e-01 -8.39524567e-01
-8.82497206e-02 -3.52475643e-01 -9.88279700e-01 -2.86198229e-01
7.23565400e-01 -1.60016865e-02 2.61263289e-02 9.48820710e-01
4.95181203e-01 1.21562723e-02 5.51239073e-01 -4.73084480e-01
-6.33980036e-01 7.78734326e-01 1.18634872e-01 3.64596128e-01
9.07097280e-01 -1.33241475e-01 1.36313689e+00 -5.26203275e-01
9.12310362e-01 1.28530455e+00 7.50392854e-01 2.77624398e-01
-1.20914531e+00 -2.85031438e-01 1.00074252e-02 5.79506978e-02
-1.02490461e+00 -6.31045341e-01 6.32188499e-01 -4.60206240e-01
1.03767776e+00 4.70484287e-01 4.94602770e-01 7.61429548e-01
2.78473943e-01 4.76287007e-01 1.23408043e+00 -9.21675503e-01
-3.03771526e-01 2.25768179e-01 4.98483032e-01 3.34894419e-01
4.67745215e-01 -3.88486058e-01 -4.57276374e-01 1.30777866e-01
4.87714075e-02 -4.96599674e-01 -3.29536021e-01 2.08145425e-01
-1.14705861e+00 7.06672251e-01 -9.68846753e-02 9.10751700e-01
-1.45181805e-01 -4.58689362e-01 7.18249261e-01 6.84904575e-01
6.67981982e-01 1.05534554e+00 -5.59840858e-01 -5.67201257e-01
-9.80830312e-01 3.13167363e-01 1.03765917e+00 5.88367641e-01
3.70652735e-01 -2.78289676e-01 -4.90498692e-01 1.01692903e+00
-1.37031212e-01 5.73062718e-01 6.09887660e-01 -9.33166087e-01
7.73341537e-01 7.26793349e-01 5.12622409e-02 -1.11255944e+00
-3.67219567e-01 -4.12965745e-01 -4.11934197e-01 -2.03507841e-01
3.46278816e-01 4.42170128e-02 -1.56209305e-01 1.79998207e+00
-2.22305488e-02 -7.50413001e-01 4.23592553e-02 4.43904251e-01
8.60620618e-01 2.82661736e-01 3.96645926e-02 -3.63231242e-01
1.42756855e+00 -7.51242101e-01 -8.33113313e-01 5.34720672e-03
1.08806062e+00 -1.21201670e+00 1.61577702e+00 3.06388497e-01
-1.20662308e+00 -6.48596048e-01 -9.13472652e-01 -3.70846897e-01
-3.02153885e-01 1.30137786e-01 3.42107475e-01 6.90706313e-01
-9.49758530e-01 9.08502698e-01 -4.17139620e-01 -5.03944159e-01
6.14410378e-02 -1.06808662e-01 -5.02719939e-01 -1.94134131e-01
-1.22819793e+00 1.50946796e+00 1.38288662e-01 -3.07377547e-01
-5.49349934e-02 -7.72230506e-01 -9.07995820e-01 6.48835599e-02
-2.45161057e-02 -6.78657830e-01 1.37230837e+00 -1.15109336e+00
-1.34214580e+00 1.06247127e+00 -2.85810739e-01 -2.67350674e-01
3.50763738e-01 -3.90087336e-01 -4.93458748e-01 -4.93609533e-02
5.20448804e-01 1.27971098e-01 3.52978706e-01 -9.14626837e-01
-5.59215724e-01 -2.65960097e-01 2.93010592e-01 4.46979731e-01
-4.90142703e-01 4.23834771e-01 -2.07094461e-01 -7.66452730e-01
-3.48429769e-01 -6.70702755e-01 2.52114832e-01 -8.16807747e-01
-2.39001319e-01 -4.04629827e-01 1.37925521e-01 -1.08348382e+00
1.71572757e+00 -2.22257495e+00 2.00104341e-01 -8.79685357e-02
2.38047197e-01 4.23594415e-01 -2.44386449e-01 9.33104157e-01
3.99212539e-03 4.80463833e-01 -8.76041949e-02 -5.97598433e-01
9.23442394e-02 -1.16440803e-01 -8.51565078e-02 3.21843654e-01
1.16828665e-01 8.28117132e-01 -1.03880668e+00 -5.52645147e-01
9.74982977e-02 3.27515244e-01 -5.18066406e-01 -2.56150737e-02
2.69289464e-01 8.88882726e-02 -2.09318012e-01 2.87665188e-01
3.33886266e-01 3.07742566e-01 1.51843935e-01 -9.64916199e-02
-5.10684431e-01 8.62624049e-01 -6.64461672e-01 1.52139199e+00
-6.93494797e-01 9.48927760e-01 -2.59044826e-01 -6.87663972e-01
6.85966492e-01 1.00345775e-01 -1.20917559e-01 -1.09650564e+00
1.69611722e-01 1.17656074e-01 2.17417851e-01 -5.16406298e-01
9.79264319e-01 -2.04925537e-01 -2.56789148e-01 4.29068416e-01
-1.24092191e-01 -4.02518153e-01 9.24381793e-01 3.48825186e-01
1.12236822e+00 -1.27944008e-01 4.40678328e-01 -5.61425984e-01
6.22644722e-01 1.05410337e-01 3.48172337e-01 7.65926659e-01
-2.22350135e-02 5.42638004e-01 6.51638925e-01 -1.45033076e-02
-1.10757792e+00 -7.59770632e-01 -3.13341051e-01 1.11382043e+00
-2.01362327e-01 -8.42448473e-01 -1.04650652e+00 -4.02797729e-01
-9.02338102e-02 1.18077123e+00 -5.76391697e-01 -1.86307162e-01
-5.98042250e-01 -3.96274060e-01 6.12909198e-01 1.17154315e-01
2.24748743e-03 -8.45667422e-01 -4.98322517e-01 2.61810850e-02
-6.27294302e-01 -1.04510939e+00 -4.36525047e-01 -3.04320604e-01
-8.79656076e-01 -1.25962400e+00 -3.49863529e-01 -6.94182038e-01
4.19023156e-01 2.71297783e-01 1.53910744e+00 4.32814538e-01
5.51400846e-03 2.50847459e-01 -7.29995549e-01 -4.90528077e-01
-8.84235024e-01 4.55861121e-01 7.06626149e-03 -6.87387884e-01
2.97778010e-01 -3.76325518e-01 -1.92274436e-01 -6.43946677e-02
-8.85846674e-01 -1.11560866e-01 8.12109590e-01 6.00954890e-01
2.47282863e-01 -4.70536947e-02 5.99057555e-01 -1.13649035e+00
1.44394672e+00 -1.62268266e-01 -1.63750738e-01 3.37050557e-01
-7.83769190e-01 7.28665665e-02 7.12063670e-01 -7.21139312e-02
-1.14565563e+00 -6.20296419e-01 -2.86253750e-01 4.69314784e-01
-1.02917753e-01 9.14221168e-01 6.04477338e-02 7.06763342e-02
6.65430307e-01 -1.82667807e-01 1.02188252e-01 -6.36344731e-01
3.06602240e-01 7.06131876e-01 5.12145579e-01 -5.16300321e-01
7.10597396e-01 -2.64105815e-02 -2.86768913e-01 -9.47358012e-01
-1.13044691e+00 -4.05989438e-01 -7.58186102e-01 -8.79702196e-02
5.64540148e-01 -7.66943514e-01 -2.97349751e-01 2.07726464e-01
-1.16908276e+00 -1.93597138e-01 -3.12305450e-01 5.02605796e-01
-1.38861209e-01 7.78021097e-01 -4.85095948e-01 -3.63373548e-01
-3.69339556e-01 -9.18712974e-01 9.95875955e-01 -7.68847689e-02
-9.27577496e-01 -1.11425006e+00 5.21208704e-01 7.72972524e-01
3.41657728e-01 1.72707036e-01 1.20053661e+00 -8.16410899e-01
1.33102864e-01 -4.59586978e-01 -9.38934926e-03 5.61554790e-01
-3.25788409e-02 3.97335052e-01 -5.93894243e-01 -1.08067729e-01
9.99368578e-02 -1.94376007e-01 5.67776680e-01 1.91836536e-01
5.92721283e-01 -7.68918917e-02 3.24842595e-02 7.74730742e-02
1.15079701e+00 -3.36289436e-01 7.60436356e-01 6.93522155e-01
5.71143389e-01 7.30284810e-01 6.51791394e-01 1.67824894e-01
6.10518396e-01 6.49862289e-01 -1.79688334e-01 7.41970390e-02
-4.91276234e-01 -1.33846566e-01 4.04863417e-01 1.42894769e+00
-4.75068651e-02 -2.59553403e-01 -9.06277478e-01 4.85533357e-01
-1.37166345e+00 -1.04313302e+00 -3.97919804e-01 2.22620082e+00
9.84605193e-01 3.93586099e-01 7.00214133e-03 4.96565133e-01
5.65513253e-01 5.67547493e-02 2.30709344e-01 -9.31447744e-01
-2.68100679e-01 3.00756872e-01 1.54962778e-01 8.01457226e-01
-8.05627286e-01 8.86237741e-01 6.65325117e+00 8.20777953e-01
-7.43980765e-01 -2.47149877e-02 3.47127438e-01 -5.99508584e-02
-6.76909268e-01 -6.74397424e-02 -6.12235844e-01 3.83917749e-01
1.06807017e+00 -4.97006476e-01 2.84111381e-01 5.19981146e-01
5.53532362e-01 -3.61707121e-01 -1.11544490e+00 4.80362684e-01
4.28933412e-01 -1.05769992e+00 1.74962804e-02 -3.36729348e-01
6.55386090e-01 6.60800189e-02 -2.08540559e-01 4.99224871e-01
-8.75583887e-02 -8.67837071e-01 9.49269593e-01 5.90314209e-01
7.73252010e-01 -6.90749526e-01 1.02489340e+00 3.76873702e-01
-6.41033232e-01 1.90001577e-01 -5.82978614e-02 -7.04792619e-01
1.61403537e-01 7.36266077e-01 -5.29304028e-01 5.88151813e-01
4.37001288e-01 5.05344450e-01 -9.81920540e-01 7.16001034e-01
-4.15894955e-01 5.85489392e-01 5.32467887e-02 -6.60572648e-02
-1.27202034e-01 -1.93200186e-01 5.81832945e-01 1.47302365e+00
-4.54567596e-02 -1.21615931e-01 -2.47027770e-01 4.18587595e-01
-2.98096184e-02 6.34879887e-01 -4.79042739e-01 -2.13195354e-01
6.65368140e-01 1.22478235e+00 -6.29325688e-01 -2.25383341e-01
-5.37209570e-01 8.16000164e-01 7.26112962e-01 -6.63454086e-02
-4.95885104e-01 -7.70407915e-01 3.35881799e-01 -3.15683410e-02
8.33173543e-02 -2.42855877e-01 -6.62178516e-01 -1.23810148e+00
3.04112047e-01 -1.09542763e+00 -4.17249277e-02 -4.34382945e-01
-1.23341799e+00 5.42812943e-01 4.14908081e-02 -9.10371542e-01
-8.76146927e-02 -2.67296731e-01 -5.47199965e-01 9.92926478e-01
-1.50021982e+00 -8.95018876e-01 -3.76786113e-01 -2.46094652e-02
4.09006149e-01 1.50363415e-01 9.51191485e-01 3.12006980e-01
-7.03523755e-01 7.26424754e-01 2.50550628e-01 -2.62818858e-02
9.19318795e-01 -1.15539980e+00 4.57728565e-01 8.79727602e-01
6.94397837e-02 8.68368149e-01 9.75663304e-01 -6.61880672e-01
-8.64372313e-01 -6.56796634e-01 1.86465800e+00 -7.32251883e-01
7.30032384e-01 -1.04979768e-01 -8.17611456e-01 2.50758290e-01
4.17445123e-01 -1.05359375e+00 9.33015585e-01 5.89276850e-01
-2.48138055e-01 -4.19598557e-02 -9.82811570e-01 6.61825895e-01
1.05364811e+00 -7.68442392e-01 -1.23316193e+00 3.93994719e-01
7.37711251e-01 -3.72172982e-01 -1.16708302e+00 3.02143097e-01
4.76657748e-01 -1.13609147e+00 6.50794268e-01 -4.62953806e-01
8.81077290e-01 1.78235531e-01 -3.69161479e-02 -1.47043765e+00
-4.57110345e-01 -4.72785950e-01 4.66747463e-01 1.55014694e+00
5.98426104e-01 -3.80470037e-01 1.77262619e-01 7.22650766e-01
-5.54691792e-01 -6.72350824e-01 -6.61971331e-01 -7.23861635e-01
3.79280835e-01 -3.89737129e-01 5.43168664e-01 7.46750236e-01
2.91422904e-01 5.87943435e-01 2.58161604e-01 -3.94889295e-01
6.98006004e-02 1.50005659e-03 7.03286767e-01 -1.27894628e+00
2.71898489e-02 -8.60741675e-01 -3.63245070e-01 -5.19284725e-01
6.10471010e-01 -9.11642253e-01 -2.00112462e-01 -1.59695697e+00
4.08235371e-01 -7.36299977e-02 1.21430561e-01 7.61977211e-02
-5.36753297e-01 1.76434100e-01 4.61193204e-01 2.23563448e-01
-5.98257720e-01 3.63422185e-01 1.09891808e+00 1.88724950e-01
-1.00860320e-01 -1.91773489e-01 -1.15872228e+00 6.84472919e-01
9.87358689e-01 -3.78374696e-01 -2.46289045e-01 -8.01842451e-01
4.33791786e-01 -3.14331740e-01 -8.69267583e-02 -9.08683360e-01
-1.59247473e-01 -5.01252674e-02 -1.98538750e-01 -1.66729584e-01
-5.49195670e-02 -3.76140505e-01 -1.20976970e-01 3.57205063e-01
-6.20072663e-01 5.31570494e-01 2.34509617e-01 1.83879018e-01
-4.47885007e-01 -6.21573091e-01 4.93589014e-01 -5.84124289e-02
-3.85448277e-01 -4.96816695e-01 -3.66471976e-01 5.68154216e-01
6.87613964e-01 -2.24801689e-01 -3.97688597e-01 -3.95507902e-01
-2.13147521e-01 -1.20069914e-01 8.20617676e-01 4.24571633e-01
7.52779916e-02 -9.49431956e-01 -9.09753442e-01 -2.84382738e-02
1.93983749e-01 -4.83491629e-01 2.02200934e-02 8.49867105e-01
-6.90953672e-01 6.84409499e-01 -3.28271911e-02 -7.71791786e-02
-1.45607960e+00 2.51120687e-01 5.61953411e-02 -4.51879263e-01
-2.97764033e-01 5.77103198e-01 -3.67981851e-01 -4.29463625e-01
-8.01166333e-03 -4.52962905e-01 -4.81357396e-01 2.60211051e-01
4.64777470e-01 6.33066654e-01 4.50851083e-01 -8.61558676e-01
-6.66124374e-02 3.72632027e-01 -1.42247006e-01 -1.09188326e-01
1.24846661e+00 -3.49921852e-01 -4.86281514e-01 6.67682469e-01
1.12979078e+00 6.32055163e-01 -3.74606252e-01 1.17607549e-01
1.50114998e-01 -5.43018401e-01 -3.79651450e-02 -8.55509579e-01
-3.18428814e-01 4.70884919e-01 -8.74901935e-02 4.47797954e-01
8.59936953e-01 -1.70942590e-01 6.53112411e-01 5.09474218e-01
1.08057642e-02 -1.36475098e+00 -2.23733261e-01 8.17018926e-01
9.29150462e-01 -1.16441810e+00 1.71370998e-01 -5.35125554e-01
-6.60510063e-01 9.59082484e-01 2.29955360e-01 2.31978204e-03
2.45470271e-01 -4.54459246e-03 1.16446525e-01 -7.00142682e-02
-5.13800740e-01 -9.17977169e-02 5.06883085e-01 4.80677217e-01
1.12461257e+00 1.70802489e-01 -1.25525212e+00 7.90507257e-01
-8.89406264e-01 -1.48514286e-01 5.88511288e-01 6.11731231e-01
-3.46157968e-01 -1.40047956e+00 -9.94973779e-02 5.72805762e-01
-5.52046597e-01 -5.09352088e-01 -6.22571051e-01 9.64706063e-01
-5.35908267e-02 1.24092543e+00 -6.89767227e-02 -2.99930900e-01
6.78418577e-01 2.61690430e-02 6.07003927e-01 -8.20606351e-01
-9.78924155e-01 -4.44214851e-01 7.57196963e-01 -2.98819393e-01
-4.20736998e-01 -9.56840277e-01 -8.24199498e-01 -6.84033453e-01
-3.90158117e-01 3.24450105e-01 5.76204419e-01 1.24418712e+00
4.87456828e-01 5.75897694e-01 4.47006851e-01 -6.16175711e-01
-6.65094495e-01 -1.26336932e+00 -3.05220604e-01 9.19271350e-01
4.20334451e-02 -4.42560703e-01 -4.94425327e-01 -1.34563586e-02] | [11.168195724487305, 10.06638240814209] |
016f462a-4c17-4fab-8fae-0412aca7a196 | combining-lexical-features-and-a-supervised | 1710.08451 | null | http://arxiv.org/abs/1710.08451v1 | http://arxiv.org/pdf/1710.08451v1.pdf | Combining Lexical Features and a Supervised Learning Approach for Arabic Sentiment Analysis | The importance of building sentiment analysis tools for Arabic social media
has been recognized during the past couple of years, especially with the rapid
increase in the number of Arabic social media users. One of the main
difficulties in tackling this problem is that text within social media is
mostly colloquial, with many dialects being used within social media platforms.
In this paper, we present a set of features that were integrated with a machine
learning based sentiment analysis model and applied on Egyptian, Saudi,
Levantine, and MSA Arabic social media datasets. Many of the proposed features
were derived through the use of an Arabic Sentiment Lexicon. The model also
presents emoticon based features, as well as input text related features such
as the number of segments within the text, the length of the text, whether the
text ends with a question mark or not, etc. We show that the presented features
have resulted in an increased accuracy across six of the seven datasets we've
experimented with and which are all benchmarked. Since the developed model
out-performs all existing Arabic sentiment analysis systems that have publicly
available datasets, we can state that this model presents state-of-the-art in
Arabic sentiment analysis. | ['Muhammad Hammad', 'Talaat Khalil', 'Samhaa R. El-Beltagy', 'Amal Halaby'] | 2017-10-23 | null | null | null | null | ['arabic-sentiment-analysis'] | ['natural-language-processing'] | [-1.95485681e-01 -4.84311283e-02 4.68249381e-01 -4.56619084e-01
-2.75673658e-01 -6.68271899e-01 6.85022056e-01 8.68167877e-01
-5.57565272e-01 5.01866043e-01 3.20444375e-01 -1.16428733e-01
-1.71622723e-01 -8.71047735e-01 4.93721366e-02 -3.68937701e-01
1.94042213e-02 2.45332614e-01 9.49511230e-02 -1.27747369e+00
7.86458552e-01 2.93125212e-01 -1.59302330e+00 4.95793015e-01
6.80192292e-01 9.36043262e-01 -1.69056952e-01 5.32408774e-01
-3.26384038e-01 8.07129383e-01 -6.77048922e-01 -7.12593138e-01
-1.09109521e-01 -4.00920302e-01 -9.00946975e-01 2.32873946e-01
-1.40304446e-01 2.57389337e-01 1.59359083e-01 6.41174555e-01
3.88761580e-01 4.03448120e-02 8.51943374e-01 -1.06414986e+00
-3.57080430e-01 5.22601366e-01 -4.94395524e-01 -3.87608297e-02
6.74630880e-01 -5.76710939e-01 9.91805196e-01 -1.05762148e+00
5.73570907e-01 1.02807224e+00 5.64896584e-01 -6.14751875e-02
-3.37473422e-01 -2.18304083e-01 -1.69970587e-01 9.38749090e-02
-1.01283920e+00 -2.05290958e-01 5.33125877e-01 -4.45570588e-01
9.44596410e-01 5.73299751e-02 5.12935281e-01 3.11804116e-01
4.79275346e-01 5.53825915e-01 1.25119913e+00 -1.03233480e+00
2.89303903e-02 5.54653823e-01 4.23653841e-01 7.15959489e-01
-1.06969967e-01 -7.27826893e-01 -3.60966980e-01 -1.21418945e-01
-3.47690135e-01 -9.26831663e-02 2.95167863e-01 3.78012162e-04
-8.73974025e-01 1.11292815e+00 6.69016987e-02 6.79444611e-01
-1.67558834e-01 -6.09512627e-01 7.15210855e-01 4.90982443e-01
6.83749855e-01 1.93363637e-01 -4.65940326e-01 -3.80787522e-01
-7.70043910e-01 2.52158046e-01 1.08339095e+00 6.38510287e-01
7.28641510e-01 -2.05563605e-01 4.77375776e-01 9.86872196e-01
5.58899283e-01 5.36712885e-01 7.99670100e-01 -6.84569255e-02
3.87490541e-01 1.24093974e+00 2.25697473e-01 -1.41831529e+00
-6.21756256e-01 1.60725981e-01 -2.80359179e-01 3.22380252e-02
4.53959018e-01 -5.42646646e-01 -6.11387730e-01 9.66893852e-01
2.68709779e-01 -7.72720754e-01 3.94301504e-01 4.75224733e-01
9.09373403e-01 8.01209629e-01 -2.42336586e-01 -1.09263502e-01
1.50232017e+00 -7.95931935e-01 -7.55761862e-01 4.42715856e-04
7.91930199e-01 -1.46902299e+00 9.70695376e-01 5.48621356e-01
-7.36130595e-01 -2.23418146e-01 -1.11925614e+00 1.74336299e-01
-1.19212818e+00 7.19607845e-02 4.05715495e-01 1.16042876e+00
-7.20182180e-01 3.10858369e-01 -5.01691103e-01 -7.51699209e-01
-1.12787522e-01 4.00893658e-01 -4.49667782e-01 2.06248671e-01
-1.07337630e+00 1.19190180e+00 1.94125012e-01 1.98205374e-02
9.34117287e-02 1.23467281e-01 -8.30864429e-01 -3.08081657e-01
7.66329020e-02 2.99157083e-01 8.42267573e-01 -1.39321983e+00
-1.43312323e+00 8.93165290e-01 -4.59444635e-02 -9.94650647e-02
1.83600843e-01 -2.05645770e-01 -1.06846905e+00 5.09609953e-02
1.34898722e-01 2.05836892e-01 4.40142453e-01 -9.05434906e-01
-7.51504183e-01 -5.30974388e-01 1.92829788e-01 1.84155479e-01
-6.60595298e-01 6.76196456e-01 -2.55857229e-01 -5.39765298e-01
1.66058745e-02 -1.24366724e+00 4.48887609e-02 -7.40457594e-01
-1.63695514e-01 -9.63223055e-02 9.45913434e-01 -7.20991075e-01
1.33694541e+00 -2.16508007e+00 3.44516560e-02 5.91107428e-01
-3.86389285e-01 2.65916020e-01 1.55594751e-01 1.16083467e+00
4.70163710e-02 -1.52745871e-02 -2.70729661e-01 -1.03747854e-02
2.41398122e-02 7.86810890e-02 -2.26927698e-01 3.43392074e-01
1.47419885e-01 1.89306438e-01 -7.44537771e-01 -4.26098824e-01
9.60057452e-02 5.42564213e-01 -2.41327643e-01 -7.19758645e-02
3.87004986e-02 1.12847209e-01 -2.41342530e-01 6.13140106e-01
4.51778859e-01 1.01531886e-01 2.10235223e-01 -6.94291387e-03
-3.53369087e-01 5.25426120e-02 -1.24996507e+00 1.15793979e+00
-3.56196076e-01 5.43584526e-01 -3.63787264e-01 -6.61610842e-01
1.13869715e+00 3.12814265e-01 4.28012937e-01 -4.66759652e-01
8.00149918e-01 3.26547742e-01 1.19618356e-01 -5.18164873e-01
1.02236545e+00 8.52637067e-02 -3.44128937e-01 7.87781358e-01
-4.97740023e-02 -3.06623787e-01 8.69206846e-01 2.40616798e-01
3.56817275e-01 -7.57318586e-02 4.09729064e-01 -4.05722201e-01
1.10064745e+00 1.46199316e-01 -2.14888871e-01 3.69483195e-02
-1.01671934e-01 5.23379862e-01 8.02349925e-01 -3.23656529e-01
-7.34446406e-01 -2.64984876e-01 -3.81023325e-02 1.26953149e+00
-2.00112745e-01 -7.21997142e-01 -9.27252948e-01 -7.31013119e-01
-2.55469650e-01 2.93490738e-01 -7.44674385e-01 3.37111682e-01
-3.55478942e-01 -1.12341261e+00 4.09429640e-01 -1.18092008e-01
3.06002021e-01 -1.24505258e+00 -4.59555864e-01 2.47258112e-01
-6.00191019e-02 -8.23031843e-01 3.19456160e-02 1.23406217e-01
-6.21909201e-01 -1.12685025e+00 -4.32610422e-01 -9.15684998e-01
5.93333602e-01 8.86845440e-02 8.12072635e-01 1.02450453e-01
-2.98931748e-02 1.67743891e-01 -1.22731066e+00 -9.46587205e-01
-5.38043499e-01 4.45594192e-01 -2.32784197e-01 2.82638758e-01
7.60044098e-01 7.07595795e-02 -1.87088013e-01 2.64224321e-01
-1.14797688e+00 -3.00380945e-01 6.23662174e-02 3.38735312e-01
-1.49272472e-01 -3.07937842e-02 8.36029828e-01 -1.22406936e+00
9.20082331e-01 -7.07109272e-01 -1.11350350e-01 7.63497800e-02
-4.80927467e-01 -9.17135030e-02 4.64044869e-01 1.75805524e-01
-7.33426094e-01 -2.30589375e-01 -4.47787374e-01 9.29900229e-01
-1.25317290e-01 1.11204851e+00 1.19071551e-01 -5.38759977e-02
6.81177557e-01 -1.23214073e-01 4.33073401e-01 -7.76228532e-02
1.05209149e-01 1.25806510e+00 -3.93785328e-01 -1.28809199e-01
3.13899338e-01 4.43250328e-01 -1.08550929e-01 -1.07651520e+00
-7.59584248e-01 -5.91160655e-01 -5.89404345e-01 -3.91786933e-01
6.19134128e-01 -5.91114581e-01 -5.55954397e-01 1.02341163e+00
-5.43460965e-01 1.42485857e-01 1.87084720e-01 2.29850888e-01
-9.49780941e-02 5.24123967e-01 -6.37287438e-01 -7.23843515e-01
-4.06622976e-01 -1.18597949e+00 4.57294285e-01 2.50370145e-01
-5.48234761e-01 -1.01751482e+00 1.46888405e-01 3.58401358e-01
4.84044999e-01 3.78807962e-01 9.56420839e-01 -7.96186686e-01
5.29501617e-01 -5.69005966e-01 4.96527329e-02 4.87218827e-01
5.05950630e-01 5.21186411e-01 -6.10890746e-01 -1.70797735e-01
-1.41497642e-01 -3.74444813e-01 3.67052436e-01 -2.33906060e-02
3.41355711e-01 1.01445206e-01 1.98047146e-01 -3.52687776e-01
1.38512707e+00 2.37805471e-01 6.34519339e-01 8.67259085e-01
1.77108824e-01 8.50000679e-01 1.06795359e+00 6.30829930e-01
6.81933880e-01 3.59158963e-01 4.03346866e-01 8.66695270e-02
5.87991774e-01 4.34378415e-01 6.29410088e-01 1.20184481e+00
-4.84552532e-02 -2.51314521e-01 -1.06531441e+00 5.33042431e-01
-1.68647659e+00 -5.07941365e-01 -4.68539476e-01 1.75163293e+00
6.40389740e-01 9.78068113e-02 4.93469656e-01 8.83916795e-01
4.27868426e-01 7.11442754e-02 2.79089808e-01 -1.10377932e+00
-3.09280276e-01 3.46740901e-01 1.71637461e-01 5.99354446e-01
-1.22439194e+00 9.39099371e-01 5.52008629e+00 3.26749504e-01
-1.31383777e+00 -9.23598409e-02 5.01546383e-01 2.41372764e-01
1.38490885e-01 -2.31696770e-01 -7.16174126e-01 3.89211178e-01
1.13299656e+00 1.14263967e-01 3.70690711e-02 6.57627583e-01
1.46992728e-01 -4.66267735e-01 -4.47774261e-01 5.02586424e-01
7.43351698e-01 -8.56916249e-01 7.80155137e-03 -1.92595840e-01
7.38189816e-01 -5.14049530e-02 1.36474714e-01 1.20773971e-01
-1.13138415e-01 -8.28987718e-01 7.72942543e-01 1.32546768e-01
1.27171263e-01 -1.26301491e+00 1.25684690e+00 -3.94731648e-02
-7.44603217e-01 -5.86242601e-02 -3.34067382e-02 -3.70080203e-01
-8.28807428e-02 3.49115700e-01 -7.01079905e-01 5.48493564e-01
7.03773320e-01 7.61881471e-01 -9.33142900e-01 6.73963010e-01
-9.97372270e-02 4.66310680e-01 -2.24901259e-01 -6.18446708e-01
6.34628594e-01 -5.05968809e-01 1.25291198e-01 1.42871141e+00
3.42928290e-01 -2.07119852e-01 -2.55885988e-01 -3.06073219e-01
4.22757983e-01 1.14881873e+00 -3.80589157e-01 -2.75704235e-01
-8.51558223e-02 1.51188111e+00 -1.12940264e+00 -1.86141476e-01
-7.59362519e-01 6.31026506e-01 -1.19158886e-01 -1.46907970e-01
-5.73783696e-01 -9.54893291e-01 9.62595940e-02 8.65164772e-02
1.28717840e-01 -1.85185477e-01 -8.26164111e-02 -9.97275531e-01
-1.80057541e-01 -1.25366187e+00 4.50163215e-01 -6.72671676e-01
-1.19072449e+00 9.79046822e-01 -2.67585754e-01 -1.15178096e+00
-2.35697433e-01 -8.93036604e-01 -3.85868758e-01 7.00805008e-01
-1.21207845e+00 -1.25898826e+00 -1.05281226e-01 6.21407032e-01
3.13875943e-01 -6.10217094e-01 1.06269526e+00 5.04229963e-01
-3.25075775e-01 3.46211761e-01 1.02171473e-01 3.36609393e-01
9.34895813e-01 -1.13764930e+00 4.04994301e-02 7.25521743e-01
-2.00285856e-02 5.81942499e-01 7.94417083e-01 -3.39269340e-01
-1.06974709e+00 -5.73791206e-01 1.36731005e+00 -5.17384112e-01
9.90567029e-01 -9.14273709e-02 -5.13022482e-01 3.88135612e-01
5.66006303e-01 -5.16508162e-01 1.26412570e+00 9.99025032e-02
-7.74080232e-02 1.27635961e-02 -1.19938385e+00 4.44454104e-01
-2.01051444e-01 -2.73023665e-01 -7.05333889e-01 2.93404609e-01
-1.71812102e-01 -2.84591049e-01 -7.67581046e-01 -6.20212685e-03
5.95674217e-01 -1.02945268e+00 3.75522196e-01 -5.45942247e-01
7.53086567e-01 -3.03546518e-01 -3.80105793e-01 -1.28666294e+00
3.42549950e-01 -2.92922020e-01 3.58211070e-01 1.29263139e+00
8.60927999e-01 -6.25260711e-01 6.14013433e-01 3.65460753e-01
1.52347863e-01 -5.54623008e-01 -3.32903892e-01 5.74172810e-02
1.94749415e-01 -1.49384901e-01 4.66718882e-01 1.09011805e+00
5.93191683e-01 3.94666195e-01 -1.96280107e-01 -1.91392064e-01
-2.42147043e-01 6.37917668e-02 7.14368284e-01 -1.12408674e+00
2.99444467e-01 -2.74255574e-01 -5.52999914e-01 -9.80912521e-02
-2.17374265e-01 -6.22339129e-01 -4.05114949e-01 -1.45047700e+00
-1.95880860e-01 -3.83093596e-01 -1.32385045e-01 2.79580772e-01
3.27081978e-02 6.35435760e-01 2.86973238e-01 -2.00155362e-01
-3.45445752e-01 -7.51625970e-02 8.16631973e-01 1.58054426e-01
-3.15541178e-01 -1.06237419e-01 -7.83795357e-01 9.10323083e-01
8.68336618e-01 -4.44066495e-01 -9.46959034e-02 -3.38208117e-02
1.11669505e+00 -3.53850365e-01 -4.47710156e-01 -8.07582259e-01
-2.29273196e-02 -1.05788736e-02 2.34658748e-01 -5.78172565e-01
1.70799524e-01 -8.94225121e-01 -2.61873901e-01 4.54052508e-01
-2.82386206e-02 5.16714990e-01 1.53015897e-01 -1.00403503e-01
-5.41474044e-01 -5.55814147e-01 7.33039677e-01 -3.31365243e-02
-6.88458204e-01 -2.58277684e-01 -9.03903186e-01 -6.93972707e-02
1.16475594e+00 -2.36029357e-01 -1.31014273e-01 -3.65806639e-01
-7.31979132e-01 -1.05715834e-01 3.68738204e-01 6.96572602e-01
1.70774564e-01 -7.61722684e-01 -7.70476580e-01 1.34336635e-01
3.91334981e-01 -5.82610130e-01 -6.97138384e-02 7.29582727e-01
-1.11495078e+00 2.25438267e-01 -5.73136687e-01 -2.21249580e-01
-1.53163731e+00 7.84448683e-02 -6.02108166e-02 -1.26850083e-01
4.41623218e-02 3.61741245e-01 -8.77324760e-01 -5.98194659e-01
-1.77990496e-01 -1.34948939e-01 -1.03374195e+00 1.03108156e+00
4.54763889e-01 4.27321434e-01 5.65524518e-01 -1.37214649e+00
-3.12845588e-01 6.03901863e-01 -2.41728142e-01 -2.75622010e-01
1.56411624e+00 -2.65641868e-01 -6.67850316e-01 7.30657399e-01
8.86098802e-01 6.12246573e-01 -2.25604311e-01 1.28340691e-01
1.54729709e-01 -2.55214483e-01 -1.23550437e-01 -8.74485910e-01
-8.71536195e-01 6.24636710e-01 6.47291422e-01 7.69936860e-01
9.32127774e-01 -4.39636379e-01 5.82229376e-01 5.07972360e-01
2.26602182e-01 -1.39632344e+00 -4.31151688e-03 1.08552468e+00
6.04939163e-01 -1.30781126e+00 6.24202713e-02 -3.64165336e-01
-9.88747001e-01 1.43743289e+00 1.44886240e-01 -2.37236977e-01
1.11269891e+00 1.12918802e-01 5.35015821e-01 -2.75469899e-01
-1.80079088e-01 -1.69148117e-01 1.40357301e-01 2.05167055e-01
1.06507790e+00 1.54672749e-02 -8.59932244e-01 6.18299246e-01
-5.90601981e-01 -1.83465064e-01 1.11275625e+00 1.43280029e+00
-6.55468345e-01 -1.42623258e+00 -3.69377553e-01 5.14081359e-01
-1.05240500e+00 -1.05289958e-01 -5.99914074e-01 9.01206613e-01
3.57497819e-02 1.39284706e+00 -1.38547435e-01 -2.27330029e-01
3.47738355e-01 1.84046000e-01 2.69814491e-01 -6.50850415e-01
-1.25592601e+00 -1.55735850e-01 2.90993899e-01 1.69266924e-01
-8.73989701e-01 -7.55341589e-01 -1.24169576e+00 -6.10813618e-01
-4.91931587e-01 4.90946591e-01 1.19166291e+00 1.17556167e+00
1.63114369e-01 6.45816103e-02 6.68082058e-01 -5.79397559e-01
4.13797721e-02 -1.31597614e+00 -6.12519085e-01 5.42687178e-01
6.45427555e-02 -4.33688223e-01 -2.92663693e-01 2.18778208e-01] | [11.06828498840332, 6.928679466247559] |
e68804c1-0002-4788-999c-40e4af7aaf5d | survey-on-sparse-coded-features-for-content | 1402.4888 | null | https://arxiv.org/abs/1402.4888v1 | https://arxiv.org/pdf/1402.4888v1.pdf | Survey on Sparse Coded Features for Content Based Face Image Retrieval | Content based image retrieval, a technique which uses visual contents of image to search images from large scale image databases according to users' interests. This paper provides a comprehensive survey on recent technology used in the area of content based face image retrieval. Nowadays digital devices and photo sharing sites are getting more popularity, large human face photos are available in database. Multiple types of facial features are used to represent discriminality on large scale human facial image database. Searching and mining of facial images are challenging problems and important research issues. Sparse representation on features provides significant improvement in indexing related images to query image. | ['D. Johnvictor', 'G. Selvavinayagam'] | 2014-02-20 | null | null | null | null | ['face-image-retrieval'] | ['computer-vision'] | [ 1.91454709e-01 -5.35071433e-01 -5.61753213e-01 -5.74740529e-01
-6.03914797e-01 -2.57764071e-01 3.54394644e-01 -7.69390464e-02
-3.52215797e-01 4.89002436e-01 3.97362530e-01 5.88570654e-01
-6.07967257e-01 -8.88411283e-01 5.25243245e-02 -5.90415359e-01
-1.49206802e-01 5.81341684e-02 2.73749560e-01 -2.32004777e-01
8.10873926e-01 9.94329154e-01 -2.41940808e+00 3.97460639e-01
-6.15301393e-02 1.33573616e+00 2.81137347e-01 4.07116145e-01
-4.54109788e-01 8.00933063e-01 -5.91631591e-01 -1.12179719e-01
2.78549105e-01 -4.48637664e-01 -6.93669558e-01 2.45636523e-01
6.23084843e-01 -3.00173104e-01 -4.48204428e-01 1.19177616e+00
7.94041276e-01 2.38177747e-01 7.55745113e-01 -1.40934253e+00
-9.73152220e-01 -2.19686016e-01 -9.45627570e-01 5.84040105e-01
4.84606713e-01 -7.20778823e-01 3.85393620e-01 -1.04417098e+00
8.25658381e-01 1.68800342e+00 1.71687275e-01 4.51683879e-01
-3.97903204e-01 -8.12204897e-01 -3.76848161e-01 8.35156977e-01
-2.00146580e+00 -8.04581344e-01 8.36597502e-01 -2.91816324e-01
1.05817735e+00 4.42441285e-01 5.44460475e-01 9.31787491e-02
1.56267285e-01 2.13382751e-01 7.43678331e-01 -5.61302722e-01
-7.36764297e-02 2.20686570e-01 -3.97248305e-02 8.74251008e-01
8.65054503e-02 -7.29860663e-02 -1.00513053e+00 -5.16178668e-01
7.26962984e-01 5.46514988e-01 8.08110237e-02 3.86804566e-02
-2.02360988e-01 1.15568876e+00 2.78401554e-01 5.68761408e-01
-5.53287208e-01 -6.28326684e-02 5.31368315e-01 4.91357327e-01
2.82862574e-01 -4.21236664e-01 1.09875925e-01 1.85905099e-01
-1.10219121e+00 3.23894769e-01 3.08574200e-01 9.26193535e-01
1.05053747e+00 -2.74622664e-02 1.90218925e-01 1.38389564e+00
5.97085834e-01 7.89893568e-01 1.17922699e+00 -9.66605127e-01
-2.44782344e-01 7.31699407e-01 -4.40928221e-01 -1.59220994e+00
3.56806405e-02 5.28027773e-01 -5.47310114e-01 -1.37367487e-01
-4.56713349e-01 4.79412407e-01 -7.87059009e-01 8.85468602e-01
5.00266016e-01 -4.54888195e-02 -1.74515456e-01 7.25103080e-01
1.25575399e+00 7.06425965e-01 1.74798131e-01 -4.39529479e-01
1.54827571e+00 -4.62716728e-01 -8.32807004e-01 1.23395741e-01
2.75408700e-02 -1.19771230e+00 3.06085110e-01 2.26988003e-01
-9.90699768e-01 -3.91577572e-01 -5.88551998e-01 -5.90014197e-02
-6.12919629e-01 6.93545640e-02 6.09777808e-01 6.43534005e-01
-1.19855702e+00 1.64550096e-01 -4.11414765e-02 -8.16405714e-01
9.53417122e-01 6.47511601e-01 -8.65657330e-01 -4.03456807e-01
-7.01200962e-01 5.05527616e-01 3.63540202e-01 -2.11315870e-01
-6.06018186e-01 -3.00926328e-01 -9.08148408e-01 -1.61899820e-01
2.03344315e-01 1.87361374e-01 6.63146496e-01 -1.20307612e+00
-1.07198453e+00 1.43496251e+00 -6.66350543e-01 -1.08932078e-01
-5.35661697e-01 2.41102055e-01 -7.72813678e-01 7.93134511e-01
3.55026573e-01 7.80553877e-01 1.13158739e+00 -4.51842457e-01
-4.42457795e-01 -8.57245445e-01 -6.65667236e-01 4.07503955e-02
-7.31717885e-01 9.43487883e-01 -5.64696729e-01 -5.86070180e-01
9.35267732e-02 -5.77875435e-01 1.99835926e-01 3.86784345e-01
2.05469742e-01 -7.02746689e-01 1.49890184e+00 -1.86232597e-01
9.69045877e-01 -1.94424176e+00 -2.40469724e-01 5.61386645e-01
-2.66208738e-01 3.76615673e-01 -3.51753265e-01 8.49385142e-01
-7.42300004e-02 1.48394093e-01 3.82829368e-01 4.16814566e-01
-6.26354396e-01 1.91002384e-01 7.07824379e-02 5.32645583e-01
-3.43056694e-02 6.16180480e-01 -2.56117433e-01 -1.11811340e+00
2.76693642e-01 6.46542549e-01 -4.75996047e-01 -6.29188269e-02
5.22794247e-01 -2.11440787e-01 -9.03035045e-01 1.19006991e+00
6.78408742e-01 -2.07315058e-01 -1.49138883e-01 -1.59952745e-01
9.24082920e-02 -4.98126000e-01 -1.08575261e+00 1.53087950e+00
-1.19153142e-01 6.84476018e-01 4.14015017e-02 -1.35650826e+00
1.01936626e+00 4.79006946e-01 8.04775774e-01 -1.10538375e+00
3.51932734e-01 -6.07813662e-03 -5.21482170e-01 -9.57982719e-01
3.06972891e-01 -1.05774090e-01 5.19976556e-01 4.85167712e-01
2.33625084e-01 2.35466555e-01 4.19021159e-01 1.80601761e-01
4.95202780e-01 -4.21369582e-01 5.67645848e-01 -4.24200088e-01
1.00420702e+00 -2.86192317e-02 2.45156541e-01 4.12092321e-02
-2.15388119e-01 3.90899658e-01 -8.24001580e-02 -7.64422059e-01
-7.58630812e-01 -4.51129019e-01 -3.89037818e-01 1.29485440e+00
-8.07518363e-02 -3.88779521e-01 -5.33887565e-01 -1.34480774e-01
2.52933979e-01 -6.77492261e-01 -7.35386312e-01 7.47312531e-02
-3.62190157e-01 -4.19561267e-01 2.40856037e-01 1.22101456e-01
7.70462930e-01 -1.44459546e+00 -5.44410646e-01 -1.97384328e-01
3.93200368e-01 -7.88508534e-01 -5.31209648e-01 -8.61104429e-01
-7.74330616e-01 -1.53628373e+00 -1.13276911e+00 -1.29900455e+00
1.02170479e+00 7.62508094e-01 8.34557354e-01 5.19107044e-01
-1.38609362e+00 8.10417175e-01 -4.69497085e-01 -5.85273504e-01
8.81391391e-02 -5.17807603e-01 1.37385920e-01 4.13101390e-02
7.91496158e-01 -2.91269332e-01 -8.85011077e-01 2.70613462e-01
-1.08480072e+00 -9.66936648e-01 3.38266075e-01 6.71741307e-01
7.00408757e-01 3.94005537e-01 6.48471653e-01 -7.49446273e-01
5.68274021e-01 -6.69326723e-01 -5.82714319e-01 4.23464924e-01
-5.32048285e-01 -3.60734701e-01 -2.12420613e-01 -1.98191494e-01
-6.98877096e-01 5.43181710e-02 -2.18577702e-02 -5.05377650e-01
2.83917058e-02 5.45642138e-01 1.80752221e-02 -5.87749958e-01
5.56548476e-01 5.60917139e-01 5.67150593e-01 -5.22718012e-01
2.90448889e-02 1.06145501e+00 7.86279738e-02 1.71790216e-02
2.86379665e-01 5.28313696e-01 3.14390153e-01 -1.40130258e+00
-1.83865279e-01 -9.59238112e-01 -3.63775611e-01 -3.33775461e-01
6.11158431e-01 -8.32382083e-01 -8.55780780e-01 3.22370261e-01
-6.39047563e-01 6.87082708e-01 2.35089526e-01 2.80873120e-01
-7.77937919e-02 3.56929243e-01 -3.41757178e-01 -8.90791893e-01
-7.08379328e-01 -8.68044913e-01 1.11416292e+00 5.33995330e-01
2.21457586e-01 -6.59588039e-01 -3.65314148e-02 4.16729331e-01
5.53563476e-01 -1.13643669e-01 6.67672217e-01 -4.37166482e-01
-3.46947372e-01 -4.86671776e-01 -3.65275234e-01 1.49395183e-01
6.83775485e-01 8.58603343e-02 -8.69087815e-01 -3.12175333e-01
-1.51566729e-01 -6.76560104e-01 7.07431972e-01 3.74332577e-01
1.49039769e+00 -7.30413914e-01 -5.52949369e-01 -7.00071678e-02
1.71924472e+00 4.77420270e-01 7.99149096e-01 -1.08757854e-01
1.93333194e-01 8.10594738e-01 5.64350963e-01 6.44084930e-01
-7.47581646e-02 4.64421213e-01 3.40630487e-03 5.34273863e-01
-1.68491919e-02 -9.39971805e-02 -1.60950348e-01 6.34521246e-01
-1.04004093e-01 -8.87072366e-03 -5.76775968e-01 4.92753118e-01
-1.40182126e+00 -1.46138644e+00 4.95429456e-01 1.93566144e+00
7.39986360e-01 -8.40722442e-01 2.62152165e-01 1.96864203e-01
8.56996000e-01 -2.52554733e-02 -4.28602189e-01 -2.40819946e-01
5.81507571e-02 4.71641392e-01 2.08445430e-01 4.14331295e-02
-9.45811272e-01 9.51520681e-01 7.11743402e+00 1.21611619e+00
-1.38386643e+00 9.74815488e-02 6.33160472e-01 -1.72406018e-01
7.45677054e-02 -3.70138824e-01 -9.06291842e-01 4.68496740e-01
7.81205118e-01 -5.32255113e-01 2.38507703e-01 1.12786973e+00
8.57775584e-02 -3.69978607e-01 -3.90148073e-01 1.75569451e+00
7.81035721e-01 -1.52205718e+00 6.06310725e-01 2.83645749e-01
7.53913641e-01 -3.20956856e-01 2.54642159e-01 -2.53577262e-01
-2.58255750e-01 -1.34812438e+00 6.49738312e-02 6.38235211e-01
9.59842682e-01 -9.96272862e-01 3.86131674e-01 -8.01240578e-02
-1.41662514e+00 -7.30611086e-02 -9.33503926e-01 1.46522328e-01
-2.50558555e-01 9.18016490e-03 -2.13868573e-01 -1.46414833e-02
1.07129955e+00 7.40851760e-01 -5.30997872e-01 1.08945680e+00
6.96158171e-01 1.24929152e-01 -2.12983876e-01 -6.62073493e-02
5.93383759e-02 -7.03007802e-02 2.60792933e-02 8.78411114e-01
4.50435907e-01 4.10045236e-01 5.83593324e-02 1.65386021e-01
-1.71596110e-01 8.50779116e-01 -1.13092017e+00 -3.92000496e-01
7.06580818e-01 1.38794112e+00 -7.68549502e-01 -3.07035923e-01
-5.38670659e-01 8.68768215e-01 -7.04453290e-02 -2.68260669e-02
-2.52253693e-02 -5.32462120e-01 5.76401412e-01 3.19888949e-01
2.90216327e-01 2.70987116e-02 9.77094591e-01 -4.89374161e-01
-4.10974652e-01 -9.76035595e-01 6.77338839e-01 -6.10208392e-01
-1.10112667e+00 5.27478576e-01 2.47964144e-01 -1.10994875e+00
-2.83297598e-01 -6.49190009e-01 -3.47604334e-01 7.43636906e-01
-1.52621102e+00 -1.14050603e+00 -2.41642892e-01 1.37306488e+00
8.42240632e-01 -9.76540029e-01 1.00657821e+00 4.83266205e-01
-1.71356976e-01 4.66824472e-01 8.76508653e-02 1.92384392e-01
7.95974314e-01 -2.31600568e-01 -7.23338664e-01 5.22737317e-02
4.79796171e-01 8.37636888e-01 9.22846422e-02 -6.41269803e-01
-1.75210834e+00 -7.50456095e-01 8.19976926e-01 2.41202861e-01
1.73581928e-01 3.09854150e-01 -5.36299765e-01 1.52204618e-01
3.78416002e-01 5.08554637e-01 1.21294236e+00 -3.16345692e-01
-2.00588331e-01 -4.65995997e-01 -1.67202497e+00 8.93497244e-02
7.30026782e-01 -5.91493368e-01 -1.63835064e-02 6.30851865e-01
-2.34238550e-01 3.13231707e-01 -8.67698908e-01 1.44463927e-02
7.43217468e-01 -7.99701631e-01 1.10475242e+00 -5.49681127e-01
4.61637862e-02 1.13081886e-02 -4.89384294e-01 -4.50007230e-01
-2.35433821e-02 -2.41530791e-01 5.06123900e-01 1.17505848e+00
-3.02286863e-01 -3.91827792e-01 9.72487450e-01 5.03543556e-01
7.63922095e-01 -4.87872571e-01 -1.00524259e+00 -4.68992829e-01
-2.90245444e-01 3.32063138e-01 5.46549261e-01 6.79927647e-01
-1.29932180e-01 3.32306772e-02 -4.23991382e-01 -4.63668555e-01
7.50358462e-01 1.81969985e-01 2.32620135e-01 -1.49630415e+00
4.37022895e-01 -4.05778408e-01 -1.12717891e+00 -2.34095141e-01
2.51062095e-01 -7.69762099e-01 -6.72000349e-01 -1.21882975e+00
6.01593554e-01 -1.27856329e-01 -3.06511372e-01 5.49082160e-01
3.44195634e-01 1.15368903e+00 4.95803691e-02 6.14312887e-01
-7.05284238e-01 3.57704997e-01 9.97126698e-01 -2.34637409e-01
2.89404750e-01 -3.92483592e-01 -5.28452516e-01 6.31083250e-01
7.65908480e-01 -3.20196897e-01 -7.58560956e-01 -1.08272016e-01
1.30895153e-01 9.20920894e-02 1.36307389e-01 -7.41134644e-01
4.15461063e-01 -4.50404793e-01 7.29950368e-01 -5.50331354e-01
6.44086003e-01 -8.00117373e-01 9.45138410e-02 2.73035526e-01
-2.36777738e-01 2.24461049e-01 1.32880628e-01 4.84148949e-01
-7.13109672e-01 -6.86082959e-01 9.30670321e-01 -6.72587514e-01
-1.29689312e+00 7.01547027e-01 -5.47506988e-01 -3.01466912e-01
1.05984211e+00 -4.03532743e-01 9.43816826e-02 -4.40535396e-01
-5.19004285e-01 -1.28118902e-01 1.15621895e-01 5.85719943e-01
1.20499325e+00 -1.62708127e+00 -5.01042724e-01 6.02305114e-01
2.81754047e-01 -8.07106078e-01 3.94733548e-01 8.19032788e-02
-7.38424659e-01 6.68189466e-01 -6.97236300e-01 -2.42438674e-01
-2.14748192e+00 5.49896777e-01 3.80543023e-02 6.42533302e-01
-2.55543947e-01 1.06925309e+00 -2.03823820e-01 4.06011671e-01
-1.28252024e-03 7.25160480e-01 -8.40824902e-01 3.77531916e-01
1.26649630e+00 4.29670483e-01 -7.04210103e-02 -1.49750173e+00
-5.50291657e-01 1.28758192e+00 -2.19899401e-01 -7.69124776e-02
1.18583190e+00 -2.25753397e-01 -6.60922050e-01 6.62064645e-03
1.94432354e+00 -1.81362763e-01 -1.94611818e-01 -3.82029712e-01
-1.33646578e-01 -1.02389991e+00 2.31362030e-01 -3.37788165e-01
-1.50286162e+00 6.94657981e-01 1.25869024e+00 6.91517815e-02
1.26540661e+00 1.61419466e-01 4.32963103e-01 6.39431179e-01
5.89646757e-01 -1.40298152e+00 3.00711721e-01 -2.51500383e-02
1.17783546e+00 -1.68665278e+00 4.42808300e-01 -3.84240448e-01
-4.45601314e-01 1.16813767e+00 2.32987970e-01 -3.86753798e-01
1.49110985e+00 9.91044268e-02 1.54583529e-01 -6.54094636e-01
-5.13732791e-01 -2.38214552e-01 3.42090368e-01 5.79694211e-01
5.18102705e-01 -5.79094708e-01 -6.37103856e-01 -1.21614829e-01
-6.28802404e-02 1.15825847e-01 -3.41513604e-01 1.35705280e+00
-9.88711238e-01 -1.47010410e+00 -5.18574774e-01 6.74433351e-01
-9.88389790e-01 -2.76462063e-02 -2.99060881e-01 3.73592049e-01
-1.35553315e-01 1.05894136e+00 1.33429021e-01 1.34027019e-01
-2.73953080e-01 5.85459247e-02 7.48389781e-01 -5.06644785e-01
-1.63259096e-02 4.76924442e-02 -5.63533664e-01 -6.47314966e-01
-8.84762466e-01 -3.53090912e-01 -1.02978659e+00 -2.84212232e-01
-2.25808159e-01 3.80206615e-01 1.02641118e+00 4.04979020e-01
5.33940434e-01 -3.34966958e-01 8.74723375e-01 -6.93989217e-01
4.03081365e-02 -9.14108396e-01 -1.01970363e+00 3.81795913e-01
7.57855847e-02 -8.26748013e-01 1.50434390e-01 4.75801826e-01] | [11.184035301208496, 0.36685362458229065] |
1166a47a-8705-4cb3-8454-fb3cb73f4cad | z-gmot-zero-shot-generic-multiple-object | 2305.17648 | null | https://arxiv.org/abs/2305.17648v1 | https://arxiv.org/pdf/2305.17648v1.pdf | Z-GMOT: Zero-shot Generic Multiple Object Tracking | Despite the significant progress made in recent years, Multi-Object Tracking (MOT) approaches still suffer from several limitations, including their reliance on prior knowledge of tracking targets, which necessitates the costly annotation of large labeled datasets. As a result, existing MOT methods are limited to a small set of predefined categories, and they struggle with unseen objects in the real world. To address these issues, Generic Multiple Object Tracking (GMOT) has been proposed, which requires less prior information about the targets. However, all existing GMOT approaches follow a one-shot paradigm, relying mainly on the initial bounding box and thus struggling to handle variants e.g., viewpoint, lighting, occlusion, scale, and etc. In this paper, we introduce a novel approach to address the limitations of existing MOT and GMOT methods. Specifically, we propose a zero-shot GMOT (Z-GMOT) algorithm that can track never-seen object categories with zero training examples, without the need for predefined categories or an initial bounding box. To achieve this, we propose iGLIP, an improved version of Grounded language-image pretraining (GLIP), which can detect unseen objects while minimizing false positives. We evaluate our Z-GMOT thoroughly on the GMOT-40 dataset, AnimalTrack testset, DanceTrack testset. The results of these evaluations demonstrate a significant improvement over existing methods. For instance, on the GMOT-40 dataset, the Z-GMOT outperforms one-shot GMOT with OC-SORT by 27.79 points HOTA and 44.37 points MOTA. On the AnimalTrack dataset, it surpasses fully-supervised methods with DeepSORT by 12.55 points HOTA and 8.97 points MOTA. To facilitate further research, we will make our code and models publicly available upon acceptance of this paper. | ['Ngan Hoang Le', 'Donald Adjeroh', 'Khoa Luu', 'Thinh Phan', 'Pha Nguyen', 'Anh Duy Le Dinh', 'Tien-Phat Nguyen', 'Kim Hoang Tran'] | 2023-05-28 | null | null | null | null | ['object-tracking', 'multiple-object-tracking', 'multi-object-tracking'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [-5.72218858e-02 -3.71640384e-01 -1.42591327e-01 -1.78395793e-01
-7.42856145e-01 -6.74520075e-01 4.96206075e-01 -2.11343896e-02
-5.97880602e-01 6.33916020e-01 -3.51458192e-01 5.47416359e-02
-7.10228235e-02 -5.58382690e-01 -9.20422673e-01 -6.94384038e-01
5.18128835e-02 6.43021703e-01 1.11874044e+00 -9.95998159e-02
-1.78660860e-03 3.16482753e-01 -1.76883566e+00 1.90807268e-01
6.29994035e-01 8.95825744e-01 2.76784539e-01 4.51300651e-01
-9.76341814e-02 4.67981786e-01 -7.42445111e-01 -3.69688720e-01
3.83394450e-01 -3.92554671e-01 -4.75573510e-01 -9.90240723e-02
9.13438201e-01 -3.07997316e-01 -8.55507851e-02 1.10278761e+00
5.49202204e-01 1.75915435e-01 3.54450077e-01 -1.60528171e+00
-4.13895905e-01 3.78021240e-01 -5.87909102e-01 2.92387187e-01
5.42694330e-03 2.97385126e-01 7.96887517e-01 -8.47654462e-01
5.45127690e-01 1.33066249e+00 1.05151379e+00 7.50460386e-01
-1.12955046e+00 -9.47287321e-01 3.05753052e-01 8.80786031e-02
-1.56575871e+00 -3.39424998e-01 2.79819697e-01 -5.11241317e-01
6.52132988e-01 3.05693835e-01 5.57692647e-01 9.56643522e-01
1.38122961e-01 8.83552253e-01 1.00559556e+00 -2.54697770e-01
8.39530826e-02 1.19814433e-01 1.61824465e-01 7.01910734e-01
4.99325454e-01 3.54632288e-01 -3.26753825e-01 3.66656519e-02
4.69553977e-01 2.42398366e-01 -2.47501321e-02 -7.25409269e-01
-1.25322020e+00 5.95512569e-01 5.28272569e-01 3.72713983e-01
1.98571049e-02 2.91609675e-01 2.79623747e-01 -3.34495381e-02
2.63958961e-01 9.92907286e-02 -2.96006382e-01 3.26215923e-02
-9.67761397e-01 1.65081888e-01 4.95328128e-01 1.16868412e+00
8.29303980e-01 5.49740903e-02 -3.93047333e-01 5.42780340e-01
3.70925218e-01 7.78883398e-01 4.48894024e-01 -5.91126084e-01
2.84553170e-01 6.67165577e-01 3.03983480e-01 -9.14915621e-01
-5.07885396e-01 -5.51403999e-01 -3.69753748e-01 2.32062906e-01
5.27826846e-01 -1.00641817e-01 -1.21964931e+00 1.73227239e+00
6.76378369e-01 3.03869903e-01 -9.63783860e-02 1.11927843e+00
9.22450960e-01 4.58123147e-01 2.18090743e-01 6.15414046e-02
1.29078972e+00 -1.11573577e+00 -6.67092443e-01 -4.07400489e-01
6.97057307e-01 -6.76903665e-01 1.08349872e+00 2.86856294e-01
-6.16058528e-01 -7.54761934e-01 -9.02396977e-01 2.92436630e-01
-5.52481949e-01 2.73359954e-01 5.56301832e-01 7.81184494e-01
-8.25495481e-01 4.26287234e-01 -9.75599170e-01 -6.25139296e-01
4.40850347e-01 4.25262541e-01 -8.24171528e-02 -1.56511217e-01
-7.67876804e-01 7.11157918e-01 7.10663974e-01 6.19831830e-02
-1.29766345e+00 -5.83080590e-01 -6.06557906e-01 -1.20757356e-01
8.14800858e-01 -3.86184603e-01 1.27587402e+00 -8.48839879e-01
-1.20014501e+00 6.62297904e-01 -1.85787696e-02 -3.99335980e-01
6.92105055e-01 -6.13313735e-01 -4.83675331e-01 -6.13904148e-02
3.83723080e-01 7.98925519e-01 6.54699504e-01 -1.33043993e+00
-8.36520016e-01 -1.92001626e-01 3.40174623e-02 5.78962825e-02
-3.50840986e-01 1.77978441e-01 -6.95369959e-01 -6.32535756e-01
-1.27758011e-01 -1.11871779e+00 -1.15358204e-01 2.74026811e-01
-3.62318397e-01 -2.37733394e-01 1.22700012e+00 -1.24743424e-01
1.02785885e+00 -2.25147986e+00 -4.35512438e-02 -3.74180496e-01
5.09418771e-02 5.88867009e-01 -1.86162576e-01 1.75727352e-01
4.17671770e-01 -1.57957286e-01 -1.35252450e-03 -5.24258137e-01
5.19323088e-02 3.75236511e-01 -9.69327539e-02 4.90786940e-01
-4.28522602e-02 9.04689848e-01 -1.02646482e+00 -7.22012639e-01
3.85617167e-01 3.03636849e-01 -4.87787813e-01 -2.82038953e-02
-4.78184074e-01 4.68918383e-01 -3.71548116e-01 9.56701219e-01
6.12238765e-01 -3.33135694e-01 -1.36108071e-01 -1.12322241e-01
-4.36217010e-01 -2.14367941e-01 -1.33630490e+00 1.59546208e+00
-2.61591803e-02 4.78161842e-01 -2.17038780e-01 -6.17920399e-01
7.82934129e-01 1.09717168e-01 6.21205509e-01 -5.75832486e-01
2.84407526e-01 1.72831997e-01 -1.66290682e-02 -1.95129171e-01
5.34361899e-01 -1.46608025e-01 -1.05517276e-01 8.00218508e-02
9.70924646e-02 2.64303476e-01 4.10473287e-01 2.31136531e-01
1.02619207e+00 3.25537384e-01 8.99716839e-02 -1.20473780e-01
2.72390842e-01 4.12445545e-01 9.12242591e-01 9.92749572e-01
-5.39181173e-01 6.66720748e-01 -1.71548560e-01 -5.01222193e-01
-5.98856151e-01 -1.00464392e+00 -7.59453028e-02 1.11715508e+00
6.77917421e-01 -5.06526709e-01 -4.71281826e-01 -8.17483962e-01
1.39056653e-01 4.64791000e-01 -4.02935386e-01 -1.95291545e-02
-6.37830198e-01 -8.17406654e-01 5.83369374e-01 5.63446879e-01
6.42124772e-01 -9.90004897e-01 -8.48387539e-01 1.95083499e-01
-2.21843958e-01 -1.34597874e+00 -4.22459871e-01 4.19971682e-02
-7.35496819e-01 -1.13345909e+00 -7.15941727e-01 -4.82160985e-01
5.30915976e-01 6.46630526e-01 8.65647972e-01 2.74224132e-01
-3.53216141e-01 3.90342504e-01 -4.26263243e-01 -6.63062632e-01
-5.76324351e-02 -1.10762633e-01 1.48566708e-01 -1.21014072e-02
4.43804830e-01 -6.97131231e-02 -5.37806690e-01 8.42761278e-01
-6.58366263e-01 5.92792295e-02 7.82399178e-01 6.13193929e-01
8.00800920e-01 -5.84026463e-02 3.34942430e-01 -5.50252676e-01
-1.61626995e-01 -3.99468958e-01 -8.99012625e-01 3.53879631e-01
-3.45376432e-01 -3.27940464e-01 5.16734183e-01 -8.64337683e-01
-7.30124950e-01 3.34759086e-01 6.20189644e-02 -7.47447968e-01
-1.71399727e-01 4.01063338e-02 -2.03058451e-01 -4.82542336e-01
6.60258174e-01 3.66986468e-02 -4.34117943e-01 -5.43808818e-01
1.80953711e-01 3.62815887e-01 5.97832501e-01 -4.28901523e-01
1.06609809e+00 6.29518509e-01 -1.49088591e-01 -6.51605248e-01
-1.17979705e+00 -7.76763439e-01 -6.63817763e-01 -5.42113543e-01
8.60652745e-01 -9.59855616e-01 -6.23955369e-01 3.63221645e-01
-8.16277862e-01 -4.11468923e-01 -2.06991196e-01 6.81896150e-01
-2.92797536e-01 3.70850354e-01 -2.29160830e-01 -8.02099943e-01
-1.66781947e-01 -9.47173774e-01 1.24052560e+00 3.81970376e-01
2.56412365e-02 -6.42203867e-01 -5.80213331e-02 2.61469424e-01
3.04402888e-01 3.82118255e-01 2.48138756e-01 -6.59520686e-01
-1.00600541e+00 -1.87167153e-01 -1.75628662e-01 -3.91399264e-02
6.36543483e-02 -1.74913242e-01 -7.44771659e-01 -7.11569786e-01
-3.46811742e-01 -3.98953289e-01 7.50973463e-01 2.65651166e-01
8.50766420e-01 1.74436197e-02 -7.96990752e-01 6.17040694e-01
1.57472646e+00 3.14565986e-01 2.38098875e-01 5.57181358e-01
7.85659134e-01 1.71251416e-01 1.11604047e+00 3.80988866e-01
2.77292728e-01 9.46177542e-01 7.56478727e-01 -1.35787334e-02
-3.26855749e-01 -2.88011461e-01 3.35568368e-01 3.95290077e-01
-7.18071833e-02 -4.77870524e-01 -1.00494063e+00 7.72821486e-01
-1.94801497e+00 -8.62886310e-01 -4.41265583e-01 2.12637854e+00
4.01497364e-01 3.16179603e-01 4.18981016e-01 -6.31918535e-02
8.24045300e-01 -8.07551816e-02 -5.72365880e-01 3.70765567e-01
-4.37560491e-02 -1.18953906e-01 6.23415589e-01 1.69238552e-01
-1.33204162e+00 1.11712921e+00 5.67623711e+00 8.02066207e-01
-1.02400959e+00 3.95657331e-01 -2.21339792e-01 -2.04462290e-01
3.22967559e-01 6.64299577e-02 -1.50942683e+00 6.05145335e-01
7.26812303e-01 4.61533070e-02 1.16158627e-01 1.00590110e+00
-1.45097449e-01 -1.57695398e-01 -8.83307576e-01 9.24446762e-01
1.43767342e-01 -8.45240414e-01 -1.15203820e-01 -1.70386359e-01
5.90210617e-01 2.67780960e-01 -8.26535746e-02 7.65564799e-01
5.04567385e-01 -5.73080778e-01 9.69903529e-01 2.55685836e-01
5.52801907e-01 -2.85796434e-01 7.56150246e-01 6.23062849e-01
-1.53470469e+00 -1.84804738e-01 -5.09917259e-01 1.81465954e-01
1.25676706e-01 2.89770097e-01 -6.61210716e-01 7.77616560e-01
1.15275431e+00 5.74430287e-01 -8.44154537e-01 1.58422947e+00
-5.40921316e-02 5.83873987e-01 -7.25746095e-01 -1.54289857e-01
3.21058363e-01 3.32085669e-01 6.87075973e-01 1.08236957e+00
3.73087376e-01 6.29232228e-02 7.21131682e-01 5.91162264e-01
1.53251469e-01 -2.61402335e-02 -3.21627349e-01 2.67529413e-02
3.65877450e-01 1.31936061e+00 -1.07239056e+00 -4.25045311e-01
-5.48952699e-01 7.32929647e-01 1.56707972e-01 1.44935876e-01
-1.24782610e+00 -1.78514361e-01 3.19851667e-01 4.53855135e-02
6.12078190e-01 -2.02785775e-01 1.63581163e-01 -1.05938244e+00
-1.47539442e-02 -7.01363981e-01 6.70776427e-01 -6.90334141e-01
-1.00623310e+00 6.73690677e-01 2.90774375e-01 -1.52940667e+00
1.63842216e-01 -5.40640473e-01 -3.66399735e-01 3.21720332e-01
-1.42121971e+00 -1.38660514e+00 -6.84666872e-01 6.70078397e-01
5.79683125e-01 -3.89966927e-02 4.82794285e-01 7.07316160e-01
-6.54999554e-01 5.94575584e-01 7.74418935e-02 1.57786801e-01
8.12064946e-01 -1.07537389e+00 2.15148151e-01 9.88818884e-01
3.08262527e-01 4.28968549e-01 7.51111925e-01 -8.65049005e-01
-1.24836683e+00 -1.42081904e+00 5.74924648e-01 -6.94680750e-01
5.96086204e-01 -4.89940524e-01 -8.95124376e-01 1.01743591e+00
-1.19341493e-01 4.69472587e-01 4.25049365e-01 9.27884504e-02
-7.37388730e-02 -1.41469628e-01 -9.89560127e-01 4.26255077e-01
1.32309043e+00 5.92585132e-02 -5.77625453e-01 4.55885142e-01
5.18108368e-01 -6.61410153e-01 -6.41151607e-01 7.08369792e-01
4.14606124e-01 -9.23242688e-01 8.93115282e-01 -3.40237051e-01
-2.90796638e-01 -8.31613123e-01 -7.07478896e-02 -9.71918046e-01
-3.48981172e-01 -4.97203380e-01 -1.84364557e-01 1.26665616e+00
5.41229248e-02 -4.35792893e-01 8.91176939e-01 1.15152925e-01
-4.69760656e-01 -4.68216538e-01 -1.03558517e+00 -1.35291731e+00
-2.17915103e-01 -2.49560848e-01 5.22065163e-01 7.95344412e-01
-6.37623668e-01 9.23286229e-02 -6.20135188e-01 3.67353886e-01
8.60714436e-01 2.57375389e-01 1.19308996e+00 -1.33904612e+00
-2.34736785e-01 -4.25568596e-02 -5.11864185e-01 -1.07866859e+00
-3.35738778e-01 -7.29305685e-01 3.65912169e-01 -1.50516236e+00
3.20220023e-01 -6.05369151e-01 -3.80331188e-01 8.33793342e-01
-2.10540891e-01 5.18097460e-01 4.98504490e-01 4.39808816e-01
-1.28312504e+00 5.80816686e-01 1.15884864e+00 -1.36482507e-01
-9.07569006e-02 1.80157612e-03 -2.77569532e-01 7.17424929e-01
7.33651817e-01 -9.62572455e-01 -2.23702222e-01 -5.27639031e-01
-2.60373086e-01 -2.72057742e-01 7.69452214e-01 -1.57979369e+00
4.18560594e-01 -1.89385787e-01 3.81738305e-01 -1.15258920e+00
4.52839017e-01 -9.34361100e-01 3.73869479e-01 6.85752869e-01
2.26955429e-01 3.94149236e-02 5.13995528e-01 6.35419428e-01
-5.59815951e-03 -1.96932003e-01 7.70432830e-01 -6.46040514e-02
-1.13021123e+00 3.63288283e-01 -7.38513023e-02 9.04699117e-02
1.41093719e+00 -4.38519865e-01 -4.24998820e-01 1.57440409e-01
-6.29823506e-01 5.15653193e-01 4.55127418e-01 6.62114918e-01
3.68814349e-01 -1.35052884e+00 -4.21128213e-01 4.66325432e-02
2.43835881e-01 5.45062833e-02 1.58134520e-01 9.13244843e-01
-1.56261548e-01 4.41222817e-01 -2.47582003e-01 -1.00837278e+00
-1.55771136e+00 7.74586618e-01 3.23474973e-01 -8.90813470e-02
-7.90683568e-01 6.15803301e-01 3.13034534e-01 -3.15076977e-01
4.13126588e-01 -3.54309231e-01 -6.84642196e-02 -1.93873629e-01
4.46904778e-01 2.36526772e-01 -5.00222482e-02 -6.67794228e-01
-5.25368690e-01 8.31546605e-01 -1.33935004e-01 1.22650504e-01
1.03458977e+00 2.42988709e-02 4.69557256e-01 4.87659693e-01
5.73039174e-01 -8.83068703e-03 -1.43310714e+00 -2.18968377e-01
2.34574974e-01 -6.71125710e-01 -2.11960450e-01 -8.69576395e-01
-9.52319145e-01 7.14940667e-01 9.29509282e-01 -1.22326054e-02
8.69740725e-01 1.07460216e-01 7.44242430e-01 4.26952451e-01
7.20714986e-01 -8.22571993e-01 3.98218960e-01 4.61001039e-01
5.37623644e-01 -1.36509573e+00 7.63497781e-03 -3.23188335e-01
-4.23916340e-01 8.14052343e-01 1.18856263e+00 1.88477054e-01
3.00743848e-01 2.71468312e-01 1.95100605e-01 -3.07349145e-01
-4.32264775e-01 -6.54577553e-01 2.98099428e-01 5.03447652e-01
2.32060049e-02 -1.55344993e-01 -1.19492523e-01 3.35173070e-01
3.61388698e-02 1.42179608e-01 4.83891107e-02 1.19409883e+00
-6.43331409e-01 -8.66985023e-01 -7.01977491e-01 1.28305614e-01
-4.62749183e-01 2.63930857e-01 -4.38709795e-01 1.27563667e+00
4.48194474e-01 9.07053292e-01 -1.43207476e-01 -3.34601194e-01
5.85030675e-01 -2.46364191e-01 4.37576085e-01 -7.39939153e-01
-6.05977893e-01 1.91971734e-01 -1.20232083e-01 -4.99750495e-01
-6.70209885e-01 -7.10657477e-01 -1.31944406e+00 -1.38598815e-01
-8.08497965e-01 1.36565447e-01 4.18317646e-01 8.62028003e-01
2.37288430e-01 6.07656598e-01 -1.06930146e-02 -8.42681229e-01
-4.01201606e-01 -9.41816151e-01 -2.05483288e-01 5.03117740e-01
3.05207998e-01 -1.21931314e+00 -2.56741613e-01 -1.59684449e-01] | [6.375926494598389, -2.02769136428833] |
a7193f7b-35c2-4007-99be-953377302844 | polarmot-how-far-can-geometric-relations-take | 2208.01957 | null | https://arxiv.org/abs/2208.01957v1 | https://arxiv.org/pdf/2208.01957v1.pdf | PolarMOT: How Far Can Geometric Relations Take Us in 3D Multi-Object Tracking? | Most (3D) multi-object tracking methods rely on appearance-based cues for data association. By contrast, we investigate how far we can get by only encoding geometric relationships between objects in 3D space as cues for data-driven data association. We encode 3D detections as nodes in a graph, where spatial and temporal pairwise relations among objects are encoded via localized polar coordinates on graph edges. This representation makes our geometric relations invariant to global transformations and smooth trajectory changes, especially under non-holonomic motion. This allows our graph neural network to learn to effectively encode temporal and spatial interactions and fully leverage contextual and motion cues to obtain final scene interpretation by posing data association as edge classification. We establish a new state-of-the-art on nuScenes dataset and, more importantly, show that our method, PolarMOT, generalizes remarkably well across different locations (Boston, Singapore, Karlsruhe) and datasets (nuScenes and KITTI). | ['Laura Leal-Taixé', 'Aljoša Ošep', 'Guillem Brasó', 'Aleksandr Kim'] | 2022-08-03 | null | null | null | null | ['3d-multi-object-tracking'] | ['computer-vision'] | [-7.97247067e-02 -2.72383213e-01 -2.53425628e-01 -1.19106174e-01
-1.36745945e-01 -8.97887588e-01 8.66484642e-01 2.46193171e-01
-1.38062954e-01 2.42532268e-01 1.40205294e-01 -2.30920330e-01
-5.08056402e-01 -6.66756213e-01 -8.39299440e-01 -4.97187555e-01
-6.03293717e-01 5.84268272e-01 3.54117692e-01 -9.62487981e-02
2.59228004e-03 9.81538951e-01 -1.36809814e+00 -1.07059255e-01
2.85740733e-01 7.61328161e-01 -1.24428883e-01 1.02188015e+00
4.05980051e-01 4.51388389e-01 -7.32711330e-02 -2.25717396e-01
5.51246703e-01 -5.10665663e-02 -5.51656604e-01 2.55984277e-01
1.11663187e+00 -1.39953550e-02 -8.52113605e-01 8.10383797e-01
2.13155925e-01 2.85124838e-01 8.36648941e-01 -1.51578546e+00
-1.30468905e+00 -4.14217860e-02 -7.61172056e-01 2.07165971e-01
3.54169607e-01 3.17101598e-01 1.27355385e+00 -8.93174291e-01
1.11631870e+00 1.36124778e+00 8.51948917e-01 2.92317361e-01
-1.29997885e+00 -2.65500069e-01 4.53962773e-01 2.31492028e-01
-1.30540395e+00 -2.58200914e-01 8.80723417e-01 -6.56290472e-01
9.97245729e-01 4.29023355e-01 1.02460456e+00 1.04138672e+00
3.41706336e-01 6.17686212e-01 5.35181284e-01 -3.52609277e-01
-1.03731737e-01 -4.49514151e-01 2.03219861e-01 9.86203432e-01
6.49895251e-01 2.76585907e-01 -5.74497819e-01 -8.65995437e-02
1.00638556e+00 1.64360136e-01 -3.38289589e-02 -1.06678617e+00
-1.51798606e+00 5.69802642e-01 9.09499168e-01 6.01301230e-02
-2.25395545e-01 6.03687048e-01 1.30341277e-01 1.17328949e-01
4.89393830e-01 2.82097101e-01 -2.53315747e-01 3.67023885e-01
-2.06317186e-01 4.79103386e-01 5.16519010e-01 1.40161705e+00
6.86873436e-01 -8.42001066e-02 -1.25270143e-01 2.66187012e-01
5.03982604e-01 7.06292808e-01 -3.51410598e-01 -1.02717578e+00
3.25670868e-01 6.48303092e-01 7.28286952e-02 -1.37179685e+00
-7.63647676e-01 -2.68992156e-01 -7.66209245e-01 2.90111125e-01
3.71356219e-01 -4.64745015e-02 -8.35275054e-01 1.89904165e+00
5.88094056e-01 5.20209908e-01 -2.87338644e-01 1.07762814e+00
8.42146456e-01 1.34489924e-01 -1.07250370e-01 -5.34217432e-02
1.33211792e+00 -8.36062491e-01 -5.09280324e-01 -2.05273226e-01
7.85338879e-01 -4.22755837e-01 6.66034400e-01 -1.11497752e-01
-9.81279910e-01 -4.94482338e-01 -1.00353205e+00 -2.23323092e-01
-5.12903690e-01 -1.60200134e-01 9.35656905e-01 2.72907376e-01
-1.27458823e+00 3.34455252e-01 -1.02207220e+00 -6.42533481e-01
3.68942112e-01 4.69875962e-01 -5.88124573e-01 6.11028895e-02
-8.54490519e-01 8.36494088e-01 2.64891833e-01 1.25844315e-01
-5.80928087e-01 -6.72134757e-01 -1.06139815e+00 -4.58294272e-01
3.81016850e-01 -1.23953986e+00 7.56707191e-01 -2.56421983e-01
-9.97594357e-01 1.07379520e+00 -1.46358222e-01 -3.51753473e-01
4.22773927e-01 -1.26069203e-01 -4.52043563e-01 9.50378180e-02
-3.65641713e-02 8.92109275e-01 5.43285549e-01 -1.40063834e+00
-5.25835812e-01 -5.83165824e-01 2.55830109e-01 4.01852459e-01
-8.70040581e-02 -1.25493422e-01 -6.93949401e-01 -5.62423110e-01
3.36442232e-01 -1.38177335e+00 -3.08557183e-01 5.55386066e-01
-5.76661825e-01 -3.31008613e-01 1.21774709e+00 -1.70793623e-01
8.33129287e-01 -1.98554099e+00 5.35710633e-01 1.54993579e-01
5.74716032e-01 -1.81287482e-01 -2.47099668e-01 2.16116399e-01
-1.48319956e-02 9.32797492e-02 7.37767294e-02 -4.65161115e-01
1.20637916e-01 3.96014810e-01 -4.29549851e-02 9.06138122e-01
4.42180872e-01 1.30135846e+00 -1.20131922e+00 -4.12821531e-01
5.04666626e-01 5.34300864e-01 -7.55934596e-01 -1.53201520e-01
-4.24397558e-01 5.37768364e-01 -6.15214050e-01 7.04314649e-01
5.91438174e-01 -4.57874030e-01 -1.15763694e-01 -3.59674305e-01
-1.89457178e-01 -3.53913121e-02 -1.13620603e+00 1.94965804e+00
-8.33870843e-02 8.51913154e-01 -6.53488785e-02 -7.06825793e-01
7.62035847e-01 6.42017508e-03 9.11833644e-01 -4.75302368e-01
6.58386722e-02 -2.10409775e-01 -1.79060712e-01 -4.54045385e-01
7.78930187e-01 4.35265869e-01 -9.76365134e-02 2.70005882e-01
-2.20768098e-02 2.33748369e-02 4.36966941e-02 3.22697192e-01
1.13912344e+00 4.24409181e-01 8.12607110e-02 -4.11010385e-01
5.56011871e-02 1.80379823e-01 4.29866046e-01 7.18836486e-01
-2.81422615e-01 5.97672105e-01 1.89465970e-01 -6.70713246e-01
-1.08733511e+00 -1.39876091e+00 -3.73701341e-02 8.75308096e-01
6.30539298e-01 -4.64250326e-01 -5.47863729e-02 -7.02253163e-01
4.17718440e-01 3.08126211e-01 -7.37154901e-01 -2.49629781e-01
-8.59235346e-01 -5.22082329e-01 2.90714592e-01 6.52827382e-01
1.74942628e-01 -5.88433444e-01 -4.53172922e-01 9.97126028e-02
1.91361308e-01 -1.36569059e+00 -7.01889038e-01 5.52002247e-03
-8.27729762e-01 -1.14046991e+00 -3.03479940e-01 -5.50825238e-01
7.68537104e-01 7.62594342e-01 1.13060212e+00 -7.97295570e-03
-4.52795118e-01 7.60663807e-01 -1.61645502e-01 -2.15423241e-01
2.37293672e-02 -1.98331222e-01 4.48776245e-01 -7.35596195e-02
2.26917997e-01 -5.92380166e-01 -6.61798656e-01 4.32692856e-01
-4.23666209e-01 1.77572906e-01 3.04391205e-01 3.99932086e-01
6.85759246e-01 -3.62497360e-01 4.82074916e-02 -5.84140420e-01
4.85068150e-02 -4.60749775e-01 -7.45529532e-01 2.68117487e-01
-7.15875551e-02 7.82644823e-02 1.46560714e-01 -4.94109094e-01
-5.96639752e-01 2.64283568e-01 4.79050308e-01 -8.51239681e-01
-1.61006421e-01 2.54900277e-01 -5.54958032e-03 -4.37829554e-01
6.74576283e-01 -3.70958567e-01 -2.92141527e-01 -3.45783889e-01
8.78061175e-01 -1.50721855e-02 6.77627742e-01 -6.31812990e-01
1.01401794e+00 8.79364014e-01 7.33108222e-01 -8.88751268e-01
-7.40655601e-01 -4.92106438e-01 -1.29872179e+00 -4.77910906e-01
1.33351386e+00 -8.49266768e-01 -1.11414659e+00 2.70845085e-01
-1.19375753e+00 -3.94715399e-01 -1.29455671e-01 5.81369877e-01
-5.89721978e-01 2.25759491e-01 -3.31229687e-01 -7.73505211e-01
1.99009135e-01 -7.43258238e-01 1.36773634e+00 5.67361116e-02
-2.76461124e-01 -1.35356808e+00 6.99800551e-02 -1.18554793e-01
-5.51064946e-02 8.28733027e-01 7.83427775e-01 -1.67878225e-01
-1.13096833e+00 -8.54774714e-02 -3.91608685e-01 -5.53465545e-01
2.75559157e-01 2.71758288e-01 -5.88821948e-01 -3.51181269e-01
-4.62064832e-01 1.45018354e-01 8.69242430e-01 5.97587645e-01
9.50180233e-01 -1.83026791e-01 -7.61683464e-01 8.36002171e-01
1.21290529e+00 -1.06472671e-01 2.33870968e-02 2.00696394e-01
1.31193292e+00 4.66973692e-01 5.05221605e-01 2.34520093e-01
6.91554964e-01 9.75525081e-01 7.67819464e-01 -1.05495550e-01
-4.66932654e-01 -2.25372374e-01 2.00051129e-01 7.42167592e-01
-4.15875256e-01 -5.04771709e-01 -1.07214844e+00 5.66094756e-01
-2.22291040e+00 -8.22879195e-01 -7.10099876e-01 2.04705620e+00
2.38826871e-01 1.73984453e-01 1.75587744e-01 -5.17795920e-01
7.79035926e-01 3.26247066e-01 -7.06715584e-01 1.68959215e-01
-3.79712492e-01 -2.51379848e-01 8.23221207e-01 4.98324335e-01
-1.43979347e+00 9.17217672e-01 6.73401785e+00 1.87773064e-01
-8.32653642e-01 -4.00768267e-03 1.40310660e-01 -1.98002130e-01
-3.31703216e-01 -1.89971812e-02 -8.76233041e-01 -8.51119235e-02
4.77501065e-01 -7.28130266e-02 2.86002606e-01 5.17385781e-01
7.06378445e-02 1.98423982e-01 -1.40869975e+00 9.94201124e-01
2.99751516e-02 -1.65725338e+00 -1.57087166e-02 2.24297881e-01
9.47116554e-01 2.63424218e-01 2.12882265e-01 -2.60928094e-01
7.52537489e-01 -7.36733794e-01 9.31923151e-01 6.25762820e-01
5.51026881e-01 -2.05118626e-01 -1.52523685e-02 6.19553914e-03
-1.80427516e+00 2.15302020e-01 -1.33245766e-01 -1.05986319e-01
3.49725425e-01 2.28681356e-01 -6.25312865e-01 9.22765076e-01
5.11542559e-01 1.22228634e+00 -6.77525461e-01 1.01053119e+00
-1.28325328e-01 -1.17224818e-02 -5.32902360e-01 -1.41350264e-02
2.60652602e-01 -3.24561954e-01 9.61557031e-01 1.05778432e+00
3.26245368e-01 8.08675364e-02 4.35154200e-01 8.62565577e-01
-3.67368646e-02 -5.21326482e-01 -1.06288028e+00 1.09067895e-01
4.18862849e-01 1.15922678e+00 -8.98777068e-01 -1.56175762e-01
-5.55940926e-01 7.65863121e-01 4.25460130e-01 6.56269789e-01
-9.43131089e-01 2.08619595e-01 1.10537004e+00 -1.39111583e-03
2.64381230e-01 -1.07249010e+00 -3.71227384e-01 -1.14608061e+00
3.01112551e-02 -1.58206806e-01 5.58784127e-01 -9.02198255e-01
-1.24210525e+00 1.43515691e-01 2.64862686e-01 -1.35403812e+00
7.10994527e-02 -9.45478022e-01 -4.82030779e-01 4.34086591e-01
-1.15329826e+00 -1.56128991e+00 -3.69778156e-01 5.96824467e-01
1.50638893e-01 1.88777044e-01 4.66505080e-01 1.70465916e-01
-3.76521528e-01 2.69026101e-01 -2.90323973e-01 2.22153157e-01
6.04092479e-01 -1.36434448e+00 8.00587416e-01 8.53349030e-01
7.18299031e-01 5.99408805e-01 5.52940130e-01 -8.34438741e-01
-2.12389469e+00 -1.41413188e+00 5.77012599e-01 -9.69521880e-01
9.69307780e-01 -6.71162069e-01 -6.85274065e-01 1.17806804e+00
-3.27224843e-02 5.99065244e-01 2.30932176e-01 3.95059139e-01
-5.34037411e-01 2.38160878e-01 -6.61516309e-01 9.14415002e-01
2.14299417e+00 -6.58466756e-01 -3.32149088e-01 5.34666598e-01
1.04632032e+00 -9.00707841e-01 -8.77502918e-01 4.55549389e-01
5.14482975e-01 -5.34876525e-01 1.42442834e+00 -9.92552757e-01
9.39924177e-03 -5.61496496e-01 -1.47950679e-01 -1.16308391e+00
-6.58049405e-01 -7.58944750e-01 -4.58085775e-01 7.99362242e-01
3.20623338e-01 -4.86054391e-01 7.39670932e-01 5.00737429e-01
-3.84881765e-01 -4.95485723e-01 -8.18563700e-01 -9.50048983e-01
-1.88906163e-01 -5.85997701e-01 5.28788567e-01 1.19782507e+00
-4.00505364e-01 1.35978490e-01 -3.00239652e-01 5.89392364e-01
8.85782719e-01 3.50523889e-01 8.80730689e-01 -1.41386080e+00
-1.27628863e-01 -6.56703889e-01 -8.77572954e-01 -1.37311375e+00
1.95212439e-01 -1.14276910e+00 -2.42706105e-01 -1.42030215e+00
1.32785179e-02 -7.06503808e-01 -1.91831589e-01 3.77102733e-01
-1.50862768e-01 5.42019188e-01 3.05040330e-01 1.87731460e-01
-9.13048923e-01 5.14031470e-01 1.52741814e+00 -1.26789123e-01
-2.35966653e-01 -3.00618231e-01 -3.64982754e-01 6.18847609e-01
3.50645959e-01 -2.56700724e-01 -1.42478406e-01 -8.62451971e-01
4.81193304e-01 -1.14581369e-01 1.01556385e+00 -8.24995637e-01
5.19076049e-01 -3.84541243e-01 3.65582168e-01 -8.57687831e-01
4.86256659e-01 -8.47395301e-01 5.56430399e-01 2.59340495e-01
-1.96458757e-01 3.13896239e-01 3.61548692e-01 9.11024511e-01
3.41249615e-01 4.40285504e-01 4.06699657e-01 1.37770444e-01
-1.03324771e+00 1.00179231e+00 7.30452761e-02 4.51090978e-03
1.19864297e+00 -4.06466246e-01 -5.04127920e-01 -1.52060136e-01
-9.41793561e-01 5.18257022e-01 6.80914819e-01 7.75137067e-01
4.35341448e-01 -1.93829930e+00 -5.37696838e-01 2.22014442e-01
3.39942098e-01 1.89373359e-01 1.56650260e-01 9.99287128e-01
-3.72570217e-01 5.11268616e-01 -2.02372342e-01 -1.23683727e+00
-1.26922035e+00 5.81119180e-01 2.69012779e-01 2.10991055e-01
-9.47639525e-01 8.62933695e-01 4.31385845e-01 -3.27370524e-01
1.39308497e-01 -6.19273245e-01 8.17606077e-02 -8.15716572e-03
3.52374390e-02 1.77130654e-01 -7.69736320e-02 -8.75272334e-01
-5.19554973e-01 1.09346700e+00 2.46158525e-01 -3.58531550e-02
1.27590251e+00 -1.62732691e-01 1.04020342e-01 5.23422062e-01
1.19705498e+00 -1.13244802e-01 -1.51394856e+00 -3.78518820e-01
2.53041480e-02 -6.04577780e-01 -5.40928617e-02 -2.61355191e-01
-9.50136125e-01 7.30783582e-01 2.53105283e-01 4.59870458e-01
6.63956285e-01 3.86125535e-01 3.83852035e-01 4.84277904e-01
3.33519131e-01 -5.50393879e-01 3.61965537e-01 7.46794343e-01
8.75079930e-01 -1.17368364e+00 -6.64925724e-02 -6.10838890e-01
-2.80201614e-01 1.06748605e+00 6.18608356e-01 -4.13237453e-01
7.84960806e-01 2.52500549e-02 -2.98810154e-01 -5.82849622e-01
-6.59456670e-01 -4.56162423e-01 7.85026908e-01 8.13758910e-01
1.63385138e-01 2.99067199e-02 3.51961613e-01 -1.13674782e-01
-8.24958980e-02 -5.43307900e-01 1.05117135e-01 8.56109917e-01
-1.94486603e-01 -8.75655115e-01 -2.75540233e-01 2.22017005e-01
2.73604453e-01 3.31760764e-01 -5.90785980e-01 1.20496142e+00
1.40190840e-01 6.18473887e-01 4.26149726e-01 -4.49009180e-01
4.43561375e-01 -2.91699499e-01 8.15734148e-01 -5.46672463e-01
-1.20080352e-01 1.72085539e-02 1.62672356e-01 -7.39500046e-01
-7.90072501e-01 -9.05281603e-01 -1.24759686e+00 -4.26945418e-01
-2.42464185e-01 -3.61405194e-01 4.18516397e-01 8.51524889e-01
6.55391753e-01 7.18564868e-01 3.40548992e-01 -1.24098349e+00
1.53947324e-01 -4.35805380e-01 -2.20748022e-01 6.30532444e-01
6.98767304e-01 -1.14910924e+00 -1.19106561e-01 1.08924635e-01] | [6.516005516052246, -2.123605728149414] |
1165ec91-7c41-45bb-a895-92d1082e016a | conditional-random-field-and-deep-feature | 1711.04483 | null | http://arxiv.org/abs/1711.04483v2 | http://arxiv.org/pdf/1711.04483v2.pdf | Conditional Random Field and Deep Feature Learning for Hyperspectral Image Segmentation | Image segmentation is considered to be one of the critical tasks in
hyperspectral remote sensing image processing. Recently, convolutional neural
network (CNN) has established itself as a powerful model in segmentation and
classification by demonstrating excellent performances. The use of a graphical
model such as a conditional random field (CRF) contributes further in capturing
contextual information and thus improving the segmentation performance. In this
paper, we propose a method to segment hyperspectral images by considering both
spectral and spatial information via a combined framework consisting of CNN and
CRF. We use multiple spectral cubes to learn deep features using CNN, and then
formulate deep CRF with CNN-based unary and pairwise potential functions to
effectively extract the semantic correlations between patches consisting of
three-dimensional data cubes. Effective piecewise training is applied in order
to avoid the computationally expensive iterative CRF inference. Furthermore, we
introduce a deep deconvolution network that improves the segmentation masks. We
also introduce a new dataset and experimented our proposed method on it along
with several widely adopted benchmark datasets to evaluate the effectiveness of
our method. By comparing our results with those from several state-of-the-art
models, we show the promising potential of our method. | ['Alan Wee-Chung Liew', 'Jun Zhou', 'Yongsheng Gao', 'Xiuping Jia', 'Jocelyn Chanussot', 'Fahim Irfan Alam'] | 2017-11-13 | null | null | null | null | ['hyperspectral-image-segmentation'] | ['computer-vision'] | [ 6.02065980e-01 -3.03561926e-01 1.80548191e-01 -5.82378566e-01
-6.95498765e-01 -4.60131824e-01 4.31423157e-01 -3.05539444e-02
-4.81575280e-01 6.91346586e-01 -2.19876423e-01 -2.71536291e-01
-3.93104255e-01 -1.06160593e+00 -6.14859879e-01 -9.70131993e-01
2.69118901e-02 3.40409189e-01 4.70829941e-02 1.22863092e-01
1.60046563e-01 8.26169610e-01 -1.50270176e+00 -7.90026262e-02
1.51067507e+00 1.23297310e+00 5.93758762e-01 1.19412355e-01
-1.33463100e-01 4.27367777e-01 -2.11006060e-01 6.61972165e-02
3.89495730e-01 -1.74376786e-01 -1.01496482e+00 4.48645949e-01
1.46226719e-01 -1.76386684e-01 4.80928309e-02 1.34355474e+00
3.13538820e-01 3.74988914e-01 7.38339126e-01 -6.51388586e-01
-4.96888131e-01 4.08467412e-01 -8.05494487e-01 -2.59657502e-01
-2.26486549e-01 3.72478552e-02 8.50266159e-01 -7.48053193e-01
1.99740782e-01 7.06800461e-01 6.95326447e-01 -2.03194231e-01
-1.11065972e+00 -6.43801749e-01 1.30418643e-01 4.88709360e-02
-1.62409508e+00 -5.84543571e-02 7.53098369e-01 -5.97313046e-01
6.21564209e-01 1.31754413e-01 6.83403492e-01 3.46122682e-01
-1.98500842e-01 5.46367228e-01 1.33413136e+00 -4.21941161e-01
1.37830317e-01 -1.37329757e-01 1.30651593e-01 5.77751279e-01
2.87097003e-02 -5.67988679e-02 7.92610645e-02 1.54794119e-02
7.64542043e-01 2.49114618e-01 -3.80657703e-01 6.35443255e-02
-8.76931310e-01 8.25594544e-01 9.83208776e-01 1.74529970e-01
-5.74909747e-01 -1.89571336e-01 -1.91583395e-01 -5.25045574e-01
6.55023754e-01 2.94968516e-01 -3.80805850e-01 6.70080006e-01
-1.13500118e+00 -8.17009658e-02 3.87882054e-01 7.20720053e-01
1.21005583e+00 -2.15290278e-01 -1.27707273e-01 9.76505101e-01
5.08107543e-01 5.47479987e-01 -1.69170052e-02 -7.02856421e-01
2.72112459e-01 8.44492257e-01 3.99861559e-02 -8.87908995e-01
-5.99205792e-01 -6.93448067e-01 -1.08803427e+00 1.14602886e-01
1.23351522e-01 -1.28193453e-01 -1.11331952e+00 1.27288938e+00
2.45732948e-01 2.96495110e-01 1.23445131e-02 9.83512461e-01
6.44449592e-01 7.04171956e-01 2.26800427e-01 -1.78988859e-01
1.14039588e+00 -8.51046801e-01 -4.53560472e-01 -4.18319181e-02
3.39536935e-01 -8.81060183e-01 6.43012464e-01 3.81811082e-01
-6.29945099e-01 -5.62400281e-01 -9.67856824e-01 8.40133727e-02
-4.75559652e-01 5.70665061e-01 7.80274272e-01 4.42495763e-01
-6.79510891e-01 6.58627748e-01 -9.43664789e-01 -1.27193943e-01
9.20825064e-01 4.94226664e-01 -1.90932989e-01 -1.61033332e-01
-1.04450822e+00 5.70444345e-01 5.96600115e-01 6.99317813e-01
-5.78585148e-01 -3.97928149e-01 -7.54507065e-01 2.97755580e-02
3.06527078e-01 -4.58948016e-01 7.37537146e-01 -8.82247329e-01
-1.38538182e+00 5.74319959e-01 -1.74070597e-01 -1.35283008e-01
1.50462389e-01 -2.53245682e-01 -2.65595168e-01 3.89180183e-01
5.98175302e-02 6.85390711e-01 5.24790704e-01 -1.40223622e+00
-4.47754443e-01 -5.69204509e-01 2.83111595e-02 3.16224724e-01
-3.72199178e-01 -2.05424428e-02 -2.14307234e-01 -3.32176775e-01
6.31198108e-01 -7.99875438e-01 -5.01380444e-01 -1.92425147e-01
-7.32029796e-01 -6.09693825e-02 6.84423804e-01 -7.91021585e-01
9.00021911e-01 -1.93844140e+00 -7.57370563e-03 5.34680784e-01
2.25041866e-01 3.35971862e-01 1.63734302e-01 2.53672421e-01
-2.92318705e-02 2.26277217e-01 -1.10949218e+00 -2.42032364e-01
-2.05862314e-01 1.36988506e-01 -1.16164796e-01 6.31640553e-01
5.37561238e-01 8.52131307e-01 -5.93166530e-01 -3.11769217e-01
3.38354111e-01 6.69134617e-01 -2.33529314e-01 3.56114179e-01
-4.45867509e-01 9.97307181e-01 -5.59199035e-01 6.33001804e-01
1.33061755e+00 -4.94502664e-01 1.21889979e-01 -3.20631534e-01
-5.02356350e-01 -1.56647041e-02 -1.07306325e+00 1.67461836e+00
-1.92572087e-01 1.51786998e-01 5.62761538e-02 -1.21837783e+00
1.08584976e+00 5.61534092e-02 5.11717796e-01 -5.27589083e-01
1.39573723e-01 1.34996548e-01 -3.33856225e-01 -6.49864435e-01
3.57268035e-01 -3.21065158e-01 4.26312000e-01 1.10893026e-01
-4.67506573e-02 -3.11591923e-01 -1.76949203e-01 -1.71553209e-01
3.78134727e-01 3.47755641e-01 2.18555346e-01 -3.76655996e-01
6.09128773e-01 1.25442490e-01 4.92191672e-01 4.41950321e-01
1.55881450e-01 6.88144684e-01 2.14354575e-01 -1.95651159e-01
-7.05039859e-01 -6.53587282e-01 -6.13884628e-01 5.81890881e-01
3.03066760e-01 1.09420456e-01 -6.71788573e-01 -4.72980320e-01
-1.58051997e-01 3.92035007e-01 -4.24275756e-01 4.37147468e-01
-1.29866734e-01 -1.41992509e+00 4.25078481e-01 5.22866189e-01
1.01574385e+00 -9.30625618e-01 -2.71530718e-01 1.28305852e-01
-3.14196110e-01 -1.28867292e+00 1.13380976e-01 3.37565690e-01
-9.27972972e-01 -1.21855378e+00 -5.57444632e-01 -7.78715789e-01
7.78207242e-01 5.05584478e-01 7.62229741e-01 9.85622182e-02
-3.50630283e-01 -2.80710876e-01 -4.59598809e-01 -1.91710234e-01
1.51695207e-01 1.36609435e-01 -6.30053699e-01 2.34217986e-01
2.25809023e-01 -6.18157804e-01 -7.94894516e-01 2.73858964e-01
-1.26441610e+00 2.75992364e-01 7.90299952e-01 7.32785404e-01
8.40931833e-01 5.43606162e-01 2.89641380e-01 -1.11128664e+00
1.29784299e-02 -5.31603754e-01 -9.38534975e-01 2.07341254e-01
-4.55077440e-01 -1.54527172e-01 5.34127116e-01 2.42969170e-01
-1.35459054e+00 4.60918069e-01 -3.26564431e-01 -2.66778632e-03
-5.07761657e-01 1.04010224e+00 -2.77799129e-01 -3.43059450e-01
3.12239677e-01 2.18722627e-01 -1.42210692e-01 -6.44700348e-01
2.58931190e-01 8.26197386e-01 4.08842027e-01 -5.50557494e-01
7.50150144e-01 8.18165302e-01 2.47913405e-01 -9.74705040e-01
-1.16324782e+00 -6.66573286e-01 -1.03935325e+00 4.72698500e-03
1.17363560e+00 -1.09034514e+00 -7.43604898e-01 7.87423849e-01
-1.10518825e+00 -3.56792420e-01 2.36733541e-01 6.27281904e-01
-1.43076673e-01 5.95996320e-01 -4.12329733e-01 -9.17189181e-01
-2.37033561e-01 -1.32254910e+00 1.09240580e+00 5.15579104e-01
7.37441540e-01 -9.82920468e-01 -1.16396919e-01 5.03709853e-01
4.12860900e-01 4.47629422e-01 7.73979843e-01 -4.58536863e-01
-8.41419518e-01 9.93099511e-02 -9.09668505e-01 6.66728258e-01
1.20713703e-01 8.88402238e-02 -1.21171820e+00 1.01563819e-02
1.68007873e-02 -2.56190002e-01 1.26353097e+00 6.09550834e-01
1.62878335e+00 1.17208585e-01 -2.97694355e-01 9.59945083e-01
1.97829664e+00 1.86358951e-02 9.14388299e-01 1.04958631e-01
9.92259383e-01 7.08541691e-01 4.36536729e-01 5.34712374e-01
4.66337383e-01 2.86718220e-01 8.06697011e-01 -6.63562834e-01
2.78442800e-01 6.80042133e-02 -3.13040286e-01 7.07842588e-01
-4.34122980e-01 -1.63318574e-01 -9.50826824e-01 2.44646102e-01
-1.75495934e+00 -7.44204283e-01 -5.67201018e-01 2.01566148e+00
6.26032948e-01 -2.48913884e-01 -3.08376640e-01 2.74365954e-03
8.02432775e-01 1.18543148e-01 -3.51238370e-01 4.80163306e-01
-3.91358525e-01 7.40839064e-01 7.02757180e-01 5.01829863e-01
-1.63773632e+00 1.02902758e+00 5.62977743e+00 7.37614036e-01
-1.21726191e+00 -1.29186690e-01 7.18817651e-01 6.63086474e-01
-2.06791624e-01 1.91537529e-01 -5.77407360e-01 1.81701392e-01
4.46126670e-01 6.40237033e-01 4.47486252e-01 4.38312590e-01
3.02588254e-01 -4.23730880e-01 -3.76395345e-01 6.42177463e-01
-1.60178527e-01 -1.18572760e+00 -9.23096240e-02 2.12075800e-01
9.50895011e-01 2.55425215e-01 -9.10848528e-02 -1.74810052e-01
1.71520248e-01 -1.41027510e+00 4.03131485e-01 8.16995978e-01
7.62592793e-01 -6.65214241e-01 9.74127412e-01 3.35062504e-01
-1.36506796e+00 -1.03779910e-02 -5.68299592e-01 -1.43380806e-01
-1.25150979e-02 1.07745326e+00 -5.85996509e-01 1.17152965e+00
5.27476966e-01 1.02410126e+00 -5.49786210e-01 1.31537652e+00
-3.57955277e-01 7.33462155e-01 -2.84933865e-01 4.03677583e-01
3.44587743e-01 -8.49826455e-01 7.25452676e-02 1.11213088e+00
3.42389047e-01 4.11496848e-01 5.75598061e-01 1.18047702e+00
-6.74253404e-02 2.36380309e-01 -2.83705562e-01 -3.43169235e-02
2.69581974e-01 1.59125543e+00 -9.69565868e-01 -1.46908835e-01
-3.99614245e-01 6.20537579e-01 1.99917391e-01 4.55067068e-01
-8.10869515e-01 -3.70782375e-01 3.16274077e-01 -2.16392279e-01
4.20770824e-01 -4.73075122e-01 -3.83447945e-01 -1.09360802e+00
-9.87234265e-02 -4.07758236e-01 1.35291785e-01 -9.22482550e-01
-1.35258138e+00 7.61929333e-01 -1.63495153e-01 -1.17111647e+00
5.25371969e-01 -7.11054862e-01 -4.85014945e-01 1.27799809e+00
-2.27803755e+00 -1.36903179e+00 -8.76938939e-01 5.21571338e-01
8.05137083e-02 3.20134938e-01 9.52771664e-01 2.80347794e-01
-7.86773503e-01 -2.14997962e-01 2.95658052e-01 2.43130311e-01
4.22875881e-01 -1.16829681e+00 -3.92297767e-02 8.54919791e-01
-1.29076138e-01 4.57388669e-01 1.20843582e-01 -5.53788543e-01
-9.73495424e-01 -1.46312523e+00 3.14494908e-01 1.21586211e-01
5.23885727e-01 -9.52781513e-02 -9.97434318e-01 4.96165156e-01
1.16971597e-01 2.28508309e-01 8.60778153e-01 -1.42196655e-01
-1.06595092e-01 -1.24004669e-01 -1.01601052e+00 -1.35471690e-02
8.48412573e-01 -4.06672686e-01 -1.84674814e-01 5.16969502e-01
6.63128078e-01 -2.40631402e-01 -9.04593706e-01 8.47494960e-01
3.31649125e-01 -1.14997447e+00 8.13111007e-01 -1.17159925e-01
5.34374237e-01 -4.85779673e-01 -2.98900247e-01 -1.24548614e+00
-3.99426311e-01 -1.52674660e-01 5.54658473e-01 1.15085852e+00
3.69730622e-01 -8.38017523e-01 5.90425849e-01 2.57246941e-01
-3.60166878e-01 -6.12564564e-01 -3.09322566e-01 -4.57621515e-01
4.40395474e-02 -3.88352036e-01 7.70731330e-01 9.71557558e-01
-4.46227431e-01 6.46298081e-02 -2.16395557e-02 7.67619073e-01
7.33896911e-01 4.22934860e-01 3.34858179e-01 -1.60522652e+00
-2.77264416e-02 -3.09554935e-01 -1.36957541e-01 -1.18136740e+00
3.50848377e-01 -1.07830524e+00 2.04226434e-01 -1.69064343e+00
3.95266235e-01 -7.67621458e-01 -3.48121315e-01 5.16420007e-01
-1.96699068e-01 4.41318870e-01 -1.26317173e-01 3.39188546e-01
-3.79871190e-01 6.91090167e-01 1.49463224e+00 -1.91445023e-01
-2.22350612e-01 8.62969551e-03 -7.39855349e-01 6.73577785e-01
9.18037951e-01 -4.24726158e-01 -1.11675791e-01 -4.87267107e-01
9.62205455e-02 -4.39119758e-03 5.99136293e-01 -9.76836503e-01
2.49619842e-01 -2.66626716e-01 5.88938177e-01 -8.14950407e-01
2.33889684e-01 -8.57127547e-01 9.38741937e-02 7.40920678e-02
-8.16235542e-02 -6.65017545e-01 1.70985073e-01 4.35593516e-01
-4.23966169e-01 -2.79322565e-01 8.54060769e-01 -1.39073983e-01
-6.10008717e-01 5.70622683e-01 7.37309456e-02 -3.48591238e-01
8.24427545e-01 4.43129018e-02 -2.21638188e-01 8.08934048e-02
-7.12581277e-01 2.60533780e-01 2.90224880e-01 -3.37001741e-01
5.37864387e-01 -8.30676436e-01 -6.73111737e-01 3.51728290e-01
1.36419863e-01 5.41780055e-01 3.88338417e-01 1.06654525e+00
-8.02295804e-01 3.92104685e-01 -1.33963913e-01 -8.68915617e-01
-9.08897460e-01 1.01804659e-01 3.66998464e-01 -2.72641540e-01
-3.81193876e-01 8.18067968e-01 1.84677318e-01 -6.05721354e-01
-7.40318969e-02 -4.26301926e-01 -5.35007894e-01 -3.17710266e-02
1.34632990e-01 2.50638962e-01 2.01000720e-01 -7.50011683e-01
-3.07222426e-01 7.50928879e-01 3.18434149e-01 1.47129565e-01
1.52642429e+00 -1.23105563e-01 -6.71621323e-01 -2.51382105e-02
1.12951338e+00 -1.23428807e-01 -1.51424193e+00 -2.91403204e-01
-1.04251429e-01 -4.27247673e-01 3.51110548e-01 -7.98471034e-01
-1.20352161e+00 1.09036362e+00 4.03719097e-01 2.43933499e-01
1.44698966e+00 -1.81901455e-01 6.08289719e-01 1.73006549e-01
2.33992875e-01 -8.05152655e-01 -3.83370936e-01 5.43012679e-01
4.79742408e-01 -1.53882813e+00 2.27034330e-01 -8.46924067e-01
-4.97514725e-01 1.31721079e+00 2.84237981e-01 -1.04146682e-01
9.49162543e-01 -1.25009954e-01 2.12812684e-02 -2.56912619e-01
1.47702008e-01 -7.57544160e-01 3.86341214e-01 4.18985158e-01
4.40232009e-01 3.88063639e-01 -7.00038373e-02 6.07391894e-01
7.63088092e-02 1.04719758e-01 1.68574244e-01 6.96786284e-01
-5.80647230e-01 -1.12297654e+00 -3.70701045e-01 4.92950112e-01
-3.40919793e-01 -2.83954173e-01 -5.27704172e-02 5.12201726e-01
4.24065202e-01 1.03114414e+00 -1.34228244e-01 -2.68725246e-01
-4.70541045e-02 -2.83596553e-02 3.07076246e-01 -7.50187218e-01
-3.73730987e-01 3.12582731e-01 -3.98113906e-01 -2.45011270e-01
-1.05495083e+00 -4.11797583e-01 -1.43357646e+00 2.43301541e-01
-7.58587599e-01 1.20553277e-01 9.36045885e-01 1.31891108e+00
1.87066808e-01 5.11464894e-01 8.60748768e-01 -1.01470447e+00
-3.67038548e-01 -1.00127554e+00 -9.03550565e-01 1.21559523e-01
1.78556561e-01 -6.92226648e-01 -2.56819516e-01 1.79018155e-01] | [9.870851516723633, -1.7479420900344849] |
1e038714-299a-4976-91f4-62897dba9636 | joint-goal-segmentation-and-goal-success | null | null | https://aclanthology.org/2022.coling-1.41 | https://aclanthology.org/2022.coling-1.41.pdf | Joint Goal Segmentation and Goal Success Prediction on Multi-Domain Conversations | To evaluate the performance of a multi-domain goal-oriented Dialogue System (DS), it is important to understand what the users’ goals are for the conversations and whether those goals are successfully achieved. The success rate of goals directly correlates with user satisfaction and perceived usefulness of the DS. In this paper, we propose a novel automatic dialogue evaluation framework that jointly performs two tasks: goal segmentation and goal success prediction. We extend the RoBERTa-IQ model (Gupta et al., 2021) by adding multi-task learning heads for goal segmentation and success prediction. Using an annotated dataset from a commercial DS, we demonstrate that our proposed model reaches an accuracy that is on-par with single-pass human annotation comparing to a three-pass gold annotation benchmark. | ['Chenlei Guo', 'Tuan-Hung Pham', 'Yu Zhang', 'Xiaohu Liu', 'Bin Guo', 'Benjamin Yao', 'Meiguo Wang'] | null | null | null | null | coling-2022-10 | ['dialogue-evaluation'] | ['natural-language-processing'] | [ 5.49605303e-02 7.19061792e-01 -9.76017267e-02 -5.22507727e-01
-1.13107979e+00 -5.77562571e-01 8.03475082e-01 3.02484810e-01
-4.97893184e-01 9.79691565e-01 6.45629287e-01 -1.22954704e-01
3.05132251e-02 -5.67053139e-01 2.63322473e-01 3.83350626e-02
1.35575473e-01 8.95691037e-01 2.22643375e-01 -7.98707664e-01
5.07333159e-01 -2.61482149e-01 -1.23870075e+00 3.39475811e-01
1.11309433e+00 8.85344446e-01 1.52499065e-01 1.31602180e+00
-2.51317561e-01 1.25284231e+00 -7.91915774e-01 -5.25007784e-01
-2.41374344e-01 -5.57505667e-01 -1.70740938e+00 1.23175435e-01
-3.46584506e-02 -4.94984806e-01 2.12969199e-01 6.76942587e-01
3.66809279e-01 3.97947460e-01 6.44410849e-01 -1.24077654e+00
-4.88490090e-02 5.37383437e-01 2.31756538e-01 -2.29030356e-01
9.97943163e-01 1.01593412e-01 1.58180237e+00 -3.05853546e-01
5.58570206e-01 1.06052840e+00 4.18947995e-01 7.96916544e-01
-1.01208794e+00 -5.22012264e-02 -1.02967612e-01 -7.32903108e-02
-7.97900438e-01 -7.30165780e-01 6.08865619e-01 -5.04792929e-01
1.13894355e+00 3.79706025e-01 3.21689546e-01 9.33021128e-01
-2.63248801e-01 9.72555697e-01 1.03809261e+00 -5.65372705e-01
3.53413522e-01 2.85773367e-01 6.01670027e-01 6.36098504e-01
-5.02872825e-01 -3.87604147e-01 -4.96516973e-01 -1.90164253e-01
3.33933979e-01 -8.05427969e-01 -1.79163605e-01 2.08551005e-01
-8.60178173e-01 1.09418714e+00 -2.96924651e-01 5.67970514e-01
-4.44510967e-01 -3.49128515e-01 6.50990546e-01 4.82758671e-01
6.53296411e-01 9.53264594e-01 -6.51289821e-01 -1.01967061e+00
-6.34328485e-01 3.34423751e-01 1.61763179e+00 9.79945004e-01
4.79583085e-01 -2.53188431e-01 -5.45336545e-01 1.16869855e+00
3.04955870e-01 7.03202263e-02 4.24794793e-01 -1.16297674e+00
2.09851578e-01 9.23394144e-01 5.85465670e-01 -6.71952367e-01
-8.33167315e-01 -7.67425895e-02 -3.48623574e-01 -1.91767782e-01
8.30127835e-01 -4.73498911e-01 -3.27561527e-01 1.60700142e+00
1.15525559e-01 -3.09128195e-01 3.48429352e-01 6.46646202e-01
1.24452174e+00 6.29618585e-01 4.49417055e-01 -5.01876116e-01
1.37799525e+00 -1.07861161e+00 -9.47153687e-01 -4.25180405e-01
1.11384964e+00 -6.99920297e-01 1.50098324e+00 5.21629453e-01
-1.16020954e+00 -5.74166656e-01 -7.16821909e-01 -1.19556896e-01
-5.65975495e-02 3.63553725e-02 4.94428456e-01 9.17531788e-01
-1.17033422e+00 3.98584038e-01 -2.29217872e-01 -5.83666146e-01
-2.60524303e-01 2.59106696e-01 -1.33556888e-01 4.59323436e-01
-1.28886771e+00 1.08739424e+00 4.34978783e-01 -6.80067182e-01
-7.57337570e-01 -2.35035643e-01 -8.68025243e-01 -7.83666223e-03
4.87306863e-01 -3.48603964e-01 1.94120228e+00 -5.64664066e-01
-2.02107573e+00 1.10163379e+00 -1.70643702e-02 -3.36900234e-01
4.11856264e-01 -3.61675113e-01 -2.01648355e-01 1.84980463e-02
2.19540507e-01 3.87009621e-01 2.61753917e-01 -1.02172589e+00
-1.27458274e+00 -3.23115945e-01 6.20231211e-01 7.09872484e-01
-4.81624663e-01 2.84755617e-01 -4.48351771e-01 1.18172087e-01
-4.17099833e-01 -5.70118010e-01 -2.95713723e-01 -8.53433490e-01
-4.28264558e-01 -8.69037986e-01 2.42970079e-01 -9.87183988e-01
1.63892150e+00 -1.66759002e+00 1.87268808e-01 -6.18185624e-02
5.07329583e-01 2.04899102e-01 -1.08084619e-01 5.05082190e-01
4.33347136e-01 2.40195170e-01 1.17715053e-01 -5.67733049e-01
1.58966586e-01 3.64808296e-03 2.80045003e-01 -9.63683277e-02
-8.31209216e-03 7.11365521e-01 -1.07905877e+00 -7.01692760e-01
2.07625419e-01 -2.18694359e-01 -5.86713016e-01 9.13942218e-01
-6.08128667e-01 6.31145179e-01 -6.38985157e-01 3.40299934e-01
-9.89068896e-02 -3.96776140e-01 3.18116367e-01 1.90514401e-01
-1.64122596e-01 7.98672736e-01 -7.90084422e-01 1.64964259e+00
-6.51877284e-01 2.06172243e-01 7.32837766e-02 -7.47753561e-01
1.21448720e+00 6.06168926e-01 5.63113034e-01 -9.20098424e-01
1.96576938e-01 1.19679712e-01 -1.47503093e-01 -5.45104265e-01
9.81488883e-01 -1.96451902e-01 -5.52041948e-01 7.00348318e-01
4.02831882e-01 -5.37103772e-01 5.27961016e-01 1.32927895e-01
1.02267540e+00 1.85495004e-01 8.12755644e-01 -2.37048849e-01
8.43396068e-01 3.99002880e-01 2.23882854e-01 8.89961720e-01
-5.53375840e-01 2.62005091e-01 9.28942800e-01 -8.84105861e-02
-9.49286640e-01 -3.92353952e-01 3.40688348e-01 1.73460889e+00
-1.37801975e-01 -6.17960453e-01 -1.04436207e+00 -9.61557865e-01
-6.05287313e-01 1.25265300e+00 -3.20342779e-01 -2.98843756e-02
-4.03866202e-01 -3.42536956e-01 7.28089213e-01 -7.17768967e-02
6.42361045e-01 -1.14624846e+00 -4.54433203e-01 6.07547700e-01
-8.68886590e-01 -1.40779209e+00 -3.06354076e-01 -1.62055373e-01
-4.76324856e-01 -1.19225574e+00 -3.68805438e-01 -6.99811161e-01
-1.78077623e-01 -1.84364617e-01 1.57712424e+00 1.55093729e-01
2.40578145e-01 6.16718650e-01 -8.06388736e-01 -2.82542199e-01
-9.61833000e-01 3.94018352e-01 -2.41285205e-01 -3.29137832e-01
5.70177555e-01 -8.99376795e-02 -2.61312693e-01 5.78896046e-01
-2.14410752e-01 2.78171331e-01 -4.68400232e-02 7.86328912e-01
-1.31888032e-01 -1.59880921e-01 9.99146402e-01 -8.52766752e-01
1.45418787e+00 -3.11776787e-01 -1.98712319e-01 3.55421215e-01
-7.29251444e-01 -1.99169114e-01 5.49819708e-01 1.60220161e-01
-1.20042181e+00 -2.18070656e-01 -8.48299921e-01 7.17793107e-01
-3.36673498e-01 4.64396536e-01 -7.49755576e-02 2.17782006e-01
7.69980669e-01 1.29210815e-01 1.37300968e-01 -2.78870374e-01
4.13256049e-01 1.15479779e+00 9.92195234e-02 -6.67423129e-01
-2.10904777e-02 -3.76043111e-01 -5.44684052e-01 -9.60431993e-01
-1.11551595e+00 -9.31802750e-01 -5.22000790e-01 -7.32228160e-01
8.95970404e-01 -6.83033407e-01 -1.18837118e+00 3.61532182e-01
-1.18035567e+00 -6.95011020e-01 -1.84653938e-01 1.25149742e-01
-1.22683716e+00 3.75920534e-01 -5.87504923e-01 -1.34150422e+00
-6.60363019e-01 -1.02564633e+00 8.94898653e-01 3.63118112e-01
-7.79831529e-01 -1.31791508e+00 7.29691461e-02 8.64118099e-01
3.96819383e-01 -1.47313133e-01 6.68077767e-01 -1.46443915e+00
2.50486493e-01 -2.37457097e-01 -8.28216225e-02 3.76589686e-01
1.52461097e-01 -3.68438363e-01 -7.42923379e-01 6.14278903e-03
2.63683647e-01 -9.16597426e-01 5.83309829e-02 2.81342596e-01
3.98744076e-01 -3.88140261e-01 1.49517938e-01 -2.55514085e-01
1.01544523e+00 3.46564710e-01 5.98080099e-01 3.67946714e-01
1.54858485e-01 9.66678977e-01 1.29922259e+00 8.10148358e-01
9.16413367e-01 8.31061602e-01 2.13483334e-01 7.70305544e-02
2.22602174e-01 -1.51420861e-01 2.90725023e-01 6.27735555e-01
-1.85622200e-01 -5.43960750e-01 -1.10767925e+00 5.37795305e-01
-1.90381396e+00 -7.61173666e-01 -3.17540914e-01 2.07953763e+00
1.13058853e+00 4.10279363e-01 8.45579565e-01 6.45027161e-02
4.06774729e-01 1.26916423e-01 2.85117775e-02 -6.22135699e-01
2.72743016e-01 5.98299541e-02 -2.39006486e-02 9.75121617e-01
-1.10803783e+00 1.24466419e+00 6.75457287e+00 6.99399352e-01
-4.11461681e-01 4.08812374e-01 9.05286491e-01 4.89263356e-01
5.49159870e-02 -2.16218472e-01 -9.15088594e-01 3.84125449e-02
1.24309766e+00 -5.75553775e-01 3.30036670e-01 8.74591410e-01
3.76370758e-01 -1.83201760e-01 -1.07379150e+00 4.74982530e-01
1.82844885e-02 -9.68773007e-01 -4.07227546e-01 -1.55091174e-02
2.22883895e-01 -3.57957929e-01 -4.64384407e-01 8.63846421e-01
8.37173462e-01 -9.99326348e-01 2.47586802e-01 4.45496887e-01
5.37678719e-01 -4.79177654e-01 8.48114192e-01 7.57768929e-01
-8.08170199e-01 -5.16152643e-02 9.64662358e-02 -3.40053678e-01
4.93562758e-01 4.74664085e-02 -1.56496370e+00 3.66481781e-01
9.59272981e-02 3.25127989e-01 -3.46669674e-01 5.78151464e-01
-2.18147382e-01 6.35555327e-01 -1.38869718e-01 -6.29852772e-01
4.15966392e-01 -2.88585097e-01 5.94073236e-01 1.25911725e+00
-1.00444665e-03 4.37906444e-01 5.16464651e-01 5.60231268e-01
1.65499244e-02 7.31324017e-01 -4.91352916e-01 -3.43757182e-01
2.09772334e-01 1.36885917e+00 -3.21209937e-01 -4.21700478e-01
-3.98613304e-01 8.79744291e-01 2.97573179e-01 -1.08638778e-01
-4.60317940e-01 -6.34464920e-02 4.96596694e-01 -1.56309098e-01
-3.74645472e-01 -1.90906450e-01 -5.55713713e-01 -6.73469245e-01
-3.33246619e-01 -1.06154025e+00 5.09703279e-01 -4.74349141e-01
-9.25818682e-01 7.17648566e-01 -3.23541999e-01 -9.91843760e-01
-6.80373847e-01 -4.25463200e-01 -5.14462292e-01 8.72486055e-01
-1.35955679e+00 -1.25208020e+00 -3.73730063e-01 3.81908357e-01
9.72574234e-01 -1.91504866e-01 1.20925915e+00 8.38922784e-02
-3.40543181e-01 3.89193833e-01 -4.42417830e-01 7.75354281e-02
7.09968388e-01 -1.63632429e+00 4.04791534e-01 3.09042603e-01
-3.97853166e-01 1.69473350e-01 1.00330961e+00 -6.66227043e-01
-8.77536356e-01 -6.96915984e-01 1.30825067e+00 -4.79495198e-01
7.49246001e-01 -6.88809976e-02 -7.13093460e-01 4.95081425e-01
4.16285664e-01 -9.78892565e-01 9.04162526e-01 6.63560569e-01
2.93998122e-01 3.99159282e-01 -1.25532448e+00 5.31849921e-01
9.58332777e-01 -2.62735575e-01 -7.30724096e-01 7.17456460e-01
7.17628300e-01 -4.49151367e-01 -1.17086196e+00 2.51469254e-01
3.18800747e-01 -1.13720381e+00 7.55035698e-01 -8.97033870e-01
5.24646342e-01 4.38594460e-01 -1.67138830e-01 -1.41739178e+00
-1.61748722e-01 -7.95744658e-01 1.09666005e-01 1.29179394e+00
5.66899478e-01 -2.04278305e-01 6.12898409e-01 1.16394377e+00
-2.78247535e-01 -4.10067707e-01 -7.20643163e-01 -3.30316663e-01
-3.21838595e-02 -4.02150005e-01 3.30469668e-01 7.77698934e-01
7.83491552e-01 1.05470657e+00 -7.15876698e-01 -2.63540000e-01
2.27667898e-01 -2.48835325e-01 1.02674448e+00 -1.53619599e+00
-1.04533978e-01 -7.32164264e-01 -2.55470313e-02 -1.37597799e+00
1.65662140e-01 -3.99981409e-01 2.54763424e-01 -1.82281435e+00
-1.92263722e-01 -5.00132263e-01 1.58373281e-01 4.84731883e-01
-3.69643122e-01 -2.01278120e-01 -4.26837131e-02 2.16977373e-02
-1.24753320e+00 4.77480203e-01 1.10631728e+00 1.50591880e-01
-7.35652804e-01 3.90487611e-01 -9.62493598e-01 7.35221446e-01
1.07886302e+00 5.72635280e-03 -2.75385559e-01 1.29655346e-01
7.55826011e-02 7.36416399e-01 -2.18940690e-01 -8.85561645e-01
2.41150677e-01 -4.47874784e-01 -5.76837420e-01 -2.51199603e-01
4.24932986e-01 -3.32423657e-01 -3.85690182e-01 2.33692959e-01
-7.21777678e-01 -3.72868836e-01 -5.64946309e-02 1.98459506e-01
-2.31646225e-01 -6.59022331e-01 7.57234633e-01 -2.89790452e-01
-8.05636466e-01 -4.18048538e-02 -5.06869435e-01 3.41028094e-01
1.06273425e+00 -1.39293790e-01 -3.35764289e-01 -9.94286895e-01
-9.62691665e-01 6.99466288e-01 1.77936286e-01 4.65790927e-01
4.52884525e-01 -8.36767793e-01 -1.00889277e+00 -2.58950680e-01
3.45510691e-01 -4.56328362e-01 3.34204338e-03 5.71408033e-01
-2.70391852e-01 6.55914128e-01 -1.54787660e-01 -4.17430192e-01
-1.39691460e+00 -2.16658160e-01 4.79264289e-01 -9.06131029e-01
-2.18366697e-01 8.13405335e-01 -2.84964532e-01 -7.03442037e-01
3.01543891e-01 3.17109495e-01 -1.09017467e+00 2.35343412e-01
5.74076653e-01 4.51902598e-01 -1.31932823e-02 -5.93531311e-01
9.03331488e-02 -9.55982879e-02 -2.34033931e-02 -4.26919371e-01
1.22506988e+00 -4.29398566e-01 3.07984594e-02 6.13278568e-01
6.86066866e-01 -2.83511013e-01 -7.69461036e-01 -3.91155392e-01
5.41281641e-01 -2.77054638e-01 2.18718164e-02 -1.21586907e+00
-2.63859451e-01 5.89700580e-01 2.60741740e-01 9.95762229e-01
9.18116391e-01 -1.27395252e-02 7.45142698e-01 3.57147098e-01
5.24498105e-01 -1.62796068e+00 5.52676916e-01 1.04674387e+00
1.01427317e+00 -1.52331102e+00 -4.91968483e-01 -3.80688667e-01
-1.28728044e+00 1.05883479e+00 9.24534798e-01 4.74587768e-01
2.43136987e-01 -1.41934663e-01 1.33952692e-01 -3.27067971e-01
-7.86726177e-01 -5.74808419e-01 1.60180196e-01 6.16158605e-01
8.28267694e-01 2.39087343e-01 -7.87555695e-01 8.90643954e-01
-3.21856827e-01 6.19773045e-02 6.04519665e-01 7.35544205e-01
-9.47926581e-01 -1.13290870e+00 1.25654951e-01 5.59283674e-01
-3.82658422e-01 -3.32256407e-02 -7.11910069e-01 4.87451822e-01
-6.12548292e-01 1.85545266e+00 -3.41766685e-01 -5.24205565e-01
8.04682672e-01 5.60258269e-01 4.33022641e-02 -1.10435390e+00
-9.66090918e-01 -8.43197480e-03 1.30169773e+00 -4.66796130e-01
-3.81964147e-01 -5.60770810e-01 -1.20239162e+00 -1.68639690e-01
-4.56851810e-01 4.70482975e-01 6.15030766e-01 1.13113773e+00
6.91474304e-02 4.38288808e-01 8.67985487e-01 -2.50404835e-01
-7.72584796e-01 -1.53234923e+00 -4.56439823e-01 5.05448699e-01
-3.33574526e-02 -3.96905094e-01 -2.17272520e-01 -1.67775378e-01] | [12.865876197814941, 8.029955863952637] |
c659d5cf-1f36-4a57-845e-a7a142942382 | exploring-the-use-of-an-unsupervised | 2008.03615 | null | https://arxiv.org/abs/2008.03615v1 | https://arxiv.org/pdf/2008.03615v1.pdf | Exploring the Use of an Unsupervised Autoregressive Model as a Shared Encoder for Text-Dependent Speaker Verification | In this paper, we propose a novel way of addressing text-dependent automatic speaker verification (TD-ASV) by using a shared-encoder with task-specific decoders. An autoregressive predictive coding (APC) encoder is pre-trained in an unsupervised manner using both out-of-domain (LibriSpeech, VoxCeleb) and in-domain (DeepMine) unlabeled datasets to learn generic, high-level feature representation that encapsulates speaker and phonetic content. Two task-specific decoders were trained using labeled datasets to classify speakers (SID) and phrases (PID). Speaker embeddings extracted from the SID decoder were scored using a PLDA. SID and PID systems were fused at the score level. There is a 51.9% relative improvement in minDCF for our system compared to the fully supervised x-vector baseline on the cross-lingual DeepMine dataset. However, the i-vector/HMM method outperformed the proposed APC encoder-decoder system. A fusion of the x-vector/PLDA baseline and the SID/PLDA scores prior to PID fusion further improved performance by 15% indicating complementarity of the proposed approach to the x-vector system. We show that the proposed approach can leverage from large, unlabeled, data-rich domains, and learn speech patterns independent of downstream tasks. Such a system can provide competitive performance in domain-mismatched scenarios where test data is from data-scarce domains. | ['Ruchao Fan', 'Vijay Ravi', 'Amber Afshan', 'Abeer Alwan', 'Huanhua Lu'] | 2020-08-08 | null | null | null | null | ['text-dependent-speaker-verification'] | ['speech'] | [ 2.19889820e-01 6.72042668e-02 -5.96868433e-02 -9.15957928e-01
-1.72737062e+00 -7.20266104e-01 7.32781887e-01 -3.34154636e-01
-4.45047438e-01 3.83905768e-01 4.91912097e-01 -5.14564216e-01
4.50917095e-01 5.60925622e-03 -6.09718978e-01 -7.86538899e-01
3.71518224e-01 7.09253669e-01 -1.86076224e-01 -1.05121419e-01
-1.30914018e-01 1.24782667e-01 -1.56444812e+00 3.80365640e-01
6.77651167e-01 9.19691145e-01 2.97230989e-01 9.03648555e-01
3.12513001e-02 4.71576333e-01 -6.92802608e-01 -6.18437171e-01
2.02353835e-01 -2.15565994e-01 -5.67105651e-01 -8.13441630e-03
7.19649196e-01 -2.82826185e-01 -2.98878074e-01 8.51550519e-01
8.85159016e-01 -1.03786200e-01 7.86269486e-01 -1.14189589e+00
-7.27039576e-01 8.05079818e-01 -3.35489392e-01 -4.56825346e-02
2.33342335e-01 2.17178449e-01 1.05831194e+00 -1.17917967e+00
4.81163949e-01 1.34746599e+00 5.72337747e-01 7.50469148e-01
-1.26201475e+00 -8.97470295e-01 3.13328169e-02 2.36164004e-01
-1.38010955e+00 -1.16069543e+00 8.00432146e-01 -4.54015255e-01
1.38566852e+00 7.26792738e-02 -1.78834960e-01 1.61829543e+00
-1.29914299e-01 8.98655534e-01 1.08167005e+00 -6.63593769e-01
3.18846673e-01 6.19139552e-01 4.23887730e-01 2.79674441e-01
-2.48407662e-01 4.94502574e-01 -1.00398374e+00 -1.16305612e-01
-7.12511912e-02 -5.86171865e-01 -1.51495486e-01 -1.85597148e-02
-9.81527925e-01 8.92578959e-01 -2.88146853e-01 1.68907955e-01
-2.02963918e-01 -4.09111261e-01 6.77476823e-01 5.75594008e-01
5.77525139e-01 -5.32775484e-02 -1.05181313e+00 -3.45783621e-01
-1.35226607e+00 1.42885614e-02 9.58526909e-01 1.05995250e+00
6.48654342e-01 5.65095961e-01 -3.46478701e-01 1.20196116e+00
7.14740217e-01 8.23572457e-01 8.62259984e-01 -6.00179791e-01
8.35837126e-01 8.28015804e-02 -3.92032474e-01 -1.19176731e-01
1.50851876e-01 -4.42160755e-01 -3.95460844e-01 6.03112541e-02
2.70593643e-01 -3.64881545e-01 -1.16500521e+00 2.01406455e+00
2.31094092e-01 1.80001557e-01 6.00441277e-01 6.27082527e-01
8.49891067e-01 7.39083529e-01 2.60804206e-01 -7.82366842e-03
1.49620163e+00 -8.90441656e-01 -8.08454216e-01 -3.67775649e-01
7.44943738e-01 -8.47574055e-01 9.01218176e-01 2.53313661e-01
-9.16911244e-01 -7.72585213e-01 -1.08967865e+00 -1.28660619e-01
-3.71751249e-01 5.81855476e-01 -4.95496452e-01 1.09065604e+00
-1.30553532e+00 -4.17442359e-02 -6.20935798e-01 -2.89861292e-01
1.09474689e-01 4.05093640e-01 -3.86928976e-01 -3.41924094e-02
-1.23728979e+00 1.02140570e+00 2.40306050e-01 -2.25201055e-01
-1.23472524e+00 -7.02161670e-01 -1.13454771e+00 7.01711774e-02
-2.45774731e-01 -7.22409040e-02 1.63159215e+00 -7.32140720e-01
-2.01030254e+00 9.81537521e-01 -6.47598028e-01 -6.74049795e-01
2.11214706e-01 -1.94644272e-01 -8.22349429e-01 -1.30230650e-01
5.34248874e-02 5.89585483e-01 1.02426589e+00 -1.01472664e+00
-5.36732376e-01 -5.21538079e-01 -8.75741839e-01 2.77036846e-01
-4.00119722e-01 2.69326091e-01 -3.07199121e-01 -4.23589975e-01
-3.06770146e-01 -8.73270810e-01 2.77416378e-01 -5.43574154e-01
-5.13067424e-01 -4.24646914e-01 1.19675052e+00 -1.35645700e+00
1.05296504e+00 -2.46465135e+00 3.73878726e-03 7.43816942e-02
-4.00858402e-01 6.27906084e-01 -2.55237490e-01 5.08518994e-01
-1.57513902e-01 -2.51444012e-01 -4.30601120e-01 -9.98821437e-01
2.96042204e-01 3.10460508e-01 -3.84823918e-01 5.21781027e-01
4.42925304e-01 6.70212388e-01 -3.97988677e-01 -3.95102322e-01
1.66869089e-01 8.46312344e-01 -5.10715961e-01 4.11611676e-01
-1.52171612e-01 4.48698580e-01 1.23718865e-01 6.21684492e-01
8.07296932e-01 3.48190665e-01 2.45299354e-01 1.19564675e-01
-1.26839593e-01 9.41405177e-01 -1.06290936e+00 1.66673732e+00
-6.20362699e-01 8.51619244e-01 3.88357908e-01 -9.30686951e-01
1.11125445e+00 8.54523361e-01 -4.04690690e-02 -7.04327762e-01
6.49222061e-02 3.34819376e-01 -6.72173277e-02 -1.25399798e-01
3.95646453e-01 -3.52290630e-01 -2.56875068e-01 3.67372453e-01
7.50779748e-01 2.06935685e-03 -3.77366155e-01 1.29690111e-01
9.80354846e-01 -1.53011428e-02 1.00342266e-01 -1.10612512e-01
8.95411313e-01 -9.24891010e-02 6.10920727e-01 4.36573088e-01
-3.33497494e-01 5.26052237e-01 2.61226177e-01 3.54778200e-01
-1.00737035e+00 -1.14611983e+00 -4.39829707e-01 1.50823152e+00
-5.93404412e-01 -3.26819539e-01 -8.49875271e-01 -7.32183576e-01
1.28757313e-01 1.21419251e+00 -2.92275697e-01 -6.87729716e-02
-5.69723487e-01 -2.30526745e-01 1.08806062e+00 3.58263552e-01
1.54372245e-01 -8.11982274e-01 2.88018525e-01 4.03684676e-01
3.81641313e-02 -1.42107272e+00 -5.93925953e-01 5.53871274e-01
-4.82186854e-01 -5.76218545e-01 -8.12340200e-01 -1.06456649e+00
2.62534112e-01 -2.14716807e-01 9.06652391e-01 -7.61654675e-01
4.14249957e-01 3.87514710e-01 -2.56449491e-01 -3.68508250e-01
-9.79754984e-01 1.07281148e-01 3.63573641e-01 1.96674377e-01
8.09680223e-01 -1.54680997e-01 -9.88088995e-02 2.98237741e-01
-3.79581392e-01 -3.37444305e-01 6.08251274e-01 1.08190966e+00
3.43781352e-01 -5.27384341e-01 8.82526219e-01 -5.93187988e-01
4.42471892e-01 -3.43753994e-01 -5.81637919e-01 9.72766578e-02
-5.81458867e-01 3.18396717e-01 4.96765673e-01 -3.61962527e-01
-1.53327155e+00 3.17937344e-01 -7.34771132e-01 -4.61165637e-01
-4.77913022e-01 3.36186260e-01 -6.27386332e-01 3.72099519e-01
4.91495252e-01 4.49980646e-01 1.61971569e-01 -8.76813769e-01
5.65959394e-01 1.53237879e+00 7.50336826e-01 -3.90695006e-01
6.01606965e-01 -1.54330507e-01 -8.31891537e-01 -9.62101340e-01
-4.31514025e-01 -6.86923623e-01 -7.54260421e-01 2.17141658e-01
8.92779589e-01 -1.41971135e+00 -3.70016217e-01 5.21887839e-01
-1.30757749e+00 -1.60812438e-01 -2.57530451e-01 6.22735739e-01
-3.52444649e-01 2.72769421e-01 -6.73384011e-01 -1.00385487e+00
-4.68834639e-01 -1.38028729e+00 1.38029325e+00 -2.37254590e-01
-2.30135858e-01 -9.32923436e-01 3.55930388e-01 6.14498496e-01
4.60063249e-01 -5.71793914e-01 6.38535857e-01 -1.49247658e+00
-2.87292916e-02 -1.39348686e-01 9.68319848e-02 1.05333805e+00
-1.36361673e-01 -1.69569165e-01 -1.72323561e+00 -3.80425006e-01
2.03572348e-01 -2.53648579e-01 7.88130879e-01 2.99253941e-01
6.46031559e-01 -4.11099762e-01 -1.94665551e-01 4.18021768e-01
1.08629405e+00 1.88622147e-01 2.58285314e-01 -8.79445672e-02
5.27758658e-01 6.07421637e-01 2.73706943e-01 2.06234768e-01
5.82002103e-01 1.00228262e+00 -2.43371204e-01 1.67389154e-01
-6.05477273e-01 -3.78765732e-01 1.19678652e+00 1.38572991e+00
6.79085493e-01 -2.37757415e-01 -1.19906473e+00 9.07841504e-01
-1.26829791e+00 -9.33634043e-01 -1.03514746e-01 2.08082652e+00
1.14545786e+00 -1.28779337e-01 3.43941413e-02 2.42204383e-01
8.03966939e-01 -5.71078286e-02 -4.59795147e-01 -7.25360751e-01
-2.77281523e-01 3.96854877e-01 3.74352902e-01 7.65500009e-01
-9.38062847e-01 9.91176307e-01 5.88255978e+00 8.61864328e-01
-1.29816937e+00 7.35575199e-01 2.83934176e-01 7.43271187e-02
-1.68989182e-01 -1.99403808e-01 -1.48202932e+00 6.74288213e-01
1.96037829e+00 -5.98512553e-02 2.44774111e-02 1.09145701e+00
4.50512618e-02 4.73880529e-01 -1.26478302e+00 9.91908312e-01
4.35701370e-01 -9.15588140e-01 -3.42455298e-01 2.20364004e-01
5.81312776e-01 6.65841341e-01 4.12425607e-01 7.25884795e-01
4.53218400e-01 -7.66401231e-01 8.18464637e-01 -7.42884129e-02
1.18625116e+00 -5.63026965e-01 7.87563622e-01 4.09568816e-01
-1.02662790e+00 5.63792214e-02 -3.36543396e-02 4.02363718e-01
2.89823681e-01 3.18602324e-01 -1.66667962e+00 4.29858536e-01
4.06630933e-01 6.62090063e-01 -3.49643350e-01 5.11353195e-01
-2.59600759e-01 1.25782490e+00 -1.80478469e-01 3.68970513e-01
7.65129924e-02 1.69940963e-01 6.27057254e-01 1.76078284e+00
3.09967220e-01 -4.61904615e-01 -3.02054971e-01 5.09426236e-01
-1.67572618e-01 5.59426211e-02 -5.84651113e-01 -6.47718692e-03
6.59283400e-01 8.15935254e-01 3.10592502e-01 -6.50854588e-01
-6.13589644e-01 1.00648320e+00 2.42842421e-01 4.57574844e-01
-7.35625267e-01 -6.05398417e-02 1.18944573e+00 -1.30750611e-01
7.35415876e-01 -2.40321472e-01 -3.81538510e-01 -1.15936780e+00
-1.19621746e-01 -1.10577261e+00 2.56867677e-01 -3.34368885e-01
-1.53533280e+00 8.85307372e-01 -1.81597084e-01 -1.30206251e+00
-8.31990123e-01 -6.87974215e-01 -4.97968018e-01 1.38936722e+00
-1.88260531e+00 -1.22992849e+00 2.26440772e-01 6.64281428e-01
8.88709843e-01 -8.29959214e-01 1.13785982e+00 4.82787102e-01
-5.56090653e-01 1.01511443e+00 5.23754120e-01 3.52844357e-01
1.13896441e+00 -1.19023716e+00 5.48447728e-01 9.60592270e-01
2.56647110e-01 4.24095690e-01 4.41097289e-01 -4.25995290e-01
-1.27179003e+00 -1.23660994e+00 1.36240542e+00 -6.33699179e-01
3.18023801e-01 -8.29062462e-01 -1.07628310e+00 8.91493559e-01
5.30669451e-01 -1.97895080e-01 1.04574466e+00 1.64547667e-01
-6.55795038e-01 -3.55271995e-02 -1.23162198e+00 5.84374927e-02
5.64030588e-01 -1.22070944e+00 -1.00344884e+00 1.00092396e-01
8.66515219e-01 -2.63546020e-01 -7.83469021e-01 2.34199371e-02
4.22804624e-01 -6.28259182e-01 7.90278673e-01 -3.50602925e-01
-1.36758499e-02 -1.17096223e-01 -7.32868254e-01 -1.46921372e+00
1.46273253e-02 -4.63487446e-01 -1.22657262e-01 1.71238720e+00
7.41426528e-01 -6.90107644e-01 4.83757138e-01 4.64772135e-01
-5.77141345e-01 1.94068495e-02 -1.45411360e+00 -8.99156928e-01
2.95512289e-01 -9.78799403e-01 5.60636878e-01 7.98682272e-01
-1.36387199e-01 5.31709135e-01 -2.82819599e-01 5.46732664e-01
6.73583269e-01 -5.05468369e-01 6.64464235e-01 -1.08208978e+00
-2.79353887e-01 -2.17615351e-01 -4.26863462e-01 -1.04836273e+00
6.94377363e-01 -1.26811779e+00 2.07921371e-01 -1.05749822e+00
-2.17842132e-01 -2.98518062e-01 -3.46998483e-01 5.29742777e-01
5.83240995e-03 -7.43392110e-02 1.08799271e-01 1.23215795e-01
-1.28045231e-01 6.54300570e-01 3.69993895e-01 -2.59827435e-01
-1.81204304e-01 -1.70419946e-01 -4.90487278e-01 3.00643146e-01
6.21694326e-01 -6.04869068e-01 -1.24194361e-01 -4.29710925e-01
-7.02438176e-01 1.79655492e-01 5.18871732e-02 -9.54589188e-01
3.89390886e-01 5.86181104e-01 2.63614446e-01 -7.23717332e-01
7.37929702e-01 -5.52558482e-01 -1.36645705e-01 2.17645392e-01
-4.89123881e-01 -2.03337207e-01 3.55867267e-01 3.28462631e-01
-5.66967964e-01 5.18048834e-03 8.80834818e-01 5.15453398e-01
-6.76941574e-01 -1.05301283e-01 -5.07641971e-01 -5.82222678e-02
6.96731031e-01 -1.60467282e-01 -2.57646590e-01 -2.86031067e-01
-6.28376663e-01 -5.77028617e-02 1.79332703e-01 6.21112287e-01
5.29723585e-01 -1.25357366e+00 -1.15365338e+00 8.73410165e-01
3.32707524e-01 -4.60598558e-01 2.83502489e-01 6.43553376e-01
1.88534006e-01 8.07341635e-01 3.30482349e-02 -9.57443297e-01
-1.49421489e+00 2.02020213e-01 1.75120607e-01 -8.03902373e-02
-2.43536934e-01 1.01367211e+00 2.14063302e-01 -1.08919334e+00
4.67074484e-01 -1.75750732e-01 1.18165165e-02 1.59589335e-01
5.26129484e-01 9.98227745e-02 5.07856786e-01 -1.29046881e+00
-7.51562357e-01 2.44986549e-01 -2.79289752e-01 -7.62002349e-01
1.37156570e+00 -2.53650039e-01 5.67515254e-01 3.77337754e-01
1.57980120e+00 2.15054289e-01 -1.19479156e+00 -5.31411707e-01
1.62171543e-01 1.73783712e-02 3.80620301e-01 -1.08177888e+00
-7.16129541e-01 1.25274491e+00 7.64243543e-01 -3.13778788e-01
8.63110662e-01 2.15579689e-01 7.60953844e-01 -6.04387671e-02
-5.47578670e-02 -1.08497167e+00 -3.78415972e-01 6.60005033e-01
7.83048630e-01 -1.35686553e+00 -6.17276967e-01 -2.55256724e-02
-1.30367219e+00 8.16371441e-01 3.90630335e-01 3.87328684e-01
7.76398242e-01 4.29394156e-01 5.05119622e-01 1.61354467e-01
-1.05704749e+00 -2.04640403e-01 4.75120783e-01 8.55244279e-01
6.25850618e-01 2.85069376e-01 2.69942969e-01 8.04599285e-01
-4.83091593e-01 -3.59205097e-01 1.33844525e-01 7.12726831e-01
-1.73026189e-01 -1.42116976e+00 -5.37187517e-01 1.73377380e-01
-3.46027642e-01 -3.34566504e-01 -1.87595055e-01 4.41217959e-01
8.87510274e-03 1.29935372e+00 1.00986235e-01 -5.04233181e-01
2.07911238e-01 8.76436234e-01 -3.66132893e-02 -6.87451124e-01
-7.82922983e-01 1.66917831e-01 2.85769314e-01 -2.40532309e-01
2.56913044e-02 -1.05545306e+00 -9.74693716e-01 5.25620691e-02
-3.13900203e-01 2.32282609e-01 1.21215463e+00 8.83218169e-01
6.23036265e-01 3.32030982e-01 6.75677836e-01 -5.23731411e-01
-9.36293960e-01 -1.26139390e+00 -4.67244774e-01 2.26549655e-01
6.91217959e-01 -4.78121519e-01 -5.53503931e-01 2.79692620e-01] | [14.358796119689941, 6.249532699584961] |
11ab5378-57ae-4bc2-83ed-8b8a6179c688 | heterogeneous-trajectory-forecasting-via-risk | 2211.00848 | null | https://arxiv.org/abs/2211.00848v2 | https://arxiv.org/pdf/2211.00848v2.pdf | Heterogeneous Trajectory Forecasting via Risk and Scene Graph Learning | Heterogeneous trajectory forecasting is critical for intelligent transportation systems, but it is challenging because of the difficulty of modeling the complex interaction relations among the heterogeneous road agents as well as their agent-environment constraints. In this work, we propose a risk and scene graph learning method for trajectory forecasting of heterogeneous road agents, which consists of a Heterogeneous Risk Graph (HRG) and a Hierarchical Scene Graph (HSG) from the aspects of agent category and their movable semantic regions. HRG groups each kind of road agent and calculates their interaction adjacency matrix based on an effective collision risk metric. HSG of the driving scene is modeled by inferring the relationship between road agents and road semantic layout aligned by the road scene grammar. Based on this formulation, we can obtain effective trajectory forecasting in driving situations, and superior performance to other state-of-the-art approaches is demonstrated by exhaustive experiments on the nuScenes, ApolloScape, and Argoverse datasets. | ['Jianru Xue', 'Hongkai Yu', 'Pu Zhang', 'Chen Zhu', 'Jianwu Fang'] | 2022-11-02 | null | null | null | null | ['trajectory-forecasting'] | ['computer-vision'] | [-1.81071103e-01 3.11615348e-01 -1.08287513e-01 -4.48171824e-01
-4.44501638e-01 -4.45319384e-01 9.64304149e-01 8.78004432e-02
-4.08078097e-02 4.51347023e-01 3.73723984e-01 -5.10063112e-01
-5.37137926e-01 -1.37210274e+00 -6.12506211e-01 -4.94401783e-01
-3.35575670e-01 9.53787804e-01 7.33421326e-01 -4.92899626e-01
1.38137087e-01 6.49369478e-01 -1.78419244e+00 -6.85716867e-02
1.34633362e+00 6.88398361e-01 3.35968494e-01 5.65038443e-01
-2.73933679e-01 1.10880935e+00 -9.70782712e-02 -5.01744568e-01
1.03934526e-01 6.23448528e-02 -5.16815126e-01 6.59174100e-02
2.99298733e-01 -1.69568077e-01 -8.43316555e-01 7.09830940e-01
5.55155464e-02 6.05971873e-01 1.15639830e+00 -1.83751965e+00
-9.60214715e-03 5.97193241e-01 -1.87677845e-01 7.79673830e-02
8.57148468e-02 5.60211390e-02 7.98302650e-01 -4.91462678e-01
8.76818120e-01 1.42350292e+00 2.40486100e-01 -2.99022812e-02
-7.01217651e-01 -4.14288640e-01 7.64515162e-01 8.67237151e-01
-1.51425803e+00 -4.16457914e-02 7.97212183e-01 -8.16720247e-01
8.12844813e-01 3.34470600e-01 5.91947556e-01 6.20828211e-01
3.38733673e-01 6.32278860e-01 6.44161284e-01 1.78249553e-01
3.03635359e-01 -2.17249766e-01 4.26938802e-01 9.41613376e-01
4.61515523e-02 2.83199430e-01 -1.11861981e-01 1.04261160e-01
1.07456066e-01 -1.30286694e-01 3.44967157e-01 -5.88190258e-01
-9.84091103e-01 6.61339700e-01 6.74440861e-01 -1.90162256e-01
-4.54250664e-01 -2.76748780e-02 2.79583305e-01 -2.42344454e-01
5.47236264e-01 -1.66589677e-01 2.03133658e-01 5.43367974e-02
-3.58197063e-01 6.42526329e-01 6.77130044e-01 1.35511351e+00
8.59596729e-01 1.30196540e-02 -4.85332310e-03 5.35384595e-01
3.33696991e-01 7.47550666e-01 -4.57561344e-01 -8.63500059e-01
9.50205803e-01 9.82813656e-01 1.20036155e-01 -1.50283372e+00
-7.62297094e-01 -7.61018470e-02 -7.62204468e-01 1.92610189e-01
4.03859854e-01 -8.77895877e-02 -7.53806293e-01 1.46881640e+00
7.14247882e-01 6.73228085e-01 1.83388129e-01 7.48078525e-01
8.20374489e-01 9.12988484e-01 3.47367525e-01 1.06898449e-01
1.03376222e+00 -1.20702040e+00 -6.23553097e-01 -2.09071144e-01
8.54567349e-01 -2.16113731e-01 7.34747946e-01 -2.38335624e-01
-8.47925484e-01 -5.07765114e-01 -7.87616432e-01 -7.27347359e-02
-8.21904361e-01 -1.38791978e-01 6.18174970e-01 3.59370381e-01
-8.82035792e-01 5.26805669e-02 -5.81856549e-01 -3.36816788e-01
2.49637201e-01 1.51685044e-01 -6.00814074e-02 -2.28689507e-01
-1.26295197e+00 1.17612255e+00 2.29138017e-01 2.55717546e-01
-1.03947532e+00 -7.48620510e-01 -1.14530504e+00 -6.43592477e-02
6.54376984e-01 -7.02052832e-01 7.10268199e-01 -6.41504163e-03
-1.30788147e+00 2.36222744e-01 -2.34836698e-01 -2.45609269e-01
8.10817897e-01 3.17417562e-01 -8.22579086e-01 -9.82205048e-02
7.40635619e-02 4.03310537e-01 4.05407876e-01 -1.51697242e+00
-1.35100615e+00 -5.11080742e-01 2.87195891e-01 6.42276168e-01
4.32071835e-01 -2.16602266e-01 -5.26498139e-01 -1.50124133e-02
-1.90822743e-02 -1.26929617e+00 -5.27316988e-01 -5.56888878e-01
-6.93736076e-01 -6.02731824e-01 9.77543056e-01 -6.15422070e-01
1.34982920e+00 -1.99356997e+00 2.34471962e-01 6.73374653e-01
3.95817876e-01 -9.42604765e-02 -2.93549091e-01 5.92699349e-01
4.07751173e-01 -1.27173319e-01 -2.03492478e-01 -3.27323489e-02
3.41274023e-01 5.88482618e-02 -2.64712363e-01 4.22878176e-01
-3.31180543e-01 8.50860059e-01 -1.03915370e+00 -5.40008903e-01
5.57022333e-01 3.45717520e-01 -1.42159030e-01 2.51680851e-01
-3.02893877e-01 1.50750339e-01 -8.56004238e-01 3.34649950e-01
7.74005830e-01 3.18975598e-01 1.87666163e-01 -2.47164771e-01
-5.31324685e-01 5.83635978e-02 -1.16238546e+00 1.03764117e+00
-3.99365574e-01 6.83608234e-01 -3.14794630e-01 -6.09515488e-01
9.52770233e-01 -1.74600437e-01 4.90658879e-01 -8.24226379e-01
-1.58044919e-01 -1.24531865e-01 -7.39743114e-02 -7.02220142e-01
7.51894653e-01 5.74989498e-01 -2.77524233e-01 3.16016406e-01
-7.43965685e-01 -1.53838322e-01 4.02835011e-01 5.75649202e-01
9.58292663e-01 7.77375177e-02 -2.35906169e-01 -3.05571526e-01
5.35500228e-01 3.31811577e-01 3.11809689e-01 5.43105185e-01
-1.44620255e-01 2.50845607e-02 5.14289498e-01 -7.80135095e-01
-9.89274025e-01 -1.30443299e+00 1.74346551e-01 8.46610725e-01
9.17119503e-01 -4.36903000e-01 -9.15899932e-01 -7.62716115e-01
7.86990896e-02 9.89276588e-01 -4.04008627e-01 -1.35447010e-01
-8.13193858e-01 -6.13097131e-01 2.08241060e-01 4.49769139e-01
5.74663460e-01 -6.46375179e-01 -1.17927521e-01 9.09487605e-02
-3.24150294e-01 -1.51922166e+00 -4.69979078e-01 -8.85362148e-01
-1.67840779e-01 -1.20850194e+00 1.43994287e-01 -6.52509868e-01
7.87814200e-01 6.80920780e-01 7.87858546e-01 2.11551096e-02
9.14100707e-02 2.18437836e-01 -4.14704829e-02 -2.02012137e-01
-3.65802050e-01 3.49604219e-01 -8.71965662e-02 2.90571660e-01
7.20780566e-02 -3.44137341e-01 -4.61739987e-01 7.82886803e-01
-4.02647167e-01 5.82398236e-01 2.15020463e-01 1.03574693e-01
5.59007108e-01 6.70550406e-01 4.57998186e-01 -9.22472656e-01
2.06930116e-01 -8.27265263e-01 -9.86481190e-01 3.78580242e-01
-5.37680089e-01 -3.12119126e-01 4.53193665e-01 8.79444778e-02
-1.39762771e+00 5.74173443e-02 1.55135781e-01 7.72235915e-02
-3.00781846e-01 6.36496007e-01 -6.44205987e-01 -9.29438975e-03
1.92948207e-01 -1.66436821e-01 -2.73023278e-01 1.77562192e-01
7.16004550e-01 4.78972048e-01 3.54441255e-01 -4.56688076e-01
1.04010844e+00 6.59121573e-01 4.36924189e-01 -1.00181735e+00
-3.68765354e-01 -3.27305198e-01 -8.11942101e-01 -9.89213765e-01
1.00203407e+00 -8.01683128e-01 -9.92313564e-01 6.04387343e-01
-9.66827095e-01 -6.48013175e-01 2.59695977e-01 6.16802633e-01
-6.59806490e-01 1.37687549e-01 -3.29940230e-01 -8.61948907e-01
4.44581896e-01 -1.39862537e+00 9.06389713e-01 2.90161576e-02
2.23359704e-01 -1.15392220e+00 -1.58638675e-02 6.20844662e-01
1.55303523e-01 4.83889669e-01 1.32225382e+00 -1.17632620e-01
-1.44906330e+00 -1.52462736e-01 -4.64360118e-01 -5.88883221e-01
-1.13955617e-01 2.50127554e-01 -5.32331049e-01 1.54074237e-01
-7.27555454e-01 5.10762513e-01 5.70597112e-01 4.59805042e-01
8.75820637e-01 -2.79410809e-01 -8.99136066e-01 4.24766421e-01
1.23020613e+00 4.38795418e-01 8.74769390e-01 2.71649331e-01
1.30049658e+00 1.31911290e+00 9.21870887e-01 2.73372456e-02
1.42934978e+00 1.06447566e+00 7.03459442e-01 -6.75284192e-02
-1.74464092e-01 -4.49866056e-01 2.09729478e-01 6.03952587e-01
-3.03713590e-01 -7.09771454e-01 -1.25383508e+00 4.93841678e-01
-2.33888531e+00 -1.16156209e+00 -8.45250309e-01 1.90719759e+00
-3.45911860e-01 2.80485395e-02 4.01766658e-01 -4.39886421e-01
9.62667942e-01 3.82335544e-01 -6.95304811e-01 -1.14769481e-01
-1.35251701e-01 -1.02567756e+00 7.81364620e-01 1.01335895e+00
-9.47165489e-01 1.23623669e+00 5.79886389e+00 9.44129288e-01
-3.78608942e-01 -5.62220812e-02 6.32742882e-01 3.26458633e-01
-6.81576729e-01 -9.80627909e-03 -1.03582108e+00 4.88645703e-01
9.75947678e-01 -2.76213706e-01 7.57124603e-01 5.51939428e-01
4.98265982e-01 -5.42017557e-02 -7.73653328e-01 5.18981278e-01
2.36842427e-02 -1.52584171e+00 1.22099519e-01 2.24905610e-01
7.12066948e-01 3.55009645e-01 1.03888279e-02 2.94399649e-01
7.09591329e-01 -7.64148891e-01 7.40910351e-01 7.44604170e-01
3.74532849e-01 -1.03202665e+00 4.23323244e-01 4.37237114e-01
-1.83360159e+00 -1.76778242e-01 -1.30937740e-01 1.60741016e-01
7.29514122e-01 1.59785092e-01 -9.34961319e-01 1.05734122e+00
2.78063089e-01 8.90413523e-01 -3.83782536e-01 8.09038639e-01
-1.97053584e-03 3.13705236e-01 -2.20601335e-01 2.10532490e-02
3.71415943e-01 -8.99000943e-01 7.27068007e-01 8.64931107e-01
3.85447025e-01 2.47643992e-01 4.22021866e-01 5.89792252e-01
2.91472763e-01 1.16116084e-01 -1.05685019e+00 3.50940794e-01
7.18059123e-01 1.27091122e+00 -7.13340580e-01 -2.67819852e-01
-4.96135235e-01 1.07082069e-01 4.67016131e-01 7.44759560e-01
-1.17357886e+00 2.61809658e-02 8.63490283e-01 2.89613247e-01
-1.52288839e-01 -5.36261499e-01 -1.84924662e-01 -8.52567732e-01
-1.23117089e-01 -1.77075639e-01 3.25446874e-01 -8.37653637e-01
-1.03580606e+00 7.86538422e-01 3.43307585e-01 -1.32866669e+00
-2.98562765e-01 -1.84702933e-01 -6.86558247e-01 5.51228464e-01
-1.76858389e+00 -1.71017992e+00 -5.61494529e-01 6.01837575e-01
4.37381655e-01 -3.28483105e-01 2.41389647e-01 3.08116823e-01
-7.92076766e-01 5.97073399e-02 1.36497185e-01 -2.78630763e-01
-1.76773071e-01 -9.68017161e-01 7.42029965e-01 7.91229725e-01
-2.49531314e-01 -1.61344036e-01 4.57821369e-01 -9.70912695e-01
-1.38599646e+00 -1.70977724e+00 1.00230777e+00 -5.42189419e-01
8.12833130e-01 -4.77657467e-01 -6.46307290e-01 8.31147134e-01
7.07178190e-02 -3.33416581e-01 2.63555974e-01 1.41829893e-01
-9.34530869e-02 -2.20862105e-01 -8.24936092e-01 1.18052638e+00
1.69548166e+00 -3.17472905e-01 -1.03675030e-01 6.08009994e-01
8.05104554e-01 -4.69588608e-01 -6.56700790e-01 4.49204206e-01
4.47506994e-01 -6.98183537e-01 1.04694045e+00 -5.73376119e-01
1.27954438e-01 -7.27629125e-01 -2.55946308e-01 -1.40786612e+00
-4.98528630e-01 -2.28947043e-01 1.56362712e-01 1.18441868e+00
5.62046349e-01 -7.51526356e-01 6.72780991e-01 8.79613161e-01
-8.06464911e-01 -4.51081485e-01 -8.40964973e-01 -7.43262768e-01
-1.97897181e-01 -6.42960310e-01 1.28324211e+00 6.91955388e-01
-1.47128001e-01 2.19207779e-01 -2.71104246e-01 8.21841419e-01
7.78915226e-01 2.02897191e-01 1.05470765e+00 -1.18829513e+00
4.28356171e-01 -5.42957366e-01 -6.46813989e-01 -7.27510989e-01
8.67764533e-01 -9.12492096e-01 -3.83567326e-02 -1.89828241e+00
-4.02658172e-02 -9.55178261e-01 2.47242346e-01 -1.22995198e-01
-2.58239597e-04 -3.20071489e-01 1.44148946e-01 4.31550443e-02
-7.60009050e-01 6.46364868e-01 1.25297081e+00 -4.41711545e-01
-2.33131051e-01 1.58336759e-01 -6.94246367e-02 7.45599091e-01
6.07189596e-01 -2.04518493e-02 -9.63639498e-01 -4.65769768e-01
2.56706625e-01 3.47973436e-01 3.00110072e-01 -7.90491164e-01
5.28833926e-01 -7.01898992e-01 -5.13823211e-01 -1.16708434e+00
3.55793238e-01 -9.88834321e-01 7.02784359e-01 1.22374989e-01
-3.35993141e-01 2.51726359e-01 -6.53460203e-03 7.64751434e-01
-3.32521610e-02 3.58876735e-01 2.44026467e-01 3.33294123e-01
-1.15437043e+00 7.52868474e-01 -6.88820839e-01 -5.89579083e-02
1.50527084e+00 -2.13034287e-01 -9.29843247e-01 -3.63115460e-01
-5.17065048e-01 7.95161307e-01 4.43916202e-01 6.48020029e-01
6.44697487e-01 -1.46624279e+00 -7.51365185e-01 1.74181327e-01
3.19515198e-01 6.38147891e-02 8.71662796e-01 6.36709154e-01
-4.61398005e-01 3.13501209e-01 -5.90852797e-02 -5.19798040e-01
-1.08923852e+00 6.52760088e-01 2.87081569e-01 -1.24747016e-01
-7.91697800e-01 2.84057230e-01 6.14232123e-01 -5.67601204e-01
-2.73442473e-02 -1.71147346e-01 -7.04726458e-01 1.39435962e-01
2.81347245e-01 1.01414728e+00 -1.43442288e-01 -1.43066061e+00
-2.31016517e-01 7.55687952e-01 1.99307248e-01 5.37685156e-02
9.65850770e-01 -5.96675515e-01 5.57400584e-02 1.69304058e-01
7.94638336e-01 -1.00232400e-01 -1.32500434e+00 3.62641253e-02
6.08249716e-02 -5.09922683e-01 -9.07069352e-03 -4.66556221e-01
-1.07061899e+00 4.92707103e-01 3.57323080e-01 4.56416637e-01
6.42364085e-01 1.15410745e-01 9.48306739e-01 2.08486512e-01
6.95898473e-01 -1.19329178e+00 -4.47224140e-01 6.16926372e-01
7.45712817e-01 -9.88958180e-01 -3.30437869e-01 -1.10356319e+00
-8.86276245e-01 6.73598289e-01 7.14833081e-01 6.13820553e-02
9.04787719e-01 2.56794691e-02 1.09041417e-02 -4.17571515e-01
-6.27736628e-01 -5.38797855e-01 3.89207155e-01 7.93020546e-01
-4.82014418e-01 5.94804823e-01 1.15674198e-01 1.96881577e-01
-2.82833487e-01 -4.22759444e-01 4.67379957e-01 2.95055240e-01
-3.65697175e-01 -8.29826117e-01 3.02644279e-02 4.32268620e-01
5.41816413e-01 3.31836969e-01 -2.04791185e-02 1.00278199e+00
1.24414325e-01 1.22345400e+00 3.60302210e-01 -6.87001944e-01
7.39069641e-01 -5.12905419e-01 3.56892236e-02 -3.38961244e-01
-9.94612798e-02 -2.62822479e-01 6.36879027e-01 -6.96818590e-01
-2.12616727e-01 -9.46417272e-01 -1.59828508e+00 -7.31862247e-01
-4.25766893e-02 1.55772790e-01 7.37828434e-01 1.21782315e+00
5.63975394e-01 5.31753540e-01 9.37907159e-01 -8.23116064e-01
3.09258074e-01 -2.82678962e-01 -6.91573441e-01 4.35358077e-01
1.23035468e-01 -9.82489765e-01 -1.81911111e-01 -2.25361213e-01] | [5.946789741516113, 0.8864409327507019] |
d51949b4-5aa4-4b45-8192-69e62b353c7a | a-collaborative-transfer-learning-framework | 2306.16425 | null | https://arxiv.org/abs/2306.16425v1 | https://arxiv.org/pdf/2306.16425v1.pdf | A Collaborative Transfer Learning Framework for Cross-domain Recommendation | In the recommendation systems, there are multiple business domains to meet the diverse interests and needs of users, and the click-through rate(CTR) of each domain can be quite different, which leads to the demand for CTR prediction modeling for different business domains. The industry solution is to use domain-specific models or transfer learning techniques for each domain. The disadvantage of the former is that the data from other domains is not utilized by a single domain model, while the latter leverage all the data from different domains, but the fine-tuned model of transfer learning may trap the model in a local optimum of the source domain, making it difficult to fit the target domain. Meanwhile, significant differences in data quantity and feature schemas between different domains, known as domain shift, may lead to negative transfer in the process of transferring. To overcome these challenges, we propose the Collaborative Cross-Domain Transfer Learning Framework (CCTL). CCTL evaluates the information gain of the source domain on the target domain using a symmetric companion network and adjusts the information transfer weight of each source domain sample using the information flow network. This approach enables full utilization of other domain data while avoiding negative migration. Additionally, a representation enhancement network is used as an auxiliary task to preserve domain-specific features. Comprehensive experiments on both public and real-world industrial datasets, CCTL achieved SOTA score on offline metrics. At the same time, the CCTL algorithm has been deployed in Meituan, bringing 4.37% CTR and 5.43% GMV lift, which is significant to the business. | ['Dong Wang', 'Xingxing Wang', 'Bo Zhang', 'Pengye Zhang', 'Wei zhang'] | 2023-06-26 | null | null | null | null | ['transfer-learning', 'click-through-rate-prediction'] | ['miscellaneous', 'miscellaneous'] | [-1.82027921e-01 -4.16762859e-01 -5.17236888e-01 -2.81092227e-01
-2.46539071e-01 -4.22849029e-01 2.45178297e-01 -2.08604187e-01
-1.97453678e-01 6.40666366e-01 -6.34143651e-02 -8.05946812e-02
-2.64334291e-01 -1.04679668e+00 -5.95540583e-01 -7.12021887e-01
2.11594209e-01 4.65499640e-01 4.50223565e-01 -6.01316631e-01
2.63018698e-01 1.98225290e-01 -1.10385001e+00 5.21837294e-01
1.24321342e+00 1.19198704e+00 6.02929831e-01 -2.26595581e-01
-6.05557859e-01 2.73829490e-01 -5.78547180e-01 -4.84300584e-01
5.42952120e-01 -1.62877351e-01 -2.69218415e-01 -2.09054083e-01
-6.29722029e-02 -2.25807071e-01 -5.08465588e-01 9.47291017e-01
4.79792327e-01 2.30212882e-01 6.52218938e-01 -1.41071928e+00
-9.23530698e-01 4.24030989e-01 -7.91151345e-01 1.06261130e-02
-1.35987639e-01 -3.05795036e-02 7.33433723e-01 -7.15374470e-01
5.12917340e-01 1.13737130e+00 4.24518198e-01 2.63628989e-01
-1.11316335e+00 -1.23728514e+00 4.93786901e-01 2.06220642e-01
-1.18521059e+00 1.24706596e-01 9.90422547e-01 -4.85073030e-01
3.65718573e-01 -2.06895396e-01 4.02521044e-01 1.08588862e+00
2.67063260e-01 6.40984714e-01 8.17838609e-01 3.92394364e-02
5.82378879e-02 8.63229632e-01 -5.52988462e-02 -1.72636211e-02
1.60615489e-01 6.27532378e-02 -4.35250372e-01 1.93169072e-01
7.85424709e-01 5.44329166e-01 -8.53779614e-02 -8.09054911e-01
-9.68574643e-01 8.22206557e-01 7.11474121e-01 3.14757645e-01
-2.63492078e-01 -7.21918225e-01 5.12219906e-01 7.83233047e-01
4.77485925e-01 1.71278447e-01 -9.71186638e-01 1.13446943e-01
-3.73945385e-01 1.78064883e-01 7.01397121e-01 1.31927335e+00
9.70703125e-01 -1.57586426e-01 -5.12111709e-02 1.07055831e+00
1.95631802e-01 4.42530423e-01 7.39077449e-01 -5.66919267e-01
1.11860108e+00 9.99120235e-01 1.20417230e-01 -1.27169383e+00
-1.71409175e-02 -6.03186309e-01 -8.88734579e-01 -5.30342944e-02
5.75481057e-01 -1.65927202e-01 -5.15641928e-01 1.75028217e+00
3.42365026e-01 -1.39745092e-02 -2.77990364e-02 9.15810406e-01
4.43991810e-01 8.74543905e-01 1.00983165e-01 -1.98029444e-01
9.45634425e-01 -1.00115728e+00 -5.10308862e-01 -2.10581988e-01
5.16161263e-01 -7.51099706e-01 1.15381885e+00 5.26471734e-01
-6.50024831e-01 -8.73487711e-01 -1.01587462e+00 3.80607277e-01
-5.34660578e-01 1.61035538e-01 4.23424184e-01 3.91911000e-01
-2.42648363e-01 6.28112912e-01 -1.54205248e-01 -2.02949822e-01
3.54520708e-01 3.28541338e-01 -7.77092800e-02 -4.74927872e-01
-1.63426840e+00 6.24500096e-01 5.61734021e-01 -1.57404557e-01
-4.03037190e-01 -1.07523799e+00 -2.34443367e-01 1.81187585e-01
3.66171777e-01 -2.25664124e-01 8.36879194e-01 -1.31926906e+00
-1.33580983e+00 1.51892781e-01 2.49824062e-01 -6.94938004e-02
8.41956139e-01 -1.68541759e-01 -1.13671446e+00 -4.21815902e-01
1.78787053e-01 2.96819121e-01 7.66297936e-01 -1.13355672e+00
-1.08379042e+00 -5.08266389e-01 -1.69197604e-01 4.22650546e-01
-8.80161047e-01 -4.06043559e-01 -6.27031147e-01 -5.80027819e-01
4.19959538e-02 -8.08601916e-01 9.43875611e-02 -2.15555534e-01
1.63130715e-01 -1.87184647e-01 1.11375725e+00 -8.09652984e-01
1.40387988e+00 -2.40046835e+00 -1.38231823e-02 4.26529914e-01
6.01642430e-02 4.06214297e-01 -3.22637886e-01 3.52401525e-01
-1.87408015e-01 -5.34064956e-02 1.96477249e-01 3.75476509e-01
-1.12996615e-01 1.40334353e-01 -3.55356544e-01 7.33131776e-03
9.27866250e-02 4.54724640e-01 -7.83148587e-01 -2.31386766e-01
8.79064798e-02 1.61301345e-01 -6.38992667e-01 1.02099270e-01
-8.78252164e-02 6.34802222e-01 -7.56468117e-01 4.41520423e-01
1.14755726e+00 -3.03843677e-01 3.65597636e-01 -3.39301318e-01
-6.50683418e-02 1.33048341e-01 -1.43410623e+00 1.59859657e+00
-5.84922373e-01 2.96682548e-02 -1.49104409e-02 -1.34793508e+00
1.28969646e+00 3.05488221e-02 8.19859385e-01 -1.34514332e+00
-1.44779980e-01 2.86989421e-01 2.87420303e-02 -2.73032725e-01
2.36340180e-01 1.22150395e-03 -1.20000146e-01 2.43518740e-01
-8.12935978e-02 5.03943384e-01 -1.59891546e-02 -2.68852748e-02
6.87014818e-01 1.31001100e-01 -8.01748857e-02 -6.21560849e-02
5.52397728e-01 -1.07496856e-02 9.05123830e-01 3.51424068e-01
-2.35306710e-01 1.25742778e-01 3.44380111e-01 -4.34402913e-01
-1.07105172e+00 -1.15117490e+00 -7.87665248e-02 1.30292714e+00
6.64219379e-01 8.35585967e-02 -1.93514973e-01 -9.05709386e-01
3.30886722e-01 6.35543168e-01 -2.02035397e-01 -6.41338766e-01
-5.05580068e-01 -5.08662820e-01 -6.99991286e-02 3.03236723e-01
8.65874767e-01 -8.22870374e-01 2.15229779e-01 4.11556870e-01
-3.82747948e-01 -7.95932353e-01 -6.36035085e-01 -6.18796758e-02
-1.05934060e+00 -8.12059462e-01 -8.65863800e-01 -7.86084175e-01
4.94518787e-01 5.00950217e-01 8.96186590e-01 -3.91222656e-01
2.68815696e-01 -1.47188485e-01 -4.82319623e-01 -3.42753619e-01
-2.31565207e-01 4.37318504e-01 1.83841735e-01 1.41722888e-01
7.43036330e-01 -4.62129861e-01 -7.06757545e-01 9.72174942e-01
-7.23036885e-01 -1.55442804e-01 6.21230662e-01 9.64126229e-01
2.63859451e-01 4.72875506e-01 1.08485472e+00 -8.77110064e-01
7.52364755e-01 -1.03555298e+00 -6.07474744e-01 1.99507251e-01
-1.07049155e+00 -3.97299081e-01 9.19517517e-01 -9.68895197e-01
-1.44962800e+00 -1.94003120e-01 4.98790532e-01 -4.14835125e-01
1.30678952e-01 4.53372389e-01 -5.49161971e-01 2.70214796e-01
5.48032641e-01 1.15448989e-01 2.20225424e-01 -5.56786418e-01
1.80426478e-01 9.27951694e-01 3.98778655e-02 -4.81878310e-01
9.27491665e-01 4.50111739e-02 -5.33625007e-01 -2.30683967e-01
-6.28702581e-01 -4.65683311e-01 -5.19421637e-01 -9.41931829e-02
4.97560352e-01 -1.03412652e+00 -3.93584371e-01 3.43430668e-01
-9.01474833e-01 2.65017319e-02 -2.20573261e-01 5.53720355e-01
-1.43371925e-01 2.54616886e-01 -2.71204829e-01 -4.55808699e-01
-6.12176433e-02 -1.02625573e+00 4.20325011e-01 5.08534491e-01
1.42460257e-01 -8.75793040e-01 -2.51362652e-01 4.50517237e-01
6.14321589e-01 -2.83481985e-01 1.06081522e+00 -8.87324452e-01
-5.26382565e-01 -2.53628701e-01 -5.19849539e-01 7.38798320e-01
4.09486234e-01 -4.31618452e-01 -5.95721126e-01 -5.36601543e-01
9.65169258e-03 -7.84507394e-02 3.38975608e-01 2.84040123e-01
1.02948332e+00 -2.07811728e-01 -4.72687483e-01 3.25769365e-01
1.30721724e+00 7.56623447e-01 6.06971025e-01 5.17771423e-01
5.65048575e-01 7.12053359e-01 1.16210115e+00 4.56996948e-01
2.01738164e-01 7.01717198e-01 9.88083631e-02 -2.62605548e-01
1.19631797e-01 -4.34438646e-01 5.19464910e-01 8.49945188e-01
9.87081826e-02 1.76696177e-03 -4.86878365e-01 3.41978967e-01
-1.87878442e+00 -8.10079694e-01 2.33122688e-02 2.39190125e+00
6.42678738e-01 2.97163785e-01 1.77094176e-01 -3.00227284e-01
1.01954675e+00 -3.83367807e-01 -9.99335885e-01 -7.44717866e-02
1.32138252e-01 -1.69562548e-01 5.99727571e-01 -2.28049845e-01
-8.83174837e-01 6.79971457e-01 5.23827505e+00 1.05369949e+00
-1.32727265e+00 1.06806494e-01 5.06308079e-01 -5.61159812e-02
-1.26825467e-01 -2.04143554e-01 -6.96823001e-01 9.83194232e-01
7.68291175e-01 -5.14245510e-01 6.80771112e-01 1.09089613e+00
1.81235731e-01 3.88501525e-01 -7.81727791e-01 9.18178380e-01
-3.06997746e-01 -8.88165474e-01 3.06917075e-03 3.45961064e-01
7.61020720e-01 1.49053812e-01 3.10561210e-01 9.43201840e-01
2.31018096e-01 -3.71983737e-01 2.51196682e-01 3.68430436e-01
7.50589311e-01 -9.58442807e-01 7.83293784e-01 5.26034236e-01
-1.21267605e+00 -5.53747594e-01 -7.32838511e-01 2.01383218e-01
-1.49208412e-01 6.60389602e-01 -7.86915123e-01 8.57231677e-01
9.98166919e-01 8.28805089e-01 -3.01587731e-01 7.90953338e-01
4.07030761e-01 4.42837775e-01 -3.65126319e-02 2.04559758e-01
-1.36270657e-01 -8.05082858e-01 3.34998310e-01 7.80195475e-01
7.08891749e-01 -4.10863012e-01 3.50154102e-01 8.51076543e-01
-2.28401408e-01 4.55405474e-01 -6.54804528e-01 -1.83052178e-02
6.36285007e-01 1.19643342e+00 -2.17154548e-02 -1.98429719e-01
-6.89305842e-01 9.70174134e-01 8.04926082e-02 4.42391574e-01
-1.09149945e+00 -5.48827350e-01 7.60634243e-01 2.10169837e-01
4.53769267e-01 1.02473535e-01 -2.86839038e-01 -1.10799658e+00
1.07938424e-01 -8.95323813e-01 4.81372118e-01 -3.88386935e-01
-1.88270462e+00 1.29444495e-01 -1.85413778e-01 -1.99745321e+00
1.67221799e-01 -4.53945547e-01 -3.57158214e-01 1.16889548e+00
-1.67376149e+00 -8.87059748e-01 -3.10671419e-01 7.96330631e-01
5.42072713e-01 -8.04553330e-01 4.17024553e-01 1.00143528e+00
-5.19543648e-01 8.14497709e-01 6.07367694e-01 -1.15802296e-01
1.27699280e+00 -8.56472075e-01 1.29702128e-02 3.99595767e-01
-5.22355795e-01 6.56020761e-01 2.74416000e-01 -5.82095742e-01
-1.14821100e+00 -1.33135903e+00 5.94074726e-01 -2.25613326e-01
7.08500385e-01 -1.68057621e-01 -1.25329232e+00 4.43138480e-01
-2.16238424e-01 -2.45529473e-01 7.04884529e-01 2.92242199e-01
-4.14280832e-01 -7.69117117e-01 -1.31337368e+00 4.34121460e-01
8.97837400e-01 -1.33514419e-01 -4.03409481e-01 2.27523834e-01
9.14373696e-01 -4.16397527e-02 -1.15594292e+00 2.96377122e-01
5.56803763e-01 -8.81179690e-01 1.00886071e+00 -6.84413433e-01
4.80787784e-01 -3.41991425e-01 -1.18694551e-01 -1.47805071e+00
-6.51672959e-01 2.67981142e-02 6.62597641e-02 1.60410035e+00
4.29211676e-01 -9.63080883e-01 7.12322831e-01 8.00370693e-01
9.16149467e-02 -2.85903841e-01 -7.96581328e-01 -9.69268203e-01
2.61754483e-01 -7.24935979e-02 8.25620830e-01 1.24283099e+00
-5.49693219e-02 5.61192572e-01 -4.08003449e-01 -9.90365744e-02
2.37743512e-01 2.59816378e-01 9.00014222e-01 -1.52557623e+00
-1.68624267e-01 -2.75679022e-01 8.61545131e-02 -1.22884059e+00
-1.07605174e-01 -1.05498385e+00 -4.37553972e-01 -1.14569771e+00
8.94018114e-02 -8.77138376e-01 -7.92182684e-01 3.40620339e-01
1.51780888e-01 -3.14007968e-01 2.65544146e-01 5.07818818e-01
-3.83494288e-01 5.27284026e-01 1.67989051e+00 -2.40704849e-01
-6.12726271e-01 2.81001985e-01 -1.05185616e+00 4.04917628e-01
8.65058661e-01 -4.21693355e-01 -7.68853486e-01 -3.17413867e-01
-2.59849131e-02 -4.42335717e-02 -1.13987125e-01 -7.84099638e-01
1.84734076e-01 -3.74386132e-01 7.15744078e-01 -5.28623939e-01
-1.08952718e-02 -1.31676996e+00 2.07209602e-01 3.22623849e-01
-7.90565610e-02 -1.17913671e-01 1.79589555e-01 8.76426697e-01
-2.27789596e-01 1.26548201e-01 7.76300311e-01 2.65433136e-02
-8.36953402e-01 4.19906825e-01 8.20932314e-02 -3.66184153e-02
1.14924419e+00 -3.48716646e-01 -2.78392613e-01 -2.71834165e-01
-4.25189644e-01 4.05971169e-01 3.20797324e-01 9.40903366e-01
2.70575643e-01 -1.61547673e+00 -5.60982347e-01 3.37839872e-01
3.08804601e-01 -9.14038792e-02 7.41274238e-01 6.14637136e-01
1.02699332e-01 1.64768189e-01 -5.37783504e-01 -4.34461415e-01
-7.45898604e-01 7.63046801e-01 1.69346794e-01 -4.77920473e-01
-3.06664079e-01 2.43388712e-01 4.86307293e-01 -8.94725800e-01
-1.88818593e-02 4.34935652e-02 -3.17990780e-01 2.97021955e-01
3.76895666e-01 5.56222081e-01 2.20738292e-01 -2.49578953e-01
-1.51829734e-01 3.42057645e-01 -4.23491359e-01 3.08762729e-01
1.23049486e+00 -4.29943591e-01 1.65819034e-01 2.99748421e-01
1.23184013e+00 -1.65068641e-01 -1.17571867e+00 -6.52222276e-01
-2.03369245e-01 -6.61000669e-01 -7.16942549e-02 -1.11142230e+00
-1.33800983e+00 6.28721595e-01 8.75156105e-01 1.48320049e-01
1.37261617e+00 -2.65260756e-01 1.02091455e+00 2.34176219e-01
4.37713534e-01 -1.70523274e+00 1.97163627e-01 4.26028341e-01
7.36972213e-01 -1.31572783e+00 -2.71970272e-01 -2.87157387e-01
-1.11276639e+00 8.84747028e-01 9.95647132e-01 -1.72754489e-02
7.02235281e-01 -2.28269741e-01 -3.72829586e-02 2.60367364e-01
-5.94977975e-01 4.41223234e-01 1.40577793e-01 7.46521652e-01
2.24400043e-01 1.22735538e-01 -2.92105108e-01 9.19957042e-01
2.00322703e-01 2.20126346e-01 5.62935248e-02 5.62688828e-01
-2.01921523e-01 -1.41035962e+00 -2.50307560e-01 5.83314955e-01
-1.28474742e-01 1.63206741e-01 -3.18448506e-02 8.79146039e-01
4.23486143e-01 9.03437853e-01 4.44845185e-02 -8.32642853e-01
7.98306882e-01 2.60026623e-02 -4.22102548e-02 -3.06799024e-01
-4.16747838e-01 3.53220820e-01 -3.25495422e-01 -3.18830907e-01
4.53019654e-03 -6.12301767e-01 -1.25563562e+00 -4.12421227e-01
-3.75044674e-01 1.07554205e-01 5.71475089e-01 5.52709877e-01
6.20700240e-01 5.95221639e-01 1.12664354e+00 -3.70933473e-01
-9.38297629e-01 -9.38036442e-01 -9.55203176e-01 7.39018798e-01
-1.39260888e-01 -9.13361073e-01 -2.22720966e-01 -9.35245976e-02] | [10.088403701782227, 5.277833938598633] |
a0de62f4-c930-4632-b2d5-0044c4b2fb92 | ddlp-unsupervised-object-centric-video | 2306.05957 | null | https://arxiv.org/abs/2306.05957v1 | https://arxiv.org/pdf/2306.05957v1.pdf | DDLP: Unsupervised Object-Centric Video Prediction with Deep Dynamic Latent Particles | We propose a new object-centric video prediction algorithm based on the deep latent particle (DLP) representation. In comparison to existing slot- or patch-based representations, DLPs model the scene using a set of keypoints with learned parameters for properties such as position and size, and are both efficient and interpretable. Our method, deep dynamic latent particles (DDLP), yields state-of-the-art object-centric video prediction results on several challenging datasets. The interpretable nature of DDLP allows us to perform ``what-if'' generation -- predict the consequence of changing properties of objects in the initial frames, and DLP's compact structure enables efficient diffusion-based unconditional video generation. Videos, code and pre-trained models are available: https://taldatech.github.io/ddlp-web | ['Aviv Tamar', 'Tal Daniel'] | 2023-06-09 | null | null | null | null | ['video-generation', 'video-prediction', 'unconditional-video-generation'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [-2.02566683e-01 4.91465256e-03 -3.57402802e-01 1.02546372e-01
-4.58467454e-01 -3.72770190e-01 8.68879616e-01 -9.49445143e-02
1.07356958e-01 5.61385751e-01 6.64804637e-01 1.17632717e-01
8.68814290e-02 -8.46248686e-01 -1.28800178e+00 -8.58673930e-01
-2.37379998e-01 6.98588192e-01 6.09205246e-01 2.47743964e-01
1.74560681e-01 3.76483440e-01 -1.66254330e+00 5.99125445e-01
5.21899283e-01 9.91661608e-01 5.36172569e-01 8.99316013e-01
1.53962106e-01 1.09517252e+00 -1.99319348e-02 -2.95907646e-01
3.14461946e-01 -2.48538882e-01 -5.49337387e-01 1.21301547e-01
3.11537802e-01 -6.15540266e-01 -8.98248076e-01 6.00624144e-01
2.75897563e-01 2.46569023e-01 6.71278477e-01 -1.17323291e+00
-9.18654561e-01 3.85806262e-01 -2.61605501e-01 2.81571835e-01
4.13407922e-01 5.06638527e-01 1.07578290e+00 -1.00515497e+00
1.03617370e+00 1.35468853e+00 6.35462880e-01 5.22232413e-01
-1.23091435e+00 -2.56417632e-01 4.15939242e-01 5.89118004e-01
-1.25229001e+00 -4.69781011e-01 8.53385389e-01 -6.69923306e-01
1.03313422e+00 2.26126641e-01 1.13648748e+00 1.51302087e+00
4.75718886e-01 1.24343038e+00 4.32243437e-01 -1.96523875e-01
2.02143997e-01 -2.91862965e-01 -4.83013123e-01 9.40366089e-01
1.18610814e-01 1.05753697e-01 -7.76764631e-01 -3.61487150e-01
1.08687317e+00 2.01868460e-01 -2.11455703e-01 -6.98660970e-01
-1.46710908e+00 7.37874210e-01 1.35858327e-01 -1.72464430e-01
-6.60890043e-01 8.34042668e-01 2.51679689e-01 -4.08568472e-01
8.65163386e-01 1.01228155e-01 -4.70890164e-01 -4.99162227e-01
-8.26136053e-01 7.36110806e-01 5.28582215e-01 1.07900620e+00
5.89947104e-01 1.74381118e-02 -4.10296321e-01 5.62652171e-01
4.84397978e-01 6.03050530e-01 2.57269084e-01 -1.49578631e+00
1.75675198e-01 2.08460897e-01 3.27215225e-01 -1.24845755e+00
-1.57731702e-03 -1.48326665e-01 -4.79807317e-01 -4.07191887e-02
-1.16448134e-01 2.21904162e-02 -9.45119202e-01 1.43913662e+00
4.90134716e-01 7.74983585e-01 -2.77450979e-01 7.94308782e-01
7.63886869e-01 1.40464091e+00 3.36292535e-02 -2.69495010e-01
8.83965492e-01 -1.45921516e+00 -4.28149462e-01 1.06451035e-01
3.34919095e-01 -7.04450369e-01 7.94038057e-01 2.70455420e-01
-1.26065588e+00 -5.93805015e-01 -4.73368436e-01 -7.86020607e-02
5.99164329e-02 -3.83131839e-02 7.32614994e-01 1.06018700e-01
-1.29606438e+00 6.55154288e-01 -1.28962040e+00 -2.04212874e-01
6.28854811e-01 1.07085891e-01 -1.57381758e-01 -1.02461040e-01
-7.20861614e-01 3.45853508e-01 1.49498910e-01 -1.57731086e-01
-1.47165906e+00 -1.12626600e+00 -6.22008502e-01 1.09875776e-01
3.41718465e-01 -1.03585255e+00 1.07819462e+00 -7.85491109e-01
-1.58497882e+00 6.14733815e-01 -4.21948463e-01 -6.20486677e-01
6.61765635e-01 -5.19691288e-01 5.74207306e-02 5.51441431e-01
5.28214872e-02 1.07371998e+00 1.20758748e+00 -1.42418480e+00
-4.96955276e-01 2.98310697e-01 -6.49580790e-04 1.81566134e-01
3.42818946e-02 -1.59250051e-01 -9.38183188e-01 -8.58209252e-01
-1.48767740e-01 -1.10742664e+00 -2.55826116e-01 4.03625786e-01
-4.12279099e-01 -2.50315577e-01 9.04484451e-01 -7.52354503e-01
1.05517530e+00 -2.03274584e+00 5.61988473e-01 -2.73298502e-01
4.27282184e-01 1.39033318e-01 -2.00757280e-01 5.35480678e-01
1.22006319e-01 7.19875246e-02 2.41649196e-01 -6.74588680e-01
1.66777298e-01 2.37356082e-01 -6.00408137e-01 4.67266470e-01
4.57255216e-03 1.15898395e+00 -1.03121758e+00 -5.56723893e-01
5.21050155e-01 7.12494135e-01 -1.08049262e+00 2.03830719e-01
-8.99129093e-01 4.54391867e-01 -4.60729480e-01 6.34796500e-01
5.47618151e-01 -5.07569194e-01 7.69652426e-03 -3.51859093e-01
-8.12779516e-02 1.23441972e-01 -6.04638100e-01 1.67609739e+00
-2.24596724e-01 8.66347075e-01 -4.16506261e-01 -6.27100587e-01
4.18224454e-01 3.43971163e-01 8.98421884e-01 -2.37870038e-01
-7.84756765e-02 -2.41512388e-01 -5.13406157e-01 -4.23390687e-01
6.17312074e-01 2.47766450e-01 3.40856522e-01 4.80616361e-01
1.74528658e-01 7.42214322e-02 2.51239896e-01 5.86967051e-01
1.17068028e+00 6.62603855e-01 -2.50495933e-02 -1.74654618e-01
3.30279097e-02 -1.46756455e-01 7.30905771e-01 7.14928746e-01
4.81081568e-03 7.97953069e-01 5.15846252e-01 -7.83696473e-01
-1.35035145e+00 -1.24585819e+00 6.22233041e-02 8.64238679e-01
2.03801766e-01 -7.89535880e-01 -4.94548589e-01 -3.91367376e-01
1.41452849e-01 7.66610920e-01 -7.00893044e-01 -8.03431496e-02
-4.95245844e-01 -4.80008960e-01 -4.62923050e-02 4.83444303e-01
1.17323697e-01 -1.05752611e+00 -3.79114896e-01 4.36390996e-01
-3.71515244e-01 -1.15671945e+00 -4.14089084e-01 -4.48124200e-01
-8.72644722e-01 -9.21258330e-01 -6.25541508e-01 -4.06650662e-01
8.08792651e-01 2.06136853e-01 1.26151586e+00 2.90018648e-01
-2.92327374e-01 7.71698177e-01 -6.80173099e-01 -1.02355517e-01
-4.11927730e-01 -2.47231841e-01 2.85309047e-01 2.71244477e-02
-1.04264379e-01 -7.14389741e-01 -1.06637847e+00 1.30204827e-01
-7.55696774e-01 7.36530781e-01 3.44715267e-01 5.88392794e-01
1.02732193e+00 1.18135236e-01 -1.81214064e-01 -7.01967716e-01
2.25897208e-02 -6.09434068e-01 -6.02609158e-01 1.60730585e-01
-2.63497949e-01 -1.20089706e-02 5.13129115e-01 -5.25358021e-01
-1.10436940e+00 -3.25613916e-02 -1.25727296e-01 -8.77447069e-01
-5.15438989e-02 1.20615229e-01 1.32749686e-02 1.59525916e-01
2.42995620e-01 5.74708641e-01 -2.04277173e-01 -4.58586454e-01
4.38547671e-01 -9.24581438e-02 2.54814327e-01 -9.43613708e-01
7.27468491e-01 9.42683339e-01 -1.59554426e-02 -5.85429132e-01
-8.60472262e-01 -1.42989635e-01 -6.62215710e-01 -4.82422680e-01
8.85604143e-01 -9.97303069e-01 -6.58998370e-01 5.66467524e-01
-1.28470218e+00 -8.58099103e-01 -5.65926135e-01 3.72440845e-01
-1.05545723e+00 3.78393412e-01 -8.90423894e-01 -6.15834355e-01
-5.35191745e-02 -1.11519015e+00 1.19439125e+00 -8.33349973e-02
-1.18296288e-01 -1.28075087e+00 2.29154184e-01 3.34871799e-01
1.27587184e-01 2.48762712e-01 8.09144437e-01 2.28867102e-02
-1.32986367e+00 -4.85096835e-02 5.58266267e-02 1.61079198e-01
-4.55390103e-02 6.32392466e-01 -5.20319939e-01 -3.18045914e-01
-3.13139707e-01 -4.36829180e-02 9.55026507e-01 8.50355148e-01
1.28607011e+00 -7.44778931e-01 -6.92542911e-01 9.61785555e-01
1.27280807e+00 -1.66211411e-01 6.99170768e-01 2.78777421e-01
8.89184356e-01 7.78498203e-02 6.29694223e-01 9.33870971e-01
5.21981001e-01 9.75694120e-01 5.05712330e-01 3.10488164e-01
-4.26857084e-01 -7.40959406e-01 4.90681291e-01 9.01256084e-01
-5.51849008e-01 -6.58694208e-01 -7.35623658e-01 5.37015855e-01
-2.19410014e+00 -1.27216065e+00 -5.08682020e-02 1.64742494e+00
5.26798904e-01 -1.26625812e-02 1.31011948e-01 -4.41439092e-01
5.25880039e-01 4.92159039e-01 -5.15687168e-01 1.22273281e-01
-1.42685682e-01 -2.44302347e-01 4.87344891e-01 6.74334466e-01
-9.84438598e-01 1.21449840e+00 6.40088320e+00 1.00127697e+00
-9.11719680e-01 4.27939117e-01 6.90325260e-01 -4.88613397e-01
-6.43961489e-01 1.95993021e-01 -1.05632949e+00 7.77969539e-01
7.80591130e-01 -1.47404507e-01 3.46057445e-01 8.40818584e-01
5.34754813e-01 -5.60499802e-02 -9.44915235e-01 1.00449395e+00
1.03436589e-01 -2.20261192e+00 6.79852545e-01 1.61961615e-01
9.26015437e-01 2.07735538e-01 3.16611767e-01 -1.11752213e-03
3.42545271e-01 -4.72612202e-01 1.30056643e+00 1.01862836e+00
5.13703525e-01 -4.78579313e-01 1.54092655e-01 1.99232385e-01
-1.20543671e+00 -9.77815986e-02 -7.62771249e-01 8.27073157e-02
5.32690525e-01 5.91732621e-01 -4.94753331e-01 1.43436983e-01
7.99193501e-01 1.27550089e+00 -2.93531686e-01 9.15672898e-01
-3.16307038e-01 8.78633320e-01 -3.51416260e-01 2.45039091e-01
6.86627328e-02 -1.50588095e-01 9.28622663e-01 1.02321923e+00
5.18823683e-01 1.77492663e-01 8.47051367e-02 9.33366954e-01
-2.07054969e-02 -1.99062169e-01 -3.20157498e-01 -1.99931964e-01
4.35013473e-01 9.55066562e-01 -8.42908859e-01 -4.50513721e-01
-2.05696926e-01 1.02578413e+00 2.66702980e-01 4.35635179e-01
-1.18571150e+00 5.42942703e-01 8.95469427e-01 5.73027790e-01
7.86178052e-01 -5.20462394e-01 2.64927030e-01 -1.47307360e+00
-1.46021441e-01 -4.54719573e-01 5.69388084e-02 -1.04369736e+00
-1.07685161e+00 4.87968236e-01 1.44654781e-01 -1.26608539e+00
-1.96332246e-01 -6.41202033e-01 -4.85351145e-01 1.21554114e-01
-1.23342025e+00 -1.49278116e+00 -1.29920170e-01 5.05988717e-01
8.61224592e-01 -2.11789146e-01 6.86657965e-01 8.73415023e-02
-3.72406662e-01 1.91074759e-01 3.86582106e-01 -2.09943071e-01
2.23256186e-01 -8.55707765e-01 6.81239247e-01 8.24703097e-01
3.64844918e-01 2.52516747e-01 9.49924588e-01 -9.28424954e-01
-1.59713185e+00 -1.22385883e+00 4.55884427e-01 -8.68339598e-01
6.98215783e-01 -4.53835636e-01 -4.32564288e-01 9.73900199e-01
-5.98430745e-02 2.17138737e-01 6.01007640e-01 -1.28944628e-02
-1.96991906e-01 6.09745532e-02 -5.69456697e-01 6.82231009e-01
1.35962021e+00 -3.04727405e-01 -1.49171188e-01 9.61564302e-01
1.01094961e+00 -4.37720478e-01 -6.85963690e-01 1.44051671e-01
5.89162290e-01 -1.11205685e+00 1.31145537e+00 -4.24346238e-01
7.74642110e-01 -1.95844471e-01 -2.10698932e-01 -1.06005037e+00
-9.20810997e-01 -7.86743879e-01 -8.66825044e-01 7.98411667e-01
1.36529863e-01 -3.14085424e-01 1.05851126e+00 5.27972639e-01
-3.38916361e-01 -1.13629651e+00 -7.38008738e-01 -6.95785761e-01
-2.82787383e-01 -5.95937908e-01 4.41904128e-01 4.93710130e-01
-3.67643982e-01 -2.00017735e-01 -8.30279052e-01 2.77846098e-01
6.83595061e-01 2.55151391e-01 8.18048179e-01 -8.51377845e-01
-6.61080539e-01 -9.54280272e-02 -6.13612771e-01 -1.51737511e+00
2.63150603e-01 -5.33099771e-01 -1.22554883e-01 -1.73892295e+00
4.06894863e-01 -5.66510320e-01 -2.12508105e-02 3.98310900e-01
-4.54988517e-02 2.80297428e-01 5.11078119e-01 5.99901676e-01
-1.16581285e+00 9.99863386e-01 1.39624250e+00 -6.12755343e-02
-1.04057550e-01 -9.02924463e-02 -8.88810679e-02 7.48894513e-01
5.31307876e-01 -4.49104995e-01 -4.40851867e-01 -7.10509837e-01
4.61748093e-01 -2.51119193e-02 6.55345559e-01 -1.04355359e+00
1.08535498e-01 -3.88789326e-01 5.29448628e-01 -7.26190269e-01
9.59315240e-01 -4.62349296e-01 5.87075651e-01 3.80670726e-01
-9.66367424e-02 -5.41443452e-02 5.50190583e-02 8.12485278e-01
6.40265346e-02 1.25896096e-01 5.54549456e-01 -3.09071392e-01
-8.43704939e-01 9.48444605e-01 -5.46169877e-01 -1.70915231e-01
1.23686922e+00 -3.10611039e-01 -3.69515419e-01 -6.41154349e-01
-8.06132734e-01 -4.92584407e-02 7.88380325e-01 4.68154550e-01
8.78772676e-01 -1.42402709e+00 -6.53710723e-01 -7.01146424e-02
-9.24897566e-02 6.05232678e-02 6.84765577e-01 6.56513929e-01
-1.02363360e+00 4.56864297e-01 -1.19796649e-01 -6.96223080e-01
-9.96967375e-01 6.91719592e-01 -1.63581432e-03 -1.20233908e-01
-1.15687406e+00 1.22489309e+00 6.12240374e-01 -1.19970642e-01
-1.00447662e-01 -1.83156610e-01 3.06501299e-01 -3.24269891e-01
4.34325814e-01 2.72932827e-01 -5.97315371e-01 -8.63439500e-01
-5.89289516e-02 3.05684596e-01 -7.43007436e-02 1.40930280e-01
1.67763865e+00 -2.00917274e-01 -3.56802605e-02 4.38114136e-01
9.97612417e-01 -2.19124854e-02 -2.13573766e+00 -1.98877659e-02
-6.81960046e-01 -8.72379065e-01 -2.05263738e-02 -3.02281380e-01
-1.01442146e+00 7.01605916e-01 2.22388148e-01 -1.48544744e-01
6.89976096e-01 3.37524891e-01 1.03327525e+00 1.13177761e-01
5.18082201e-01 -9.29828763e-01 3.50109279e-01 4.48071092e-01
9.43049908e-01 -8.11131597e-01 3.20338517e-01 -5.19800425e-01
-5.43181241e-01 8.76797259e-01 4.21215028e-01 -2.72420496e-01
9.22709286e-01 1.75902277e-01 -3.57664555e-01 -1.21786416e-01
-1.42481363e+00 2.82128364e-01 1.84581712e-01 6.11855149e-01
-2.55207741e-03 6.70718104e-02 2.39535660e-01 1.72762394e-01
-7.44868889e-02 1.95372105e-02 4.12178427e-01 9.00061965e-01
-3.69977981e-01 -1.04941750e+00 -4.44238484e-02 5.17405093e-01
-1.49349168e-01 -1.50377378e-01 2.35296845e-01 4.70047593e-01
2.49911726e-01 3.16061139e-01 3.98359329e-01 -3.29777539e-01
-2.63675809e-01 -3.57131571e-01 7.11237133e-01 -6.25770390e-01
2.21324563e-01 1.36587739e-01 -1.50232509e-01 -9.06180561e-01
-5.70334673e-01 -8.92252207e-01 -1.08619976e+00 -6.47657812e-01
-4.61733751e-02 3.51210050e-02 3.02456111e-01 8.68221521e-01
7.60877609e-01 2.97845095e-01 2.98216909e-01 -1.49037766e+00
7.05966875e-02 -6.54207051e-01 -5.02875686e-01 4.74211365e-01
1.90461189e-01 -9.59633470e-01 -2.76601136e-01 6.14567339e-01] | [10.57587718963623, -0.4247100353240967] |
4a27b212-affd-4907-9563-7bce5afc63c0 | deep-fully-connected-networks-for-video | 1603.04930 | null | http://arxiv.org/abs/1603.04930v2 | http://arxiv.org/pdf/1603.04930v2.pdf | Deep Fully-Connected Networks for Video Compressive Sensing | In this work we present a deep learning framework for video compressive
sensing. The proposed formulation enables recovery of video frames in a few
seconds at significantly improved reconstruction quality compared to previous
approaches. Our investigation starts by learning a linear mapping between video
sequences and corresponding measured frames which turns out to provide
promising results. We then extend the linear formulation to deep
fully-connected networks and explore the performance gains using deeper
architectures. Our analysis is always driven by the applicability of the
proposed framework on existing compressive video architectures. Extensive
simulations on several video sequences document the superiority of our approach
both quantitatively and qualitatively. Finally, our analysis offers insights
into understanding how dataset sizes and number of layers affect reconstruction
performance while raising a few points for future investigation.
Code is available at Github: https://github.com/miliadis/DeepVideoCS | ['Michael Iliadis', 'Aggelos K. Katsaggelos', 'Leonidas Spinoulas'] | 2016-03-16 | null | null | null | null | ['video-compressive-sensing'] | ['computer-vision'] | [ 1.71407446e-01 3.36835235e-02 -2.11891383e-01 -1.16399683e-01
-8.12219024e-01 -3.00381929e-01 4.79495525e-01 -3.84619713e-01
-1.45126790e-01 6.22149706e-01 4.62323248e-01 -3.65996301e-01
-1.79312468e-01 -3.69408637e-01 -1.02574730e+00 -5.25472522e-01
-5.81954539e-01 -2.68121630e-01 -5.78647070e-02 9.54175666e-02
2.39730179e-01 4.14490104e-01 -1.19301498e+00 2.93007851e-01
3.14108908e-01 1.10956550e+00 4.23580289e-01 8.08100224e-01
5.80524802e-01 1.27149749e+00 -1.16763681e-01 -1.24066457e-01
6.98765695e-01 -3.60671729e-01 -6.08641565e-01 3.76589745e-01
5.58851540e-01 -1.01718545e+00 -9.99177814e-01 9.71374989e-01
3.57068717e-01 -4.78590950e-02 1.06853291e-01 -9.91062164e-01
-5.86546242e-01 7.96225309e-01 -3.49487722e-01 5.29102445e-01
3.45768660e-01 1.72375351e-01 9.81145024e-01 -1.08968937e+00
4.23358083e-01 8.92322361e-01 7.16196954e-01 3.29782546e-01
-7.81504631e-01 -6.50727749e-01 3.99430022e-02 5.01700521e-01
-1.50284588e+00 -1.13329625e+00 8.04347634e-01 -2.32900426e-01
8.80805373e-01 -3.30155157e-02 5.95499337e-01 1.04414666e+00
1.18363552e-01 9.46907163e-01 7.83035934e-01 -3.57235283e-01
5.96807972e-02 -3.34837049e-01 -2.56583452e-01 7.75675416e-01
3.99627596e-01 1.78507939e-01 -5.79133511e-01 1.67436108e-01
1.08982122e+00 1.01854719e-01 -7.40352869e-01 -3.85724753e-01
-1.23475802e+00 7.60502398e-01 3.16234201e-01 4.66391623e-01
-5.18978417e-01 6.69931650e-01 3.01914543e-01 5.27911305e-01
1.94263071e-01 7.75805116e-02 -1.03492253e-01 -1.49020001e-01
-1.40259206e+00 4.30768467e-02 6.57756567e-01 9.82923329e-01
4.59522307e-01 5.99255264e-01 1.10060155e-01 5.67456245e-01
2.29847312e-01 3.11772197e-01 2.01484472e-01 -1.52133560e+00
4.26469445e-01 -3.24957997e-01 7.62882382e-02 -1.23146677e+00
-6.97029009e-02 -5.14328122e-01 -8.68279994e-01 4.88599250e-03
1.64128706e-01 -4.86003131e-01 -3.78219038e-01 1.63582313e+00
-3.20858240e-01 7.37670243e-01 1.64409846e-01 1.15196359e+00
6.00311756e-01 6.54468715e-01 -5.41546047e-01 -2.16764644e-01
7.81733394e-01 -9.85857546e-01 -5.90883970e-01 -1.12509601e-01
3.87840539e-01 -6.82867587e-01 5.48002958e-01 5.15016794e-01
-1.37808740e+00 -5.63746452e-01 -1.23386633e+00 2.93879192e-02
4.16515678e-01 2.74298877e-01 5.50611734e-01 4.04513508e-01
-1.65559518e+00 8.85575712e-01 -1.12348175e+00 -3.21048081e-01
7.25842834e-01 3.84955674e-01 -1.47219613e-01 -3.74487668e-01
-9.68955517e-01 3.05743188e-01 3.47766727e-01 1.96283847e-01
-1.37598467e+00 -4.61169988e-01 -8.58691633e-01 2.88907081e-01
5.00660539e-01 -6.56224728e-01 1.23297334e+00 -9.42521691e-01
-1.39972675e+00 3.97762239e-01 -1.46041378e-01 -9.20112193e-01
5.65779388e-01 -4.72873926e-01 -2.40549192e-01 7.42425740e-01
-1.88521296e-01 6.59977317e-01 9.27160680e-01 -1.16466224e+00
-3.90731096e-01 8.03602785e-02 4.26048428e-01 1.65941399e-02
-4.13917899e-01 -2.40752637e-01 -6.12178028e-01 -7.84447908e-01
1.78781018e-01 -9.31057692e-01 -1.65068805e-01 8.75657797e-02
-2.44894534e-01 3.02783728e-01 6.83146656e-01 -7.39613652e-01
1.00389183e+00 -2.12584734e+00 2.48274356e-01 -7.21342415e-02
4.96215641e-01 1.89437345e-01 -3.17598969e-01 7.20768154e-01
-1.70045167e-01 -1.57638818e-01 -2.45002165e-01 -3.80978286e-01
-1.01711236e-01 -1.45953134e-01 -3.86662692e-01 8.03897679e-01
8.48283246e-03 1.06975055e+00 -6.75182700e-01 -8.76632109e-02
4.51774508e-01 8.60932589e-01 -7.97149956e-01 1.48915425e-01
6.11288324e-02 5.68967462e-01 -3.47790837e-01 7.27223396e-01
6.75192595e-01 -5.94643235e-01 4.81859744e-01 -3.68782043e-01
1.24026798e-01 3.17156106e-01 -8.88731837e-01 1.90495920e+00
-3.66012543e-01 1.21911120e+00 2.48871624e-01 -1.36205614e+00
4.58878011e-01 4.37065065e-01 7.51541972e-01 -5.18621683e-01
4.48809505e-01 5.42245917e-02 -6.61928281e-02 -5.65949202e-01
4.84571368e-01 -1.86880171e-01 4.21251863e-01 4.34812546e-01
1.09712623e-01 2.77167827e-01 4.87168208e-02 3.72135133e-01
1.09833384e+00 1.56193078e-01 3.16429347e-01 -3.63594621e-01
3.57927233e-01 -3.61690611e-01 2.71320224e-01 8.75427544e-01
-1.34876654e-01 5.73830009e-01 3.06746393e-01 -2.58435249e-01
-1.19048762e+00 -9.50899839e-01 -8.24262649e-02 6.16578281e-01
1.74568206e-01 -5.11430621e-01 -6.16895378e-01 -1.80996001e-01
-3.16043288e-01 2.91184902e-01 -3.41763049e-01 1.24306925e-01
-5.83654940e-01 -2.84592927e-01 6.19780779e-01 5.62822580e-01
6.22530043e-01 -8.09082925e-01 -6.89171851e-01 -3.41363214e-02
-3.31769437e-01 -1.62200475e+00 -1.54362589e-01 -1.66899636e-01
-1.13829660e+00 -1.03627670e+00 -8.66340518e-01 -8.42546642e-01
3.99549723e-01 7.44016469e-01 9.73523915e-01 4.76562917e-01
1.44873530e-01 6.99023247e-01 -5.64705849e-01 1.63880974e-01
-2.78724968e-01 -2.32634488e-02 1.78026915e-01 -5.82042001e-02
2.27661908e-01 -9.45098519e-01 -9.80116785e-01 -1.43517718e-01
-1.15392041e+00 5.38841970e-02 6.30014598e-01 5.55581808e-01
2.67623872e-01 6.35595471e-02 4.24038768e-01 -6.26660883e-01
4.31909055e-01 -7.82278001e-01 -6.29778087e-01 -2.40685791e-01
-4.15460348e-01 -2.10118264e-01 6.94666266e-01 3.26333344e-02
-5.32527745e-01 -1.81346178e-01 -1.92480415e-01 -9.50486243e-01
1.21659460e-02 6.03258669e-01 2.32292533e-01 -1.61586314e-01
3.01764488e-01 4.28025961e-01 8.11282173e-02 -1.99958220e-01
2.82269597e-01 5.85366309e-01 5.49099922e-01 -4.72556651e-01
7.77641892e-01 8.60149145e-01 -1.89657360e-01 -9.40258980e-01
-5.78602493e-01 -2.63636887e-01 -5.09572208e-01 -2.51961648e-01
5.02321839e-01 -1.51944506e+00 -6.35907173e-01 3.13141942e-01
-9.15288329e-01 -5.98209381e-01 -8.03316385e-02 8.00365269e-01
-7.61919260e-01 7.93257296e-01 -1.11594141e+00 -5.13761580e-01
-1.27444863e-01 -1.23371148e+00 9.11600888e-01 -1.29036203e-01
3.79198641e-02 -1.29711497e+00 -1.29081473e-01 4.93985564e-01
4.32304561e-01 1.28098980e-01 3.35584015e-01 -1.88510582e-01
-1.11682260e+00 6.32110909e-02 -1.95879400e-01 5.42799354e-01
2.82405969e-02 -4.13005680e-01 -1.00424278e+00 -9.18894053e-01
2.65566975e-01 -4.13118094e-01 1.22222793e+00 6.43419445e-01
1.28917742e+00 -3.77622932e-01 -2.31215656e-02 9.95119154e-01
1.77335989e+00 -8.73905122e-02 8.31237793e-01 2.34823063e-01
6.43407643e-01 1.25712484e-01 2.21961260e-01 7.50480533e-01
2.15945765e-01 4.32600021e-01 6.22154176e-01 1.50937468e-01
-3.31120700e-01 -9.08422619e-02 5.21969318e-01 1.13165021e+00
-2.61931032e-01 -5.34805179e-01 -6.11215115e-01 6.79508030e-01
-1.74291885e+00 -1.13280690e+00 1.91096291e-01 1.98628283e+00
2.78033286e-01 -1.67724080e-02 -4.10382636e-02 1.90027863e-01
5.36166966e-01 6.03358865e-01 -4.49750453e-01 3.21586244e-02
-2.22827569e-01 1.89980000e-01 5.55751801e-01 7.12482631e-01
-1.00017154e+00 7.59018481e-01 6.27898502e+00 5.35360456e-01
-1.26160324e+00 1.74054846e-01 6.99013412e-01 -5.28076708e-01
-2.37387866e-01 1.98410898e-02 -2.38849476e-01 2.76040196e-01
1.05768287e+00 -1.09103918e-01 6.96515977e-01 4.88892496e-01
4.59896028e-01 6.72193915e-02 -1.25749075e+00 1.25774908e+00
2.79846847e-01 -1.85519135e+00 6.08700477e-02 2.86421150e-01
9.27359939e-01 4.83310938e-01 1.98640749e-01 -6.02068715e-02
9.04535353e-02 -1.01831412e+00 7.87224233e-01 2.55139470e-01
9.21409130e-01 -4.68776822e-01 6.50001228e-01 8.27087685e-02
-1.04509974e+00 -2.41284788e-01 -5.80000222e-01 -5.29006600e-01
3.19956928e-01 5.10515273e-01 -4.87747729e-01 5.07656455e-01
7.25036144e-01 1.21703792e+00 -3.05036724e-01 9.38379943e-01
-5.90678267e-02 7.89841771e-01 -4.24931757e-02 4.82971966e-01
3.31030250e-01 -7.08671212e-02 5.62008679e-01 1.24079072e+00
6.78827286e-01 2.13013709e-01 3.85291725e-02 6.23753488e-01
-3.71665806e-01 -2.84687191e-01 -8.50659072e-01 -9.80157703e-02
4.72663105e-01 9.36226726e-01 -6.37215137e-01 -3.28606367e-01
-7.45903850e-01 1.00204229e+00 2.64937133e-01 6.70888603e-01
-8.29476714e-01 8.37603733e-02 5.21154523e-01 3.11497241e-01
8.07640672e-01 -5.66302955e-01 -2.00767927e-02 -1.49594295e+00
2.20916405e-01 -1.01773691e+00 4.47571129e-02 -6.71066403e-01
-8.64422321e-01 4.65618938e-01 -9.87686515e-02 -1.48359752e+00
-3.39820862e-01 -4.65953588e-01 -3.34740937e-01 2.77551085e-01
-1.71398330e+00 -6.62360728e-01 -4.08060730e-01 5.84255993e-01
7.37752914e-01 -3.58681142e-01 3.85167450e-01 5.56744158e-01
-3.95976961e-01 5.74719012e-01 3.34422350e-01 2.54662156e-01
2.76370406e-01 -5.78530848e-01 4.16066051e-01 1.37435031e+00
3.52726877e-01 5.57747006e-01 6.66476548e-01 -3.64813209e-01
-1.60399401e+00 -9.67283607e-01 3.56741041e-01 -4.43871617e-02
7.58307457e-01 -1.97240531e-01 -8.43340456e-01 1.04727304e+00
7.22544551e-01 1.61972612e-01 6.53307557e-01 -3.98093641e-01
-2.42454171e-01 1.58305541e-01 -8.05239618e-01 4.57131982e-01
1.07846010e+00 -7.04332769e-01 -9.97482166e-02 3.54529709e-01
5.71410239e-01 -4.20122236e-01 -7.83647358e-01 2.90045649e-01
6.38008356e-01 -1.52378786e+00 9.00926948e-01 -1.97506204e-01
1.00572562e+00 -1.05171561e-01 -6.76946521e-01 -9.42140222e-01
-4.10470009e-01 -8.38711500e-01 -4.75308925e-01 4.10602361e-01
6.96509480e-02 -4.68094110e-01 9.00458634e-01 1.65298432e-01
-1.36840954e-01 -8.81429791e-01 -7.45853245e-01 -7.51321912e-01
1.28146559e-01 -7.37135291e-01 3.98583770e-01 1.01080573e+00
-6.29213527e-02 -4.38429192e-02 -8.66981208e-01 4.89928752e-01
7.63876200e-01 1.88218385e-01 5.50924778e-01 -5.20337462e-01
-8.31476152e-01 -2.16601104e-01 -5.45827270e-01 -1.80367374e+00
1.26416907e-01 -9.70380425e-01 -2.66236603e-01 -1.37225246e+00
2.41891965e-01 -1.02084868e-01 -5.35858512e-01 1.46974862e-01
3.33381921e-01 6.98521376e-01 6.08653665e-01 3.61776680e-01
-7.91062593e-01 5.60431480e-01 1.05543220e+00 -3.67447622e-02
2.46404797e-01 -2.08975956e-01 -8.77870440e-01 6.09858871e-01
9.95971024e-01 -1.68400720e-01 -5.21391094e-01 -1.14547527e+00
1.94895584e-02 6.10200524e-01 5.76387584e-01 -1.27745783e+00
2.24686861e-01 1.42699182e-01 3.74066681e-01 -1.70428157e-01
4.97087955e-01 -8.48053634e-01 3.81668583e-02 6.44028902e-01
-4.09222841e-01 1.03194872e-02 1.56618327e-01 6.03235781e-01
-2.32113093e-01 -8.58923569e-02 7.08131850e-01 -1.43777698e-01
-7.41328478e-01 4.11671579e-01 -3.30949813e-01 -9.43120718e-02
7.50536740e-01 -2.55980402e-01 -3.31251286e-02 -1.02378726e+00
-7.35690534e-01 2.79740095e-02 5.22565424e-01 3.26573402e-01
9.42469597e-01 -1.27320993e+00 -8.47565532e-01 2.27060214e-01
-2.68350810e-01 -3.42427641e-01 2.97978550e-01 9.59806561e-01
-8.73321354e-01 6.97430849e-01 -1.43816799e-01 -7.11118639e-01
-1.04006565e+00 4.14542824e-01 3.32060814e-01 1.69809088e-01
-7.83914804e-01 7.86764264e-01 1.83880493e-01 1.51918381e-01
3.34750980e-01 -3.50559890e-01 6.39705285e-02 -4.37821835e-01
5.74185193e-01 4.12231714e-01 -2.12509200e-01 -7.63088703e-01
-1.67710438e-01 3.98298532e-01 -4.21252511e-02 3.10745016e-02
1.53381979e+00 -4.57147896e-01 1.42618299e-01 2.76126087e-01
1.42478216e+00 -1.48581387e-02 -1.70555556e+00 -2.95134157e-01
-3.11861753e-01 -6.68959498e-01 2.59206504e-01 -1.24470152e-01
-1.65234232e+00 7.70757437e-01 6.12037182e-01 3.17587554e-02
1.29102361e+00 4.59990539e-02 9.79050517e-01 4.74285185e-01
4.29931879e-01 -5.47071099e-01 2.25096151e-01 2.50863552e-01
8.03237557e-01 -1.38670945e+00 2.48342484e-01 -1.80136025e-01
-2.89503217e-01 1.08507645e+00 1.08567484e-01 -5.66591501e-01
7.61172295e-01 3.20525795e-01 3.20605896e-02 -2.64075369e-01
-7.85289049e-01 9.72245038e-02 -9.52051058e-02 3.82565558e-01
5.71912706e-01 -1.18000597e-01 -9.62732434e-02 -6.52310029e-02
-2.29080364e-01 3.07914674e-01 9.38736439e-01 8.57869506e-01
-4.10274178e-01 -8.45992923e-01 -1.78943202e-01 3.76735538e-01
-6.93826377e-01 -3.13503236e-01 1.07441172e-01 6.87409878e-01
-3.06193024e-01 1.04676092e+00 1.68759167e-01 -5.01733482e-01
-2.27352217e-01 -4.37586218e-01 6.97595716e-01 -3.49885970e-01
3.88502963e-02 3.03504646e-01 -3.27416584e-02 -8.22236001e-01
-8.61677706e-01 -7.54723370e-01 -1.09303248e+00 -6.60757363e-01
9.25530940e-02 -1.83120929e-02 2.78909594e-01 8.56878340e-01
4.17523980e-01 4.49095190e-01 6.81254029e-01 -1.02015448e+00
-4.65595543e-01 -6.75448060e-01 -4.45298225e-01 1.19168997e-01
7.81240582e-01 -2.63097674e-01 -3.58076602e-01 2.53352135e-01] | [11.164856910705566, -2.0739054679870605] |
237b5544-c24b-4296-ad9a-3342705bed6d | dynamic-loss-balancing-and-sequential | 2211.04165 | null | https://arxiv.org/abs/2211.04165v1 | https://arxiv.org/pdf/2211.04165v1.pdf | Dynamic loss balancing and sequential enhancement for road-safety assessment and traffic scene classification | Road-safety inspection is an indispensable instrument for reducing road-accident fatalities contributed to road infrastructure. Recent work formalizes road-safety assessment in terms of carefully selected risk factors that are also known as road-safety attributes. In current practice, these attributes are manually annotated in geo-referenced monocular video for each road segment. We propose to reduce dependency on tedious human labor by automating recognition with a two-stage neural architecture. The first stage predicts more than forty road-safety attributes by observing a local spatio-temporal context. Our design leverages an efficient convolutional pipeline, which benefits from pre-training on semantic segmentation of street scenes. The second stage enhances predictions through sequential integration across a larger temporal window. Our design leverages per-attribute instances of a lightweight bidirectional LSTM architecture. Both stages alleviate extreme class imbalance by incorporating a multi-task variant of recall-based dynamic loss weighting. We perform experiments on the iRAP-BH dataset, which involves fully labeled geo-referenced video along 2,300 km of public roads in Bosnia and Herzegovina. We also validate our approach by comparing it with the related work on two road-scene classification datasets from the literature: Honda Scenes and FM3m. Experimental evaluation confirms the value of our contributions on all three datasets. | ['Siniša Šegvić', 'Marko Ševrović', 'Marin Kačan'] | 2022-11-08 | null | null | null | null | ['scene-classification'] | ['computer-vision'] | [ 4.35211778e-01 -1.25765115e-01 -4.49937820e-01 -7.72603095e-01
-1.28460741e+00 -3.30203176e-01 6.03085399e-01 -4.59502749e-02
-7.72777498e-01 6.09337449e-01 3.50962579e-01 -5.62493861e-01
-3.15661937e-01 -1.06104374e+00 -1.09077001e+00 -4.88220781e-01
-5.17340265e-02 1.18407495e-01 4.09872562e-01 -1.49371475e-01
1.33911923e-01 6.55245483e-01 -1.78573179e+00 3.95342261e-01
6.63722873e-01 1.21159256e+00 -1.58141240e-01 7.61983454e-01
2.38285556e-01 9.47965562e-01 -1.30249811e-02 -7.58902371e-01
5.78041077e-01 2.09154099e-01 -9.48541939e-01 7.61743337e-02
1.09326673e+00 -4.28590506e-01 -5.48329771e-01 5.75464964e-01
3.02016526e-01 2.71185756e-01 6.48773670e-01 -1.34482658e+00
-1.43350378e-01 3.60781103e-02 -5.21905601e-01 2.48451665e-01
-8.84926766e-02 4.33752209e-01 1.16882050e+00 -7.29147196e-01
3.18348527e-01 1.08423066e+00 1.09402776e+00 2.62345709e-02
-1.05288565e+00 -7.15506196e-01 1.52047679e-01 6.73931062e-01
-1.43477345e+00 -5.64724982e-01 4.90075856e-01 -6.62511230e-01
1.07419157e+00 1.87040880e-01 4.16598916e-01 9.21292484e-01
2.63222624e-02 6.60705984e-01 8.09632301e-01 -2.15049200e-02
-1.51534140e-01 -4.85464744e-02 4.87655886e-02 6.88094795e-01
9.54277590e-02 2.90125340e-01 -4.07037258e-01 1.81652844e-01
2.36427575e-01 -9.26046968e-02 2.56509870e-01 -2.68770576e-01
-1.00557029e+00 8.03306222e-01 4.16488469e-01 -4.11853701e-01
-4.13389891e-01 4.09287125e-01 6.89170122e-01 1.91498965e-01
8.59187603e-01 -1.05004750e-01 -5.69061518e-01 -3.09582055e-02
-9.36438143e-01 9.99206007e-02 2.98210680e-01 8.97387743e-01
1.12397146e+00 -6.70191869e-02 -1.71389088e-01 9.93727744e-01
7.14941546e-02 6.27103329e-01 -3.01998079e-01 -1.35523856e+00
7.97496438e-01 4.40333217e-01 -7.06693456e-02 -8.44407201e-01
-4.43204015e-01 -2.85425872e-01 -5.87524295e-01 3.61545295e-01
3.32944840e-01 -1.02717265e-01 -1.06851375e+00 1.67723167e+00
4.44218844e-01 2.78845191e-01 -3.59884977e-01 6.08809233e-01
5.02962828e-01 3.80669057e-01 6.92908645e-01 3.92723978e-01
1.19260561e+00 -9.60839331e-01 -2.80557364e-01 -2.01031715e-01
7.85546958e-01 -4.61165130e-01 1.11107111e+00 1.48448780e-01
-7.75488257e-01 -6.19485915e-01 -7.57803559e-01 -2.36907065e-01
-6.60004556e-01 3.02177817e-01 3.77224088e-01 6.60204649e-01
-1.00350368e+00 5.10318041e-01 -4.01126057e-01 -3.79703671e-01
9.76729572e-01 2.68743813e-01 -6.04835153e-01 -2.29734145e-02
-1.44288754e+00 1.01256657e+00 1.75307497e-01 1.04234777e-01
-8.80974352e-01 -1.01882339e+00 -1.09482586e+00 -1.16723605e-01
3.83625746e-01 -6.18203282e-01 1.04914248e+00 -7.73504078e-01
-1.03149784e+00 1.12351847e+00 -3.44869316e-01 -5.99940836e-01
6.58430457e-01 -6.67662144e-01 -4.20730233e-01 2.23014861e-01
3.56457502e-01 8.07659864e-01 7.72020340e-01 -1.08566022e+00
-1.20546818e+00 -1.52626127e-01 1.08545423e-01 1.46404162e-01
-2.74180532e-01 1.09934382e-01 -4.68388379e-01 -4.75172371e-01
-3.58723313e-01 -9.81574714e-01 -2.77414888e-01 -1.40071716e-02
-3.97476077e-01 -7.04767928e-02 7.84408867e-01 -9.57650661e-01
1.13573730e+00 -1.99446785e+00 -6.62938774e-01 2.83902258e-01
1.74664091e-02 3.99125129e-01 -1.64657682e-01 1.89577773e-01
-1.99461982e-01 1.05746374e-01 -3.57019633e-01 -4.22110051e-01
-1.87940568e-01 1.78761825e-01 -5.51552415e-01 6.52302921e-01
5.20633698e-01 8.75026703e-01 -7.65998781e-01 -6.53570116e-01
6.01319969e-01 5.22566676e-01 -3.96420628e-01 6.73713982e-02
1.52441949e-01 3.26171964e-01 -2.82139450e-01 6.21642351e-01
6.96078181e-01 1.70587033e-01 -5.28826773e-01 -2.07462296e-01
-3.85817319e-01 6.36923075e-01 -8.30730736e-01 1.44012272e+00
-6.40827000e-01 8.62361908e-01 -3.13215524e-01 -1.12158036e+00
6.59177840e-01 2.84914374e-01 5.00160217e-01 -1.02914929e+00
-1.52769536e-01 7.74220824e-02 -5.30039608e-01 -8.53106260e-01
5.81034303e-01 1.34427667e-01 -1.19519748e-01 1.90462768e-01
-2.05926105e-01 3.55152078e-02 1.69461593e-01 6.48381636e-02
1.15490580e+00 4.51732397e-01 -4.77451943e-02 -7.01817349e-02
6.61164999e-01 3.07938904e-01 5.97363710e-01 8.29793394e-01
-3.78096074e-01 8.01907122e-01 4.32734728e-01 -8.04299474e-01
-1.39085221e+00 -1.14205194e+00 -2.37646997e-01 1.41834354e+00
-1.91853851e-01 5.34978248e-02 -6.82788253e-01 -8.99605393e-01
-1.80235617e-02 7.74207711e-01 -7.56137252e-01 -1.22434780e-01
-9.11942065e-01 -8.20318043e-01 9.41573918e-01 7.82225966e-01
6.29127502e-01 -7.81516671e-01 -6.09194458e-01 3.42194289e-02
-3.18702489e-01 -1.35646474e+00 -1.32254884e-01 -1.13630518e-01
-5.26791275e-01 -1.13803375e+00 -4.78517264e-01 -3.41527283e-01
2.57146925e-01 5.93657792e-01 1.22330070e+00 -1.93317667e-01
-2.92191535e-01 4.68823344e-01 -1.16960563e-01 -4.28270310e-01
-3.91842574e-02 4.97834474e-01 -2.37964928e-01 3.04013610e-01
5.88786840e-01 -5.07986426e-01 -5.57667434e-01 4.76898700e-01
-4.16270733e-01 1.21017851e-01 6.70886636e-01 5.98520577e-01
5.51037014e-01 -1.90962955e-01 6.27921999e-01 -8.82353723e-01
-1.23392925e-01 -8.28566909e-01 -4.90628779e-01 3.70667167e-02
-4.13297027e-01 -1.77443549e-01 3.00906062e-01 4.32707891e-02
-1.41298652e+00 3.47862512e-01 -4.70058590e-01 -2.78522670e-01
-5.98770320e-01 1.36978760e-01 -2.54250109e-01 9.38712656e-02
4.89141732e-01 -1.03400409e-01 -3.67670804e-01 -1.82094291e-01
3.89214575e-01 6.56729460e-01 7.97528505e-01 -5.77578664e-01
1.02854049e+00 7.82574832e-01 2.79181302e-01 -9.78959680e-01
-1.25779903e+00 -7.87505746e-01 -7.40317225e-01 -5.15891433e-01
1.09154415e+00 -1.28346622e+00 -7.99607933e-01 5.64838767e-01
-9.94450450e-01 -5.03549218e-01 -2.56530762e-01 5.08953273e-01
-6.37583315e-01 8.55983123e-02 -3.71749818e-01 -7.83773899e-01
-1.71201691e-01 -8.48433495e-01 1.26051295e+00 -1.49437785e-01
-1.30161270e-01 -7.17481315e-01 4.96776104e-02 6.66951537e-01
4.39394802e-01 3.31807494e-01 7.64790297e-01 -4.20378387e-01
-7.11653948e-01 -1.68668881e-01 -6.37362301e-01 5.05571485e-01
-2.64892399e-01 1.13048077e-01 -1.52937257e+00 2.43591592e-01
-4.31178808e-01 -4.10673261e-01 1.52960503e+00 4.39176232e-01
1.40947342e+00 -1.21734560e-01 -1.65540799e-01 7.32548535e-01
1.21040118e+00 -1.00219160e-01 9.24375892e-01 7.01520503e-01
1.13098359e+00 1.20490849e+00 9.50179636e-01 9.50306803e-02
9.41617668e-01 5.90904176e-01 6.15244091e-01 -4.91625071e-01
-2.04436898e-01 -1.88844711e-01 1.66531429e-01 1.51819929e-01
-2.80084997e-01 -6.39647692e-02 -1.20539343e+00 1.15355599e+00
-1.81482947e+00 -1.22443259e+00 -6.06090724e-01 2.18402004e+00
5.25232613e-01 2.40231723e-01 2.36341268e-01 3.58650476e-01
6.44052148e-01 3.44386697e-01 -3.73677939e-01 -3.23503524e-01
-1.23869896e-01 -2.02113315e-02 1.21097159e+00 5.92343688e-01
-1.77888632e+00 9.68229949e-01 5.82673836e+00 1.02422273e+00
-9.03794765e-01 2.37784564e-01 1.04678416e+00 -2.18674138e-01
-1.15918420e-01 -4.98339720e-02 -8.66787136e-01 2.36523122e-01
1.25444400e+00 3.60381514e-01 1.87163949e-02 8.06411922e-01
4.43168849e-01 -2.49041826e-01 -8.84005189e-01 5.44777632e-01
-1.91431999e-01 -1.28635859e+00 -8.23872536e-02 -1.94057394e-02
5.91987014e-01 6.32034600e-01 2.07853869e-01 2.63759673e-01
3.67956132e-01 -1.14248061e+00 8.86704564e-01 5.11235833e-01
1.05655193e+00 -9.52180028e-01 6.15458727e-01 -5.75588234e-02
-1.35136545e+00 -1.54231414e-01 -6.90087080e-02 6.78054765e-02
3.26410413e-01 7.35864818e-01 -8.36708426e-01 5.39680600e-01
1.17307663e+00 8.72933269e-01 -5.18727422e-01 9.11463439e-01
-3.05685580e-01 9.16495442e-01 -2.55351603e-01 6.63540483e-01
5.58535516e-01 -1.62301771e-02 4.64395821e-01 1.60512197e+00
1.73349380e-01 -8.77195671e-02 -1.90745294e-01 4.09576267e-01
-3.03253271e-02 -1.64185330e-01 -9.61404741e-01 6.75984263e-01
4.15612131e-01 1.20617342e+00 -2.81632125e-01 -3.10818911e-01
-7.31073380e-01 4.15343821e-01 2.38496006e-01 3.77351522e-01
-1.11312318e+00 -3.87448639e-01 1.00741792e+00 3.95164818e-01
2.68797219e-01 -1.21362343e-01 -7.99489975e-01 -6.79442763e-01
7.41441846e-02 -3.99324417e-01 4.36081469e-01 -6.90811217e-01
-9.99443710e-01 4.30719167e-01 1.33819371e-01 -1.37172568e+00
-1.49453981e-02 -4.54047263e-01 -6.00829542e-01 9.27872658e-01
-2.13009810e+00 -1.62200880e+00 -4.22625005e-01 5.87804854e-01
6.22353435e-01 3.01020145e-02 3.78816634e-01 8.63457382e-01
-8.26822162e-01 4.29265141e-01 -3.33877802e-01 1.93039000e-01
8.34277451e-01 -1.01313400e+00 1.09280455e+00 9.79126692e-01
-2.55909026e-01 7.74865970e-02 4.73866045e-01 -5.24517596e-01
-6.64531648e-01 -1.79129362e+00 1.37321436e+00 -7.22801745e-01
6.67668521e-01 -2.03216463e-01 -9.28075969e-01 7.88543820e-01
-1.40537977e-01 -9.72937513e-03 5.90967476e-01 1.57220155e-01
-4.52262551e-01 -4.61086482e-01 -9.67400253e-01 4.00381386e-01
1.35275126e+00 -6.97224855e-01 -3.74436647e-01 1.40356466e-01
6.58513546e-01 -9.67137292e-02 -6.97093606e-01 6.43315852e-01
6.13468468e-01 -9.75143850e-01 1.28440404e+00 -5.81481218e-01
6.17287159e-01 -2.90847927e-01 -2.04726785e-01 -9.83829141e-01
-2.41114914e-01 -2.93581605e-01 3.31919044e-01 1.19105828e+00
4.69725519e-01 -4.41961884e-01 8.42192531e-01 8.15775275e-01
-4.93590444e-01 -7.36413717e-01 -1.06409860e+00 -7.10323215e-01
-1.14268316e-02 -1.12889075e+00 6.91037357e-01 7.72339165e-01
-8.13273370e-01 1.36777431e-01 -6.23037517e-01 2.42854685e-01
7.24065661e-01 -4.20755237e-01 9.72867489e-01 -1.30135190e+00
2.77477622e-01 -3.42306137e-01 -3.32273751e-01 -7.42760003e-01
4.76290911e-01 -5.92823446e-01 3.31126824e-02 -1.44411170e+00
8.11476484e-02 -6.55244291e-01 -3.00540388e-01 7.41420329e-01
-1.63293481e-01 5.80770850e-01 -2.65062004e-01 7.00115487e-02
-5.92996776e-01 5.12834668e-01 6.97340369e-01 -1.68971404e-01
1.18505672e-01 5.98077178e-02 -4.71476197e-01 9.16351795e-01
9.39540684e-01 -7.71780252e-01 -3.60059202e-01 -7.13673711e-01
3.77656311e-01 -3.82591158e-01 8.11488748e-01 -1.05660951e+00
2.66081356e-02 -4.00762647e-01 2.22501725e-01 -9.21743631e-01
3.72720182e-01 -8.41484785e-01 -1.81552488e-02 3.05154622e-01
-4.34373379e-01 -3.89647409e-02 2.53167897e-01 5.41631758e-01
-1.91736177e-01 2.03513116e-01 8.56592536e-01 2.31797576e-01
-9.89919364e-01 4.69555557e-01 -5.46559751e-01 1.18711479e-01
1.07524216e+00 -3.83286327e-01 -4.30018365e-01 -2.70140827e-01
-2.16472581e-01 4.68602300e-01 2.51891226e-01 4.55752552e-01
4.46659088e-01 -1.30485535e+00 -1.03052473e+00 2.84783661e-01
3.23622167e-01 -5.31600975e-02 4.10770029e-01 8.78124058e-01
-5.47932208e-01 4.87979144e-01 -2.33213171e-01 -6.29666865e-01
-1.09771764e+00 3.03559333e-01 2.47106165e-01 -1.73657149e-01
-5.57551146e-01 7.08579659e-01 3.06593418e-01 -4.68190402e-01
1.58884406e-01 -3.74016725e-02 -3.59758258e-01 2.76214719e-01
3.70741218e-01 9.08707678e-01 2.99438626e-01 -1.03353512e+00
-3.71008188e-01 7.25087225e-01 1.85340971e-01 -1.55627429e-01
1.35391879e+00 -4.20603007e-01 2.88589954e-01 2.57825494e-01
1.36842704e+00 -2.53829002e-01 -1.68486321e+00 -2.32808098e-01
2.05100402e-01 -5.25795937e-01 1.09515764e-01 -5.98523915e-01
-1.14714313e+00 1.10144126e+00 5.39847732e-01 -2.67860651e-01
1.20169401e+00 -1.87590167e-01 1.14511061e+00 6.00145698e-01
1.10550992e-01 -1.32240891e+00 -2.73032933e-01 7.44717658e-01
5.56672871e-01 -1.49015200e+00 -1.45903692e-01 -3.80365998e-01
-8.45816612e-01 6.95295036e-01 5.58734834e-01 1.05591142e-03
5.77635705e-01 -8.98211636e-03 1.02806374e-01 -6.39996082e-02
-6.56762362e-01 -6.52674913e-01 4.51070517e-01 5.61523080e-01
1.20141745e-01 -6.55589178e-02 3.27840745e-02 1.71144873e-01
3.10915373e-02 -1.07288405e-01 3.25958937e-01 5.96455634e-01
-5.11018813e-01 -7.84476459e-01 -1.92827925e-01 5.81703782e-01
-6.67941928e-01 -1.85563639e-01 2.49358304e-02 7.77299047e-01
3.84004593e-01 1.08580565e+00 3.33906472e-01 -5.39911389e-01
4.48057234e-01 -9.82188955e-02 -1.02329895e-01 -3.12789887e-01
-4.66654271e-01 -4.31109607e-01 7.00704277e-01 -9.48560238e-01
-5.27326167e-01 -8.70955169e-01 -9.39971149e-01 -5.38146257e-01
2.54564047e-01 -3.65966111e-01 6.58765912e-01 9.99335825e-01
3.64091456e-01 5.27019262e-01 7.86866367e-01 -9.94867265e-01
-5.51093966e-02 -6.10560536e-01 -8.10366776e-03 4.72465485e-01
7.11356878e-01 -7.93095708e-01 -2.20887557e-01 2.98421770e-01] | [6.687402725219727, 0.6077577471733093] |
9a0876f2-362e-41e0-924f-46a48e5ced4c | effective-transfer-of-pretrained-large-visual | 2306.16186 | null | https://arxiv.org/abs/2306.16186v1 | https://arxiv.org/pdf/2306.16186v1.pdf | Effective Transfer of Pretrained Large Visual Model for Fabric Defect Segmentation via Specifc Knowledge Injection | Fabric defect segmentation is integral to textile quality control. Despite this, the scarcity of high-quality annotated data and the diversity of fabric defects present significant challenges to the application of deep learning in this field. These factors limit the generalization and segmentation performance of existing models, impeding their ability to handle the complexity of diverse fabric types and defects. To overcome these obstacles, this study introduces an innovative method to infuse specialized knowledge of fabric defects into the Segment Anything Model (SAM), a large-scale visual model. By introducing and training a unique set of fabric defect-related parameters, this approach seamlessly integrates domain-specific knowledge into SAM without the need for extensive modifications to the pre-existing model parameters. The revamped SAM model leverages generalized image understanding learned from large-scale natural image datasets while incorporating fabric defect-specific knowledge, ensuring its proficiency in fabric defect segmentation tasks. The experimental results reveal a significant improvement in the model's segmentation performance, attributable to this novel amalgamation of generic and fabric-specific knowledge. When benchmarking against popular existing segmentation models across three datasets, our proposed model demonstrates a substantial leap in performance. Its impressive results in cross-dataset comparisons and few-shot learning experiments further demonstrate its potential for practical applications in textile quality control. | ['Ying Qu', 'Jinpiao Liao', 'Zuofeng Zhong', 'Wai Keung Wong', 'Zhewei Chen'] | 2023-06-28 | null | null | null | null | ['few-shot-learning', 'benchmarking', 'benchmarking'] | ['methodology', 'miscellaneous', 'robots'] | [ 6.15812182e-01 -2.37676337e-01 -2.46997878e-01 -3.07100415e-01
-5.94923139e-01 -5.37881076e-01 6.09998554e-02 1.25993297e-01
-2.73453984e-02 3.90504271e-01 5.22822365e-02 1.99195463e-02
-1.97203666e-01 -7.61713564e-01 -6.29503489e-01 -5.42473614e-01
1.93795994e-01 7.73421004e-02 2.72931725e-01 -3.21990848e-01
3.02125216e-01 3.57543617e-01 -1.50474811e+00 3.16752076e-01
1.16233158e+00 1.17373610e+00 5.55254221e-01 5.38378358e-01
-1.48147866e-01 1.82367757e-01 -4.50591832e-01 -1.75604880e-01
4.03528243e-01 -2.36208484e-01 -7.96422303e-01 5.71269095e-01
4.88114715e-01 -4.44310576e-01 2.10153013e-02 7.68253863e-01
6.57074749e-01 -1.63134187e-01 3.71999055e-01 -7.50814855e-01
-9.85314369e-01 1.09517254e-01 -5.05325198e-01 8.36786106e-02
1.47140875e-01 6.31381035e-01 1.10690796e+00 -4.66476262e-01
7.39906967e-01 9.74574268e-01 9.41925287e-01 4.29014862e-01
-1.64314878e+00 -1.81034058e-01 2.17494175e-01 -1.91005394e-01
-9.77919400e-01 -1.48804739e-01 1.11065495e+00 -5.76859355e-01
6.77846849e-01 2.00309187e-01 1.17627490e+00 1.16964316e+00
1.90360159e-01 7.46946692e-01 1.26144254e+00 -2.85126925e-01
3.58636945e-01 -3.32244560e-02 6.37814850e-02 8.30299616e-01
3.03165853e-01 4.60647158e-02 -2.85874605e-01 1.98971331e-01
1.25304854e+00 1.96084492e-02 -7.96209499e-02 -6.74350381e-01
-1.16645086e+00 5.93799114e-01 6.00913584e-01 1.20841041e-02
-2.76499778e-01 1.20071828e-01 7.07662463e-01 2.07503378e-01
5.81686497e-01 7.93679893e-01 -5.56891322e-01 -6.62159026e-02
-9.59794223e-01 1.12957180e-01 6.20899856e-01 8.75299215e-01
7.37380385e-01 1.32360294e-01 -2.36798510e-01 1.13800991e+00
5.21640480e-02 3.46562833e-01 2.54686810e-02 -1.02156603e+00
4.50877622e-02 9.39074099e-01 -4.03043777e-02 -1.17270458e+00
-3.82049620e-01 -4.49906409e-01 -5.82909822e-01 2.28639916e-01
3.10447276e-01 1.49177521e-01 -1.31555605e+00 1.44147015e+00
4.22961980e-01 -2.16364950e-01 -4.04842019e-01 1.01543331e+00
6.36512101e-01 2.03765184e-01 3.22633505e-01 1.55327737e-01
1.22650480e+00 -9.12540972e-01 -4.09187526e-01 -2.72849858e-01
2.77692765e-01 -8.77850592e-01 1.49067986e+00 3.36718172e-01
-7.45323837e-01 -7.71111667e-01 -1.04027843e+00 -1.09689154e-01
-3.20871741e-01 2.02430338e-01 1.14678335e+00 6.67522728e-01
-6.32767439e-01 5.97377598e-01 -7.73204505e-01 -5.95173120e-01
9.62648451e-01 3.83554220e-01 -2.19963282e-01 -3.70440513e-01
-7.00818777e-01 5.18955112e-01 4.04106557e-01 3.37298900e-01
-9.21667993e-01 -1.18105912e+00 -8.06564212e-01 -2.72119939e-01
5.25329113e-01 -9.00609374e-01 8.98578465e-01 -1.06628585e+00
-1.41498613e+00 9.00248826e-01 2.56137073e-01 -6.22104853e-03
5.41333199e-01 -3.14578712e-01 -1.54737279e-01 2.95206755e-01
1.33507505e-01 7.68936753e-01 7.91061163e-01 -1.67442894e+00
-3.99958462e-01 -2.00743586e-01 3.78572106e-01 -9.20006558e-02
-3.23408037e-01 -3.92530233e-01 -5.65104008e-01 -9.06967938e-01
5.50759695e-02 -7.77314901e-01 -2.90845245e-01 5.33244371e-01
-4.89701986e-01 1.23588681e-01 8.73594165e-01 -7.66838968e-01
1.00197971e+00 -1.93989789e+00 1.97182640e-01 3.46230865e-02
1.82306901e-01 3.22682887e-01 -3.29310745e-01 4.87707913e-01
3.11678797e-01 1.22634307e-01 -4.33313131e-01 -1.74446732e-01
-1.20290294e-01 5.72620034e-01 1.90020680e-01 3.19141835e-01
5.12192130e-01 1.18302131e+00 -1.05080450e+00 -5.87783694e-01
6.00838542e-01 3.08348060e-01 -7.33429790e-01 1.03481427e-01
-5.44223964e-01 6.25282645e-01 -4.36788678e-01 1.17046881e+00
5.73000968e-01 -3.27480763e-01 5.52163906e-02 -8.41905296e-01
7.33152851e-02 -2.43757308e-01 -1.06904542e+00 2.23265266e+00
-5.11246681e-01 2.43171960e-01 1.21070594e-01 -9.75127339e-01
8.44900429e-01 7.97940232e-03 7.19840586e-01 -8.06495070e-01
1.71370745e-01 1.24752901e-01 -1.70709893e-01 -9.30835724e-01
2.32957438e-01 -2.18455851e-01 -1.90885812e-01 2.95287251e-01
3.19782197e-01 -3.31246197e-01 2.65324652e-01 -6.86273575e-02
7.34607041e-01 4.19901997e-01 -2.70828661e-02 -4.09585774e-01
5.92353605e-02 1.65519595e-01 7.32143402e-01 7.11650193e-01
-3.38899642e-01 7.60699034e-01 2.80306727e-01 -5.60279191e-01
-1.29850841e+00 -1.24876750e+00 -4.54873592e-02 9.32188869e-01
4.35717553e-01 -2.23954111e-01 -6.29191101e-01 -7.47080743e-01
3.40863794e-01 5.34950718e-02 -1.04778659e+00 -3.30098540e-01
-4.53328669e-01 -8.68473291e-01 3.07717055e-01 6.67932987e-01
7.01124191e-01 -8.97210658e-01 -7.75109470e-01 2.45040804e-01
-1.86856180e-01 -9.06918526e-01 -4.03264195e-01 -1.35063333e-02
-9.23498809e-01 -1.20164657e+00 -6.81834638e-01 -8.51371527e-01
6.72174454e-01 2.91074753e-01 1.14440775e+00 7.57955015e-02
-1.14109957e+00 6.81962371e-01 -3.01337242e-01 -3.44005287e-01
-3.74080926e-01 1.04587100e-01 -3.15672994e-01 -1.26052558e-01
-8.84681102e-03 -5.46739936e-01 -1.04062319e+00 2.32120559e-01
-8.82832527e-01 1.66986570e-01 7.74871647e-01 1.04257798e+00
8.34277630e-01 -6.93144575e-02 7.52916455e-01 -8.60414922e-01
5.86938262e-01 -2.28110954e-01 -1.70980021e-01 2.88867742e-01
-6.40196204e-01 -3.49230260e-01 2.79289186e-01 -5.74668288e-01
-1.05805206e+00 -9.76708904e-02 -3.13079320e-02 -1.98029220e-01
-2.81694919e-01 3.67558688e-01 -2.49440983e-01 -3.89309615e-01
4.56678361e-01 -3.97941805e-02 2.81534880e-01 -6.04573250e-01
4.70894784e-01 3.31534505e-01 4.56016213e-01 -9.57046151e-01
4.22138482e-01 4.80850101e-01 -2.49118954e-01 -8.01350296e-01
-1.13502991e+00 -4.27293569e-01 -8.68992269e-01 -4.48277861e-01
9.26999807e-01 -8.45705509e-01 -5.70826292e-01 7.10430205e-01
-8.19037676e-01 -3.94612372e-01 -7.10761726e-01 -2.76936423e-02
-5.48799813e-01 5.35894692e-01 -7.90516257e-01 -5.02040863e-01
-4.11685795e-01 -1.23395991e+00 1.24525774e+00 1.57487109e-01
-2.13114798e-01 -1.14486527e+00 -1.43707961e-01 7.14070439e-01
5.28910935e-01 7.23801434e-01 1.16315448e+00 7.60038197e-02
-5.55701792e-01 -1.19500741e-01 -4.71492618e-01 9.47832346e-01
7.22373605e-01 -3.10928728e-02 -8.68142545e-01 -2.45360285e-01
-2.51833886e-01 -4.98771727e-01 7.82481551e-01 4.15390193e-01
1.17875063e+00 1.25391349e-01 -1.81303829e-01 5.05874157e-01
1.71140921e+00 -1.73377141e-01 3.58240902e-01 4.19396251e-01
1.02981544e+00 5.95786154e-01 6.25850260e-01 4.29251671e-01
4.09918308e-01 5.30212343e-01 4.68362123e-01 -6.82969093e-01
-4.90447640e-01 -1.42292842e-01 -1.64785132e-01 8.78647685e-01
-2.09138587e-01 1.06950536e-01 -8.36906433e-01 7.74687052e-01
-1.62474263e+00 -7.02807784e-01 1.54047742e-01 1.78762662e+00
1.04707277e+00 1.51383147e-01 1.53619587e-01 3.14078070e-02
4.37486589e-01 7.46273473e-02 -9.51567411e-01 -3.07427227e-01
-1.41434610e-01 3.01140010e-01 3.33151191e-01 -5.50496951e-02
-1.17140996e+00 9.64733243e-01 6.95373249e+00 7.18033969e-01
-1.14375854e+00 -2.41143350e-02 4.49688226e-01 -4.36725803e-02
-2.79295951e-01 -2.01627061e-01 -3.16278279e-01 4.66062844e-01
4.41714644e-01 2.38930583e-01 3.98000777e-01 6.56636834e-01
3.05247933e-01 -1.72090545e-01 -9.35410619e-01 7.56958723e-01
2.72284951e-02 -1.48038507e+00 1.63818672e-01 -2.32205186e-02
8.14990044e-01 -2.24716231e-01 7.22292587e-02 1.38977483e-01
1.68536931e-01 -8.31231415e-01 7.94333160e-01 5.51736057e-01
7.19233930e-01 -5.89118898e-01 5.89481771e-01 -3.02284896e-01
-1.08734584e+00 -3.10460180e-01 -4.28096831e-01 -1.08336483e-03
2.79498070e-01 7.90083766e-01 -5.46410024e-01 8.09112191e-01
6.65963709e-01 8.97999823e-01 -6.69235051e-01 1.06041634e+00
1.69467390e-01 6.74229026e-01 -1.24059133e-01 4.92808372e-01
1.38135478e-01 -1.84795678e-01 3.07776034e-01 1.25224674e+00
-5.41752614e-02 -3.82451296e-01 6.09834552e-01 1.20260811e+00
3.63314524e-02 -1.97478873e-03 -3.19933921e-01 -2.34498233e-01
1.59735098e-01 1.20195568e+00 -9.60419774e-01 -8.97818431e-02
-4.02529746e-01 1.15120506e+00 1.45665169e-01 3.21359932e-01
-8.23001087e-01 -1.56501114e-01 8.75499010e-01 1.08501896e-01
4.38220829e-01 -3.76939714e-01 -7.69873798e-01 -7.78424323e-01
2.08395794e-01 -9.83192503e-01 3.78491059e-02 -4.42639649e-01
-1.74488163e+00 1.48989916e-01 -1.65166840e-01 -1.03466058e+00
6.59999728e-01 -6.28584385e-01 -4.21183079e-01 5.80947638e-01
-1.54484761e+00 -1.98206627e+00 -4.87230569e-01 3.52150947e-01
7.63563395e-01 2.05715388e-01 7.66624987e-01 3.91647786e-01
-7.36904681e-01 3.92236352e-01 -1.02513246e-01 -7.13353083e-02
6.21894896e-01 -1.18787289e+00 2.89285392e-01 6.37044787e-01
-2.65091717e-01 5.94846010e-01 6.33441508e-01 -8.16215277e-01
-1.50558949e+00 -1.24250102e+00 -6.36490658e-02 -4.81707811e-01
5.14419556e-01 -5.60595870e-01 -8.44225049e-01 3.28059971e-01
-1.69955060e-01 1.09861389e-01 9.17840898e-01 1.53915912e-01
-3.35685492e-01 -1.64595440e-01 -1.26449490e+00 4.55695152e-01
1.23269796e+00 -4.90040213e-01 -4.38690186e-01 2.65386254e-01
8.84780526e-01 -3.18229735e-01 -1.50545740e+00 6.71587646e-01
7.84864068e-01 -7.28943348e-01 1.18224919e+00 -4.97786880e-01
7.38212824e-01 -1.92648143e-01 -1.48137003e-01 -1.08315551e+00
-3.92408460e-01 -2.60493308e-01 -3.89853343e-02 1.48321843e+00
-3.89084709e-03 -2.49854833e-01 6.59557819e-01 5.48487723e-01
-3.45774889e-01 -1.08006942e+00 -6.29767597e-01 -6.98579252e-01
-5.60596362e-02 -3.38385820e-01 4.30433810e-01 9.16393638e-01
-4.83205050e-01 -9.89602208e-02 -4.05506760e-01 3.41099724e-02
6.42822862e-01 4.36842233e-01 6.57893121e-01 -1.42582297e+00
-2.60349929e-01 -3.09687883e-01 -4.58459377e-01 -8.72232139e-01
-1.62688017e-01 -8.90646756e-01 1.10155977e-01 -1.88544619e+00
2.65021622e-01 -8.47005069e-01 -3.67901087e-01 5.75398803e-01
-1.93800911e-01 5.56729376e-01 1.81110799e-01 -3.97548825e-03
-5.16360402e-01 6.33461058e-01 1.78330207e+00 -4.57387179e-01
-2.58631796e-01 -2.32581437e-01 -9.16565537e-01 4.47297931e-01
7.85340905e-01 -1.15405247e-01 -4.84140694e-01 -6.30588055e-01
-2.40452632e-01 -5.19573808e-01 6.40394330e-01 -9.94790494e-01
-2.18688533e-01 -2.18685240e-01 4.75284129e-01 -4.69048060e-02
1.08139038e-01 -6.61057949e-01 4.89051566e-02 4.32482779e-01
-7.61850625e-02 -2.77699292e-01 5.18448949e-01 8.62832010e-01
4.93554324e-02 1.90656602e-01 8.50475252e-01 -3.78627449e-01
-1.04593766e+00 3.00431073e-01 -1.36633366e-01 -6.74162284e-02
1.10103476e+00 -7.91288733e-01 -1.33332118e-01 4.31859940e-01
-7.35400736e-01 1.24588571e-01 8.98160100e-01 7.70298719e-01
5.64565599e-01 -1.11921513e+00 -2.65371919e-01 3.99455905e-01
3.49067271e-01 1.24317802e-01 7.21986234e-01 7.77298152e-01
-4.66041803e-01 -1.31878043e-02 -6.70845330e-01 -8.90985727e-01
-9.57277179e-01 4.97910738e-01 1.96024269e-01 -1.07742205e-01
-8.88913274e-01 7.48730063e-01 5.24795577e-02 -6.02952838e-01
-8.22204575e-02 -4.82178509e-01 3.00187230e-01 -1.55560330e-01
5.96083328e-03 4.72629040e-01 1.77589934e-02 -1.78944528e-01
-4.14313078e-02 8.35434675e-01 -1.71198517e-01 5.41300476e-01
1.51410902e+00 -3.82990539e-01 -1.12331681e-01 6.84131384e-01
9.38030839e-01 -1.67582542e-01 -1.71832514e+00 9.49079171e-03
-5.84471300e-02 -5.32044709e-01 5.18325977e-02 -1.15410399e+00
-1.23634493e+00 7.55723596e-01 8.99954975e-01 -1.50649309e-01
1.18132389e+00 -1.09327212e-02 1.11746264e+00 8.12308863e-03
7.04325676e-01 -1.31578314e+00 3.69046986e-01 4.42409627e-02
7.37649381e-01 -1.38410437e+00 -2.83008106e-02 -6.64340913e-01
-6.07581019e-01 1.04179311e+00 6.93028927e-01 -2.04043388e-01
4.48530734e-01 2.59830326e-01 2.08077684e-01 -4.55537796e-01
-2.58903921e-01 -8.95954967e-02 2.82854587e-01 8.69667351e-01
2.24289164e-01 -2.76154513e-03 -1.81515425e-01 3.95541579e-01
2.40631953e-01 2.88318008e-01 1.31776944e-01 1.30527091e+00
-3.58196706e-01 -1.18995678e+00 1.83234215e-01 6.68572605e-01
-1.64032191e-01 2.11438224e-01 -8.81872177e-02 7.58513451e-01
5.87721229e-01 7.16246247e-01 -2.16729686e-01 -2.75897175e-01
5.39908111e-01 -2.31965438e-01 7.84036279e-01 -6.61475956e-01
-6.98753417e-01 -1.84138808e-02 -1.20538816e-01 -8.26069534e-01
-5.18530190e-01 -4.39783245e-01 -8.57164860e-01 -4.45061214e-02
-3.49575192e-01 -5.44588983e-01 6.15989387e-01 8.63552690e-01
5.82837343e-01 9.81704175e-01 3.36468607e-01 -9.60088551e-01
-2.73318440e-01 -5.94919145e-01 -6.67919338e-01 7.24043369e-01
2.17597604e-01 -9.19470489e-01 1.56657442e-01 4.20149982e-01] | [10.23140811920166, -0.16905878484249115] |
2019ffc2-37c6-4dcb-a240-b8dfcbde62cb | selective-transfer-machine-for-personalized | null | null | http://openaccess.thecvf.com/content_cvpr_2013/html/Chu_Selective_Transfer_Machine_2013_CVPR_paper.html | http://openaccess.thecvf.com/content_cvpr_2013/papers/Chu_Selective_Transfer_Machine_2013_CVPR_paper.pdf | Selective Transfer Machine for Personalized Facial Action Unit Detection | Automatic facial action unit (AFA) detection from video is a long-standing problem in facial expression analysis. Most approaches emphasize choices of features and classifiers. They neglect individual differences in target persons. People vary markedly in facial morphology (e.g., heavy versus delicate brows, smooth versus deeply etched wrinkles) and behavior. Individual differences can dramatically influence how well generic classifiers generalize to previously unseen persons. While a possible solution would be to train person-specific classifiers, that often is neither feasible nor theoretically compelling. The alternative that we propose is to personalize a generic classifier in an unsupervised manner (no additional labels for the test subjects are required). We introduce a transductive learning method, which we refer to Selective Transfer Machine (STM), to personalize a generic classifier by attenuating person-specific biases. STM achieves this effect by simultaneously learning a classifier and re-weighting the training samples that are most relevant to the test subject. To evaluate the effectiveness of STM, we compared STM to generic classifiers and to cross-domain learning methods in three major databases: CK+ [20], GEMEP-FERA [32] and RU-FACS [2]. STM outperformed generic classifiers in all. | ['Fernando de la Torre', 'Wen-Sheng Chu', 'Jeffery F. Cohn'] | 2013-06-01 | null | null | null | cvpr-2013-6 | ['action-unit-detection', 'facial-action-unit-detection'] | ['computer-vision', 'computer-vision'] | [ 3.40325058e-01 -1.61926568e-01 -2.75560200e-01 -8.19924712e-01
-5.30397177e-01 -5.45024872e-01 6.10323012e-01 -2.95326561e-01
-3.71628463e-01 8.25299263e-01 1.07519254e-01 3.83929312e-01
3.69381234e-02 -6.09018385e-01 -3.75169724e-01 -8.83966029e-01
-7.69679025e-02 2.49353305e-01 -6.51429817e-02 -2.55966187e-01
8.29711780e-02 6.35659516e-01 -1.92842567e+00 6.09479547e-01
5.12351811e-01 1.04825187e+00 -3.00325245e-01 3.28140199e-01
8.92545804e-02 3.78721327e-01 -7.01069117e-01 -6.86194181e-01
1.17603101e-01 -7.05064178e-01 -7.82784104e-01 4.93870735e-01
7.54825175e-01 -6.20232373e-02 2.57630706e-01 1.10993230e+00
3.89768869e-01 2.70652808e-02 1.12856054e+00 -1.29282820e+00
-5.32664359e-01 1.70178786e-01 -6.89146876e-01 4.70801517e-02
4.11642939e-01 2.11339165e-02 7.28313625e-01 -8.52592349e-01
7.00046062e-01 1.43276191e+00 6.86892390e-01 1.22763455e+00
-1.36534381e+00 -9.96014237e-01 1.27082556e-01 9.57982764e-02
-1.61705768e+00 -8.68391633e-01 9.64759707e-01 -5.88698208e-01
3.03420603e-01 3.36699724e-01 5.46573699e-01 1.48231435e+00
2.10259743e-02 7.33260691e-01 1.54135489e+00 -3.78998280e-01
2.59756684e-01 4.39655721e-01 2.35356949e-03 8.08016121e-01
7.09820762e-02 -8.04468915e-02 -7.12423146e-01 -3.96636963e-01
4.89725858e-01 -2.13403627e-01 -2.72330344e-01 -3.04254055e-01
-6.97044075e-01 8.92396212e-01 2.65010986e-02 3.50706398e-01
-3.84259403e-01 -2.38827214e-01 5.93433022e-01 4.84435499e-01
5.96579134e-01 4.38050121e-01 -3.77013505e-01 2.29964163e-02
-8.62576067e-01 3.66125941e-01 6.03169620e-01 6.26908064e-01
9.84533012e-01 -5.46403714e-02 -1.83381408e-01 1.06217003e+00
8.97462666e-02 2.85430551e-01 7.15638280e-01 -9.82492447e-01
-1.52854681e-01 4.92651850e-01 -1.14271112e-01 -1.00465465e+00
-1.92022681e-01 1.00795338e-02 -5.75741470e-01 5.02840042e-01
4.25848186e-01 -3.48947585e-01 -7.72480071e-01 2.03718066e+00
3.07036519e-01 -2.29542982e-02 1.58697981e-02 8.49361002e-01
8.55010629e-01 3.24382395e-01 4.28626120e-01 -3.34521264e-01
1.35005879e+00 -6.50742829e-01 -5.61540782e-01 -1.87792256e-01
5.24830043e-01 -6.27157569e-01 8.63936126e-01 5.56997478e-01
-8.24893296e-01 -4.74419326e-01 -8.12952876e-01 3.73407215e-01
-4.28592175e-01 2.06182808e-01 6.81318521e-01 1.08628738e+00
-8.60958278e-01 6.77284479e-01 -1.73278317e-01 -7.26286769e-01
7.74961710e-01 5.45513213e-01 -8.56401742e-01 1.62181824e-01
-1.02683246e+00 7.77673066e-01 8.07908773e-02 -6.79524019e-02
-7.13044286e-01 -3.62122923e-01 -8.29449475e-01 -3.27602893e-01
2.42006585e-01 -6.01248741e-01 1.19545698e+00 -2.02717566e+00
-1.81925857e+00 1.52913713e+00 -2.51518250e-01 -7.80491754e-02
3.49847466e-01 -2.87420191e-02 -5.39954364e-01 1.35504290e-01
-7.81787112e-02 8.80524337e-01 1.48688185e+00 -1.30016971e+00
-4.77480590e-01 -5.75499535e-01 -3.13755088e-02 1.34107739e-01
-4.39999670e-01 3.54589760e-01 -1.34032127e-02 -7.02785432e-01
-2.85578877e-01 -1.01054871e+00 1.80505633e-01 3.04308772e-01
1.02462918e-02 -5.50647140e-01 9.17803407e-01 -4.33482647e-01
9.72789943e-01 -2.22955346e+00 2.81032205e-01 1.43773898e-01
4.17221561e-02 4.49958891e-01 -2.66580582e-01 4.41565998e-02
-2.20318004e-01 -3.94306593e-02 -1.27572954e-01 -1.95510954e-01
-1.66239887e-01 2.61308134e-01 1.10132299e-01 6.04548454e-01
3.97316873e-01 6.58475816e-01 -8.19118619e-01 -6.70950234e-01
5.08696679e-03 3.42495382e-01 -5.76629996e-01 1.28171563e-01
-1.55550092e-02 4.13205594e-01 -4.43814397e-01 1.01922703e+00
5.38773179e-01 1.23885609e-01 1.86412036e-01 -1.89820603e-01
1.85754642e-01 -2.89265931e-01 -1.03492403e+00 1.26302981e+00
-2.73239851e-01 6.22727931e-01 3.55792433e-01 -1.05456614e+00
1.12511158e+00 4.56748813e-01 3.36690724e-01 -3.61835450e-01
3.29913378e-01 1.81998789e-01 -1.02012075e-01 -6.45231307e-01
1.30875364e-01 -6.81023359e-01 3.07413358e-02 1.13059781e-01
2.57978112e-01 1.05993070e-01 -8.92399624e-02 -3.47340614e-01
6.31772459e-01 1.98979750e-01 5.70956230e-01 -3.73378724e-01
7.37690091e-01 -3.71636510e-01 7.14608848e-01 3.14490885e-01
-6.14650071e-01 5.33207536e-01 4.41055655e-01 -4.09395874e-01
-3.49663615e-01 -1.04056191e+00 -2.04030871e-01 1.56661224e+00
-2.50468135e-01 -2.94980347e-01 -9.51325297e-01 -9.55675483e-01
-3.65010253e-03 2.92322963e-01 -1.00845122e+00 -2.56001294e-01
-2.46203169e-01 -5.51504195e-01 4.79811460e-01 3.35591376e-01
3.67316902e-01 -1.26225328e+00 -5.44363439e-01 -1.21720701e-01
-1.16982095e-01 -8.15980196e-01 -4.11905110e-01 -1.96220968e-02
-6.79208159e-01 -8.49239409e-01 -8.76204133e-01 -6.71781719e-01
8.05507004e-01 4.73033711e-02 1.07838428e+00 -5.07785752e-02
-3.77475619e-01 5.13944805e-01 -3.79112512e-01 -5.44731975e-01
-1.87617913e-01 -1.76273033e-01 3.04804444e-01 7.67911375e-01
7.21527100e-01 -3.81499380e-01 -4.26535517e-01 5.21069288e-01
-6.27625823e-01 -2.95251250e-01 5.06185591e-01 9.21906710e-01
3.73505741e-01 -1.63125113e-01 6.05219185e-01 -1.10545707e+00
5.71500123e-01 -4.19630527e-01 -1.18668959e-01 1.44812599e-01
-3.08443695e-01 -3.58357012e-01 2.93030649e-01 -7.38691449e-01
-1.09505045e+00 3.51404011e-01 -1.41478125e-02 -5.86783707e-01
-5.53130805e-01 1.69220001e-01 -3.14361393e-01 -2.92436808e-01
8.11013520e-01 6.60193190e-02 2.55298764e-01 -2.06255764e-01
6.04453534e-02 7.80161917e-01 4.63719159e-01 -8.51536989e-01
6.05640769e-01 5.09423614e-01 -3.06997038e-02 -1.21517503e+00
-9.31533813e-01 -3.22127193e-01 -8.38513792e-01 -6.20364666e-01
9.69110966e-01 -7.87915230e-01 -6.38154685e-01 5.99453866e-01
-8.97701919e-01 -1.29241392e-01 -1.27053484e-01 2.43831784e-01
-5.84483922e-01 2.51635402e-01 -2.64148623e-01 -7.53214121e-01
-1.83833554e-01 -8.63848388e-01 1.14097273e+00 3.17815095e-01
-8.33554506e-01 -9.08583701e-01 -5.54441940e-05 3.13222021e-01
3.53594720e-01 6.86203778e-01 6.26773179e-01 -5.38829863e-01
1.11650698e-01 -2.86634833e-01 -1.96398585e-03 6.07193887e-01
5.48088908e-01 2.95233458e-01 -1.36192214e+00 -1.82633832e-01
-1.00953914e-01 -7.03567266e-01 7.26737082e-01 1.45106271e-01
1.26697683e+00 -2.73211479e-01 -3.22211653e-01 4.65055197e-01
9.67072070e-01 1.96160644e-01 6.07924998e-01 1.44925281e-01
3.82018805e-01 1.14643598e+00 5.09276748e-01 3.73385131e-01
-6.14133589e-02 7.91916132e-01 -2.79695466e-02 -6.37049899e-02
3.18922754e-03 -7.72251114e-02 6.20050430e-01 2.74721026e-01
-5.35762310e-01 9.15057138e-02 -5.05282938e-01 3.18788677e-01
-1.44198787e+00 -1.25555515e+00 2.04253659e-01 2.01624012e+00
8.48858595e-01 -1.46808907e-01 4.99405444e-01 1.72349587e-01
7.89088249e-01 -3.32348049e-02 -2.84180492e-01 -5.83842337e-01
-2.82156229e-01 3.92286420e-01 2.98164058e-02 2.81101167e-01
-1.24357092e+00 9.84504104e-01 6.54898262e+00 8.03047597e-01
-1.45995545e+00 1.45855233e-01 7.01749921e-01 -1.48448935e-02
1.81041241e-01 -3.81148249e-01 -8.42431307e-01 3.92847329e-01
7.82849669e-01 -1.70878097e-01 1.75995588e-01 8.94273341e-01
-3.65255363e-02 9.81463566e-02 -1.30278087e+00 1.22940016e+00
4.50871140e-01 -1.01376557e+00 1.57401443e-01 -4.37873080e-02
5.63605666e-01 -5.19253671e-01 1.99371189e-01 4.08643544e-01
-2.31296375e-01 -1.18231463e+00 5.41463196e-01 6.20146871e-01
8.11848223e-01 -6.87619448e-01 5.36905348e-01 -8.26386884e-02
-8.67690086e-01 2.60841213e-02 -2.53062904e-01 -4.41843010e-02
-2.62248755e-01 5.44694774e-02 -5.47023535e-01 1.60380602e-01
8.78494203e-01 6.46363318e-01 -6.37485743e-01 6.11768484e-01
-9.76264626e-02 5.04037261e-01 -1.76211193e-01 -9.61784720e-02
6.13544062e-02 -1.78195208e-01 4.32845473e-01 1.34284914e+00
8.42801481e-02 1.94524363e-01 -6.42924756e-02 5.02764463e-01
6.08597435e-02 3.84592503e-01 -6.42270684e-01 2.07800288e-02
4.44017164e-02 1.46420097e+00 -4.16345835e-01 -1.71177253e-01
-5.13878107e-01 1.21131349e+00 3.16219181e-01 2.93644786e-01
-4.86495048e-01 -2.80875731e-02 8.72998416e-01 3.62705171e-01
2.96200216e-01 2.50697166e-01 1.49934024e-01 -1.01614964e+00
-1.40571877e-01 -1.20200980e+00 7.10598767e-01 -5.68817616e-01
-1.67658198e+00 6.99962974e-01 1.19854093e-01 -1.26409769e+00
-2.65162200e-01 -8.23444068e-01 -5.71299851e-01 7.28141010e-01
-1.32763672e+00 -1.33527827e+00 -4.28543895e-01 1.03514624e+00
4.09988523e-01 -4.27373469e-01 9.88455772e-01 2.93489218e-01
-4.85571533e-01 9.13008571e-01 -4.44821805e-01 2.56467968e-01
1.13199508e+00 -8.12373877e-01 -3.42140526e-01 2.42851838e-01
-4.80355360e-02 5.82703948e-01 6.83160782e-01 -2.91802824e-01
-1.17031455e+00 -1.04474139e+00 9.02537405e-01 -3.84159237e-01
5.01087785e-01 -4.29913372e-01 -9.85089540e-01 6.39814258e-01
7.92632401e-02 1.87461942e-01 9.35519218e-01 3.52042049e-01
-6.46921515e-01 -3.09171230e-01 -1.32915664e+00 6.57554567e-01
1.05172932e+00 -4.72853988e-01 -5.13452172e-01 2.02246532e-01
-2.89793193e-01 4.37098704e-02 -7.52451181e-01 4.45688069e-01
7.93473721e-01 -1.21521473e+00 6.74612343e-01 -7.45830178e-01
1.83957070e-01 -1.76826850e-01 -2.14262128e-01 -1.27071476e+00
-4.29151386e-01 -7.38236248e-01 2.05365360e-01 1.34738195e+00
2.40056396e-01 -7.23634124e-01 9.64892030e-01 5.87170601e-01
2.22639069e-01 -6.94639802e-01 -9.46859419e-01 -8.35820794e-01
1.81533486e-01 -2.01323051e-02 4.62366462e-01 1.20037913e+00
1.05141923e-02 4.01210994e-01 -5.34445643e-01 -2.06708670e-01
5.59518874e-01 2.83943806e-02 9.78946388e-01 -1.46131098e+00
-1.41616002e-01 -7.01908410e-01 -7.32308984e-01 -4.21646863e-01
6.70671582e-01 -9.08512235e-01 -1.91007093e-01 -6.31023109e-01
2.70085424e-01 -2.21321955e-01 -2.78085440e-01 6.36618972e-01
-2.70803124e-02 5.03184676e-01 -6.82141110e-02 3.55062596e-02
-4.17543948e-01 4.37820673e-01 1.18172133e+00 -1.19389080e-01
-1.12724714e-01 1.05320521e-01 -7.71672845e-01 9.78439987e-01
7.21063733e-01 -2.92600185e-01 -2.41998449e-01 4.63524722e-02
-2.96820462e-01 -1.14157327e-01 3.35516155e-01 -8.85081172e-01
-1.49341866e-01 -2.96628416e-01 5.70160389e-01 3.35704461e-02
5.11732876e-01 -4.74140823e-01 8.09174925e-02 2.74385542e-01
-1.68317780e-01 -2.54624963e-01 1.62655368e-01 3.51948649e-01
-3.23485434e-01 -3.35205019e-01 1.28370905e+00 -1.74810842e-01
-8.97094250e-01 4.53637511e-01 -5.93482316e-01 -1.13595240e-01
1.22836196e+00 -5.45332611e-01 -1.31194461e-02 -5.24252117e-01
-8.49617958e-01 -2.64222115e-01 5.44865787e-01 5.94551146e-01
5.09768426e-01 -1.34547913e+00 -8.83632123e-01 2.59303451e-01
5.08067250e-01 -6.43871427e-01 2.15332359e-02 6.29492044e-01
-4.59649228e-02 4.69214395e-02 -6.18571997e-01 -6.29015803e-01
-1.81350183e+00 5.02149642e-01 5.36480427e-01 3.61164480e-01
-1.59842417e-01 1.02929950e+00 4.28476036e-01 -1.67044997e-01
4.13660184e-02 2.00794294e-01 -4.00396824e-01 4.48419750e-01
5.78628600e-01 6.97972104e-02 -1.28955841e-01 -1.05957270e+00
-3.63267094e-01 7.52289891e-01 -1.06684387e-01 1.47300139e-01
1.15875649e+00 1.49247542e-01 -1.84340641e-01 5.56202173e-01
1.27766597e+00 1.88576579e-02 -9.92757499e-01 -2.14330643e-01
-9.71962065e-02 -6.26378775e-01 -1.64433718e-01 -5.00366867e-01
-1.08702946e+00 6.82004452e-01 7.22441852e-01 -9.52767730e-02
1.37111497e+00 1.22513719e-01 2.11114764e-01 1.94250375e-01
4.08307195e-01 -1.13572180e+00 2.21167877e-01 1.81627452e-01
1.06501472e+00 -1.32296240e+00 2.92053204e-02 -5.50698757e-01
-7.64031947e-01 1.14288652e+00 7.34378040e-01 -2.02190895e-02
6.31715000e-01 -4.08095941e-02 2.93817282e-01 -2.60623068e-01
-7.12947845e-01 -1.48928329e-01 5.00684738e-01 8.14489782e-01
5.93495548e-01 1.08738661e-01 -3.58395964e-01 5.30739844e-01
-1.55612484e-01 6.89845160e-02 1.90464780e-01 9.46062624e-01
-2.37835839e-01 -1.10021651e+00 -5.03106892e-01 4.14773911e-01
-5.22481024e-01 4.49132174e-01 -7.56635904e-01 8.20272267e-01
5.02392650e-01 7.92738914e-01 3.23747732e-02 -4.44287837e-01
2.29373962e-01 3.59419465e-01 8.78316343e-01 -7.39305854e-01
-5.48787653e-01 -6.14127889e-03 2.45216638e-01 -3.83674890e-01
-8.55358481e-01 -1.17065966e+00 -7.49961793e-01 -2.13942274e-01
-7.19142929e-02 -1.00209974e-01 3.69307309e-01 8.67716908e-01
6.62433580e-02 -1.39644086e-01 6.61113620e-01 -8.84216905e-01
-2.57278472e-01 -8.55827153e-01 -6.66180968e-01 7.62898922e-01
2.33082622e-01 -1.03476787e+00 -4.78501499e-01 3.13320398e-01] | [13.56940746307373, 1.6920770406723022] |
9cfd4256-0fa4-405c-99c8-cb5f5ece757a | viewsynth-learning-local-features-from-depth | 1911.10248 | null | https://arxiv.org/abs/1911.10248v4 | https://arxiv.org/pdf/1911.10248v4.pdf | ViewSynth: Learning Local Features from Depth using View Synthesis | The rapid development of inexpensive commodity depth sensors has made keypoint detection and matching in the depth image modality an important problem in computer vision. Despite great improvements in recent RGB local feature learning methods, adapting them directly in the depth modality leads to unsatisfactory performance. Most of these methods do not explicitly reason beyond the visible pixels in the images. To address the limitations of these methods, we propose a framework ViewSynth, to jointly learn: (1) viewpoint invariant keypoint-descriptor from depth images using a proposed Contrastive Matching Loss, and (2) view synthesis of depth images from different viewpoints using the proposed View Synthesis Module and View Synthesis Loss. By learning view synthesis, we explicitly encourage the feature extractor to encode information about not only the visible, but also the occluded parts of the scene. We demonstrate that in the depth modality, ViewSynth outperforms the state-of-the-art depth and RGB local feature extraction techniques in the 3D keypoint matching and camera localization tasks on the RGB-D datasets 7-Scenes, TUM RGBD and CoRBS in most scenarios. We also show the generalizability of ViewSynth in 3D keypoint matching across different datasets. | ['Jan-Michael Frahm', 'Kuan-Chuan Peng', 'Peri Akiva', 'Rajat Vikram Singh', 'Spondon Kundu', 'Jisan Mahmud'] | 2019-11-22 | null | null | null | null | ['camera-localization'] | ['computer-vision'] | [ 1.65956885e-01 -1.15910545e-01 -1.72633976e-01 -5.50563574e-01
-8.66516471e-01 -6.43952310e-01 7.15592265e-01 -1.96715862e-01
-3.15365791e-01 6.38596341e-02 6.07775524e-02 9.00724679e-02
-9.74160284e-02 -7.68712223e-01 -7.96982586e-01 -6.94816709e-01
4.81632531e-01 2.54684061e-01 2.91672111e-01 4.24973257e-02
4.40931708e-01 1.11466849e+00 -1.79439092e+00 1.40001342e-01
3.05222869e-01 1.38042724e+00 1.03257194e-01 6.26422763e-01
-7.33555481e-02 7.70738840e-01 -7.58448616e-02 -2.64139697e-02
8.22674453e-01 -2.27827922e-01 -4.74802792e-01 3.33347142e-01
1.31288123e+00 -7.39018083e-01 -8.55202675e-01 9.61074889e-01
6.72168851e-01 7.24143675e-03 3.58991444e-01 -1.40143788e+00
-3.51884454e-01 -2.82078177e-01 -6.03355527e-01 -7.20093176e-02
8.59530449e-01 -6.25743568e-02 1.04775572e+00 -1.26375747e+00
8.95880640e-01 1.27619803e+00 5.47383249e-01 4.70172077e-01
-9.26258326e-01 -4.83565778e-01 1.44718438e-01 1.88253373e-01
-1.23540068e+00 -4.85799164e-01 1.08363950e+00 -2.84358710e-01
1.15527725e+00 1.31657824e-01 7.78466403e-01 6.96630239e-01
2.88798869e-01 7.68227458e-01 1.21815491e+00 -5.16976714e-01
2.03466222e-01 -2.74922010e-02 -3.19293529e-01 8.47175837e-01
-1.18038997e-01 5.15423000e-01 -1.00080431e+00 -9.27851498e-02
1.12688398e+00 5.13291895e-01 -3.14473659e-01 -1.24075580e+00
-1.32411802e+00 7.47892737e-01 7.06487536e-01 -1.55836970e-01
-1.98722169e-01 3.48846763e-01 -4.44233827e-02 3.65752667e-01
4.33764189e-01 1.86183482e-01 -5.63583195e-01 1.49092395e-02
-6.25497699e-01 1.82164416e-01 4.92424488e-01 1.28376746e+00
1.29324758e+00 -4.62320536e-01 4.53070730e-01 2.44030774e-01
4.11573231e-01 8.63576233e-01 1.85654879e-01 -1.22743642e+00
6.95443094e-01 9.19994593e-01 -4.35566390e-03 -9.32244897e-01
-4.92081553e-01 1.13767222e-01 -4.56158966e-01 6.03956759e-01
2.31372684e-01 3.50085884e-01 -9.33623850e-01 1.28659821e+00
5.78536928e-01 -1.77465916e-01 -3.76028344e-02 9.33261991e-01
9.10369098e-01 2.81357944e-01 -6.90448523e-01 2.70935327e-01
9.76992667e-01 -7.47307003e-01 -3.39034408e-01 -3.76900852e-01
3.83500040e-01 -7.73654938e-01 9.05605733e-01 3.18882883e-01
-1.03250706e+00 -5.13333261e-01 -1.06347477e+00 -6.53685153e-01
-4.27333504e-01 -3.18552814e-02 8.60149264e-01 4.38227862e-01
-1.08563638e+00 3.14444631e-01 -9.12410617e-01 -5.72308600e-01
1.52308613e-01 4.51960474e-01 -8.46113205e-01 -6.22257888e-01
-7.10149884e-01 7.54454076e-01 -4.73483466e-02 -3.04961540e-02
-5.59287310e-01 -8.44015241e-01 -1.18752444e+00 -3.58706892e-01
2.87187397e-01 -9.55613136e-01 9.03954804e-01 -6.01850867e-01
-1.42850482e+00 1.29878545e+00 -2.77012497e-01 1.53948441e-01
6.96934938e-01 -4.51253802e-01 1.59760669e-01 4.77554321e-01
4.17659767e-02 7.46953249e-01 8.87317479e-01 -1.19855392e+00
-9.61514533e-01 -9.57658648e-01 2.50976712e-01 5.63255548e-01
2.80369490e-01 -4.82825726e-01 -7.07687497e-01 -1.79873884e-01
9.81154799e-01 -9.01342452e-01 -1.11973479e-01 8.21219027e-01
-2.48513862e-01 8.76341853e-03 7.93149710e-01 -1.05295539e-01
3.15750659e-01 -2.15430951e+00 3.53234798e-01 2.28366092e-01
2.26707771e-01 -4.19076502e-01 2.77950056e-02 3.78435105e-01
3.48065011e-02 -5.92058718e-01 1.62859410e-01 -6.06223166e-01
-4.08504391e-03 4.57392573e-01 -2.91271091e-01 1.05425632e+00
-1.53540075e-01 7.72340000e-01 -8.83036733e-01 -1.59767523e-01
9.40056682e-01 7.19161749e-01 -5.63267887e-01 4.38434422e-01
1.16574630e-01 4.26124692e-01 -4.30612952e-01 1.09397662e+00
8.97659123e-01 -1.13345407e-01 -4.99556720e-01 -4.94801342e-01
-2.41814166e-01 1.68572351e-01 -1.17136490e+00 2.49331689e+00
-5.29028356e-01 6.96205497e-01 2.56926175e-02 -5.75250506e-01
6.80638611e-01 -3.98494974e-02 8.32880378e-01 -9.98585165e-01
-3.89797352e-02 3.40918869e-01 -6.92056119e-01 -1.94294736e-01
3.40494454e-01 -2.17550229e-02 -4.33884934e-02 2.34391049e-01
1.66048914e-01 -7.23152995e-01 -5.71868122e-01 1.03363447e-01
1.18378031e+00 2.58126110e-01 3.82109523e-01 1.59723654e-01
3.35597605e-01 -2.40216121e-01 1.69778451e-01 7.06084967e-01
-1.30392730e-01 1.09352565e+00 7.74563476e-02 -5.17998576e-01
-1.02339876e+00 -1.26765895e+00 -1.96099624e-01 6.44641638e-01
5.08065462e-01 -3.04836661e-01 -1.88936010e-01 -6.36397302e-01
5.39834857e-01 -1.24031290e-01 -6.33181512e-01 -4.01980318e-02
-4.49698895e-01 9.36422125e-02 1.63185030e-01 5.46760798e-01
6.46963298e-01 -4.36774433e-01 -1.11950791e+00 -1.79588303e-01
3.96908894e-02 -1.41169906e+00 -3.99531871e-01 4.69430417e-01
-9.61968660e-01 -1.27294993e+00 -6.94482803e-01 -4.35538560e-01
7.32577503e-01 7.11423635e-01 9.70098734e-01 -2.72403777e-01
-5.49468577e-01 1.13054395e+00 -3.16647589e-01 6.72186632e-03
2.04651862e-01 -1.78607911e-01 2.45774668e-02 -2.00107262e-01
4.88402992e-01 -5.56591690e-01 -9.02415454e-01 4.00121510e-01
-6.98443413e-01 -1.71487257e-02 5.27774692e-01 6.78763449e-01
7.84673631e-01 -3.92147839e-01 -3.91506284e-01 -5.44134915e-01
-4.05713141e-01 9.75667089e-02 -8.83502066e-01 7.21397400e-02
-7.10487485e-01 1.23355851e-01 -2.06995048e-02 -5.35551086e-02
-5.98455071e-01 7.07539737e-01 -1.73359722e-01 -7.70194352e-01
-9.86955464e-02 -7.79048279e-02 -4.57670480e-01 -7.44973302e-01
4.00071144e-01 2.30443090e-01 -1.63355116e-02 -3.65108520e-01
6.14465356e-01 3.76653790e-01 4.90359962e-01 -3.03127021e-01
9.93198037e-01 1.07428181e+00 4.33375716e-01 -7.95243025e-01
-6.58201277e-01 -1.05288708e+00 -1.03305244e+00 -7.40316883e-02
8.58785510e-01 -1.29308116e+00 -7.30736136e-01 6.14888012e-01
-1.04088485e+00 -7.33562931e-02 -4.05677438e-01 4.96017784e-01
-9.54778910e-01 3.00531000e-01 -2.88525641e-01 -4.81773823e-01
-1.21722840e-01 -1.23395109e+00 1.76556695e+00 1.41233012e-01
2.51165360e-01 -9.14560854e-01 3.99313197e-02 3.88617069e-02
1.27768561e-01 4.65814084e-01 6.12249911e-01 1.11864910e-01
-1.21301651e+00 -4.49331671e-01 -3.86301458e-01 2.50562608e-01
3.56602550e-01 -3.18048209e-01 -1.32925367e+00 -2.48858109e-01
9.09204315e-03 -3.60745281e-01 8.81571352e-01 3.21262807e-01
7.02698112e-01 1.72858447e-01 -3.03884983e-01 1.41392934e+00
1.76197541e+00 8.96378532e-02 5.53509474e-01 6.34222269e-01
1.06153715e+00 6.49154782e-01 5.63157916e-01 5.16856909e-01
5.66132188e-01 8.70099068e-01 7.41696537e-01 -3.25650901e-01
-2.18698025e-01 -3.84916425e-01 2.14112774e-01 3.66890639e-01
7.49073699e-02 2.03045577e-01 -7.64077246e-01 2.83148080e-01
-1.72696447e+00 -5.69777071e-01 1.97778225e-01 2.21488190e+00
4.81354743e-01 -1.90270036e-01 -2.02082440e-01 2.10595921e-01
8.44412670e-02 2.63638884e-01 -7.69825101e-01 5.59312813e-02
-4.60288376e-01 2.77621895e-01 9.23093438e-01 5.67395210e-01
-1.01925910e+00 7.28721976e-01 5.84802771e+00 1.24229126e-01
-1.08808792e+00 -5.69773987e-02 8.30364525e-02 -1.13841601e-01
-3.98640305e-01 1.97712660e-01 -8.50577295e-01 -1.11004181e-01
1.04009815e-01 2.62902319e-01 1.87048465e-01 1.03993964e+00
-3.62774223e-01 -3.17869365e-01 -1.68976700e+00 1.72991884e+00
3.88591141e-01 -1.11496019e+00 3.26994471e-02 2.80537069e-01
7.22313583e-01 3.07128340e-01 9.79197770e-02 -2.27368340e-01
-2.61022300e-01 -7.09905386e-01 7.92802930e-01 4.78053927e-01
8.28788519e-01 -6.57394350e-01 4.49730039e-01 9.58484635e-02
-1.23261046e+00 -1.01551130e-01 -5.96934974e-01 7.06982613e-02
-5.84575199e-02 4.56706971e-01 -3.70709985e-01 5.48128963e-01
8.52333605e-01 1.13819468e+00 -6.62853777e-01 8.97851825e-01
-2.91641325e-01 -3.97409886e-01 -4.94355887e-01 5.02519846e-01
1.73605159e-01 -7.86223859e-02 3.03954184e-01 6.66516244e-01
3.58738512e-01 1.13917574e-01 2.65589859e-02 7.37353265e-01
9.23298374e-02 -3.50597680e-01 -9.58757520e-01 3.78362596e-01
2.80901313e-01 1.01238644e+00 -4.47627455e-01 8.66998211e-02
-7.58147597e-01 1.41570294e+00 8.76699612e-02 4.19205308e-01
-3.15009892e-01 -4.36203778e-01 9.55433130e-01 3.09097975e-01
3.43794912e-01 -4.21245575e-01 -1.67367950e-01 -1.37630510e+00
1.83119908e-01 -5.59718132e-01 1.75725326e-01 -1.01492321e+00
-1.16127551e+00 3.10059488e-01 4.38270159e-02 -1.53730774e+00
-3.94010663e-01 -1.08184099e+00 2.40993090e-02 9.62929130e-01
-1.96612036e+00 -1.37648094e+00 -8.63438308e-01 1.15462387e+00
3.96133423e-01 4.59005348e-02 6.81310773e-01 2.10289761e-01
2.65754193e-01 5.04709065e-01 5.63791506e-02 -1.88158657e-02
8.84428263e-01 -1.24221671e+00 5.27341425e-01 6.59067690e-01
2.34664381e-01 5.45096695e-01 4.02095288e-01 -1.45719841e-01
-2.16542935e+00 -5.69483578e-01 5.04201174e-01 -8.67164969e-01
2.00239733e-01 -7.28979349e-01 -3.40102255e-01 6.50357485e-01
-1.92971572e-01 5.58453381e-01 1.96917474e-01 -2.68768370e-01
-6.34974182e-01 -3.93306285e-01 -1.18319547e+00 1.74134105e-01
1.21566665e+00 -1.10100043e+00 -3.71750504e-01 2.21278608e-01
6.04423165e-01 -9.19625998e-01 -7.86717355e-01 3.99130523e-01
8.69979143e-01 -1.35202384e+00 1.44644094e+00 -2.11980846e-02
2.14822814e-01 -1.86106205e-01 -7.97982097e-01 -9.31398034e-01
2.48575415e-02 -3.53941530e-01 -2.34642416e-01 8.61502469e-01
-1.94365367e-01 -6.35314703e-01 1.09915876e+00 7.77798891e-01
-8.58759210e-02 -4.51387346e-01 -1.36287200e+00 -4.56859499e-01
-2.55833179e-01 -4.15995419e-01 4.25903708e-01 6.97867393e-01
-3.21889043e-01 6.81765527e-02 -2.44500026e-01 3.21732193e-01
9.89178598e-01 3.74746710e-01 1.24266076e+00 -1.19136679e+00
-1.96298629e-01 -2.10206240e-01 -1.07771075e+00 -1.55317724e+00
2.61109117e-02 -4.58486825e-01 -4.08865772e-02 -1.62527263e+00
-9.68490094e-02 -4.60443437e-01 -3.43987532e-02 3.09607327e-01
1.29990920e-01 2.89575011e-01 6.58892542e-02 2.95588315e-01
-3.66093904e-01 4.32705253e-01 1.24930358e+00 -2.15181406e-03
3.24179381e-02 7.00519886e-03 -2.06888273e-01 6.45829022e-01
5.38339280e-02 -2.73888767e-01 -4.75792199e-01 -5.17260373e-01
5.53572595e-01 8.74691382e-02 6.22533977e-01 -9.64150250e-01
5.08292139e-01 -1.07351184e-01 8.15718770e-01 -1.02724087e+00
8.31680834e-01 -1.25664830e+00 3.04227695e-02 2.44747177e-01
4.28292751e-02 3.38865966e-01 -6.53710067e-02 5.78460038e-01
-2.93018967e-01 3.29379082e-01 5.95379055e-01 -4.24834311e-01
-1.14459908e+00 6.60591781e-01 3.25928807e-01 -1.76372156e-01
8.99108887e-01 -7.62666345e-01 -3.28462034e-01 -3.71952921e-01
-3.18838060e-01 -1.24251172e-01 1.09109008e+00 4.55454201e-01
1.10608852e+00 -1.44067550e+00 -1.73059314e-01 8.13217342e-01
7.09030926e-01 4.16582853e-01 2.41635248e-01 7.44329333e-01
-7.85174191e-01 5.36233187e-01 -1.93243891e-01 -1.09937263e+00
-1.11050928e+00 5.06113887e-01 5.75458467e-01 3.66089255e-01
-9.98568594e-01 9.59242165e-01 5.27035952e-01 -5.05983889e-01
5.02829731e-01 -7.30713785e-01 3.91493440e-01 -5.04661677e-03
4.29924637e-01 2.09040821e-01 3.97364467e-01 -7.70730317e-01
-6.59801364e-01 1.34848237e+00 1.71855897e-01 -2.00158626e-01
1.14171064e+00 -5.83757162e-01 -7.66499490e-02 5.13026357e-01
1.63096654e+00 1.01598747e-01 -1.59330261e+00 -5.00825226e-01
-3.12805772e-01 -1.05334413e+00 2.61256754e-01 -4.11422908e-01
-1.03575361e+00 1.03841352e+00 9.42867339e-01 -3.25664997e-01
1.29640961e+00 2.75066882e-01 5.60080230e-01 4.56456691e-01
8.38532984e-01 -7.25676715e-01 1.77471906e-01 4.96886581e-01
6.72558844e-01 -1.62605774e+00 3.62524599e-01 -1.90818056e-01
-1.64308861e-01 1.24798226e+00 3.89719099e-01 -2.50357479e-01
7.04197049e-01 9.47738960e-02 3.47856909e-01 -4.37755972e-01
-3.41270596e-01 -2.30038419e-01 4.49131966e-01 8.10822904e-01
7.57550150e-02 -3.72945607e-01 5.93220830e-01 -2.55893081e-01
1.68060977e-02 -4.03490752e-01 2.49118984e-01 1.27504909e+00
-2.10981429e-01 -1.06284547e+00 -3.46420199e-01 -9.63132232e-02
-6.86407238e-02 -1.64622124e-02 -5.73828101e-01 1.11411130e+00
6.19450137e-02 6.70581281e-01 7.68215135e-02 -4.04936522e-01
6.38597906e-01 -2.88572997e-01 1.10033476e+00 -6.12027705e-01
-2.27383211e-01 2.80889049e-02 -5.80450058e-01 -1.33513987e+00
-8.36995304e-01 -6.88417673e-01 -9.62043762e-01 -2.14591116e-01
-3.31794322e-01 -3.63513440e-01 9.45936859e-01 7.26777077e-01
2.22629681e-01 1.53196277e-02 7.63437092e-01 -1.32278562e+00
-3.76965463e-01 -4.30894911e-01 -8.22292030e-01 3.59096020e-01
9.65500712e-01 -7.64815509e-01 -6.33712232e-01 -2.36137196e-01] | [7.948269844055176, -2.7145605087280273] |
67e6b046-e89e-4a58-b014-cce51bb627b2 | evaluation-of-a-region-proposal-architecture | 2106.11797 | null | https://arxiv.org/abs/2106.11797v1 | https://arxiv.org/pdf/2106.11797v1.pdf | Evaluation of a Region Proposal Architecture for Multi-task Document Layout Analysis | Automatically recognizing the layout of handwritten documents is an important step towards useful extraction of information from those documents. The most common application is to feed downstream applications such as automatic text recognition and keyword spotting; however, the recognition of the layout also helps to establish relationships between elements in the document which allows to enrich the information that can be extracted. Most of the modern document layout analysis systems are designed to address only one part of the document layout problem, namely: baseline detection or region segmentation. In contrast, we evaluate the effectiveness of the Mask-RCNN architecture to address the problem of baseline detection and region segmentation in an integrated manner. We present experimental results on two handwritten text datasets and one handwritten music dataset. The analyzed architecture yields promising results, outperforming state-of-the-art techniques in all three datasets. | ['Enrique Vidal', 'Lorenzo Quirós'] | 2021-06-22 | null | null | null | null | ['document-layout-analysis'] | ['computer-vision'] | [ 6.57373846e-01 -3.90516490e-01 -1.48661062e-01 -1.93649247e-01
-4.27053392e-01 -8.71046484e-01 6.80947244e-01 5.85911274e-01
-2.59473264e-01 3.64584953e-01 1.89467873e-02 -4.13608670e-01
-3.88070941e-01 -5.06679475e-01 -2.80872285e-01 -5.95786393e-01
2.40013435e-01 3.85400087e-01 2.82236248e-01 -4.09339443e-02
9.94858444e-01 1.23811471e+00 -1.47789836e+00 6.43796206e-01
4.85479265e-01 1.10778558e+00 2.96095788e-01 8.43747139e-01
-7.10824132e-01 7.37691760e-01 -8.79661500e-01 6.46348894e-02
1.18949547e-01 -4.08579379e-01 -7.43516386e-01 4.57844615e-01
4.44985360e-01 -2.66272008e-01 -1.66626170e-01 8.94073009e-01
2.75493175e-01 1.19964041e-01 7.08414853e-01 -5.79262435e-01
-2.73474842e-01 9.48545516e-01 -7.52967834e-01 1.09567389e-01
9.98416245e-02 -4.05285329e-01 8.24595928e-01 -9.62045789e-01
6.29697084e-01 8.60754073e-01 3.61173511e-01 5.56695834e-02
-9.70168769e-01 -1.17258005e-01 2.08444536e-01 2.54130904e-02
-1.15815544e+00 -3.45961094e-01 1.16556978e+00 -3.85099858e-01
8.56419563e-01 4.55287814e-01 1.36436880e-01 6.67907417e-01
-1.02556810e-01 1.11667359e+00 8.13746095e-01 -9.56362426e-01
2.31041178e-01 5.76579161e-02 7.16892183e-01 4.99158651e-01
4.41541262e-02 -5.48705041e-01 -4.97333825e-01 3.14218909e-01
7.14914560e-01 -1.92820013e-01 -2.08263963e-01 -1.22965224e-01
-1.08226919e+00 3.05874705e-01 1.26226351e-01 8.25400174e-01
-3.48295718e-01 -8.68545100e-02 4.65568453e-01 -4.30726334e-02
8.20799395e-02 8.58037233e-01 -1.11742280e-01 -2.24626377e-01
-1.48135030e+00 -1.78275560e-03 7.61442840e-01 8.46502244e-01
4.38765854e-01 7.83391520e-02 -4.54072148e-01 9.36003149e-01
1.38115108e-01 6.31621927e-02 3.33097279e-01 -4.20063823e-01
6.31536424e-01 1.05055404e+00 -5.42632453e-02 -1.10153210e+00
-3.41251761e-01 -5.54831684e-01 -7.38565862e-01 1.41309738e-01
6.29765332e-01 1.76002383e-01 -1.08165455e+00 6.82949066e-01
-1.36333078e-01 -5.98526657e-01 -2.42326811e-01 6.43284082e-01
6.96870267e-01 7.08900571e-01 -5.68679512e-01 6.24607466e-02
1.34694684e+00 -1.21381295e+00 -8.34925234e-01 -2.87019879e-01
3.53155732e-01 -1.12676060e+00 8.84243011e-01 8.28427255e-01
-7.79611707e-01 -6.75176740e-01 -1.32194722e+00 -2.43519202e-01
-4.81064945e-01 9.39319193e-01 2.66911328e-01 5.50360084e-01
-6.10018909e-01 5.15503228e-01 -4.55796927e-01 -2.03160703e-01
3.76322299e-01 4.44600731e-01 -1.90199748e-01 -4.99202721e-02
-3.44581693e-01 4.54259902e-01 4.98288542e-01 5.97457111e-01
-4.21084195e-01 -2.98171639e-01 -3.43933940e-01 3.68276507e-01
5.99350452e-01 2.60077506e-01 9.79641318e-01 -9.07044113e-01
-1.36464155e+00 5.60797751e-01 -3.46131884e-02 -3.04265976e-01
5.89422166e-01 -4.13479298e-01 -3.49594593e-01 2.07654491e-01
-3.36098850e-01 3.93470258e-01 8.87234807e-01 -1.18045115e+00
-7.73103476e-01 -3.23696196e-01 -2.60818928e-01 -6.19477732e-03
-4.00864959e-01 1.11352004e-01 -8.83762777e-01 -8.40237558e-01
3.01263303e-01 -4.62716609e-01 8.55458081e-02 -1.82201117e-01
-8.03716242e-01 1.60338551e-01 1.46107244e+00 -9.61548567e-01
1.42576993e+00 -2.17930269e+00 -7.90126920e-02 6.90009475e-01
-1.84851110e-01 4.45679218e-01 -1.34399325e-01 6.34334803e-01
1.97601430e-02 2.12021880e-02 -2.13499859e-01 -1.04318298e-01
-9.51206684e-02 -2.18774855e-01 -6.01003706e-01 2.63387293e-01
2.57657409e-01 7.56667972e-01 -4.77052152e-01 -3.85995090e-01
2.92776793e-01 2.40076125e-01 -1.05986409e-01 7.72866830e-02
-4.34122980e-01 2.74707139e-01 -3.59220445e-01 7.61366725e-01
5.13217270e-01 2.36139074e-02 3.24051648e-01 -3.92499000e-01
-4.43809450e-01 3.41455266e-02 -1.33760798e+00 1.53630567e+00
-1.37865782e-01 1.06565201e+00 -1.47585630e-01 -9.66544807e-01
1.37909102e+00 -9.75505114e-02 3.82829577e-01 -6.55924737e-01
2.89205581e-01 1.38717562e-01 1.08905315e-01 -2.57659107e-01
1.05431247e+00 6.66933775e-01 -1.74601134e-02 6.98958874e-01
-1.93875611e-01 1.16947286e-01 5.47857344e-01 -3.93804694e-05
9.15870190e-01 2.33475670e-01 2.79081315e-01 -3.25413257e-01
7.38255441e-01 -7.76134431e-02 -1.12772889e-01 9.62340415e-01
1.66847393e-01 6.91089094e-01 8.40956569e-01 -3.39725971e-01
-1.04992294e+00 -5.83072245e-01 6.64402544e-02 9.56474900e-01
1.43076992e-03 -2.42281884e-01 -7.57724047e-01 -6.93210900e-01
-2.25112274e-01 5.19661307e-01 -6.29999816e-01 3.98372352e-01
-8.07193696e-01 -3.61258835e-01 7.97620416e-01 7.32413948e-01
6.41425431e-01 -1.06822145e+00 -8.96494806e-01 2.38081932e-01
7.45619237e-02 -1.10683572e+00 -3.38677138e-01 6.13745689e-01
-9.06299174e-01 -1.01633275e+00 -8.89247119e-01 -9.70424831e-01
9.07676578e-01 7.83088282e-02 5.96329689e-01 9.34739262e-02
-6.18868887e-01 2.60174006e-01 -4.87164527e-01 -2.94930726e-01
-3.41648251e-01 4.27091509e-01 -5.10918319e-01 1.53343499e-01
4.29086611e-02 -1.08878195e-01 -2.50890166e-01 2.79861391e-01
-1.07389235e+00 -5.08199520e-02 8.76031399e-01 7.44598627e-01
5.16290784e-01 4.54746366e-01 2.52939105e-01 -8.88683736e-01
9.11979616e-01 3.35354298e-01 -8.26068461e-01 6.70346141e-01
-4.04078603e-01 2.31510922e-01 6.65013492e-01 -3.99733752e-01
-1.18974125e+00 2.95418978e-01 -4.97357324e-02 5.18418625e-02
-5.30758083e-01 7.37012744e-01 -3.14801574e-01 1.09684721e-01
3.72560143e-01 3.67903680e-01 -3.56059790e-01 -8.55502307e-01
3.01256657e-01 9.13908064e-01 9.08294380e-01 -6.34158611e-01
4.21431869e-01 3.75899047e-01 2.02045277e-01 -1.41404259e+00
-3.47624272e-01 -7.81814218e-01 -1.22643650e+00 -2.94206709e-01
4.41200525e-01 1.66705381e-02 -4.77663815e-01 5.76989353e-01
-1.21848989e+00 -2.77690589e-01 -4.08007503e-01 -2.72463891e-03
-1.91135958e-01 7.19231844e-01 -4.02034461e-01 -1.04484069e+00
-2.64298528e-01 -1.13008916e+00 1.16994488e+00 2.57390678e-01
-2.02366441e-01 -6.65824413e-01 -1.84705555e-01 6.94620311e-02
8.15695524e-02 9.46990699e-02 1.37984383e+00 -8.51162136e-01
-6.38132811e-01 -6.41358256e-01 -4.53050494e-01 3.03716063e-01
3.29484940e-01 4.68656152e-01 -1.20480490e+00 5.46620600e-02
-3.43354285e-01 1.21133588e-01 9.66929734e-01 1.84639290e-01
1.57511675e+00 -5.84741235e-02 -3.00135642e-01 3.09776008e-01
1.28210795e+00 5.72787166e-01 8.67016613e-01 4.35645014e-01
5.79081953e-01 8.19205821e-01 6.05052650e-01 5.79027176e-01
-5.30354559e-01 6.66710198e-01 1.69627488e-01 -1.37395024e-01
-2.11952850e-01 -7.69375116e-02 -5.54978498e-04 7.17874587e-01
1.02684721e-01 -5.13181508e-01 -1.15177798e+00 4.28035975e-01
-1.77161956e+00 -7.51359761e-01 -2.76235670e-01 1.90012217e+00
4.18478489e-01 3.09386104e-01 5.30183017e-02 8.61373365e-01
8.46949041e-01 1.52654812e-01 -2.33157471e-01 -6.43970668e-01
-2.06832021e-01 2.23158419e-01 3.43819976e-01 1.55670732e-01
-1.26788592e+00 8.78366053e-01 6.20908737e+00 7.91688085e-01
-1.32343614e+00 -6.81422591e-01 6.97547615e-01 2.77398586e-01
2.20905274e-01 -2.50676513e-01 -9.15275335e-01 6.30757771e-03
3.38098645e-01 3.87782812e-01 4.11210477e-01 7.34792054e-01
-1.95938405e-02 -4.85192388e-01 -1.42404974e+00 9.18802798e-01
1.54792860e-01 -1.41937006e+00 1.53553054e-01 1.60878047e-01
7.30057538e-01 -6.68458104e-01 8.23499933e-02 -9.88754705e-02
-5.01313508e-01 -1.05390322e+00 8.38482499e-01 5.60508966e-01
7.42126822e-01 -9.51084614e-01 8.15433681e-01 2.66551137e-01
-1.23113692e+00 -3.29979956e-01 -1.46105960e-01 1.62584245e-01
-1.15315706e-01 5.74513614e-01 -1.10021436e+00 4.93716538e-01
3.02836061e-01 4.01068121e-01 -8.45701575e-01 1.29692209e+00
-3.11098874e-01 3.50695372e-01 -8.76601636e-02 -3.69899213e-01
5.11521339e-01 -1.25430450e-01 2.87004709e-01 1.57101405e+00
2.54087031e-01 -3.02803814e-01 -1.46617979e-01 9.14866209e-01
-1.46242514e-01 3.90466660e-01 -4.58286911e-01 -5.18603086e-01
3.56239408e-01 1.44310045e+00 -1.69485497e+00 -8.18399563e-02
-9.65333730e-02 1.00009084e+00 6.59113303e-02 3.03593129e-01
-3.86930227e-01 -9.71721053e-01 -2.34463587e-01 1.35115860e-02
7.62283802e-01 -4.60470557e-01 -6.78162634e-01 -4.83932912e-01
1.67086571e-01 -1.14987516e+00 8.19858760e-02 -6.10255539e-01
-7.65231133e-01 3.97440195e-01 -3.46989274e-01 -9.96219456e-01
-4.37618084e-02 -1.08690786e+00 -5.74774265e-01 6.91458583e-01
-1.05079162e+00 -1.21482396e+00 -3.91861707e-01 6.40569925e-02
7.87812650e-01 -2.58386254e-01 7.02607870e-01 1.60837799e-01
-6.41719162e-01 5.41547894e-01 4.66737986e-01 5.93202233e-01
5.70802748e-01 -1.22029877e+00 3.57424170e-01 1.17554080e+00
5.64332724e-01 7.12012112e-01 2.94849426e-01 -6.88925624e-01
-1.47032464e+00 -8.90654266e-01 6.52788103e-01 -1.38063475e-01
4.25560057e-01 -5.65924525e-01 -7.50421226e-01 3.16750795e-01
2.11644456e-01 -5.29489458e-01 4.33688045e-01 -6.85385317e-02
-1.49410278e-01 -1.92510188e-01 -7.13335156e-01 5.80462515e-01
5.85080743e-01 -5.38736463e-01 -5.06332576e-01 -1.27737746e-01
2.30212472e-02 -3.76432836e-01 -4.20514643e-01 1.88068464e-01
8.60147834e-01 -7.58948207e-01 7.24907219e-01 -4.32410151e-01
6.37244582e-01 -4.37562674e-01 -2.23270386e-01 -7.69512594e-01
-1.46680966e-01 -5.40341675e-01 -2.30412439e-01 1.55602276e+00
4.69925106e-01 1.00543508e-02 9.12200034e-01 3.38177025e-01
-9.86708999e-02 -6.30092323e-01 -5.69864869e-01 -6.77579522e-01
-4.15725172e-01 -3.60810339e-01 5.70027947e-01 6.35168135e-01
-9.55287442e-02 3.42929691e-01 -1.29472315e-01 -1.71356052e-02
2.72480845e-01 3.76966327e-01 8.71519983e-01 -1.21779716e+00
-2.72058785e-01 -9.90799665e-01 -3.22047561e-01 -1.30777705e+00
7.57403150e-02 -6.66190684e-01 3.15056562e-01 -1.70514166e+00
-8.76549166e-03 -2.28903428e-01 -1.32739291e-01 2.92925537e-01
3.05461705e-01 1.11904807e-01 1.72291964e-01 1.04192108e-01
-4.54261422e-01 -2.58308053e-02 8.27632248e-01 -4.97122020e-01
-4.63202566e-01 1.97900413e-03 -5.02298057e-01 5.63711405e-01
5.64277709e-01 -4.68443613e-03 -3.31320494e-01 -3.04638445e-01
1.05852187e-01 -2.72009224e-01 1.61666758e-02 -9.68592227e-01
4.79329586e-01 1.36102453e-01 7.94768989e-01 -1.41030312e+00
-2.03365162e-02 -9.11220908e-01 -4.42214191e-01 1.89287111e-01
-5.44668794e-01 1.01584636e-01 4.74938273e-01 2.40750298e-01
-2.88699985e-01 -7.75314033e-01 5.29294968e-01 5.19571528e-02
-9.91095841e-01 -4.21588093e-01 -6.95035100e-01 -3.76757920e-01
7.59028673e-01 -5.65874934e-01 -6.45584464e-01 -2.16596335e-01
-4.92965072e-01 -2.20208511e-01 2.20444188e-01 3.94687951e-01
8.14715922e-01 -8.37087691e-01 -2.68036276e-01 3.66490990e-01
1.04309782e-01 -1.05191484e-01 -6.10080436e-02 6.24733925e-01
-9.96325254e-01 9.09896553e-01 -3.38253886e-01 -5.49983680e-01
-1.45932031e+00 5.43028593e-01 5.01665026e-02 -4.21167880e-01
-4.86783743e-01 6.91557109e-01 -6.49191439e-02 1.55080333e-01
7.18282580e-01 -6.13600016e-01 -4.50792193e-01 2.84693658e-01
6.58003747e-01 7.46677041e-01 4.17732507e-01 -2.86085844e-01
-2.35821635e-01 5.84195673e-01 -2.63867557e-01 -1.59898922e-01
1.26334274e+00 7.04581439e-02 -2.81814367e-01 4.40310448e-01
7.55730331e-01 2.48596460e-01 -9.74472165e-01 7.85756409e-02
7.16378510e-01 -3.12222391e-01 8.70582983e-02 -1.03828919e+00
-7.42020726e-01 1.11421704e+00 4.95916635e-01 2.85263896e-01
1.24028409e+00 -4.25600797e-01 1.91779524e-01 9.88137960e-01
1.55254394e-01 -1.42962706e+00 3.01212937e-01 5.69111049e-01
1.00183797e+00 -7.57842660e-01 2.87667066e-01 -2.82719195e-01
-2.90381283e-01 1.83284461e+00 4.66172874e-01 1.21075571e-01
3.88936847e-01 6.38432503e-01 1.18373372e-02 -1.08187422e-01
-4.98039238e-02 7.28563666e-02 8.96589637e-01 3.19838941e-01
7.33690023e-01 -3.36737365e-01 -1.10716850e-01 5.13027251e-01
1.40382230e-01 -3.09014976e-01 5.74671447e-01 1.27346528e+00
-5.01464367e-01 -1.32081449e+00 -6.44111991e-01 5.17445803e-01
-5.26262641e-01 -3.30836624e-02 -1.19098747e+00 8.48517954e-01
-1.00982092e-01 6.75414145e-01 1.79761186e-01 -1.24031685e-01
2.52073407e-01 1.64174408e-01 6.59285486e-01 -5.26582897e-01
-7.40416765e-01 5.37075102e-01 6.82813227e-02 -1.71041980e-01
-2.29461014e-01 -4.96902794e-01 -1.08208108e+00 2.63591230e-01
-5.96827090e-01 -4.62367460e-02 9.65067625e-01 8.26852202e-01
-2.53588744e-02 9.24232185e-01 4.06452835e-01 -7.13832259e-01
-3.04971337e-01 -9.25506413e-01 -6.86352670e-01 1.47397933e-03
2.55468458e-01 -1.02236263e-01 3.85936797e-01 2.80699074e-01] | [11.771750450134277, 2.6640100479125977] |
dc836120-b862-4a2b-aeb2-e817f062c44d | contrastive-deep-graph-clustering-with | 2212.03559 | null | https://arxiv.org/abs/2212.03559v1 | https://arxiv.org/pdf/2212.03559v1.pdf | Contrastive Deep Graph Clustering with Learnable Augmentation | Graph contrastive learning is an important method for deep graph clustering. The existing methods first generate the graph views with stochastic augmentations and then train the network with a cross-view consistency principle. Although good performance has been achieved, we observe that the existing augmentation methods are usually random and rely on pre-defined augmentations, which is insufficient and lacks negotiation between the final clustering task. To solve the problem, we propose a novel Graph Contrastive Clustering method with the Learnable graph Data Augmentation (GCC-LDA), which is optimized completely by the neural networks. An adversarial learning mechanism is designed to keep cross-view consistency in the latent space while ensuring the diversity of augmented views. In our framework, a structure augmentor and an attribute augmentor are constructed for augmentation learning in both structure level and attribute level. To improve the reliability of the learned affinity matrix, clustering is introduced to the learning procedure and the learned affinity matrix is refined with both the high-confidence pseudo-label matrix and the cross-view sample similarity matrix. During the training procedure, to provide persistent optimization for the learned view, we design a two-stage training strategy to obtain more reliable clustering information. Extensive experimental results demonstrate the effectiveness of GCC-LDA on six benchmark datasets. | ['En Zhu', 'Xinwang Liu', 'Siwei Wang', 'Sihang Zhou', 'Yue Liu', 'Xihong Yang'] | 2022-12-07 | null | null | null | null | ['graph-clustering'] | ['graphs'] | [-6.74758479e-02 1.20079301e-01 -2.16179624e-01 -3.08474988e-01
-6.56059265e-01 -5.70465446e-01 5.41968346e-01 -2.31367245e-01
1.05431668e-01 3.05826634e-01 1.31233707e-01 1.17486775e-01
-9.24167037e-02 -7.83020377e-01 -6.28311753e-01 -1.12850344e+00
1.54217646e-01 6.36105955e-01 -7.81323537e-02 1.08690802e-02
-7.89275095e-02 9.49889794e-02 -1.04509115e+00 8.43482837e-02
1.05381989e+00 7.01680422e-01 -2.26780102e-02 2.77216583e-01
-8.21905732e-02 7.42192924e-01 -2.80978322e-01 -4.40954864e-01
4.59312141e-01 -5.97854078e-01 -6.30764186e-01 5.80347717e-01
3.21127921e-01 5.69131710e-02 -3.85452986e-01 1.31511366e+00
4.42325830e-01 1.27509087e-01 6.24639630e-01 -1.60992455e+00
-8.57188761e-01 6.18205845e-01 -9.00151551e-01 -2.56056726e-01
2.10313685e-02 2.19513074e-01 1.07043064e+00 -1.01123333e+00
6.39389992e-01 1.08762693e+00 5.12181818e-01 6.63271487e-01
-1.36022782e+00 -7.35727012e-01 5.04751325e-01 1.79029167e-01
-1.45637965e+00 -8.22049230e-02 1.39775121e+00 -4.88605559e-01
1.78420871e-01 9.04003158e-02 8.09092581e-01 1.13630009e+00
-9.57660601e-02 6.08084023e-01 1.21030843e+00 -9.87132266e-02
2.44515061e-01 1.48279190e-01 -7.37820715e-02 1.00478399e+00
1.37156039e-01 -1.01573139e-01 -3.06589097e-01 -2.41739884e-01
7.05233932e-01 2.78937817e-01 -3.34702015e-01 -1.08711386e+00
-1.09821069e+00 9.11767006e-01 8.37016761e-01 7.59802908e-02
-2.75919825e-01 -6.93136826e-02 4.25805032e-01 1.28018096e-01
5.32490790e-01 1.97204232e-01 -1.57253906e-01 5.16187489e-01
-6.26043916e-01 -2.32333109e-01 4.59347427e-01 1.04348159e+00
9.12836254e-01 2.08670497e-01 -5.28715216e-02 7.64626145e-01
4.21584398e-01 3.26832116e-01 4.89596605e-01 -7.43216991e-01
6.69270098e-01 1.09591413e+00 -4.37075883e-01 -1.30287671e+00
-1.74331963e-01 -8.44825447e-01 -1.48997259e+00 2.30269894e-01
1.07026614e-01 -1.46441096e-02 -1.10283840e+00 1.86884606e+00
7.18920052e-01 4.27109838e-01 7.34011605e-02 9.03404355e-01
5.80669880e-01 5.08261800e-01 -7.48965219e-02 -2.79838115e-01
9.77489293e-01 -1.14660239e+00 -7.82240391e-01 -1.36480421e-01
5.13062775e-01 -4.21627164e-01 1.22190714e+00 2.75798798e-01
-9.12070632e-01 -6.84125364e-01 -1.37454200e+00 1.44551277e-01
-2.66421795e-01 4.15257514e-02 5.52804589e-01 3.99841040e-01
-7.87763000e-01 3.69598448e-01 -8.80577803e-01 -2.74789874e-02
3.67558122e-01 3.24187756e-01 -5.27038276e-01 -1.74811020e-01
-1.00855386e+00 2.54200578e-01 5.48869252e-01 1.03180692e-01
-7.97432899e-01 -4.16186273e-01 -8.74006629e-01 1.30348094e-02
5.93776822e-01 -8.22076142e-01 2.10241020e-01 -9.53006268e-01
-1.31930017e+00 7.94544756e-01 3.35436404e-01 -2.71711666e-02
5.16318083e-01 1.92581967e-01 -3.38432491e-01 1.84656247e-01
2.73313195e-01 4.49704736e-01 1.17453706e+00 -1.87400603e+00
-2.49324396e-01 -5.90251386e-01 -8.61649588e-02 5.06919026e-01
-4.13154393e-01 -6.87238038e-01 -8.54700148e-01 -9.17634606e-01
6.34857833e-01 -1.21115208e+00 -3.48658919e-01 -3.04415077e-01
-6.10996783e-01 9.14751664e-02 8.74868214e-01 -6.47320807e-01
1.08278394e+00 -2.22381639e+00 6.30104661e-01 5.93300343e-01
5.35233498e-01 6.85733184e-02 -1.35820985e-01 2.50474304e-01
-3.63433808e-01 6.71757460e-02 -5.86588621e-01 -4.95439559e-01
-9.00444388e-02 1.91791356e-01 -7.63196945e-02 6.22950137e-01
5.29738218e-02 7.54081249e-01 -9.18961465e-01 -4.85396415e-01
2.13807479e-01 5.43984056e-01 -6.26557291e-01 5.42907000e-01
-9.20725241e-02 7.98645854e-01 -4.60728407e-01 3.71401966e-01
8.37143898e-01 -7.03413427e-01 4.68304217e-01 -5.25796056e-01
4.90017176e-01 -1.22835577e-01 -1.33889163e+00 1.89168048e+00
-2.33056352e-01 -2.92133570e-01 1.20895967e-01 -1.09715390e+00
8.93581986e-01 1.40630201e-01 5.96238971e-01 -2.26041734e-01
8.78176913e-02 -7.02222660e-02 8.15435648e-02 -2.46186882e-01
6.31653666e-02 -6.40945584e-02 5.30037805e-02 5.61023355e-01
3.86452451e-02 4.70262058e-02 -1.78705543e-01 5.99714518e-01
8.41284037e-01 1.01413816e-01 -3.76901068e-02 -1.94390669e-01
8.47432911e-01 -2.47944996e-01 6.67581856e-01 1.10948935e-01
3.05887535e-02 7.94652820e-01 5.20597041e-01 -2.38831684e-01
-1.00936782e+00 -8.98547173e-01 3.33042324e-01 7.73520589e-01
3.28970194e-01 -4.83241618e-01 -9.65192258e-01 -1.29711783e+00
-3.31693411e-01 2.63010532e-01 -8.05107594e-01 -6.22904837e-01
-4.01106566e-01 -8.53120744e-01 9.60298479e-02 5.60626924e-01
7.64417768e-01 -7.26289570e-01 4.58690703e-01 3.38542508e-03
-2.72237331e-01 -1.06754148e+00 -7.99247622e-01 -3.25085223e-02
-7.40252256e-01 -1.22943652e+00 -3.59021127e-01 -9.73258734e-01
1.24425805e+00 3.97427469e-01 8.00886333e-01 5.16854167e-01
2.28151843e-01 2.34098986e-01 -2.99787134e-01 2.96790600e-01
-4.15666670e-01 2.56404042e-01 9.81955230e-02 4.35955882e-01
6.83012083e-02 -1.00070357e+00 -6.74283266e-01 2.93258429e-01
-1.01354980e+00 2.54640937e-01 6.35841131e-01 1.17948020e+00
7.96177864e-01 1.19411960e-01 6.04672968e-01 -1.24339628e+00
3.35709363e-01 -4.63077009e-01 -5.23150861e-01 2.20033884e-01
-1.01645780e+00 1.79294080e-01 8.72183979e-01 -3.96374285e-01
-7.91712224e-01 2.63330549e-01 -5.08570448e-02 -1.00180328e+00
1.64664850e-01 6.30496204e-01 -7.66775489e-01 -2.01241523e-02
3.72722149e-01 4.25417960e-01 1.75391406e-01 -2.28234380e-01
6.78653896e-01 1.60616145e-01 5.45506656e-01 -4.43754286e-01
1.26272738e+00 7.14020669e-01 -2.50646565e-02 -1.43768251e-01
-8.70921433e-01 -3.94321173e-01 -8.65289390e-01 -1.82430029e-01
9.32396650e-01 -1.12553430e+00 -4.25480247e-01 5.44523954e-01
-5.98066986e-01 -2.20927581e-01 -6.50627539e-02 3.37960571e-01
-4.59112406e-01 5.83059132e-01 -5.23406625e-01 -3.28289807e-01
-3.68650556e-01 -1.08369529e+00 9.42452610e-01 -9.86694768e-02
3.44334245e-01 -1.12755263e+00 3.57967168e-02 4.48791862e-01
-5.64639121e-02 4.84116584e-01 9.67008173e-01 -7.06512690e-01
-6.89538360e-01 -1.43220901e-01 -1.38868883e-01 4.94711071e-01
3.78160745e-01 -8.11290517e-02 -7.30954289e-01 -6.94275677e-01
2.27148563e-01 -3.88469338e-01 6.92251444e-01 2.20690295e-02
1.38802016e+00 -4.29849207e-01 -4.63502288e-01 9.02611852e-01
1.45448852e+00 -2.63453200e-02 4.74500328e-01 1.86283827e-01
1.47289157e+00 6.16221547e-01 4.31801945e-01 2.32530981e-01
6.10143483e-01 4.20685172e-01 5.43498337e-01 -1.87837586e-01
-9.40673426e-02 -5.23016810e-01 1.15278922e-01 1.40850794e+00
-4.97325277e-03 -1.24053396e-01 -7.11618841e-01 3.67085934e-01
-1.96480405e+00 -8.26777279e-01 3.96647416e-02 2.10984063e+00
6.97265625e-01 1.95623890e-01 3.33063789e-02 1.36719361e-01
9.55376506e-01 4.12080258e-01 -6.74578428e-01 3.97544622e-01
6.76857606e-02 -2.95578986e-01 1.69400081e-01 5.50633311e-01
-9.93217707e-01 9.11775887e-01 5.02704859e+00 6.98093891e-01
-1.19813848e+00 7.12692440e-02 7.18127668e-01 1.98775828e-01
-5.56812167e-01 1.33761629e-01 -3.13988090e-01 5.47326207e-01
2.55716592e-01 2.81850129e-01 5.71998715e-01 9.38713312e-01
-1.80404514e-01 5.11345029e-01 -9.61654127e-01 9.25760984e-01
2.28722885e-01 -1.14200175e+00 2.28472754e-01 2.22690925e-01
9.99829888e-01 -2.72199482e-01 1.66348800e-01 3.92758518e-01
4.82407242e-01 -6.99535191e-01 2.59351522e-01 4.45925802e-01
7.51980484e-01 -1.06593978e+00 5.05510211e-01 3.40200514e-01
-1.29016995e+00 1.71144083e-01 -3.33962291e-01 4.13661212e-01
-7.00755343e-02 4.45895165e-01 -5.14737368e-01 9.63854849e-01
5.32682121e-01 9.44065571e-01 -7.91976690e-01 3.63516450e-01
-3.52862149e-01 4.70971435e-01 1.76564287e-02 5.17258525e-01
2.58866131e-01 -8.90486121e-01 5.53670347e-01 5.37487984e-01
-3.09642274e-02 3.28816473e-02 5.95159471e-01 8.89419794e-01
-2.62024105e-01 8.04619938e-02 -6.77449882e-01 9.07418653e-02
4.84039605e-01 1.62390471e+00 -8.15001488e-01 -3.72224838e-01
-3.28930199e-01 1.16100132e+00 7.57440865e-01 3.75471443e-01
-8.87269020e-01 -5.03464080e-02 2.54847318e-01 2.55068112e-02
3.19480240e-01 -1.92839473e-01 -1.90821216e-01 -1.39066613e+00
1.98316246e-01 -1.19886220e+00 5.46857297e-01 -5.46811640e-01
-1.63899302e+00 6.83980167e-01 -1.54096112e-01 -1.43899930e+00
-1.72033578e-01 -1.45822972e-01 -6.87637210e-01 7.05877900e-01
-1.17167401e+00 -1.67464459e+00 -6.07184350e-01 8.89171660e-01
1.65457755e-01 -4.42687392e-01 5.74940681e-01 3.81773293e-01
-8.00920010e-01 6.99783206e-01 8.29543844e-02 3.27896446e-01
7.41486967e-01 -1.46973443e+00 2.49003217e-01 9.16225731e-01
8.35173428e-02 7.52695501e-01 3.71116400e-01 -8.24857533e-01
-1.39077508e+00 -1.48660350e+00 -7.08419010e-02 -4.00766522e-01
6.48673415e-01 -6.15127325e-01 -1.25437975e+00 8.18117201e-01
2.60389775e-01 3.14995706e-01 8.02345932e-01 5.05517311e-02
-5.17994583e-01 -1.45271659e-01 -1.01994109e+00 5.98720074e-01
1.07608140e+00 -6.14288449e-01 -2.63821810e-01 3.72909963e-01
9.09437656e-01 -3.87755960e-01 -9.86763120e-01 5.99592149e-01
1.56301826e-01 -9.38229442e-01 8.07668447e-01 -6.29938781e-01
2.60566354e-01 -6.23547733e-01 -1.98877286e-02 -1.50080574e+00
-5.26546299e-01 -3.81003916e-01 -1.66218266e-01 1.66878963e+00
1.96949333e-01 -6.28780246e-01 1.04420137e+00 3.52855533e-01
-2.04972014e-01 -9.46686625e-01 -5.35671771e-01 -6.73011422e-01
-1.31015643e-01 1.23124078e-01 7.18029439e-01 1.58263469e+00
-2.43712842e-01 8.37563097e-01 -5.08457243e-01 5.47945917e-01
8.05745065e-01 2.91397929e-01 9.79144633e-01 -9.92209792e-01
-3.42806935e-01 -1.21193238e-01 -2.57667810e-01 -7.18473077e-01
3.19044560e-01 -1.07315505e+00 -1.55563340e-01 -1.49137056e+00
4.11026895e-01 -5.75816393e-01 -3.39229882e-01 2.56740093e-01
-6.52383149e-01 6.63385093e-02 -6.41835183e-02 4.26154494e-01
-5.82610309e-01 9.80012119e-01 1.44059551e+00 -2.49886706e-01
-1.84308261e-01 -1.02646008e-01 -7.64797926e-01 7.03980863e-01
6.26511455e-01 -4.40864682e-01 -9.26681578e-01 -1.64186776e-01
1.28619209e-01 -1.17514737e-01 2.68387496e-01 -9.57036316e-01
2.43942291e-01 3.44215636e-03 3.27607661e-01 -6.79625452e-01
2.93665677e-01 -1.01375425e+00 2.81387150e-01 3.85709137e-01
-2.01746732e-01 1.05168082e-01 -3.30204546e-01 1.03358495e+00
-2.15388626e-01 1.79084688e-01 8.81179690e-01 -1.10940009e-01
-3.05073619e-01 9.11456645e-01 2.36284837e-01 1.39028952e-01
9.54235017e-01 -4.20239903e-02 -2.37501353e-01 -2.66400278e-01
-8.08990598e-01 5.21489084e-01 7.53992856e-01 3.61621827e-01
5.78798354e-01 -1.84863484e+00 -5.38801432e-01 3.18792433e-01
2.50820160e-01 4.03047323e-01 3.02314997e-01 8.05253565e-01
-2.51357585e-01 -2.72779733e-01 -1.99472055e-01 -7.33996630e-01
-9.87304211e-01 1.14998710e+00 3.07514906e-01 -5.08619726e-01
-5.89597523e-01 7.05376267e-01 4.87544656e-01 -7.60096490e-01
1.80543900e-01 2.54365891e-01 -2.97478527e-01 -9.63940471e-03
1.09690621e-01 8.82160068e-02 -1.42614365e-01 -6.06759667e-01
-2.25065723e-01 5.88015020e-01 -2.28368193e-01 1.23485019e-02
1.29768836e+00 -3.19275618e-01 -3.73193949e-01 3.84944439e-01
1.34118140e+00 2.15319768e-01 -1.36267745e+00 -2.79067338e-01
-3.41211289e-01 -3.56187731e-01 -6.64218068e-02 -2.86124796e-01
-1.53869188e+00 7.68429756e-01 6.63909316e-01 3.65854383e-01
1.11737716e+00 -1.28852561e-01 6.73446238e-01 6.61740229e-02
8.82748514e-02 -9.26277757e-01 6.21174157e-01 6.13681003e-02
8.46067131e-01 -1.38090563e+00 1.98283866e-01 -8.06488514e-01
-8.28530073e-01 6.28382385e-01 1.13913715e+00 -3.36902708e-01
7.08941281e-01 -9.49929878e-02 9.74974632e-02 -6.41300619e-01
-5.58662832e-01 -7.74984062e-02 4.56992447e-01 5.77142298e-01
1.11667767e-01 -1.69537202e-01 -4.39334363e-02 5.16970456e-01
1.64383166e-02 -7.24014044e-01 2.56687492e-01 5.45588553e-01
-3.95974256e-02 -1.15811467e+00 -1.11966126e-01 3.05673033e-01
-2.53280729e-01 5.11109503e-03 -6.06878996e-01 7.77075171e-01
1.17894217e-01 6.11485124e-01 -2.12223873e-01 -7.87676513e-01
7.27832392e-02 -4.72370945e-02 2.80482799e-01 -6.84640527e-01
-3.92964453e-01 2.45475486e-01 -2.43746623e-01 -4.09097999e-01
-5.96376479e-01 -4.55526680e-01 -1.23090291e+00 -1.41710579e-01
-5.72815061e-01 3.78853261e-01 2.37910658e-01 7.96951175e-01
4.45088625e-01 6.56448960e-01 1.19447362e+00 -5.03020763e-01
-4.07888681e-01 -8.30510795e-01 -6.47618413e-01 7.48266459e-01
6.98572323e-02 -6.25894904e-01 -4.36552256e-01 1.73774436e-01] | [7.486689567565918, 5.947271347045898] |
ad033e16-0556-43ed-a001-0786d5d4592a | hierarchical-semantic-aggregation-for | 2012.02733 | null | https://arxiv.org/abs/2012.02733v2 | https://arxiv.org/pdf/2012.02733v2.pdf | Seed the Views: Hierarchical Semantic Alignment for Contrastive Representation Learning | Self-supervised learning based on instance discrimination has shown remarkable progress. In particular, contrastive learning, which regards each image as well as its augmentations as an individual class and tries to distinguish them from all other images, has been verified effective for representation learning. However, pushing away two images that are de facto similar is suboptimal for general representation. In this paper, we propose a hierarchical semantic alignment strategy via expanding the views generated by a single image to \textbf{Cross-samples and Multi-level} representation, and models the invariance to semantically similar images in a hierarchical way. This is achieved by extending the contrastive loss to allow for multiple positives per anchor, and explicitly pulling semantically similar images/patches together at different layers of the network. Our method, termed as CsMl, has the ability to integrate multi-level visual representations across samples in a robust way. CsMl is applicable to current contrastive learning based methods and consistently improves the performance. Notably, using the moco as an instantiation, CsMl achieves a \textbf{76.6\% }top-1 accuracy with linear evaluation using ResNet-50 as backbone, and \textbf{66.7\%} and \textbf{75.1\%} top-1 accuracy with only 1\% and 10\% labels, respectively. \textbf{All these numbers set the new state-of-the-art.} | ['Qi Tian', 'Hongkai Xiong', 'Lingxi Xie', 'Hao Li', 'Xiaopeng Zhang', 'Haohang Xu'] | 2020-12-04 | null | null | null | null | ['self-supervised-image-classification'] | ['computer-vision'] | [ 5.56318760e-01 2.19911098e-01 -3.80345106e-01 -2.71873176e-01
-1.06349969e+00 -5.15612245e-01 6.80289507e-01 3.22820425e-01
-3.67031068e-01 5.97009003e-01 -3.70254405e-02 1.16529599e-01
-1.50562659e-01 -7.07703590e-01 -1.01571500e+00 -6.84606135e-01
-3.47475708e-02 3.22747409e-01 2.93320060e-01 -6.51794076e-02
2.74699051e-02 5.34322143e-01 -1.78388822e+00 6.48582578e-01
5.36004007e-01 1.34646308e+00 9.77276713e-02 4.09487724e-01
1.03086689e-02 8.04751098e-01 -4.82524008e-01 -4.76018965e-01
4.43810672e-01 -3.68449181e-01 -8.44828606e-01 2.15471625e-01
1.39000940e+00 -7.77494609e-02 -1.22124940e-01 1.08102715e+00
3.51649284e-01 4.76522893e-02 8.11203599e-01 -1.46787822e+00
-7.97734380e-01 3.43805790e-01 -8.20890725e-01 2.93611795e-01
1.09240323e-01 7.74405971e-02 1.15898907e+00 -9.71515954e-01
6.58375323e-01 1.22740984e+00 5.18995345e-01 6.19709194e-01
-1.58316588e+00 -8.65593493e-01 5.54854512e-01 1.07757591e-01
-1.23175907e+00 -5.15036404e-01 7.95188487e-01 -4.83384490e-01
9.72271919e-01 3.04434538e-01 4.58916664e-01 1.21140802e+00
-1.30886376e-01 8.86107683e-01 1.58682704e+00 -4.86938059e-01
1.59379050e-01 1.98498383e-01 1.76733702e-01 8.20154428e-01
2.65397459e-01 1.11105829e-01 -4.78863806e-01 1.77394852e-01
5.90585768e-01 1.70598164e-01 -1.39342323e-01 -6.03744984e-01
-1.03362167e+00 6.84637725e-01 8.73807609e-01 3.18180025e-01
-1.10412754e-01 1.74394473e-01 3.08230162e-01 2.64309347e-01
6.20695472e-01 3.56292337e-01 -2.70239204e-01 5.18500865e-01
-1.05174911e+00 1.09243572e-01 3.48061115e-01 9.02844608e-01
7.96578586e-01 2.50070959e-01 -2.09834799e-01 1.07945228e+00
9.44402814e-02 5.05074799e-01 2.64290094e-01 -1.03875852e+00
4.90974516e-01 7.67306387e-01 -2.46991366e-01 -1.09057510e+00
-4.26329732e-01 -1.16118205e+00 -1.17004955e+00 5.85034013e-01
4.86200005e-01 3.71803731e-01 -1.20326877e+00 2.09538150e+00
7.87010565e-02 8.37173089e-02 8.00906494e-02 6.36379957e-01
8.90313804e-01 4.03290510e-01 1.80056021e-01 -1.49981260e-01
1.34817433e+00 -1.23207259e+00 -1.50948241e-01 -4.03102309e-01
3.65262657e-01 -5.70433617e-01 1.23199666e+00 4.51846600e-01
-1.27828395e+00 -7.69704759e-01 -1.27698338e+00 6.49155602e-02
-4.28692609e-01 4.04460095e-02 2.95108050e-01 5.32236040e-01
-1.28529477e+00 5.32106102e-01 -4.24640864e-01 -2.67890900e-01
8.16234469e-01 2.53762335e-01 -5.87665677e-01 -3.79485607e-01
-7.35463679e-01 6.70770824e-01 3.90963137e-01 -3.07276517e-01
-1.01462138e+00 -7.46988475e-01 -9.17344749e-01 2.12203324e-01
3.09327811e-01 -7.42117584e-01 6.45989537e-01 -1.37258148e+00
-9.16911066e-01 1.23830187e+00 -5.81933223e-02 -5.15116811e-01
5.44605792e-01 4.89424020e-02 -4.85053271e-01 4.28383410e-01
2.78904110e-01 1.17065322e+00 1.02614915e+00 -1.76369023e+00
-6.05464756e-01 -5.50022900e-01 2.55048782e-01 2.69722283e-01
-4.87800986e-01 -1.72292039e-01 -3.47773433e-01 -8.51940870e-01
3.03852499e-01 -8.94421875e-01 -1.56571604e-02 2.43990377e-01
-2.71101773e-01 -1.90040320e-02 4.76669997e-01 -5.92395544e-01
9.06754434e-01 -2.15707588e+00 6.43413365e-02 2.35856846e-01
3.78145278e-01 1.59785956e-01 -4.41898376e-01 1.60801917e-01
-2.81440228e-01 9.56401452e-02 -4.46884245e-01 -4.65704709e-01
5.22929877e-02 1.84892621e-02 -6.87506869e-02 3.59448999e-01
3.75400424e-01 7.95407355e-01 -8.44534397e-01 -4.32474941e-01
3.02009970e-01 4.46464151e-01 -5.51602960e-01 5.99711435e-03
9.50274896e-03 3.36940080e-01 1.38087561e-02 6.62525415e-01
8.13003540e-01 -5.04207790e-01 5.49796708e-02 -3.22232723e-01
2.00696588e-01 -2.08406672e-01 -1.19131064e+00 1.67871785e+00
-4.33236957e-01 4.93738592e-01 9.55830589e-02 -1.26599205e+00
8.94062340e-01 1.13447448e-02 4.43967015e-01 -1.13841999e+00
-2.77256221e-01 1.74753293e-01 -2.55898386e-01 -1.33266300e-01
2.00814158e-01 -1.33901834e-01 8.38427339e-03 1.26575693e-01
3.44015807e-01 1.29312947e-01 1.92783639e-01 1.69070482e-01
9.24251437e-01 1.77708343e-01 3.11422765e-01 -3.11542243e-01
5.05847633e-01 -1.47583261e-01 4.25243795e-01 9.38633442e-01
-1.56119153e-01 8.34069371e-01 2.79548943e-01 -2.96159804e-01
-9.76357937e-01 -1.31104982e+00 -2.47337997e-01 1.27979851e+00
2.92965263e-01 -2.67775178e-01 -7.29694068e-01 -8.83951545e-01
-3.08838990e-02 5.06515682e-01 -8.42324138e-01 -2.41384327e-01
-4.28620100e-01 -6.83360517e-01 4.04894590e-01 7.01344490e-01
7.91693151e-01 -9.48278844e-01 -3.71580005e-01 -6.85913935e-02
-1.46227390e-01 -1.14174461e+00 8.68808199e-03 2.73931623e-01
-8.17322314e-01 -1.12640584e+00 -8.30946147e-01 -9.45016205e-01
7.93340921e-01 4.46812034e-01 1.39009011e+00 1.22497573e-01
-2.62136728e-01 4.02292401e-01 -3.20585102e-01 -1.40093625e-01
-2.55678415e-01 -1.85037125e-02 -2.01529518e-01 1.44479617e-01
-1.11036329e-02 -6.21038735e-01 -9.20071781e-01 3.49491745e-01
-1.04730928e+00 1.06075443e-01 6.18992448e-01 9.01750088e-01
8.32175672e-01 -1.80670902e-01 5.74699759e-01 -9.58405256e-01
1.59593485e-03 -4.02830631e-01 -1.49767026e-01 2.93471038e-01
-7.53345072e-01 -1.30950093e-01 7.66178846e-01 -2.00151026e-01
-7.63618886e-01 -2.12129969e-02 2.53722537e-02 -5.99436641e-01
-4.63349104e-01 4.38565128e-02 -1.07091114e-01 -9.13290307e-02
8.23595822e-01 3.20431858e-01 -8.98890719e-02 -4.55314040e-01
4.35939282e-01 3.51646811e-01 5.06222904e-01 -5.34032643e-01
8.01988602e-01 6.45690620e-01 8.00772011e-02 -5.98434627e-01
-1.19846261e+00 -4.59454954e-01 -6.43316090e-01 -2.07843274e-01
9.07068014e-01 -1.07546735e+00 -4.38067019e-01 2.65381128e-01
-5.59239626e-01 -2.80990511e-01 -5.15277863e-01 2.08853394e-01
-6.09354436e-01 3.78793210e-01 -2.80753493e-01 -5.77726185e-01
-3.16945195e-01 -1.10995889e+00 1.08494925e+00 6.01499118e-02
-4.87566888e-02 -8.17328870e-01 -2.51959324e-01 7.05691576e-01
4.56779301e-01 4.04245317e-01 1.01050079e+00 -7.58319736e-01
-5.04692435e-01 -1.22103980e-02 -5.96193194e-01 7.58934855e-01
-3.87337916e-02 -2.35679239e-01 -1.34976792e+00 -8.15401375e-01
-2.09634885e-01 -5.17407238e-01 1.24338448e+00 2.66762555e-01
1.31097841e+00 -3.30616415e-01 -2.18119308e-01 5.72014570e-01
1.69137037e+00 -9.80931986e-03 4.75628376e-01 4.61844176e-01
6.94339931e-01 6.08538866e-01 2.27530152e-01 1.04775675e-01
3.33141387e-01 7.31448770e-01 8.77840757e-01 -5.35307884e-01
-5.70175469e-01 -5.21719158e-02 3.00024897e-01 4.41551954e-01
-1.27353340e-01 -2.62232602e-01 -6.70841157e-01 4.06961292e-01
-1.60957086e+00 -1.15210235e+00 9.34283510e-02 2.37680030e+00
4.63887304e-01 2.38575354e-01 1.96168557e-01 2.26395041e-01
7.41754830e-01 3.66676092e-01 -4.87900048e-01 -1.31965071e-01
-3.67808223e-01 4.83833224e-01 4.11694109e-01 2.01058626e-01
-1.43655849e+00 7.99456418e-01 5.68852425e+00 9.97004271e-01
-1.18164790e+00 1.95370615e-01 9.45286870e-01 -5.33424281e-02
-1.85986698e-01 -6.85113445e-02 -6.50628090e-01 4.35071200e-01
5.33376932e-01 2.86185294e-01 2.72766888e-01 7.99699485e-01
-3.11408907e-01 -7.64823630e-02 -1.14415145e+00 1.06444824e+00
4.60222483e-01 -1.38047361e+00 4.19347346e-01 -6.60659373e-02
8.57009411e-01 -2.77923346e-02 4.40372467e-01 3.62520605e-01
1.89470366e-01 -1.22139907e+00 8.11950207e-01 4.33512300e-01
8.80638063e-01 -5.95073819e-01 6.30250692e-01 1.76118568e-01
-1.30295086e+00 -1.69408768e-01 -2.42677316e-01 1.85659915e-01
-3.31688583e-01 2.69793034e-01 -4.25929368e-01 7.54649520e-01
9.91369903e-01 8.25536430e-01 -9.62890208e-01 9.34383750e-01
9.14865211e-02 4.56551820e-01 -3.81056927e-02 3.87253314e-01
2.51507401e-01 -8.72581080e-03 3.47931176e-01 1.35900044e+00
1.33948386e-01 -3.67599905e-01 5.77884793e-01 6.98497474e-01
-3.18163723e-01 1.71932563e-01 -4.80939418e-01 4.22711313e-01
2.51676440e-01 1.16046202e+00 -7.82884300e-01 -5.34208000e-01
-3.41890633e-01 1.03588784e+00 6.36276960e-01 4.43387091e-01
-7.26789951e-01 -8.72831792e-02 4.03605551e-01 2.09887803e-01
2.53553301e-01 1.26757205e-01 -5.67296632e-02 -1.09303570e+00
2.45554537e-01 -9.60268378e-01 7.44067013e-01 -7.66671062e-01
-1.48402274e+00 7.75980055e-01 6.58444837e-02 -1.50669944e+00
2.40255017e-02 -7.02975869e-01 -1.80906087e-01 5.65188646e-01
-1.76075661e+00 -1.33513999e+00 -6.54335141e-01 6.77937508e-01
5.60756624e-01 -2.15993136e-01 9.29903090e-01 3.21797401e-01
-3.88976544e-01 9.79980767e-01 7.89255649e-02 1.47714511e-01
6.90073788e-01 -1.33228314e+00 4.29500528e-02 6.65361345e-01
4.73524243e-01 3.31760347e-01 3.52355361e-01 -1.36351496e-01
-7.76786506e-01 -1.26664805e+00 4.71789479e-01 -2.98604548e-01
4.01781768e-01 -3.48929465e-01 -1.04376733e+00 5.35552442e-01
2.44793966e-01 3.98065716e-01 5.61535597e-01 -5.94910383e-02
-9.43122745e-01 -4.66563463e-01 -1.31673574e+00 5.45569122e-01
1.26726687e+00 -6.30115509e-01 -3.85507137e-01 4.09556061e-01
5.02405822e-01 -1.52611449e-01 -9.65443075e-01 7.25911736e-01
3.51272881e-01 -1.29629946e+00 1.30156279e+00 -6.05962455e-01
5.05856693e-01 -1.59577012e-01 -4.99778450e-01 -9.90108430e-01
-5.37505567e-01 -1.35834202e-01 -4.52669933e-02 1.25977480e+00
3.88776749e-01 -5.93395174e-01 8.11305583e-01 7.88268521e-02
-1.04456797e-01 -7.14228988e-01 -9.88900125e-01 -9.29798543e-01
2.44433075e-01 -3.47631305e-01 1.87742755e-01 1.19620442e+00
-4.77435529e-01 4.27562058e-01 -2.58738697e-01 1.55351698e-01
8.67060304e-01 1.19618610e-01 5.66594779e-01 -1.26541007e+00
-3.60607415e-01 -7.14376926e-01 -6.23227298e-01 -7.97685146e-01
1.72634080e-01 -1.42578554e+00 -3.31685811e-01 -1.51176119e+00
6.66255057e-01 -5.99426746e-01 -8.43958139e-01 5.15093207e-01
-1.09934434e-01 8.19591999e-01 4.75783408e-01 2.35135332e-01
-6.80936098e-01 3.19469243e-01 1.05484557e+00 -5.19686222e-01
1.98205650e-01 -1.85572147e-01 -7.81948805e-01 7.38247275e-01
7.35574543e-01 -3.35221648e-01 -3.76184195e-01 -3.43802571e-01
-5.60201257e-02 -2.33452648e-01 7.69058049e-01 -1.29464746e+00
-7.06632286e-02 1.78638712e-01 8.80322874e-01 -3.66231173e-01
4.95558083e-01 -8.64730835e-01 8.39100704e-02 4.35168415e-01
-6.49028182e-01 1.04023479e-01 1.82371721e-01 7.36630082e-01
-3.24349046e-01 -1.58079967e-01 1.08759928e+00 -3.43719006e-01
-7.61662006e-01 2.58719862e-01 1.45051777e-01 3.40343535e-01
1.05654228e+00 -5.95220387e-01 -5.24444461e-01 -1.40153870e-01
-1.05309355e+00 2.08135638e-02 5.48696339e-01 3.16444278e-01
5.40005088e-01 -1.25334203e+00 -7.52959669e-01 2.23242432e-01
5.15972257e-01 -1.25344917e-01 5.11480451e-01 7.29386091e-01
-2.23694056e-01 6.34485409e-02 -4.47006792e-01 -9.04818892e-01
-1.39942527e+00 6.15033567e-01 4.47008729e-01 -3.25077266e-01
-5.52501023e-01 8.82201552e-01 5.67918360e-01 -3.18941504e-01
4.08589691e-01 -8.95907581e-02 -2.90599197e-01 2.01703876e-01
3.76467198e-01 2.98679918e-01 1.70184314e-01 -8.45901847e-01
-3.83601993e-01 8.51656258e-01 -2.18720660e-01 8.26531500e-02
1.27609992e+00 1.05105834e-02 -2.45824140e-02 4.01585668e-01
1.45561039e+00 -6.92561939e-02 -1.45006692e+00 -4.02737081e-01
-5.64258024e-02 -3.85841727e-01 -1.59450829e-01 -1.14665854e+00
-1.32744670e+00 8.95190239e-01 9.32106733e-01 1.14333771e-01
1.18265843e+00 1.05133429e-01 3.10382426e-01 9.96738002e-02
4.16662812e-01 -8.70989382e-01 5.32718062e-01 1.96470872e-01
9.17718172e-01 -1.42559326e+00 9.26514715e-02 -4.30126309e-01
-5.23422003e-01 8.62203419e-01 8.23120713e-01 -3.50257784e-01
3.04375708e-01 -6.65623322e-02 -5.00019034e-03 -8.01592916e-02
-5.93188286e-01 -3.29374999e-01 5.99152088e-01 6.22670352e-01
3.07779670e-01 2.31300294e-02 1.64792806e-01 1.99985236e-01
7.06905648e-02 -3.54160726e-01 6.73852935e-02 7.31843054e-01
-3.46021950e-01 -8.86585414e-01 -2.30858058e-01 5.18664837e-01
-5.38771033e-01 -1.97517246e-01 -3.51520747e-01 9.84006703e-01
2.76194274e-01 8.90724599e-01 2.44228408e-01 -3.13701779e-01
4.02385086e-01 5.09493500e-02 6.20334208e-01 -5.16274333e-01
-8.16925228e-01 -3.92074771e-02 2.10477803e-02 -6.63771987e-01
-6.12779796e-01 -3.87398690e-01 -1.09430742e+00 1.03803881e-01
4.35615629e-02 -2.04680040e-01 3.39094251e-01 7.03379571e-01
4.19323921e-01 6.06480300e-01 7.51515985e-01 -8.47668648e-01
-4.56686586e-01 -6.28770471e-01 -4.15465623e-01 8.60573292e-01
3.63499463e-01 -7.48396218e-01 -4.14998353e-01 -7.59973302e-02] | [9.480894088745117, 2.436263084411621] |
7f8216ac-b0bb-4f3c-912e-35e0c10015c2 | semantic-visual-simultaneous-localization-and | 2209.06428 | null | https://arxiv.org/abs/2209.06428v1 | https://arxiv.org/pdf/2209.06428v1.pdf | Semantic Visual Simultaneous Localization and Mapping: A Survey | Visual Simultaneous Localization and Mapping (vSLAM) has achieved great progress in the computer vision and robotics communities, and has been successfully used in many fields such as autonomous robot navigation and AR/VR. However, vSLAM cannot achieve good localization in dynamic and complex environments. Numerous publications have reported that, by combining with the semantic information with vSLAM, the semantic vSLAM systems have the capability of solving the above problems in recent years. Nevertheless, there is no comprehensive survey about semantic vSLAM. To fill the gap, this paper first reviews the development of semantic vSLAM, explicitly focusing on its strengths and differences. Secondly, we explore three main issues of semantic vSLAM: the extraction and association of semantic information, the application of semantic information, and the advantages of semantic vSLAM. Then, we collect and analyze the current state-of-the-art SLAM datasets which have been widely used in semantic vSLAM systems. Finally, we discuss future directions that will provide a blueprint for the future development of semantic vSLAM. | ['ShengYong Chen', 'Ruyu Liu', 'Qiyi Tong', 'Jialing Liu', 'Jianhua Zhang', 'Kaiqi Chen'] | 2022-09-14 | null | null | null | null | ['simultaneous-localization-and-mapping'] | ['computer-vision'] | [-1.12051964e-01 -3.15692008e-01 -1.62590355e-01 -5.66727102e-01
-3.47858638e-01 -4.90595818e-01 7.03982711e-01 1.02560930e-01
-5.39794743e-01 6.50049567e-01 -1.49332657e-01 -6.18026517e-02
-2.00505078e-01 -8.87609065e-01 -3.50626022e-01 -4.16381359e-01
1.24780640e-01 5.54504633e-01 5.95453680e-01 -3.57003361e-01
5.39973736e-01 7.46095300e-01 -1.95739353e+00 -1.59685582e-01
9.75182474e-01 9.51785624e-01 1.00270355e+00 2.06641927e-01
-4.94410664e-01 4.13575113e-01 -4.56926048e-01 4.22657549e-01
-2.88776471e-03 -2.94866472e-01 -6.97804987e-01 -1.42100528e-01
3.96812141e-01 1.95091873e-01 -2.03996077e-01 1.27854180e+00
4.16920006e-01 3.88938785e-01 2.51573116e-01 -1.51935887e+00
-5.68606257e-01 9.76390317e-02 -3.94539624e-01 -1.14814267e-01
6.50748312e-01 -3.77316922e-01 8.77226710e-01 -1.18539155e+00
7.88200736e-01 1.42338061e+00 7.11364090e-01 2.24298030e-01
-8.25596452e-01 -5.91367602e-01 2.57862926e-01 5.64232290e-01
-1.81045401e+00 -3.83104622e-01 6.64938271e-01 -3.12007248e-01
1.01890779e+00 -1.08993994e-02 6.72676682e-01 7.06515551e-01
2.38225281e-01 8.75632823e-01 1.17970753e+00 -4.39446658e-01
2.31262356e-01 2.37643793e-01 1.69740200e-01 7.98565328e-01
3.50190669e-01 3.65843698e-02 -8.06416452e-01 -2.42845509e-02
7.79098511e-01 -2.46436819e-01 -2.12852404e-01 -1.19367838e+00
-1.41779971e+00 1.00854707e+00 8.57870817e-01 4.10790950e-01
-1.15844876e-01 2.13162035e-01 2.53423274e-01 -8.47285986e-02
1.33651227e-01 4.95475560e-01 -2.51118511e-01 1.07749745e-01
-8.43375564e-01 2.31713638e-01 4.24815714e-01 1.37048662e+00
1.10172319e+00 4.85383719e-02 7.08526015e-01 8.95293176e-01
6.03332400e-01 1.06131494e+00 3.54414433e-01 -8.83736968e-01
2.08042234e-01 5.21276832e-01 2.05477476e-01 -1.25075853e+00
-7.30260849e-01 2.81421444e-03 -3.50195378e-01 2.46887982e-01
-1.00622304e-01 2.75026590e-01 -9.41960454e-01 1.38707161e+00
6.35885820e-02 6.60649836e-02 4.48860109e-01 1.11319566e+00
1.19661093e+00 5.24499059e-01 -7.35435560e-02 7.51856044e-02
1.22595835e+00 -1.15799963e+00 -9.96481895e-01 -8.83527994e-01
6.09677315e-01 -7.80718386e-01 8.72601211e-01 1.15052834e-01
-4.19127256e-01 -5.64589202e-01 -1.18194497e+00 -3.12025011e-01
-7.36900926e-01 2.26694331e-01 9.93917108e-01 2.36067072e-01
-1.24494541e+00 1.64075702e-01 -1.00839949e+00 -1.22734857e+00
9.40312222e-02 1.39035866e-01 -6.84644401e-01 -2.52894431e-01
-1.22017205e+00 1.39209747e+00 6.27529144e-01 1.57714516e-01
-4.75990146e-01 -2.75415089e-02 -1.41765606e+00 -5.02876699e-01
4.61822212e-01 -5.22097230e-01 9.06454444e-01 -4.80191082e-01
-1.16194642e+00 8.48129272e-01 -5.82908094e-01 -4.21937317e-01
3.48286033e-01 -3.85683894e-01 -4.09197181e-01 -2.13411748e-02
6.66141570e-01 8.21209550e-01 4.70679812e-02 -1.43352139e+00
-1.01798451e+00 -5.96155107e-01 -1.82527855e-01 5.28174102e-01
2.92360485e-01 -1.48093730e-01 -6.75680220e-01 8.76566954e-03
1.04437578e+00 -1.04439604e+00 -3.60938340e-01 2.14442126e-02
-8.18388984e-02 -2.39369154e-01 9.28560674e-01 -2.10074976e-01
7.38685489e-01 -2.20260787e+00 2.65397221e-01 1.34054020e-01
-1.12891488e-01 -2.29078438e-02 6.37218282e-02 6.34695113e-01
5.94619691e-01 -3.05939227e-01 -1.25296980e-01 -3.87223989e-01
-9.39896256e-02 5.72646499e-01 -3.72168690e-01 7.01556087e-01
-6.35800302e-01 9.01148856e-01 -1.21376824e+00 -5.69947779e-01
9.59519923e-01 4.45856601e-01 -1.05799347e-01 -1.61489964e-01
3.95847484e-02 4.50819522e-01 -4.80261683e-01 8.16000938e-01
6.60236955e-01 -3.30731533e-02 1.39137447e-01 -8.90084449e-03
-6.19816720e-01 1.57784626e-01 -1.15468502e+00 2.33992600e+00
-4.67277855e-01 1.05768144e+00 1.71919182e-01 -8.78763497e-01
1.29052949e+00 -1.47153050e-01 3.70782644e-01 -9.29371238e-01
-5.39990291e-02 7.73029983e-01 -4.26883966e-01 -2.98946619e-01
1.02714610e+00 6.06506020e-02 -2.31616959e-01 -2.85994858e-01
5.95803484e-02 -4.94724900e-01 -4.01673503e-02 1.04698747e-01
6.30652964e-01 3.53254974e-01 6.96171820e-01 -2.98501343e-01
7.65313089e-01 7.23452628e-01 5.31978011e-01 8.24815214e-01
-3.96990836e-01 4.35530514e-01 -2.99603969e-01 -3.41436982e-01
-6.12923622e-01 -1.10832524e+00 -1.78010046e-01 7.49393463e-01
1.25243235e+00 -4.13804352e-01 -1.08244076e-01 -4.11312610e-01
2.93880671e-01 6.86366141e-01 -2.95574635e-01 1.52986258e-01
-4.30534512e-01 -3.76665384e-01 4.16411817e-01 5.39456069e-01
9.11668360e-01 -1.17859530e+00 -8.84652972e-01 7.92506933e-02
-3.77481818e-01 -1.32943368e+00 4.92362022e-01 8.31198320e-03
-7.55185425e-01 -1.13388515e+00 -2.15630531e-01 -1.02544391e+00
6.15560949e-01 1.13757443e+00 7.86678731e-01 -7.15799779e-02
-1.95649937e-01 3.65732938e-01 -6.63119435e-01 -4.07462627e-01
2.50243638e-02 9.40743983e-02 3.70391786e-01 -5.53449929e-01
5.89661658e-01 -3.10120106e-01 -2.27677822e-01 4.86530274e-01
-2.50907600e-01 1.08103581e-01 4.24038053e-01 4.96024281e-01
7.61116564e-01 -4.48666438e-02 3.43787342e-01 -6.58951521e-01
1.92251757e-01 -3.54382306e-01 -8.71581137e-01 1.15059629e-01
-7.06561029e-01 -1.98499992e-01 1.25665694e-01 2.37885773e-01
-8.55278432e-01 1.38985887e-01 -3.41461122e-01 -2.18724430e-01
-2.55740285e-01 6.33990943e-01 -2.05552340e-01 -5.65418780e-01
5.35813570e-01 4.43832546e-01 7.98193291e-02 -6.51972771e-01
5.85925698e-01 8.13753545e-01 7.66482592e-01 -2.58999765e-01
5.29977918e-01 9.32209134e-01 4.89866324e-02 -9.81694520e-01
-8.99070323e-01 -1.02243018e+00 -9.07214403e-01 -1.53971598e-01
7.19492376e-01 -9.70867097e-01 -2.86042213e-01 4.91229862e-01
-1.04388118e+00 -9.09233019e-02 -2.12463252e-02 5.99470496e-01
-8.38930726e-01 4.15897429e-01 -1.10764906e-01 -7.15442538e-01
3.40927616e-02 -1.47037077e+00 1.06724823e+00 3.50327641e-01
-1.47860631e-01 -1.10009778e+00 2.51665711e-04 2.80974001e-01
4.02192324e-01 4.56016231e-03 4.49267745e-01 -4.78500158e-01
-6.38273954e-01 -1.06530704e-01 -4.40226555e-01 -7.67012984e-02
2.70759553e-01 -5.47219992e-01 -7.10679471e-01 -4.45998132e-01
-3.02801102e-01 -6.50405735e-02 7.70854950e-01 3.28028560e-01
5.21831751e-01 5.17246723e-01 -1.02014673e+00 8.41006517e-01
1.70705533e+00 4.80367929e-01 4.92040813e-01 8.73197258e-01
7.66383231e-01 6.86996818e-01 1.40614700e+00 1.84135651e-03
8.35748792e-01 8.62662733e-01 7.51178324e-01 5.88745810e-02
-1.08684927e-01 -4.24862862e-01 -1.86744668e-02 7.74381638e-01
1.27590835e-01 -1.04508437e-01 -1.15772557e+00 7.21723974e-01
-2.20736933e+00 -4.23928767e-01 -4.27686036e-01 1.98234570e+00
-1.34788722e-01 -3.89905393e-01 -6.38906598e-01 -2.39792347e-01
6.26424909e-01 3.88355196e-01 -3.98141086e-01 -4.92029861e-02
-4.02223766e-01 -3.16962600e-01 7.73774385e-01 6.16701722e-01
-1.20670068e+00 1.75546718e+00 6.51902819e+00 5.51441252e-01
-1.12559128e+00 2.23044544e-01 -7.00588226e-01 5.18599391e-01
-1.93631519e-02 3.35871875e-01 -9.51165497e-01 1.81653351e-01
3.55869800e-01 4.60616406e-03 2.82816380e-01 1.25493920e+00
7.58002847e-02 -7.30586350e-01 -7.23195910e-01 1.40962839e+00
3.48168015e-01 -1.36474049e+00 -7.46404305e-02 -1.57162733e-02
6.24847710e-01 5.71796298e-01 -3.24617386e-01 1.86296999e-01
3.01475853e-01 -8.95844877e-01 9.83998299e-01 1.71494305e-01
7.10772157e-01 -6.06865942e-01 1.17036331e+00 4.99703467e-01
-1.46941316e+00 -2.05460824e-02 -6.10875428e-01 -1.05728388e-01
4.60268676e-01 2.99483955e-01 -6.67747498e-01 9.72262383e-01
8.12103808e-01 1.14687276e+00 -4.60479051e-01 1.16628063e+00
-6.37098014e-01 -3.96259427e-01 -2.89739102e-01 -6.51018415e-03
4.32789803e-01 -3.97253543e-01 6.18523657e-01 9.80203509e-01
5.78662813e-01 -2.26190776e-01 6.14818633e-01 6.84092641e-01
6.04108632e-01 9.13252383e-02 -9.38616753e-01 4.61448431e-02
9.58701253e-01 9.02100205e-01 -8.73955190e-01 -2.17459679e-01
-5.08170247e-01 1.06749642e+00 2.75610179e-01 2.84734935e-01
-5.32511175e-01 -5.76985598e-01 8.72152746e-01 -2.05656931e-01
-6.44990057e-02 -9.36451554e-01 -4.23368096e-01 -1.02896857e+00
-2.08828121e-01 -1.11693613e-01 9.65780914e-02 -1.19498587e+00
-7.53283918e-01 6.22147501e-01 1.48686972e-02 -1.17751741e+00
-3.90031151e-02 -5.46758771e-01 9.39941406e-02 8.89258981e-01
-1.68499696e+00 -1.31858146e+00 -8.44757438e-01 3.04704875e-01
8.02484393e-01 -1.55510738e-01 8.42890382e-01 3.02314125e-02
9.51597351e-04 -2.47569308e-01 1.07978329e-01 -1.58522204e-02
7.88045049e-01 -9.36925054e-01 5.56495905e-01 8.83897483e-01
2.53015012e-01 6.73454583e-01 8.87466431e-01 -8.42846751e-01
-1.58300149e+00 -9.90229249e-01 1.02724099e+00 -4.67068017e-01
6.36439085e-01 -3.90884668e-01 -6.31664336e-01 1.06133544e+00
-1.86041057e-01 1.04764774e-01 6.62232637e-02 -5.54738613e-03
-8.53509456e-02 2.43731290e-01 -1.11825359e+00 4.40983832e-01
1.30951571e+00 -4.83903050e-01 -7.68669188e-01 1.05089337e-01
6.36390269e-01 -6.55021250e-01 -3.80294293e-01 6.74089730e-01
4.56082880e-01 -1.07300985e+00 1.04558039e+00 3.12276572e-01
-3.69083524e-01 -8.14705193e-01 -6.54098928e-01 -1.48205197e+00
-2.44355977e-01 2.57464916e-01 3.25996190e-01 9.33929145e-01
-1.49921507e-01 -1.09435201e+00 7.21732378e-01 -2.14273915e-01
-4.44497526e-01 -3.37126672e-01 -9.78818953e-01 -9.97486413e-01
-3.96737456e-01 -5.28297186e-01 4.29456204e-01 7.76641130e-01
-3.42835374e-02 1.42662987e-01 -2.52248287e-01 5.03728151e-01
7.30435014e-01 3.49338681e-01 9.30176079e-01 -1.50986648e+00
6.09193087e-01 -2.00297385e-01 -1.01859581e+00 -1.39178860e+00
3.88289303e-01 -8.59209061e-01 5.27882636e-01 -2.45625210e+00
-1.49425883e-02 -9.75954294e-01 -3.48840430e-02 2.94003755e-01
2.51710355e-01 3.26159626e-01 2.73852050e-01 5.69397926e-01
-8.63908172e-01 6.19959593e-01 9.13478792e-01 1.53894573e-01
-2.31013998e-01 -2.97333390e-01 -4.33750987e-01 9.31789100e-01
6.37750685e-01 -7.44566619e-02 -5.41791975e-01 -5.62285721e-01
1.92538559e-01 -1.42160743e-01 2.48018995e-01 -1.08045912e+00
5.16698480e-01 -3.29029858e-01 -2.69920565e-02 -1.19422245e+00
5.87328970e-01 -1.00220764e+00 1.73621759e-01 4.02731538e-01
2.66736388e-01 7.29452223e-02 6.38701543e-02 6.93551421e-01
-5.42168021e-01 -1.33801565e-01 7.91503847e-01 -3.60977113e-01
-2.06790280e+00 -7.77322873e-02 -4.63039547e-01 -3.16867352e-01
1.17658055e+00 -3.56283695e-01 -3.69552791e-01 -1.49173707e-01
-5.58943510e-01 6.60340071e-01 9.66585159e-01 9.70855832e-01
9.36822832e-01 -1.29864836e+00 -7.51219541e-02 3.67493421e-01
6.16507113e-01 2.17055455e-01 1.73064947e-01 8.64624023e-01
-8.90141249e-01 8.81808579e-01 -3.30808192e-01 -9.09188509e-01
-1.38516438e+00 5.94292998e-01 9.98779833e-02 4.40809399e-01
-9.21945512e-01 7.47730136e-01 2.16711104e-01 -8.92050505e-01
8.13838318e-02 8.56378749e-02 -4.36856300e-01 -1.01274617e-01
3.00563842e-01 5.01494527e-01 -1.87631436e-02 -1.18601394e+00
-9.78843570e-01 9.53864217e-01 4.22342896e-01 -1.99923664e-01
9.32569861e-01 -8.11602354e-01 -5.54803252e-01 7.99908698e-01
8.12003970e-01 -4.55575399e-02 -6.98017120e-01 -3.51227641e-01
3.16931188e-01 -6.22386634e-01 5.91433644e-02 -4.68267083e-01
-6.81095123e-01 8.64394486e-01 4.92658794e-01 -1.38422430e-01
6.61038041e-01 3.67970884e-01 5.68172157e-01 3.21210444e-01
1.63429785e+00 -1.02990782e+00 -2.91618586e-01 9.63438570e-01
7.58395433e-01 -1.40867996e+00 1.41079277e-01 -7.94230163e-01
-5.89243174e-01 8.95595491e-01 6.62031233e-01 -4.59313206e-02
3.94708991e-01 4.54670750e-02 4.43051696e-01 -3.81631702e-01
-1.00769967e-01 -4.62636143e-01 -3.93274203e-02 8.98558736e-01
3.44434716e-02 2.35143110e-01 -3.10341179e-01 8.86447430e-02
-4.14389968e-01 -1.39049977e-01 2.83678263e-01 1.13509798e+00
-1.04806244e+00 -9.91313696e-01 -5.31281233e-01 -5.13704680e-02
2.72584409e-01 5.95747232e-02 -3.61859381e-01 1.02670109e+00
2.32754886e-01 1.04974937e+00 1.67625844e-01 -4.51150537e-01
3.44538242e-01 -1.75798893e-01 4.74711061e-01 -7.47281969e-01
2.49915287e-01 -2.70909786e-01 1.92014854e-02 -9.70481932e-01
-5.16120315e-01 -6.23040199e-01 -1.80595422e+00 -8.72172266e-02
-3.69129777e-01 3.50300938e-01 1.33190358e+00 1.06791997e+00
3.43105376e-01 3.72264355e-01 4.67327759e-02 -1.06985557e+00
1.44325376e-01 -7.20114350e-01 -7.85323441e-01 -4.16069478e-02
2.77481198e-01 -1.34336174e+00 -3.33198488e-01 -5.80408454e-01] | [7.3114824295043945, -2.1383025646209717] |
59cc45c8-dad5-4b18-94c8-16a4699e7c12 | inside-outside-net-detecting-objects-in | 1512.04143 | null | http://arxiv.org/abs/1512.04143v1 | http://arxiv.org/pdf/1512.04143v1.pdf | Inside-Outside Net: Detecting Objects in Context with Skip Pooling and Recurrent Neural Networks | It is well known that contextual and multi-scale representations are
important for accurate visual recognition. In this paper we present the
Inside-Outside Net (ION), an object detector that exploits information both
inside and outside the region of interest. Contextual information outside the
region of interest is integrated using spatial recurrent neural networks.
Inside, we use skip pooling to extract information at multiple scales and
levels of abstraction. Through extensive experiments we evaluate the design
space and provide readers with an overview of what tricks of the trade are
important. ION improves state-of-the-art on PASCAL VOC 2012 object detection
from 73.9% to 76.4% mAP. On the new and more challenging MS COCO dataset, we
improve state-of-art-the from 19.7% to 33.1% mAP. In the 2015 MS COCO Detection
Challenge, our ION model won the Best Student Entry and finished 3rd place
overall. As intuition suggests, our detection results provide strong evidence
that context and multi-scale representations improve small object detection. | ['Kavita Bala', 'C. Lawrence Zitnick', 'Sean Bell', 'Ross Girshick'] | 2015-12-14 | inside-outside-net-detecting-objects-in-1 | http://openaccess.thecvf.com/content_cvpr_2016/html/Bell_Inside-Outside_Net_Detecting_CVPR_2016_paper.html | http://openaccess.thecvf.com/content_cvpr_2016/papers/Bell_Inside-Outside_Net_Detecting_CVPR_2016_paper.pdf | cvpr-2016-6 | ['small-object-detection'] | ['computer-vision'] | [ 1.81360021e-01 -2.39918530e-01 -2.29726568e-01 -2.68727094e-01
-9.98295546e-01 -7.31671274e-01 5.35732865e-01 2.34776020e-01
-7.97122955e-01 6.69919401e-02 1.86996497e-02 -1.00946695e-01
2.99641401e-01 -4.31851983e-01 -1.11585951e+00 -4.18732405e-01
-4.23178747e-02 -2.58767545e-01 6.92133307e-01 -1.40131220e-01
3.15836906e-01 7.83708334e-01 -1.68710518e+00 9.05531347e-01
4.56623375e-01 1.12023830e+00 2.87767798e-01 1.02066767e+00
1.72845706e-01 6.98989749e-01 -1.00189698e+00 -1.05993479e-01
2.69746244e-01 1.28512820e-02 -5.72679102e-01 -3.33888203e-01
1.26170886e+00 -1.15672462e-01 -4.27143633e-01 9.72672880e-01
6.20241761e-01 -1.42346788e-02 4.56808984e-01 -9.07827139e-01
-7.12840199e-01 4.38625306e-01 -9.98138428e-01 8.92562807e-01
-1.44716591e-01 2.65216947e-01 9.73625481e-01 -1.33220756e+00
8.16650331e-01 1.05432677e+00 8.87410522e-01 3.46412599e-01
-1.08641517e+00 -6.24295354e-01 4.97852355e-01 3.17428917e-01
-1.44804227e+00 -2.49803677e-01 1.68305054e-01 -3.78495097e-01
1.50425136e+00 2.31255665e-01 6.08531654e-01 9.43296909e-01
2.11143240e-01 1.37200308e+00 1.17076814e+00 -5.41942716e-01
-2.70824045e-01 2.14910150e-01 4.11563993e-01 5.73321939e-01
3.10982615e-01 1.03583731e-01 -7.51261532e-01 3.37031186e-01
7.38256752e-01 2.62899436e-02 1.73291072e-01 -3.35514426e-01
-1.16492367e+00 6.63483918e-01 1.08462119e+00 2.99551278e-01
-3.49819362e-02 3.68853241e-01 3.67455006e-01 -1.58111900e-01
3.01958740e-01 5.87778091e-01 -1.26653582e-01 -1.53711950e-02
-1.00918090e+00 1.11818053e-01 3.28393757e-01 8.49605620e-01
3.20833772e-01 -2.14667656e-02 -4.11206931e-01 8.66448700e-01
-1.47725791e-01 4.74467188e-01 3.70061874e-01 -7.13119566e-01
5.25215328e-01 7.03704357e-01 1.02575101e-01 -8.25845718e-01
-5.25588095e-01 -1.07818472e+00 -3.46804082e-01 4.67341065e-01
5.75583279e-01 4.94256541e-02 -1.30845475e+00 1.41778994e+00
9.33990926e-02 -4.73099295e-04 -2.78211266e-01 9.91743028e-01
1.04505050e+00 5.93379319e-01 2.52734154e-01 4.91836369e-01
1.58851707e+00 -1.15210998e+00 -4.72661525e-01 -6.57621562e-01
5.32377779e-01 -8.84025276e-01 8.99437487e-01 4.28804249e-01
-9.91746485e-01 -7.07643032e-01 -1.37608850e+00 -2.00563475e-01
-8.45241249e-01 8.88563693e-01 6.07137501e-01 5.77385306e-01
-1.00740552e+00 4.76960391e-01 -8.91812742e-01 -5.20921528e-01
6.43636286e-01 2.61367559e-01 -3.58488858e-01 -1.07921503e-01
-7.21948981e-01 1.08420432e+00 2.11343199e-01 8.47954750e-02
-9.03025746e-01 -8.08125734e-01 -6.64574504e-01 -1.37789845e-02
5.05837679e-01 -2.36165702e-01 1.24622917e+00 -5.05311310e-01
-6.87532544e-01 1.02015972e+00 -2.25580886e-01 -8.00364554e-01
4.70831275e-01 -7.40486860e-01 -4.18527454e-01 2.45364040e-01
2.57736266e-01 9.61473227e-01 7.12094486e-01 -8.24712217e-01
-1.06407988e+00 -4.43326980e-01 -1.07437059e-01 2.05867529e-01
-2.16661096e-01 4.46345478e-01 -7.05062628e-01 -7.72467256e-01
5.18678054e-02 -7.85848320e-01 -9.41964332e-03 -6.00938452e-03
-3.21734130e-01 -4.05352890e-01 9.73908603e-01 -7.06720591e-01
1.06952512e+00 -2.14489436e+00 -8.55218321e-02 -8.74184594e-02
3.48801613e-01 4.20829237e-01 -1.32532150e-01 7.11379275e-02
-9.01494175e-02 2.25968078e-01 2.33427942e-01 -3.93669456e-01
-3.77879217e-02 -4.23262298e-01 -3.75697285e-01 4.60112333e-01
6.09169602e-01 1.08088541e+00 -5.88208258e-01 -1.05746321e-01
2.52641231e-01 7.11182952e-01 -3.60935956e-01 -1.17352344e-01
1.83238924e-01 -8.32682773e-02 -1.58864900e-01 8.84066164e-01
4.13310587e-01 -3.16327721e-01 -3.58507425e-01 -2.95449674e-01
-4.87750977e-01 3.83260489e-01 -1.30001903e+00 1.60859191e+00
-1.04949705e-01 1.21753395e+00 3.11613698e-02 -6.61392808e-01
5.99558771e-01 -1.96753040e-01 -1.22898683e-01 -8.54241669e-01
-1.45160645e-01 9.89300236e-02 -8.99099559e-03 -2.37808190e-02
8.30970228e-01 5.20612895e-01 -6.41460577e-03 2.57091073e-04
1.33137345e-01 1.12066604e-01 2.27429509e-01 4.38490659e-01
1.09468734e+00 7.39549324e-02 3.05703163e-01 -2.90218651e-01
1.03031248e-01 -5.76081388e-02 4.90206391e-01 1.32107472e+00
-4.88650173e-01 8.52160573e-01 4.56588864e-01 -4.73755449e-01
-9.83939528e-01 -1.07522225e+00 -2.08200321e-01 1.49569607e+00
1.66667695e-03 -5.52833140e-01 -7.15997994e-01 -6.86521828e-01
8.58081281e-02 3.77380908e-01 -1.05227613e+00 5.84222935e-03
-6.37509048e-01 -7.09219635e-01 7.81227648e-01 1.14470899e+00
4.10217583e-01 -1.02605903e+00 -8.39758217e-01 -5.39395846e-02
2.26193860e-01 -1.38482320e+00 -4.83796686e-01 3.33353698e-01
-4.67275679e-01 -8.99952173e-01 -8.98046553e-01 -6.90037608e-01
3.73984039e-01 6.30076408e-01 9.58409965e-01 -2.59869099e-01
-1.14854288e+00 2.98131943e-01 -1.56670406e-01 -7.55028307e-01
7.32879862e-02 2.97699273e-01 -4.98207510e-02 -1.58765808e-01
3.56401056e-01 1.16483048e-01 -7.48303771e-01 4.29229617e-01
-6.44726992e-01 -1.75004169e-01 8.82594705e-01 6.51251018e-01
5.73428810e-01 -5.47562778e-01 3.39355618e-01 -5.34406483e-01
1.46439537e-01 6.93409294e-02 -6.58233523e-01 3.44700605e-01
-2.58432090e-01 -8.50949585e-02 5.97365759e-02 -3.58159781e-01
-7.58156419e-01 2.22772971e-01 2.53053725e-01 -4.30241793e-01
-1.03663392e-01 -1.56819180e-01 1.83754504e-01 -2.71186322e-01
1.07141864e+00 -3.25615262e-03 -4.79497105e-01 -4.99068916e-01
4.71513391e-01 4.06799108e-01 8.19673061e-01 -1.87467933e-01
4.25721228e-01 6.06881976e-01 -1.48009449e-01 -7.05270290e-01
-1.07515693e+00 -9.19681251e-01 -6.59022152e-01 -1.24264523e-01
9.91068423e-01 -1.27957141e+00 -6.45508170e-01 4.70340699e-01
-1.02063596e+00 -2.32066229e-01 -2.21069351e-01 3.61675113e-01
-8.66248310e-02 -1.10331133e-01 -7.49479890e-01 -7.75038600e-01
-2.87800372e-01 -1.23415673e+00 1.18824947e+00 5.48664570e-01
-1.60379186e-02 -3.20679247e-01 -4.09261614e-01 3.75627875e-01
5.11057198e-01 1.81322917e-01 1.02427475e-01 -5.12157083e-01
-6.96089208e-01 -4.93883252e-01 -8.15403163e-01 1.91531464e-01
-3.37887943e-01 5.84049821e-02 -1.30304837e+00 -4.81370509e-01
-5.04976690e-01 -3.76791537e-01 1.57580590e+00 5.48346162e-01
1.24579108e+00 2.29461163e-01 -5.83181322e-01 4.90022719e-01
1.24650252e+00 -8.41338858e-02 4.45290238e-01 5.64787447e-01
7.62175202e-01 4.22049940e-01 5.85258245e-01 1.46754429e-01
1.90820366e-01 8.75396013e-01 3.46698046e-01 -1.22247353e-01
-5.84296644e-01 -1.54187664e-01 4.53763872e-01 -4.03989963e-02
6.00301102e-02 1.93781734e-01 -1.10786188e+00 8.38752806e-01
-1.74424374e+00 -9.63680148e-01 4.45075184e-02 2.17036796e+00
5.53076744e-01 5.34078896e-01 4.90912139e-01 -2.17589676e-01
9.34508979e-01 1.62639767e-01 -6.43877864e-01 -4.29288626e-01
-4.38016236e-01 2.02513665e-01 9.77922499e-01 3.41376692e-01
-1.60007989e+00 1.19903052e+00 6.47883511e+00 1.02854466e+00
-1.16899252e+00 1.42259598e-01 6.55152977e-01 -6.43693328e-01
5.22653103e-01 -3.33303779e-01 -1.57937956e+00 -8.57107118e-02
7.74384618e-01 3.16854864e-01 1.44280821e-01 1.01259518e+00
-1.51489019e-01 -2.51120150e-01 -9.56765950e-01 1.11401927e+00
3.39405477e-01 -1.63411105e+00 -3.05221707e-01 -1.07531287e-01
8.54140878e-01 6.27950609e-01 4.46008533e-01 4.75593179e-01
1.76839113e-01 -1.40144444e+00 9.17538226e-01 3.97432446e-01
6.46969497e-01 -7.14656591e-01 5.06671607e-01 2.40050271e-01
-1.45881271e+00 -2.38454074e-01 -2.43142381e-01 6.07962981e-02
-3.23853195e-01 3.80953401e-01 -9.72528279e-01 1.47246286e-01
1.20705879e+00 6.81836903e-01 -1.30833924e+00 1.36099625e+00
-3.84353787e-01 3.79131228e-01 -5.57824075e-01 -1.37949437e-01
4.76054877e-01 5.38044155e-01 6.29928946e-01 1.92453265e+00
-5.68972491e-02 -1.41691998e-01 9.51510593e-02 9.55459654e-01
-1.93643928e-01 -1.03780746e-01 -5.16711056e-01 -2.70745661e-02
3.77024740e-01 1.41023266e+00 -9.00973856e-01 -4.27711874e-01
-2.31639534e-01 8.94048870e-01 5.60850322e-01 3.41336161e-01
-7.50313878e-01 -7.80708790e-01 4.38866526e-01 -1.09420918e-01
7.83135712e-01 -1.94064021e-01 -5.14788210e-01 -9.75393891e-01
2.60622710e-01 -7.02885032e-01 4.78459001e-01 -6.93865597e-01
-8.29191208e-01 3.22654605e-01 -3.05631399e-01 -9.53945160e-01
1.78859085e-01 -9.39913929e-01 -6.14933670e-01 8.39688420e-01
-1.35260510e+00 -1.19785345e+00 -2.97717094e-01 -3.51853184e-02
6.04581475e-01 -9.04134437e-02 3.80776703e-01 2.41329432e-01
-7.59091198e-01 8.14493597e-01 -1.06896739e-02 4.80209410e-01
7.88411200e-01 -1.30337775e+00 8.46308827e-01 1.15901983e+00
5.34812033e-01 7.61422515e-01 4.50939626e-01 -4.81379002e-01
-1.28526795e+00 -1.04612172e+00 6.56747818e-01 -8.41391087e-01
5.33226490e-01 -5.69860280e-01 -7.41803706e-01 6.93035185e-01
7.38423765e-02 5.56506217e-01 2.27188677e-01 2.68783748e-01
-8.96931171e-01 -6.79946765e-02 -9.29964185e-01 4.78207767e-01
9.53402460e-01 -7.95091808e-01 -6.47685111e-01 2.65134335e-01
6.91793442e-01 -6.84821188e-01 -6.29609048e-01 6.82211578e-01
6.92033947e-01 -9.27055657e-01 1.28170025e+00 -6.42028213e-01
2.59633929e-01 -3.96519244e-01 -2.56144941e-01 -9.64710414e-01
-2.76116431e-01 -3.92764688e-01 -9.91278216e-02 7.91151285e-01
6.37273967e-01 -3.41953665e-01 8.35876763e-01 1.96247250e-01
-1.55601859e-01 -8.46636355e-01 -8.08561325e-01 -8.27299356e-01
2.14979686e-02 -6.10096514e-01 -2.97277961e-02 5.74020982e-01
-2.11994946e-01 3.36415797e-01 3.18349563e-02 2.40421727e-01
5.59805691e-01 1.72871560e-01 6.79996014e-01 -7.39321172e-01
-2.71567225e-01 -8.14544559e-01 -5.85452497e-01 -1.09023106e+00
-4.27827090e-01 -7.53617644e-01 -1.73008829e-01 -1.51588798e+00
3.82639766e-01 7.09722638e-02 -6.75589383e-01 5.22017121e-01
-2.72377163e-01 8.68122280e-01 6.91343307e-01 1.01341359e-01
-1.04148352e+00 6.98999316e-02 8.46043825e-01 -2.13468924e-01
-3.21622528e-02 -2.47374147e-01 -6.60898805e-01 7.39965558e-01
6.19736671e-01 -3.22118104e-01 2.56165892e-01 -4.27670449e-01
1.28767997e-01 -5.20036221e-01 6.78197980e-01 -1.26473582e+00
3.59629542e-01 3.07421476e-01 9.96625423e-01 -1.05904484e+00
5.14968753e-01 -1.62352413e-01 -6.98386669e-01 5.48556924e-01
-5.65472960e-01 4.52631488e-02 8.33412945e-01 6.19196951e-01
-8.25797915e-02 1.01321086e-01 8.96695077e-01 -1.76792648e-02
-1.29279983e+00 -7.75267780e-02 -1.58354983e-01 1.13546163e-01
9.13925469e-01 -1.51188225e-01 -6.59418166e-01 -7.20554590e-02
-1.04325914e+00 2.10475758e-01 1.09982729e-01 8.08333039e-01
5.25803864e-01 -1.18308043e+00 -6.12539589e-01 -4.87289717e-03
3.97354364e-01 -3.06019753e-01 2.07439959e-01 7.50364959e-01
-3.31894219e-01 7.59541750e-01 -9.33089405e-02 -9.41405118e-01
-1.56905794e+00 2.56967634e-01 4.20483053e-01 -4.06692661e-02
-6.11851096e-01 1.31812489e+00 3.00592929e-01 -2.81280488e-01
5.93200088e-01 -6.01920843e-01 -3.05940688e-01 2.13195547e-01
9.82733428e-01 4.53783810e-01 2.79636443e-01 -4.82118875e-01
-6.18395388e-01 6.62590802e-01 -4.41163689e-01 -5.03295325e-02
1.21132433e+00 2.05531329e-01 2.56105900e-01 4.75181520e-01
1.21316612e+00 -4.83124554e-02 -1.48405993e+00 -2.44081855e-01
1.30560920e-01 -3.48477036e-01 1.27153441e-01 -1.21060312e+00
-7.49109924e-01 1.11083448e+00 9.86447573e-01 7.88596924e-03
7.27697074e-01 1.69654444e-01 3.80504459e-01 4.73375767e-01
-6.90938234e-02 -1.23915362e+00 1.89358294e-01 6.02401853e-01
1.01448417e+00 -1.37138999e+00 1.10393733e-01 -3.19654524e-01
-6.19346321e-01 8.86722207e-01 8.50067735e-01 -3.32065880e-01
1.44944444e-01 2.83512682e-01 -9.50730518e-02 -2.69370139e-01
-6.05200648e-01 -3.78192127e-01 8.71023357e-01 2.70328045e-01
5.04735410e-01 2.19656661e-01 2.28059590e-01 4.08213317e-01
-3.58378887e-02 -3.81450057e-01 3.53052258e-01 9.32394624e-01
-8.45136583e-01 -6.40324950e-01 -5.04422426e-01 4.07210171e-01
-6.12723887e-01 -1.73200816e-01 -6.85035944e-01 9.69545186e-01
8.71853158e-03 8.18899453e-01 9.40438211e-02 -1.08005218e-01
4.94078517e-01 4.83272187e-02 4.81643558e-01 -7.28206813e-01
-7.65534759e-01 1.32802159e-01 3.94531991e-03 -9.01406586e-01
7.96569884e-02 -6.81869209e-01 -1.34363735e+00 9.29347873e-02
-4.28637713e-01 -3.29664469e-01 1.00398064e+00 7.09068120e-01
6.52678728e-01 7.85041571e-01 8.32692906e-02 -9.08168554e-01
-5.21914840e-01 -9.38312352e-01 -2.19692230e-01 8.88857841e-02
5.92995942e-01 -4.70215648e-01 -1.42386317e-01 -3.16858888e-01] | [9.049690246582031, 0.4735760986804962] |
d6c099f5-92dd-408f-aef8-245f35a983e8 | immune-system-approaches-to-intrusion | 1305.7144 | null | http://arxiv.org/abs/1305.7144v1 | http://arxiv.org/pdf/1305.7144v1.pdf | Immune System Approaches to Intrusion Detection - A Review (ICARIS) | The use of artificial immune systems in intrusion detection is an appealing
concept for two reasons. Firstly, the human immune system provides the human
body with a high level of protection from invading pathogens, in a robust,
self-organised and distributed manner. Secondly, current techniques used in
computer security are not able to cope with the dynamic and increasingly
complex nature of computer systems and their security. It is hoped that
biologically inspired approaches in this area, including the use of
immune-based systems will be able to meet this challenge. Here we collate the
algorithms used, the development of the systems and the outcome of their
implementation. It provides an introduction and review of the key developments
within this field, in addition to making suggestions for future research. | ['Jamie Twycross', 'Julie Greensmith', 'Uwe Aickelin'] | 2013-05-30 | null | null | null | null | ['computer-security'] | ['miscellaneous'] | [ 2.72115707e-01 -3.51292491e-01 2.72676554e-02 1.20800614e-01
6.32812381e-01 -4.65275168e-01 6.50465608e-01 5.95999956e-01
-5.45691431e-01 8.40626180e-01 -3.01347822e-01 -3.23505580e-01
-1.42128587e-01 -9.65437889e-01 1.06636882e-01 -7.51894653e-01
-4.29101199e-01 5.21547794e-01 5.65869331e-01 -6.69994295e-01
6.43351197e-01 9.55391884e-01 -1.95389581e+00 -1.23059817e-01
4.01220173e-01 8.81469727e-01 -1.57662719e-01 9.10543621e-01
-1.93489894e-01 5.71176052e-01 -9.09680545e-01 1.01982556e-01
2.74035148e-02 -7.89769232e-01 -6.23641729e-01 -3.72689784e-01
-5.56105614e-01 1.82162911e-01 5.67372628e-02 4.76202160e-01
5.78531682e-01 1.83263168e-01 7.92123199e-01 -1.10859883e+00
-1.16973564e-01 -2.99601644e-01 -2.21275017e-01 4.78268713e-01
5.97346663e-01 2.81621958e-03 1.13880984e-01 -4.85952020e-01
4.80151087e-01 1.21926379e+00 4.94347900e-01 1.01658940e+00
-9.54576671e-01 -4.94732201e-01 -6.39325902e-02 3.30070183e-02
-1.18976033e+00 -4.20863152e-01 5.26100338e-01 -2.48479024e-01
1.36958170e+00 5.77511430e-01 9.58921611e-01 7.20486641e-01
7.21474290e-01 3.35763067e-01 1.26066089e+00 -6.38195157e-01
7.29674280e-01 3.76802146e-01 1.18688829e-01 4.40376192e-01
5.64714611e-01 4.81866330e-01 -3.46820146e-01 -5.61862648e-01
4.99639571e-01 -7.27917999e-02 3.57843488e-02 -3.07624996e-01
-7.34433055e-01 8.81848216e-01 1.83353692e-01 9.98927951e-01
-6.39576674e-01 -1.21078759e-01 5.55224538e-01 2.80265719e-01
1.15679830e-01 5.65696597e-01 -4.51650530e-01 8.72151554e-03
-4.41589296e-01 2.07737014e-01 9.71745789e-01 -3.85430269e-02
1.87181503e-01 2.94409335e-01 5.93087971e-01 5.59110165e-01
4.66162086e-01 3.24752301e-01 5.38643420e-01 -5.39842248e-01
-6.57039523e-01 7.77178407e-01 -3.14196974e-01 -1.20454693e+00
-5.73649228e-01 -2.16116637e-01 -9.28863704e-01 7.70923972e-01
2.89272368e-01 -5.92507944e-02 -6.80700660e-01 1.42490888e+00
7.53195405e-01 -1.20241962e-01 2.36608759e-01 3.52618873e-01
3.29780132e-01 4.81610835e-01 3.20191175e-01 -2.96565175e-01
1.55254221e+00 -2.66503870e-01 -6.46499991e-01 -9.53070372e-02
4.55772094e-02 -7.35004544e-01 1.78630501e-01 3.94775242e-01
-8.00037980e-01 -2.24618167e-01 -1.16620469e+00 8.00521910e-01
-8.75573874e-01 -9.82653022e-01 9.38857079e-01 1.34845507e+00
-9.20298576e-01 2.59616524e-01 -1.00795352e+00 -1.09979510e+00
6.53784573e-02 6.63642108e-01 -1.44904315e-01 1.95727974e-01
-1.32373595e+00 1.19234693e+00 4.34795678e-01 -1.81178525e-02
-2.84978718e-01 -1.50997460e-01 -6.28824830e-01 -4.31622982e-01
6.40146807e-03 -8.85294318e-01 5.60158551e-01 -6.65235817e-01
-1.39011860e+00 8.80433440e-01 -3.44897620e-02 -6.99739695e-01
2.75731623e-01 3.71615708e-01 -2.65747726e-01 1.59994841e-01
-5.57874501e-01 4.28748488e-01 5.02780855e-01 -1.07322848e+00
-6.18257880e-01 -5.62314808e-01 -2.89641470e-01 -2.07681824e-02
-3.04880083e-01 4.32379186e-01 1.96202978e-01 -5.14805377e-01
-1.05876468e-01 -7.34094262e-01 -6.66031182e-01 -2.49718912e-02
1.58259913e-01 -1.69072196e-01 9.97305155e-01 -1.21752068e-01
1.01720095e+00 -1.79165900e+00 -1.24324448e-01 3.23086917e-01
9.37911496e-02 8.69772613e-01 2.10756093e-01 1.15850127e+00
6.22387901e-02 2.83003330e-01 -2.48612940e-01 2.43533939e-01
-3.78467530e-01 3.31740081e-01 -2.29536965e-01 4.94272590e-01
2.65629858e-01 3.95712882e-01 -6.78205609e-01 -2.72535831e-01
4.32344496e-01 7.41128743e-01 -8.42198581e-02 1.37659207e-01
1.71534449e-01 5.41694283e-01 -6.46356881e-01 8.64331484e-01
3.26998830e-01 2.73690552e-01 1.46361411e-01 6.76810801e-01
-2.17449084e-01 -7.78296813e-02 -1.05739939e+00 9.05204892e-01
-7.38408044e-02 2.63856739e-01 3.72902900e-01 -1.07913125e+00
1.08347905e+00 6.28246427e-01 6.65761709e-01 -8.62266243e-01
2.53367573e-01 2.06009805e-01 2.83954322e-01 -2.58259922e-01
-1.22558929e-01 -4.64586735e-01 4.51908931e-02 8.69019091e-01
-3.80070388e-01 -1.09455779e-01 1.22556373e-01 9.24952421e-03
1.52239251e+00 -2.72124320e-01 8.05789232e-01 -2.31829852e-01
1.04225361e+00 3.47312093e-02 5.47078490e-01 5.76514304e-01
-6.66438520e-01 -1.75607815e-01 3.40337381e-02 -9.81418014e-01
-7.08996415e-01 -7.43587434e-01 -3.04713011e-01 5.45312524e-01
7.02642202e-02 -6.06186986e-02 -9.08012211e-01 -2.20304921e-01
4.16471064e-02 3.01491290e-01 -6.73714519e-01 -4.22008157e-01
-4.17908579e-01 -1.20392060e+00 6.66370451e-01 3.38011328e-03
5.07816911e-01 -1.62136781e+00 -1.31838155e+00 5.48745215e-01
3.27653408e-01 -5.20646036e-01 6.71050489e-01 3.06319475e-01
-1.08891189e+00 -1.22656357e+00 -3.54717970e-01 -5.77702641e-01
5.90082705e-01 1.79324627e-01 8.41635883e-01 9.61326778e-01
-8.99966776e-01 3.64508033e-01 -3.11538041e-01 -8.19321275e-01
-7.11285412e-01 -3.50503892e-01 3.62144172e-01 -3.61353219e-01
6.22887194e-01 -5.54522455e-01 -6.28621280e-01 2.54513830e-01
-1.17941904e+00 -5.80681145e-01 3.81713867e-01 9.40440714e-01
8.40720087e-02 6.69505537e-01 8.89662743e-01 -6.73831880e-01
7.61284292e-01 -4.57680464e-01 -1.90288395e-01 1.05896637e-01
-8.27654123e-01 -3.56368989e-01 4.28978175e-01 -2.71449029e-01
-6.72167301e-01 -7.68525619e-03 -3.36619705e-01 4.42117214e-01
-5.18155694e-01 1.06138639e-01 -7.62302801e-02 -6.60689592e-01
8.38318408e-01 3.51193249e-01 4.60879236e-01 -1.61419079e-01
-1.99500173e-01 7.06615984e-01 2.18708083e-01 -4.38721478e-01
6.91259563e-01 6.45727813e-01 2.71212637e-01 -1.31131637e+00
1.61567688e-01 -5.19134820e-01 -3.91487062e-01 -3.40868324e-01
5.19450784e-01 -3.07704397e-02 -7.53541768e-01 8.44732463e-01
-9.60595429e-01 3.05606909e-02 -1.28181666e-01 9.42745954e-02
-3.82509142e-01 4.11015272e-01 -6.45959496e-01 -1.57709312e+00
-6.16893828e-01 -7.52165735e-01 4.55957979e-01 5.17791450e-01
-5.56596756e-01 -1.17403758e+00 5.87252975e-01 3.31721365e-01
9.55436289e-01 7.33406246e-01 9.18109775e-01 -9.64410722e-01
-8.96539465e-02 -7.04000950e-01 4.03861225e-01 5.19540813e-03
2.71918714e-01 2.64094651e-01 -8.54461491e-01 -3.24647993e-01
5.43272674e-01 -2.55262434e-01 6.14411354e-01 4.96131144e-02
2.34108046e-01 -9.43791717e-02 -6.83747470e-01 -5.05589359e-02
1.39853382e+00 8.62108827e-01 8.10166419e-01 7.26845741e-01
-3.28349233e-01 1.26409209e+00 6.92643344e-01 3.07600707e-01
-1.38345003e-01 3.30356508e-01 6.77684128e-01 -1.89251617e-01
2.67317235e-01 4.32459921e-01 3.04214540e-03 5.01224518e-01
3.46404202e-02 -4.29875672e-01 -1.09216964e+00 2.24941120e-01
-1.58169973e+00 -1.04692447e+00 6.13447465e-02 2.32605982e+00
4.89692539e-01 3.26305598e-01 3.32800210e-01 7.34053016e-01
7.91009665e-01 -4.48242903e-01 -5.48048794e-01 -1.00367856e+00
-2.18437508e-01 5.68176985e-01 2.71778852e-02 3.39831740e-01
-9.27430212e-01 7.74693191e-01 7.70052671e+00 3.02937955e-01
-1.16676128e+00 -2.22589746e-01 4.62906152e-01 3.66683751e-01
3.79989326e-01 -9.29430500e-02 -4.01141614e-01 5.12793064e-01
1.37960088e+00 -8.35573450e-02 3.18769813e-01 2.97478825e-01
1.35702550e-01 -5.81602812e-01 -2.53710598e-01 3.12056452e-01
7.33026564e-02 -1.04624832e+00 -2.30299771e-01 3.12081784e-01
1.17928050e-01 -2.22520113e-01 -2.17060506e-01 -6.45151585e-02
2.94556748e-02 -9.58495677e-01 8.51767212e-02 3.64997119e-01
-6.84118941e-02 -1.04773283e+00 9.58681583e-01 5.76835930e-01
-7.91552782e-01 -1.04312502e-01 -9.19665024e-02 -5.92545271e-01
5.65767810e-02 4.08981383e-01 -6.74334347e-01 3.69355083e-01
6.33780122e-01 -5.93817756e-02 -4.48803157e-01 1.31113243e+00
1.63335845e-01 2.33926535e-01 -5.34982145e-01 -5.13350129e-01
-7.93333501e-02 -4.51958291e-02 5.80648363e-01 1.01368523e+00
-1.98581904e-01 2.13699475e-01 7.43615851e-02 2.87517548e-01
9.00506020e-01 2.63599098e-01 -7.54042208e-01 -2.76601225e-01
4.78458554e-01 1.11428213e+00 -1.43757737e+00 -2.99834520e-01
6.71605021e-02 7.07091153e-01 -1.64244801e-01 -3.65328901e-02
-3.37991983e-01 -6.34582937e-01 9.65862155e-01 7.13234320e-02
-1.97342515e-01 -2.71154404e-01 -3.55005354e-01 -5.44848442e-01
-4.37377453e-01 -1.15338135e+00 5.66753268e-01 -3.79588678e-02
-1.11989450e+00 8.95436466e-01 -2.96194553e-01 -6.16353512e-01
-3.37351978e-01 -6.03695512e-01 -6.61892891e-01 5.73827803e-01
-9.50622737e-01 -8.34350765e-01 9.12641436e-02 2.62824953e-01
1.96377039e-01 -2.52572209e-01 1.50458622e+00 -6.32600933e-02
-6.28710985e-01 1.48988783e-01 -1.37915062e-02 -2.80984342e-01
4.92771357e-01 -8.56998682e-01 3.24499995e-01 6.84750736e-01
-1.83491290e-01 1.12670851e+00 1.14100027e+00 -7.23224640e-01
-1.50815749e+00 -4.58760858e-01 7.16173828e-01 -3.53115320e-01
4.81718928e-01 -4.78437841e-01 -8.96157980e-01 -1.96336031e-01
2.88218230e-01 -3.66014630e-01 1.05313933e+00 -8.13249126e-02
3.70349102e-02 1.63389027e-01 -1.78736460e+00 5.12093186e-01
3.34930688e-01 3.46401259e-02 -5.72150767e-01 2.27189690e-01
6.80342019e-02 1.73436433e-01 -6.98141038e-01 4.51577067e-01
7.39705265e-01 -1.18356502e+00 9.68317807e-01 -5.32915950e-01
-4.39058512e-01 -4.50474977e-01 2.57099211e-01 -7.48481691e-01
-1.53741494e-01 -7.07892954e-01 -1.31499916e-01 1.08294308e+00
6.91912919e-02 -1.37914705e+00 9.29521739e-01 4.95311260e-01
4.93457884e-01 -1.02273107e+00 -1.16079021e+00 -6.49651885e-01
1.10100515e-01 -9.42238420e-02 3.49391431e-01 7.92304873e-01
3.12295496e-01 1.67734772e-01 -1.96491107e-02 -2.07879275e-01
7.39000320e-01 -3.07820201e-01 6.10565901e-01 -1.56211591e+00
-3.18671837e-02 -5.34088433e-01 -1.18053520e+00 1.58438042e-01
-1.80625081e-01 -2.60924369e-01 -1.81022614e-01 -1.33327448e+00
-6.48938641e-02 -4.93590176e-01 -4.79212672e-01 2.89634854e-01
1.08097628e-01 5.87633729e-01 -7.59866312e-02 2.14061245e-01
-1.47008017e-01 5.99622838e-02 5.82472682e-01 2.20350012e-01
-2.21736237e-01 5.56543795e-03 -6.76934838e-01 6.98103249e-01
1.22823977e+00 -5.80657780e-01 -1.65446386e-01 4.86197084e-01
-8.00996833e-03 -2.39082009e-01 -2.92872805e-02 -1.25430453e+00
3.82253647e-01 -3.00294369e-01 6.62049234e-01 -2.30769694e-01
4.65460390e-01 -1.00330353e+00 3.44767183e-01 1.44930029e+00
2.71078408e-01 2.86651224e-01 3.72019529e-01 6.84255004e-01
-1.29988745e-01 -2.30069488e-01 1.14840317e+00 -2.41927743e-01
-4.90764529e-01 -2.37913698e-01 -1.34505260e+00 -5.66479743e-01
1.65422058e+00 -7.27269530e-01 -1.45662457e-01 -7.71178156e-02
-5.79696834e-01 2.73476809e-01 1.06041372e+00 5.02708018e-01
5.53236365e-01 -6.59103274e-01 -2.63280392e-01 4.95399743e-01
-9.66045931e-02 -6.41467452e-01 1.17136478e-01 5.79367876e-01
-9.55410540e-01 6.98453009e-01 -8.48891616e-01 -2.99917996e-01
-1.50655556e+00 9.07337725e-01 5.01425266e-01 -1.94616497e-01
-3.76044750e-01 5.66826761e-01 -5.70931621e-02 -2.92784870e-01
6.38841391e-02 6.83910668e-01 -5.81797719e-01 -2.43560538e-01
9.28512156e-01 2.12138012e-01 9.43080559e-02 -6.30374312e-01
-9.27292705e-01 4.22605246e-01 4.87610213e-02 -1.30928025e-01
1.23291576e+00 -6.94855154e-02 -6.95028663e-01 2.63694167e-01
4.25291419e-01 -1.78504586e-01 -3.84734601e-01 3.51338834e-01
3.18739146e-01 -2.32403323e-01 -1.87517628e-01 -8.31038117e-01
-7.14599669e-01 6.73326254e-01 7.80475497e-01 5.01367748e-01
1.39355505e+00 -5.27376235e-01 6.71697438e-01 1.71435952e-01
6.41367733e-01 -7.84941614e-01 2.53553763e-02 3.35404843e-01
4.77176249e-01 -7.03554630e-01 9.27220136e-02 -3.22621167e-01
-1.31164521e-01 1.13697910e+00 4.92017090e-01 -3.96205544e-01
6.56172752e-01 5.33519924e-01 2.93602645e-01 -2.35746145e-01
-1.16646266e+00 -1.90941811e-01 -2.59077609e-01 1.17838085e+00
5.95812917e-01 -3.96096259e-01 -8.58611107e-01 -1.98214918e-01
1.44743666e-01 -2.59788454e-01 4.23691213e-01 1.66464627e+00
-8.25139940e-01 -1.79434860e+00 -8.90672088e-01 -4.35545715e-03
-6.92885578e-01 4.69517946e-01 -1.20829833e+00 7.48300016e-01
4.00668532e-02 1.33182681e+00 -2.20992342e-01 -3.79050583e-01
1.40666530e-01 1.67193204e-01 3.60798180e-01 -4.66950864e-01
-1.02913690e+00 -8.60522762e-02 -1.17788292e-01 -3.98360074e-01
-3.90448213e-01 -6.95528984e-01 -1.32695770e+00 -5.50818384e-01
-2.69261837e-01 5.28639257e-01 1.19387460e+00 7.73337543e-01
3.93995583e-01 4.18774903e-01 5.28188527e-01 -6.44660771e-01
-9.00052935e-02 -6.31644845e-01 -5.82703769e-01 -1.40144765e-01
2.21663956e-02 -7.21810281e-01 -4.15315926e-01 -2.96697021e-01] | [5.70191764831543, 4.102051258087158] |
22dbae09-21fe-4a74-bfb2-610ace161146 | multi-robot-coordination-and-layout-design | 2305.06436 | null | https://arxiv.org/abs/2305.06436v2 | https://arxiv.org/pdf/2305.06436v2.pdf | Multi-Robot Coordination and Layout Design for Automated Warehousing | With the rapid progress in Multi-Agent Path Finding (MAPF), researchers have studied how MAPF algorithms can be deployed to coordinate hundreds of robots in large automated warehouses. While most works try to improve the throughput of such warehouses by developing better MAPF algorithms, we focus on improving the throughput by optimizing the warehouse layout. We show that, even with state-of-the-art MAPF algorithms, commonly used human-designed layouts can lead to congestion for warehouses with large numbers of robots and thus have limited scalability. We extend existing automatic scenario generation methods to optimize warehouse layouts. Results show that our optimized warehouse layouts (1) reduce traffic congestion and thus improve throughput, (2) improve the scalability of the automated warehouses by doubling the number of robots in some cases, and (3) are capable of generating layouts with user-specified diversity measures. We include the source code at: https://github.com/lunjohnzhang/warehouse_env_gen_public | ['Jiaoyang Li', 'Stefanos Nikolaidis', 'Varun Bhatt', 'Matthew C. Fontaine', 'Yulun Zhang'] | 2023-05-10 | null | null | null | null | ['layout-design', 'multi-agent-path-finding'] | ['computer-vision', 'playing-games'] | [-4.92613018e-01 2.08596244e-01 3.09634864e-01 -2.12872684e-01
-4.05702651e-01 -9.96393800e-01 1.44733325e-01 8.53361964e-01
-1.09123558e-01 9.21545446e-01 3.10607366e-02 -4.02095914e-01
-5.55239439e-01 -1.36885667e+00 -7.10890770e-01 -4.61884648e-01
-3.14605564e-01 1.19875526e+00 3.77333343e-01 -6.79618299e-01
2.38760009e-01 5.66071272e-01 -1.33973718e+00 -1.43008158e-01
1.03624189e+00 5.23414254e-01 7.58782506e-01 7.11292028e-01
-2.54710436e-01 1.73033103e-01 -9.04066026e-01 -5.10185733e-02
5.73239028e-01 -3.24880034e-01 -8.19995761e-01 -1.75859556e-02
-6.33978307e-01 -2.36281216e-01 -7.70851411e-03 7.22864330e-01
4.60344195e-01 4.04696614e-02 4.33216810e-01 -1.92293191e+00
-5.18612087e-01 1.12766123e+00 -7.53000557e-01 -2.84341246e-01
2.73546159e-01 5.27307577e-02 5.36784291e-01 -2.51505941e-01
7.37138212e-01 1.28577769e+00 2.29099408e-01 -1.33619905e-01
-1.27557874e+00 -3.46396595e-01 7.53699690e-02 7.38592297e-02
-1.36488914e+00 -1.61385417e-01 4.33823884e-01 1.98158681e-01
8.52640092e-01 5.36405519e-02 6.72341824e-01 4.74524856e-01
3.70241582e-01 5.53146362e-01 7.05212474e-01 -4.12620157e-01
4.60984886e-01 -8.81583691e-02 -2.62465566e-01 7.59784877e-01
8.47759902e-01 -4.44073796e-01 -1.96058139e-01 6.93773478e-02
7.54947364e-01 -1.74762473e-01 3.18491720e-02 -7.25238979e-01
-1.75870407e+00 1.11276352e+00 6.22762501e-01 1.10355623e-01
-7.29586959e-01 4.90072705e-02 3.19370449e-01 2.08079427e-01
6.14508009e-03 1.06519938e+00 -5.78177392e-01 -9.83094126e-02
-2.56482631e-01 5.90172112e-01 1.23739576e+00 1.46128213e+00
9.04822528e-01 -4.71305221e-01 1.33292720e-01 5.29014528e-01
-1.10350288e-01 7.16348469e-01 -5.56754433e-02 -1.25896633e+00
6.48268044e-01 8.54674876e-01 6.42665207e-01 -1.08729708e+00
-1.02548909e+00 1.73769936e-01 -7.01919496e-01 1.09189101e-01
5.67917109e-01 -4.20791030e-01 -4.95938212e-01 1.25464046e+00
4.23783004e-01 -8.01724494e-01 1.02959029e-01 7.40333498e-01
1.18144222e-01 7.51126528e-01 -2.86216378e-01 -1.88252464e-01
1.38407910e+00 -1.30995154e+00 -8.09757233e-01 1.59416571e-02
9.62679327e-01 -1.04479933e+00 1.07354784e+00 2.92188019e-01
-1.08168626e+00 -1.58692867e-01 -9.68590498e-01 3.88012379e-01
-7.85818279e-01 -6.30710199e-02 1.05350232e+00 5.23382127e-01
-1.20872402e+00 3.04236621e-01 -8.90711606e-01 -8.27147841e-01
9.84245762e-02 3.83140832e-01 -3.39361370e-01 -4.55698907e-01
-7.15667605e-01 1.07474315e+00 4.56530482e-01 8.52311254e-02
-7.53823161e-01 -5.76236308e-01 -7.24869728e-01 1.68193921e-01
7.70571768e-01 -7.19487548e-01 1.03571534e+00 -9.10089724e-03
-1.40525305e+00 -7.25219920e-02 4.43448238e-02 -1.01253010e-01
4.71447885e-01 4.19248976e-02 6.42977431e-02 -8.54393374e-03
3.43189508e-01 1.09665966e+00 -2.09100828e-01 -1.53379083e+00
-8.80887508e-01 -1.04739502e-01 2.93281227e-01 2.60003716e-01
-1.70257479e-01 -3.77259314e-01 1.63255826e-01 3.04569770e-02
1.01466095e-02 -1.00261283e+00 -8.83008599e-01 -3.81834418e-01
-8.41771662e-01 -1.76471576e-01 5.53788781e-01 -1.52140126e-01
6.14686430e-01 -1.67804337e+00 6.43503070e-02 2.78857648e-01
-1.68151423e-01 -2.55625993e-01 -5.41068971e-01 1.07221949e+00
5.03323138e-01 2.34713517e-02 1.57956406e-01 8.94799083e-02
3.65182102e-01 4.76812929e-01 9.87035185e-02 2.77092248e-01
-6.19935170e-02 9.61755753e-01 -1.07965219e+00 -2.40484610e-01
4.66518998e-01 -1.91730745e-02 -4.08268660e-01 1.35695428e-01
-4.34807450e-01 2.91634947e-01 -5.30914605e-01 5.82657456e-01
8.88870656e-01 9.67007503e-03 4.13134992e-01 -3.39922532e-02
-4.27592456e-01 6.58246875e-02 -1.44561911e+00 1.76885223e+00
-4.85096723e-01 1.35418370e-01 1.83910728e-01 -7.71077871e-01
1.10915160e+00 -3.26221764e-01 6.81816161e-01 -6.13721251e-01
1.02530159e-01 1.93752751e-01 1.59511361e-02 -2.99175024e-01
9.06279624e-01 5.69481254e-01 -3.98639977e-01 7.58278430e-01
-3.92822415e-01 -2.12969437e-01 9.17351544e-01 4.70371656e-02
1.39599681e+00 -6.24126084e-02 2.20101267e-01 -5.17559886e-01
-1.32090796e-03 6.64036334e-01 1.92674413e-01 6.82638884e-01
-7.13331699e-02 1.07086144e-01 7.08965540e-01 -4.55500722e-01
-1.17717457e+00 -1.20848453e+00 4.98032629e-01 9.87339020e-01
7.88505793e-01 -2.56241292e-01 -6.73087895e-01 -3.62968266e-01
2.19819993e-01 1.18339980e+00 -1.70307145e-01 2.94344425e-01
-5.78101397e-01 -8.02475274e-01 1.90029055e-01 3.01016480e-01
5.08495688e-01 -9.50329006e-01 -1.09980357e+00 4.41511542e-01
-4.11692530e-01 -1.04028451e+00 -2.53442317e-01 3.62775058e-01
-6.06031120e-01 -1.16413462e+00 -6.93929195e-01 -7.11360753e-01
1.13765514e+00 8.72904658e-01 8.84410679e-01 -8.29093605e-02
-4.20659125e-01 1.45319074e-01 -7.34261096e-01 -7.70666897e-01
-4.63332981e-01 7.92857170e-01 4.89369705e-02 -8.04820955e-01
1.13518964e-02 -5.55761218e-01 -5.51048338e-01 8.46745551e-01
-6.56021535e-01 1.67723417e-01 7.50673890e-01 3.37869525e-01
4.11432743e-01 7.24480867e-01 8.90378654e-01 -4.55420494e-01
9.20746148e-01 -4.60788041e-01 -1.00464845e+00 3.93651783e-01
-6.86415732e-01 3.80882740e-01 6.70573831e-01 -8.14016238e-02
-7.57449508e-01 7.94189051e-02 2.52323180e-01 1.12782575e-01
-4.50012714e-01 3.45729858e-01 -1.49294391e-01 1.50297388e-01
3.52470517e-01 -2.46429801e-01 2.56979465e-01 -1.62394390e-01
7.43617237e-01 8.35277587e-02 -1.59749258e-02 -5.64951479e-01
8.68045330e-01 3.69145066e-01 2.94590890e-01 -4.07070607e-01
-7.13654459e-02 -1.77746415e-01 -8.37777019e-01 -2.00124279e-01
5.94997048e-01 -3.27697694e-01 -1.20122361e+00 1.38733089e-02
-1.10049522e+00 -7.07451880e-01 -2.36560240e-01 2.33963743e-01
-7.35621154e-01 -1.78313002e-01 -3.35645437e-01 -8.29348326e-01
2.96442979e-03 -1.41427612e+00 1.08986247e+00 3.05501997e-01
-2.46767342e-01 -8.41090679e-01 -1.94498196e-01 -4.67220731e-02
7.62400925e-01 4.89826709e-01 1.12080407e+00 -5.18172681e-01
-9.90321815e-01 7.97831193e-02 -1.05843194e-01 -6.36743724e-01
5.09919524e-01 -1.81850821e-01 -3.60426567e-02 -2.32718885e-01
-7.74755955e-01 1.12355977e-01 3.07189189e-02 5.28901935e-01
4.64710504e-01 -3.64877731e-01 -7.44017959e-01 7.25086853e-02
1.43468404e+00 4.42766458e-01 5.21253467e-01 8.77511084e-01
9.91855040e-02 1.10521662e+00 1.17935765e+00 6.03206575e-01
8.29586267e-01 7.45768070e-01 7.39647686e-01 -9.85921249e-02
3.03698391e-01 -3.59710246e-01 2.09256843e-01 7.44122148e-01
4.34133075e-02 -5.69082022e-01 -9.06984150e-01 7.72570729e-01
-2.37933040e+00 -5.15827715e-01 -2.66784936e-01 1.84011650e+00
2.21213937e-01 -1.72608629e-01 4.98583198e-01 6.73135044e-03
6.82875514e-01 -1.89482510e-01 -2.86496162e-01 -4.66001481e-01
-3.67573984e-02 -3.93375099e-01 1.08003330e+00 4.61019337e-01
-6.96200132e-01 9.40389693e-01 6.14081287e+00 3.35526466e-01
-4.13885295e-01 -9.05461982e-02 3.93490344e-01 -2.06016675e-01
-2.59122849e-01 1.28167588e-02 -6.86548233e-01 2.90251523e-01
9.64030266e-01 -5.24630964e-01 8.44005108e-01 8.88951302e-01
4.79069263e-01 -5.32275140e-01 -8.51213694e-01 4.73920107e-01
-3.54277849e-01 -1.39313066e+00 -1.84983209e-01 3.30990762e-01
5.61626315e-01 -3.13390046e-01 -4.36534017e-01 1.42230317e-01
8.94599795e-01 -7.84232497e-01 7.69854665e-01 1.55863151e-01
1.78839073e-01 -1.17756164e+00 1.07846308e+00 3.57454807e-01
-1.34216702e+00 -2.11493418e-01 -7.31615365e-01 -4.33143713e-02
8.62012386e-01 6.88326597e-01 -1.44841242e+00 9.99062717e-01
8.25229168e-01 -1.07308708e-01 -4.64536518e-01 1.24685454e+00
-1.46108314e-01 -1.64390668e-01 -6.57415628e-01 -6.27701402e-01
2.47308835e-01 -4.15145725e-01 3.95786107e-01 7.19299972e-01
6.81577921e-01 -2.71096706e-01 4.98239905e-01 8.40801239e-01
3.74651819e-01 5.49871437e-02 -8.89412403e-01 7.29628354e-02
7.95839727e-01 1.46188688e+00 -1.44154990e+00 2.64154822e-01
2.11297542e-01 7.51031876e-01 4.96641807e-02 1.87376976e-01
-9.55593646e-01 -8.32572758e-01 6.74768031e-01 2.34229490e-02
3.40662837e-01 -6.18303359e-01 -4.47050035e-01 -2.28477836e-01
-9.09224972e-02 -3.52727413e-01 9.95311365e-02 -9.94054914e-01
-8.94748509e-01 3.03333104e-01 3.17204297e-01 -7.75070548e-01
-2.78115332e-01 -6.02813303e-01 -3.14797074e-01 3.91451240e-01
-1.42733610e+00 -1.19254661e+00 -4.82764035e-01 2.22263798e-01
6.61318719e-01 -1.09692961e-01 1.00537777e+00 1.79449096e-01
-4.02467161e-01 2.17321172e-01 1.57415584e-01 -3.19094270e-01
6.48349226e-01 -1.13502109e+00 4.01032358e-01 5.40792823e-01
-3.66700083e-01 4.21015471e-01 1.05890310e+00 -6.44029558e-01
-1.92620182e+00 -1.16414094e+00 5.91578007e-01 -2.53350466e-01
6.04612648e-01 -6.24355018e-01 -2.51173526e-01 6.61132216e-01
5.16581476e-01 -5.80761254e-01 2.28864804e-01 1.90968305e-01
2.63355792e-01 -3.31506938e-01 -1.40668809e+00 8.53584528e-01
1.18798590e+00 7.37537861e-01 4.13187593e-02 7.37834096e-01
9.21567857e-01 -3.20867002e-01 -9.71817672e-01 8.05492401e-02
3.66338849e-01 -7.98916399e-01 7.62202442e-01 -2.21344784e-01
1.65333435e-01 -5.70416927e-01 -1.26185164e-01 -1.98492229e+00
-4.83244121e-01 -7.58631885e-01 4.19831127e-01 1.34323835e+00
7.18161106e-01 -1.04052949e+00 6.17813647e-01 3.66757184e-01
-2.31162995e-01 -6.05326772e-01 -5.58459699e-01 -1.00368738e+00
-2.58940786e-01 1.71891689e-01 1.29930794e+00 4.81880188e-01
3.08307678e-01 2.77061343e-01 1.26982599e-01 4.62628752e-01
5.65623820e-01 2.44436637e-01 1.06437886e+00 -1.00266981e+00
2.55101770e-01 -3.54290783e-01 -8.65802318e-02 -6.74633980e-01
-1.13162920e-01 -5.66861808e-01 4.30763423e-01 -2.36169052e+00
-1.92874372e-01 -9.54101503e-01 2.15662032e-01 6.64744020e-01
3.10850114e-01 -3.01890492e-01 4.57603782e-01 3.19444239e-02
-7.85409272e-01 2.11226597e-01 1.32520306e+00 -2.38334611e-02
-3.93905848e-01 -1.42963320e-01 -8.59981239e-01 2.33704701e-01
1.17860281e+00 -4.31687236e-01 -6.25455141e-01 -6.14016354e-01
3.97234619e-01 8.05577636e-02 -2.36902520e-01 -8.29007804e-01
3.45750868e-01 -5.00589490e-01 2.08830476e-01 -6.30567908e-01
7.91438818e-02 -8.21946740e-01 4.60987061e-01 8.15348208e-01
-8.17640200e-02 9.48745191e-01 2.10415110e-01 4.18177456e-01
2.05982059e-01 -2.98843145e-01 2.80145854e-01 -3.49549443e-01
-5.17714322e-01 -8.93308446e-02 -5.96945643e-01 -5.64517021e-01
1.56844378e+00 8.54779482e-02 -9.57445800e-01 -3.74718040e-01
-2.96918415e-02 8.36572468e-01 7.21258342e-01 4.57325935e-01
3.19760233e-01 -1.04599082e+00 -6.89949751e-01 -5.21642044e-02
-5.38126156e-02 2.63073921e-01 -4.43080254e-02 8.35345685e-01
-1.11974394e+00 5.77782452e-01 -5.75763404e-01 -4.67077345e-01
-7.36099005e-01 8.53303194e-01 -2.09398448e-01 -5.09256303e-01
-5.89513302e-01 4.25794870e-01 -3.05907652e-02 -5.71757913e-01
3.38713452e-02 -2.86556691e-01 1.23183630e-01 7.71413222e-02
3.74778956e-01 7.98484623e-01 -1.79077834e-02 -5.79488575e-02
-4.74395603e-01 2.29856387e-01 1.52505532e-01 -1.93303421e-01
1.70034480e+00 -2.78657198e-01 -2.13049367e-01 7.60644898e-02
4.71282333e-01 -9.85975489e-02 -8.76516819e-01 6.65227652e-01
1.44236475e-01 -4.82805908e-01 -6.04761660e-01 -8.63537073e-01
-9.90536213e-01 2.82527983e-01 2.37956166e-01 7.34446704e-01
1.03321576e+00 -3.21676396e-02 8.07454526e-01 7.03525841e-01
1.29912019e+00 -8.79172027e-01 6.36233762e-02 4.01923329e-01
7.21011817e-01 -8.37322474e-01 -4.72113825e-02 -6.88410223e-01
-8.18519831e-01 1.11605823e+00 5.02255976e-01 -1.34427235e-01
2.62087494e-01 8.68848503e-01 1.21566966e-01 -2.22700104e-01
-6.69194996e-01 -3.50914627e-01 -9.76972222e-01 1.19556105e+00
2.50295363e-02 3.43755215e-01 -2.30583623e-01 2.62711823e-01
-5.75750589e-01 -2.50278831e-01 1.09065747e+00 1.00674117e+00
-6.83544517e-01 -1.26488113e+00 -5.47985315e-01 3.90333235e-01
3.30929011e-01 4.62708533e-01 -1.57882363e-01 1.10482192e+00
-2.06898957e-01 1.36506283e+00 2.16463715e-01 -4.62967344e-02
8.43999028e-01 -2.47843668e-01 5.54725587e-01 -3.09956521e-01
-3.82424414e-01 -1.90751195e-01 5.33451498e-01 -6.53045654e-01
-1.04406185e-01 -5.49240053e-01 -1.40121508e+00 -7.05071867e-01
-2.71090537e-01 3.30261290e-01 9.91105676e-01 4.56038892e-01
8.02324355e-01 7.60734499e-01 4.86640334e-01 -1.11912870e+00
-3.48104656e-01 -8.56490552e-01 -6.79618716e-01 1.61601845e-02
-3.62810344e-01 -9.39432681e-01 1.15458660e-01 -2.74043173e-01] | [4.955246925354004, 1.7139307260513306] |
1a61ccf6-45a0-49b1-b64b-e48b19bcfa63 | tractable-data-enriched-distributionally | 2207.03286 | null | https://arxiv.org/abs/2207.03286v1 | https://arxiv.org/pdf/2207.03286v1.pdf | Tractable Data Enriched Distributionally Robust Chance-Constrained CVR | This paper proposes a tractable distributionally robust chance-constrained conservation voltage reduction (DRCC-CVR) method with enriched data-based ambiguity set in unbalanced three-phase distribution systems. The increasing penetration of distributed renewable energy not only brings clean power but also challenges the voltage regulation and energy-saving performance of CVR by introducing high uncertainties to distribution systems. In most cases, the conventional robust optimization methods for CVR only provide conservative solutions. To better consider the impacts of load and PV generation uncertainties on CVR implementation in distribution systems and provide less conservative solutions, this paper develops a data-based DRCC-CVR model with tractable reformulation and data enrichment method. Even though the uncertainties of load and photovoltaic (PV) can be captured by data, the availability of smart meters (SMs) and micro-phasor measurement units (PMUs) is restricted by cost budget. The limited data access may hinder the performance of the proposed DRCC-CVR. Thus, we further present a data enrichment method to statistically recover the high-resolution load and PV generation data from low-resolution data with Gaussian Process Regression (GPR) and Markov Chain (MC) models, which can be used to construct a data-based moment ambiguity set of uncertainty distributions for the proposed DRCC-CVR. Finally, the nonlinear power flow and voltage dependant load models and DRCC with moment-based ambiguity set are reformulated to be computationally tractable and tested on a real distribution feeder in Midwest U. S. to validate the effectiveness and robustness of the proposed method. | ['Zhaoyu Wang', 'Yi Guo', 'Fankun Bu', 'Qianzhi Zhang'] | 2022-07-07 | null | null | null | null | ['gpr', 'gpr'] | ['computer-vision', 'miscellaneous'] | [-3.29692394e-01 -4.92240906e-01 5.69038186e-03 -8.95105228e-02
-9.82877672e-01 -9.81864035e-01 3.18043798e-01 2.85860777e-01
3.70868474e-01 1.39569998e+00 2.44856589e-02 -4.36969936e-01
-8.02891970e-01 -9.71800327e-01 -3.86262387e-01 -1.13884139e+00
-1.89728945e-01 5.00691533e-01 -4.05717194e-01 2.24341303e-02
6.55694678e-02 7.82547176e-01 -1.13826239e+00 -4.15966839e-01
1.66761458e+00 1.06156206e+00 2.16757715e-01 1.57533720e-01
3.42022061e-01 2.84100324e-01 -1.04547572e+00 1.72597647e-01
2.09124088e-01 3.19218710e-02 6.63744807e-02 -4.77269739e-01
-7.73862958e-01 -3.69910985e-01 -1.28032848e-01 1.31390560e+00
9.11798775e-01 4.27640676e-01 9.42309141e-01 -1.53367114e+00
-6.08460546e-01 6.74041271e-01 -9.75987017e-01 2.68691063e-01
2.97422707e-01 1.05270967e-01 5.94498575e-01 -6.24255359e-01
-4.08627167e-02 1.04809618e+00 3.82687211e-01 -3.42414416e-02
-1.44825828e+00 -5.22268593e-01 1.08859748e-01 1.38143331e-01
-1.77759933e+00 1.55218810e-01 8.34008694e-01 -4.47584897e-01
1.12177372e+00 4.00392711e-01 8.51570845e-01 4.44651425e-01
3.16168994e-01 4.12640423e-01 1.39852226e+00 -4.15463373e-02
6.10930920e-01 -2.34371990e-01 1.37522370e-01 -4.18323964e-01
8.58139336e-01 7.28683174e-02 -9.03705433e-02 -3.77561599e-01
3.25705469e-01 -3.00680995e-01 -5.73015571e-01 -1.74554467e-01
-4.37591463e-01 8.24834526e-01 2.16998056e-01 1.49112374e-01
-4.55032885e-01 -2.00507224e-01 1.61693275e-01 -2.15410024e-01
2.59233862e-01 -5.45278378e-03 -4.99491692e-01 -9.52331647e-02
-1.15540600e+00 1.28075078e-01 7.29274213e-01 1.28792393e+00
1.41288549e-01 1.06667566e+00 -3.68657798e-01 4.24130291e-01
8.03977311e-01 1.45782816e+00 2.44300738e-01 -4.40248519e-01
5.02052844e-01 1.44619867e-01 4.82384354e-01 -5.27991235e-01
-4.41984951e-01 -6.84157312e-01 -1.17384815e+00 3.21049631e-01
1.21706799e-01 -6.07099414e-01 -7.32095063e-01 1.33468032e+00
1.56485364e-01 -2.26209283e-01 2.51772881e-01 6.45353198e-01
2.74139553e-01 1.21042216e+00 -1.40979365e-01 -8.86727393e-01
1.27771425e+00 -3.46084163e-02 -1.15758550e+00 4.61773336e-01
4.27974164e-02 -5.09614229e-01 3.11463892e-01 6.48059607e-01
-1.06826448e+00 -1.71643808e-01 -1.21165645e+00 4.60568458e-01
-2.34127089e-01 -7.91485142e-03 -6.66820332e-02 8.27977657e-01
-4.89458203e-01 5.17285943e-01 -7.92518556e-01 3.22075099e-01
3.59664053e-01 -8.03128406e-02 1.04123786e-01 1.26477867e-01
-1.26475596e+00 1.33697271e+00 3.84005427e-01 8.02729309e-01
-8.01806986e-01 -9.52835262e-01 -7.94434547e-01 4.09999669e-01
2.79830724e-01 -3.01315695e-01 8.91304672e-01 -8.18702579e-02
-1.51066983e+00 -6.64751291e-01 -9.56113711e-02 -6.20390952e-01
4.76407528e-01 5.69989730e-04 -6.99676752e-01 5.46084717e-02
-1.22380055e-01 -4.94910866e-01 5.07978082e-01 -1.26839197e+00
-5.31988740e-01 -3.41784984e-01 -7.09432244e-01 1.36835203e-01
4.45676237e-01 -2.52404094e-01 4.76974279e-01 -7.35643089e-01
4.18328494e-02 -3.91275883e-01 -3.82268965e-01 -8.04990411e-01
-4.87697631e-01 -4.17567343e-01 7.01049030e-01 -1.28434277e+00
1.05311143e+00 -1.65552306e+00 -4.62249964e-02 8.49526644e-01
-5.39525509e-01 5.09794876e-02 3.20358604e-01 7.53243446e-01
-3.12673777e-01 3.01381171e-01 -4.46616590e-01 1.76911175e-01
4.36076254e-01 4.92069513e-01 -3.76953602e-01 7.31140196e-01
1.35414898e-01 7.36122131e-01 -7.52055168e-01 1.23421125e-01
7.26333439e-01 5.04465699e-01 8.33201706e-02 -1.37128159e-01
-1.85989410e-01 4.76335645e-01 -2.71449983e-01 6.98778987e-01
1.33078432e+00 3.30644429e-01 1.73681825e-01 -5.73874533e-01
-3.20028543e-01 -2.69307733e-01 -1.89419270e+00 1.05883026e+00
-4.46962088e-01 2.83590525e-01 2.53247738e-01 -1.17430866e+00
1.02112210e+00 1.42012000e-01 5.35426617e-01 -7.54211009e-01
-1.57309338e-01 1.48199409e-01 1.76679358e-01 -2.17880473e-01
3.60940814e-01 -1.98369890e-01 -3.59778851e-02 7.83930346e-02
1.67936131e-01 -5.40656567e-01 3.39619368e-01 -6.56820014e-02
4.34331954e-01 1.76596314e-01 4.15646344e-01 -7.37931311e-01
5.33473611e-01 -4.09545988e-01 1.40630960e+00 2.23297745e-01
1.19325042e-01 2.37491190e-01 4.24836963e-01 5.76731563e-01
-8.57938945e-01 -1.39840269e+00 -6.16734087e-01 -2.56124169e-01
5.50025962e-02 1.09858692e-01 -3.93220522e-02 -1.94722727e-01
4.84646440e-01 1.83771706e+00 2.55297348e-02 1.75741181e-01
-2.34212503e-01 -1.44285941e+00 1.01308301e-01 6.11803591e-01
3.39013129e-01 -1.53371850e-02 3.19250785e-02 3.66079241e-01
9.14138108e-02 -6.02994442e-01 9.76769626e-02 4.28634703e-01
-6.50251627e-01 -8.61787081e-01 -6.60632014e-01 -8.75355676e-02
7.07637310e-01 -3.82604688e-01 7.57822871e-01 -6.43543482e-01
-9.75955054e-02 3.05737257e-01 -1.62416846e-01 -5.91880918e-01
-5.71096689e-02 -5.49518526e-01 3.09293270e-01 -3.77525508e-01
-2.47131109e-01 -7.17777312e-01 -5.04280210e-01 1.77157491e-01
-6.24573648e-01 -5.47612011e-01 1.14709757e-01 7.85716951e-01
8.12279940e-01 1.09885299e+00 1.37154341e+00 -2.97840059e-01
8.21394444e-01 -6.84547126e-01 -1.54247129e+00 5.10434568e-01
-9.94876564e-01 -3.57654579e-02 7.56307244e-01 -1.99358732e-01
-1.49346364e+00 -1.45381600e-01 -1.40453279e-01 -2.84011155e-01
2.36810111e-02 7.74816036e-01 -7.48789787e-01 5.48804812e-02
-2.45841946e-02 1.84325397e-01 -5.33208787e-01 -6.63405240e-01
4.16838974e-01 6.03908241e-01 4.35908288e-01 -8.49866509e-01
1.28460169e+00 1.48937613e-01 5.14454484e-01 -6.92030311e-01
-3.54751796e-02 -2.01203942e-01 -1.62689388e-01 -3.78640369e-02
4.73573297e-01 -1.13109827e+00 -1.06088758e+00 6.69268072e-01
-7.74981916e-01 7.62087181e-02 -6.75109863e-01 7.77993381e-01
-3.42027128e-01 6.98077142e-01 -2.59315282e-01 -1.59514892e+00
-5.05965829e-01 -1.08971238e+00 3.01722080e-01 5.86970210e-01
3.68279219e-01 -8.89576793e-01 -6.39256537e-02 -1.08765112e-03
4.39491808e-01 7.86494911e-01 9.10391271e-01 -5.46150208e-01
-6.67521596e-01 -5.11483960e-02 2.02780534e-02 8.01198781e-01
5.19465357e-02 3.09179962e-01 -4.80411977e-01 -5.70122123e-01
2.43234858e-01 4.13397223e-01 -9.80824754e-02 7.39492297e-01
8.44108522e-01 -3.42803389e-01 -1.78262323e-01 4.70837444e-01
2.26880884e+00 7.33170092e-01 6.91915929e-01 -1.09450676e-01
2.77839571e-01 2.65126735e-01 6.98829532e-01 8.23306322e-01
5.13125360e-01 1.76621363e-01 4.08874810e-01 3.72495234e-01
5.78329563e-01 -1.42789543e-01 1.14417955e-01 8.76315773e-01
-5.63263446e-02 -6.27239347e-01 -7.40480185e-01 7.17437685e-01
-1.76372361e+00 -9.34511244e-01 -6.03008270e-01 2.36930537e+00
5.48109889e-01 -1.43985316e-01 -3.30456823e-01 2.54219264e-01
8.02456200e-01 -1.67357773e-01 -5.65002143e-01 -4.93325144e-01
-5.54505050e-01 2.70774275e-01 8.98228765e-01 5.28711140e-01
-3.94864380e-01 -3.87125909e-01 5.04818249e+00 1.37620938e+00
-5.78579366e-01 1.45937526e-03 4.04621512e-01 4.45071794e-02
-7.12176979e-01 1.05077282e-01 -8.12472224e-01 1.03284049e+00
9.75350916e-01 -9.86974835e-01 6.06975198e-01 6.24118030e-01
9.69541490e-01 -7.85432160e-01 -8.32945585e-01 9.89656508e-01
-3.36591482e-01 -8.97858143e-01 -2.33711213e-01 8.51284787e-02
1.22549748e+00 7.34377354e-02 -3.37680757e-01 3.59292924e-01
5.25967062e-01 -9.44621861e-01 5.67806482e-01 9.70003724e-01
2.67929941e-01 -1.32070422e+00 1.08536577e+00 4.50303525e-01
-1.20191419e+00 -5.48835397e-01 -2.74267703e-01 1.74529105e-01
9.90276873e-01 1.27583361e+00 -2.28055239e-01 1.62058079e+00
7.88353860e-01 2.96774983e-01 -8.38919207e-02 1.34230328e+00
-5.56586564e-01 8.01237881e-01 -7.55597711e-01 1.07356779e-01
-4.20065045e-01 -8.43460977e-01 7.73927629e-01 6.49857283e-01
8.81344378e-01 2.74028957e-01 -1.41949892e-01 8.88505220e-01
2.39419729e-01 -1.47810310e-01 -2.54371971e-01 3.16274375e-01
9.58694756e-01 1.38651073e+00 -2.58468419e-01 -4.10650820e-02
-3.64894420e-01 1.08528502e-01 -4.84335721e-01 8.76591384e-01
-8.75907063e-01 -4.34937894e-01 4.58046198e-01 -1.53316602e-01
2.67159253e-01 -2.46582389e-01 -5.89117050e-01 -1.19267893e+00
1.45985276e-01 -5.68568408e-01 5.60387373e-01 -1.16706371e+00
-1.92701674e+00 2.21569762e-02 4.63552207e-01 -1.08873367e+00
-6.62104130e-01 -1.95773587e-01 -8.90621722e-01 1.74700344e+00
-1.79811013e+00 -8.23158622e-01 1.51984841e-01 4.26247954e-01
1.84437588e-01 -2.94613261e-02 6.65066838e-01 2.63741165e-01
-7.81650782e-01 9.57822502e-02 1.22363830e+00 -1.50335476e-01
-8.51200521e-02 -1.46885979e+00 -2.90712386e-01 1.23475349e+00
-5.12830198e-01 2.44410887e-01 8.78724575e-01 -8.73411238e-01
-1.61143017e+00 -9.46700275e-01 3.98933254e-02 9.21458527e-02
7.35880494e-01 -2.73549259e-01 -1.00775671e+00 3.63514274e-01
6.46476269e-01 -1.18252844e-01 4.72799957e-01 -4.12386209e-01
2.79867023e-01 -5.38976967e-01 -1.89785528e+00 3.22929382e-01
1.15727272e-03 -2.80263603e-01 -8.06694567e-01 1.08166806e-01
2.46231318e-01 -2.04624981e-01 -1.56885409e+00 8.29762816e-01
5.71339615e-02 -6.08364828e-02 8.02557707e-01 1.00081198e-01
-6.14274323e-01 -1.04728949e+00 -4.40492362e-01 -2.01696730e+00
-4.34974879e-02 -8.67189229e-01 -4.83122647e-01 1.99836290e+00
1.24002300e-01 -1.27697301e+00 4.00991440e-02 8.13009918e-01
-3.82677205e-02 -4.11798954e-01 -1.64413905e+00 -1.09449494e+00
4.06306207e-01 -3.61936927e-01 1.07031119e+00 7.65870035e-01
1.97662383e-01 -1.65308565e-01 4.52200584e-02 1.02478588e+00
1.24374759e+00 2.53903151e-01 1.99800357e-01 -1.03644252e+00
-6.18813634e-02 -1.21482342e-01 -1.63657125e-02 -3.88268381e-01
-2.31390104e-01 -5.73265433e-01 -3.54627557e-02 -2.10864329e+00
-3.12919497e-01 -4.50236410e-01 -1.32932454e-01 -2.31206194e-02
4.86323014e-02 -4.04402256e-01 2.46210188e-01 -4.17558476e-02
3.79255295e-01 1.16429555e+00 6.39349461e-01 -6.79697767e-02
5.16181514e-02 3.16191733e-01 -1.77623481e-01 6.45062804e-01
9.69847560e-01 -1.82153270e-01 -6.48969591e-01 1.15728952e-01
3.74132842e-01 5.56099355e-01 5.25721721e-03 -8.64155471e-01
2.26002887e-01 -2.08650380e-01 7.62151062e-01 -1.58842278e+00
4.19711471e-02 -1.30685627e+00 8.80576193e-01 3.94814640e-01
4.77883160e-01 1.31947845e-01 3.92966777e-01 7.05422163e-01
-1.23289280e-01 -5.34580410e-01 6.22509301e-01 3.45385790e-01
-2.20101178e-01 2.36080214e-03 -6.86047733e-01 -1.67609408e-01
1.33098173e+00 4.99892309e-02 -7.94266403e-01 -4.33064014e-01
-8.62805605e-01 1.04380298e+00 7.13125467e-02 6.75061122e-02
2.54597902e-01 -1.28009069e+00 -7.30332077e-01 -4.13830020e-03
-5.56739092e-01 3.01999032e-01 6.68400824e-01 5.21333814e-01
-1.97476625e-01 4.33359802e-01 1.70321837e-01 -6.32070184e-01
-6.27326369e-01 6.59551561e-01 5.05680561e-01 -2.62745231e-01
-1.20202266e-01 8.66151939e-04 -4.77220178e-01 -2.32409239e-01
-1.61867201e-01 -6.52729154e-01 -1.45230606e-01 5.29698372e-01
1.47394255e-01 1.04977703e+00 2.48518050e-01 -3.99122387e-01
-2.91413486e-01 3.05633754e-01 6.84430718e-01 -1.33374497e-01
1.35797584e+00 -5.84280074e-01 -1.37789398e-01 3.31996530e-01
7.47323811e-01 2.01243103e-01 -1.09518075e+00 2.95081824e-01
-1.74406067e-01 -2.94296175e-01 6.01941608e-02 -1.29173970e+00
-1.14044702e+00 5.93967855e-01 6.17387414e-01 5.07292628e-01
1.20812488e+00 -7.57860720e-01 1.62769347e-01 -6.03234395e-03
6.66597486e-01 -1.39456725e+00 -9.63851452e-01 4.77565825e-02
8.96340132e-01 -6.39869034e-01 3.02463323e-01 -2.01195523e-01
-4.50950921e-01 8.32837522e-01 1.69081941e-01 -8.25123563e-02
9.26148415e-01 7.38008916e-01 -5.88147700e-01 4.09854621e-01
-4.73968267e-01 -3.27319801e-02 3.12352721e-02 9.95378852e-01
-2.54791945e-01 3.11401427e-01 -6.90854430e-01 1.05913496e+00
1.79488529e-02 -7.34519819e-03 8.45784009e-01 8.96871209e-01
-3.23158242e-02 -7.45767653e-01 -7.02144921e-01 6.29888952e-01
-5.14734626e-01 -2.02517118e-02 1.02410769e+00 7.12908745e-01
1.02445990e-01 1.28434515e+00 3.00786812e-02 5.61490595e-01
7.06975937e-01 -1.17682144e-01 2.87564129e-01 -1.62341148e-01
-1.70868054e-01 3.27223092e-01 1.16668738e-01 6.88530877e-02
-2.54913978e-02 -9.68358159e-01 -1.38338006e+00 -3.52288336e-01
-8.46699715e-01 7.22199440e-01 9.04789805e-01 7.67300725e-01
4.60094260e-03 6.53840661e-01 9.63704884e-01 -5.56256235e-01
-1.04119503e+00 -9.58574116e-01 -1.26950777e+00 -3.21083575e-01
-7.67238885e-02 -5.88822365e-01 -7.35308826e-01 -5.10768116e-01] | [5.707006931304932, 2.608750104904175] |
21126e35-812d-4115-b5c8-dbb7472267f1 | abn-agent-aware-boundary-networks-for | 2203.08942 | null | https://arxiv.org/abs/2203.08942v1 | https://arxiv.org/pdf/2203.08942v1.pdf | ABN: Agent-Aware Boundary Networks for Temporal Action Proposal Generation | Temporal action proposal generation (TAPG) aims to estimate temporal intervals of actions in untrimmed videos, which is a challenging yet plays an important role in many tasks of video analysis and understanding. Despite the great achievement in TAPG, most existing works ignore the human perception of interaction between agents and the surrounding environment by applying a deep learning model as a black-box to the untrimmed videos to extract video visual representation. Therefore, it is beneficial and potentially improve the performance of TAPG if we can capture these interactions between agents and the environment. In this paper, we propose a novel framework named Agent-Aware Boundary Network (ABN), which consists of two sub-networks (i) an Agent-Aware Representation Network to obtain both agent-agent and agents-environment relationships in the video representation, and (ii) a Boundary Generation Network to estimate the confidence score of temporal intervals. In the Agent-Aware Representation Network, the interactions between agents are expressed through local pathway, which operates at a local level to focus on the motions of agents whereas the overall perception of the surroundings are expressed through global pathway, which operates at a global level to perceive the effects of agents-environment. Comprehensive evaluations on 20-action THUMOS-14 and 200-action ActivityNet-1.3 datasets with different backbone networks (i.e C3D, SlowFast and Two-Stream) show that our proposed ABN robustly outperforms state-of-the-art methods regardless of the employed backbone network on TAPG. We further examine the proposal quality by leveraging proposals generated by our method onto temporal action detection (TAD) frameworks and evaluate their detection performances. The source code can be found in this URL https://github.com/vhvkhoa/TAPG-AgentEnvNetwork.git. | ['Ngan Le', 'Akihiro Sugimoto', 'Minh-Triet Tran', 'Sang Truong', 'Kashu Yamazaki', 'Khoa Vo'] | 2022-03-16 | null | null | null | null | ['temporal-action-proposal-generation'] | ['computer-vision'] | [ 5.29624373e-02 -2.99178988e-01 -1.40115857e-01 5.37351705e-02
-2.74740577e-01 -4.36484605e-01 8.39689493e-01 -3.25454742e-01
-4.33040053e-01 5.65201938e-01 3.69032085e-01 3.54833975e-02
-1.74068157e-02 -8.22654784e-01 -6.52217507e-01 -7.78975546e-01
-5.60498536e-01 1.52879328e-01 6.95108891e-01 -1.97061390e-01
-3.28655764e-02 3.48084152e-01 -1.44858277e+00 3.95431012e-01
4.46612328e-01 9.99500751e-01 1.07669972e-01 8.68381143e-01
3.36673170e-01 1.41490424e+00 -6.48821771e-01 1.90109000e-01
3.23955834e-01 -5.13922513e-01 -3.29798430e-01 2.86282122e-01
-3.17946523e-02 -8.58546078e-01 -8.37636232e-01 9.05602098e-01
3.87394488e-01 2.82902092e-01 4.17424023e-01 -1.73131061e+00
-2.27745011e-01 4.65180784e-01 -5.10570109e-01 5.21777093e-01
5.35438478e-01 8.22414219e-01 5.67292511e-01 -6.54331267e-01
8.40283155e-01 1.42313504e+00 3.28295827e-01 3.38433355e-01
-7.82231867e-01 -6.85262024e-01 6.70858502e-01 7.52487361e-01
-1.14489746e+00 -4.99706358e-01 7.18809307e-01 -3.42544585e-01
9.82933640e-01 1.39554087e-02 8.74648988e-01 1.48164856e+00
1.89229205e-01 9.17845964e-01 7.33862162e-01 6.70853630e-02
2.11035147e-01 -3.61575574e-01 -2.82153457e-01 6.70151651e-01
-2.32459486e-01 4.25190002e-01 -6.76153839e-01 1.24064147e-01
1.16382563e+00 1.02313444e-01 -4.49373662e-01 -1.70246124e-01
-1.60899794e+00 4.86934304e-01 4.28244203e-01 5.81765808e-02
-7.69082487e-01 4.90317136e-01 5.70282400e-01 2.59182841e-01
3.31354767e-01 -1.36644810e-01 -1.95985690e-01 -4.36349243e-01
-4.89576191e-01 2.77082860e-01 5.63035548e-01 8.00109267e-01
4.28798974e-01 3.25077474e-01 -5.36064565e-01 3.69909346e-01
4.02593255e-01 4.44253176e-01 2.41299495e-01 -1.34714234e+00
4.14776206e-01 6.20337546e-01 2.58950949e-01 -1.17050314e+00
-3.59701574e-01 -3.17247301e-01 -7.50172615e-01 5.98011971e-01
3.97697806e-01 -3.72206658e-01 -8.06522667e-01 1.89893246e+00
6.00359380e-01 7.73613632e-01 8.13820288e-02 1.15699542e+00
7.90715396e-01 9.42385793e-01 6.68376386e-02 -3.24609727e-01
1.33306110e+00 -1.37997365e+00 -7.39363313e-01 -1.65056482e-01
4.30309951e-01 -4.73658681e-01 6.06319249e-01 3.82669091e-01
-1.04980826e+00 -7.00890183e-01 -1.05972433e+00 2.63015211e-01
-3.30020636e-01 1.80739313e-01 4.20891792e-01 -1.15439117e-01
-1.04443741e+00 3.21939945e-01 -9.91518736e-01 -4.89266932e-01
3.63053858e-01 2.00345382e-01 -3.54705989e-01 -4.31407131e-02
-1.36563110e+00 5.66928685e-01 4.35111970e-01 2.53134042e-01
-1.64072394e+00 -2.80613154e-01 -8.91522169e-01 3.81340086e-02
9.21401739e-01 -6.10152185e-01 1.16486406e+00 -1.07148731e+00
-1.52906537e+00 9.89331603e-02 1.22507848e-02 -6.82277858e-01
9.39764559e-01 -8.69247913e-02 -4.78122830e-01 6.33898377e-01
1.62116572e-01 6.95015430e-01 8.07608545e-01 -9.98551548e-01
-8.49538386e-01 -1.39121428e-01 4.28058922e-01 5.23313940e-01
-1.98942404e-02 1.89408846e-02 -8.77528191e-01 -7.54386008e-01
-3.48477095e-01 -8.24146807e-01 -1.76071495e-01 2.85546750e-01
-1.70706660e-01 -3.02645117e-01 1.09083259e+00 -7.00279951e-01
1.19579387e+00 -2.10700679e+00 1.62449092e-01 -1.31075710e-01
2.61082500e-01 3.94055605e-01 -4.98312056e-01 5.86372554e-01
3.35483328e-02 -2.55037874e-01 1.14858709e-01 -3.28969240e-01
1.57591216e-02 1.35521218e-01 -1.49160370e-01 5.26858032e-01
1.26973018e-02 7.42843807e-01 -1.12681067e+00 -6.04114890e-01
5.68432868e-01 6.67542875e-01 -3.88205647e-01 3.13096374e-01
-3.52279574e-01 4.13282245e-01 -5.62770069e-01 6.76410496e-01
5.09615242e-01 -1.10781759e-01 1.43855512e-01 -3.21142137e-01
-1.50193825e-01 8.83530304e-02 -1.24295819e+00 1.51925480e+00
-8.35374296e-02 5.94978511e-01 2.18743205e-01 -8.25247824e-01
5.10746539e-01 6.45684123e-01 7.69406855e-01 -8.14651132e-01
6.09711409e-02 -2.50873983e-01 1.39308602e-01 -6.88700676e-01
1.93160146e-01 4.56701308e-01 3.09329317e-03 5.01643419e-01
-4.83954474e-02 6.26091957e-01 6.15311265e-01 4.65409219e-01
1.43013656e+00 5.13279974e-01 3.38974744e-01 3.68413240e-01
5.93026340e-01 -5.15953898e-02 9.73675072e-01 6.99114084e-01
-6.71707332e-01 2.12639183e-01 6.66051626e-01 -6.78522229e-01
-6.93198264e-01 -1.02346396e+00 5.88485658e-01 1.09014404e+00
4.86703038e-01 -4.39491361e-01 -6.63892746e-01 -8.89801204e-01
-2.47651249e-01 4.91034240e-01 -7.47274160e-01 -1.48323327e-01
-6.66725516e-01 -4.67963606e-01 4.17350024e-01 8.04515541e-01
9.18640971e-01 -1.44304562e+00 -8.70587230e-01 4.44324046e-01
-3.72725487e-01 -1.45642745e+00 -6.07716084e-01 -3.76065522e-01
-4.64259028e-01 -1.36105621e+00 -3.34160119e-01 -2.32491612e-01
3.77238035e-01 3.71144891e-01 7.72053897e-01 1.09207757e-01
-1.00990396e-03 4.32857960e-01 -4.83761042e-01 -1.65503815e-01
-3.95153016e-01 -3.55431139e-01 8.53034630e-02 2.54789263e-01
3.40904325e-01 -6.88097119e-01 -8.53535414e-01 6.95929110e-01
-7.55909264e-01 2.88622051e-01 5.76317191e-01 3.74608278e-01
2.91852981e-01 1.84618533e-01 5.71547806e-01 -3.18009049e-01
3.24387461e-01 -6.82055712e-01 -6.23213887e-01 1.40043303e-01
-2.26287711e-02 -3.69675964e-01 5.71407914e-01 -5.67409277e-01
-1.16874313e+00 -8.96782205e-02 1.55424759e-01 -7.56424606e-01
-3.28445852e-01 2.60528028e-01 -1.61469251e-01 3.61493438e-01
3.39993268e-01 3.89810950e-01 1.42689899e-01 -7.21892947e-03
1.85200676e-01 3.38181496e-01 5.54601490e-01 -3.02658975e-01
6.61842525e-01 8.98146987e-01 -1.48471549e-01 -6.16311669e-01
-4.06616032e-01 -4.77037042e-01 -2.62048334e-01 -8.75832677e-01
1.11346734e+00 -9.51487541e-01 -9.84356403e-01 7.41335988e-01
-1.30831110e+00 -6.85384512e-01 -7.20963851e-02 6.81306839e-01
-6.63934588e-01 3.39322776e-01 -7.44430304e-01 -6.85743690e-01
-1.36095867e-01 -1.21903300e+00 9.76207614e-01 1.25351325e-01
3.82895023e-02 -9.16141152e-01 1.16922684e-01 4.23607200e-01
2.55639195e-01 5.19000173e-01 3.42508078e-01 -4.38717008e-01
-1.01810241e+00 6.41747862e-02 -2.34419376e-01 2.26973027e-01
1.30035385e-01 1.51126519e-01 -6.25306189e-01 -3.17133695e-01
-1.29002526e-01 -6.57836720e-02 8.31367731e-01 5.67955852e-01
8.10880601e-01 -4.15616900e-01 -3.76942545e-01 5.13207078e-01
1.12153149e+00 6.53891146e-01 7.62832165e-01 2.83226162e-01
6.36028230e-01 3.71127337e-01 8.41850519e-01 7.08241642e-01
3.84903520e-01 8.48152220e-01 9.42553759e-01 -6.76065534e-02
-2.55040765e-01 -1.04642414e-01 9.46149528e-01 4.67135280e-01
-5.29334843e-01 -6.57186508e-01 -5.54764390e-01 4.56305623e-01
-2.41407824e+00 -1.38777077e+00 1.35223836e-01 1.74028921e+00
2.52183646e-01 2.35518590e-01 3.33366901e-01 -1.87639907e-01
8.12558830e-01 6.58656061e-01 -8.61815095e-01 3.11532635e-02
-6.38464047e-03 -5.96171141e-01 1.22790799e-01 3.75447422e-01
-1.18963063e+00 8.15670729e-01 4.85014534e+00 7.73324788e-01
-8.11196029e-01 1.73891276e-01 4.83716577e-01 -3.39579254e-01
4.21349794e-01 -1.86066002e-01 -5.00516653e-01 6.82114065e-01
6.11818373e-01 3.05528995e-02 5.08575857e-01 5.78527570e-01
9.08645153e-01 -1.67027563e-01 -1.07487905e+00 7.34270990e-01
2.33127810e-02 -1.31231809e+00 -8.27839319e-03 1.72820881e-01
4.09786135e-01 1.44279793e-01 -1.47952378e-01 3.92467737e-01
3.52893740e-01 -7.90638566e-01 8.56118798e-01 6.39163971e-01
4.09949690e-01 -5.09408832e-01 8.32177222e-01 2.78175116e-01
-1.78051507e+00 -2.04410166e-01 6.03449829e-02 -2.49604210e-01
5.25570631e-01 5.05365506e-02 -7.15401769e-01 5.02018154e-01
7.26750135e-01 1.21598423e+00 -2.71781176e-01 8.65334213e-01
-3.69291216e-01 5.36081970e-01 -2.16596052e-01 2.32387826e-01
5.63194931e-01 -1.96539760e-01 9.81306314e-01 1.07401419e+00
3.43264312e-01 1.48326010e-01 5.18464804e-01 8.32913280e-01
6.49617016e-02 -5.07106960e-01 -6.12680614e-01 -3.21849505e-03
4.86219585e-01 1.23288858e+00 -8.03194463e-01 -6.50089979e-01
-4.77877498e-01 9.17847514e-01 1.45157710e-01 7.80556858e-01
-1.35017407e+00 -1.76015049e-02 9.05333340e-01 9.24266279e-02
4.82337296e-01 -3.64408582e-01 6.00428879e-01 -1.04718375e+00
-1.06054395e-01 -1.05650640e+00 5.75396776e-01 -9.64017451e-01
-8.11722279e-01 5.17536998e-01 1.51982129e-01 -1.60149944e+00
-4.37151372e-01 -4.32789147e-01 -8.16091716e-01 3.53629529e-01
-1.15029001e+00 -1.09403956e+00 -5.82823873e-01 8.31362963e-01
9.90195632e-01 -1.24714851e-01 2.36121863e-01 2.28884250e-01
-8.35090399e-01 2.56524924e-02 -3.25357795e-01 3.79627258e-01
5.45072317e-01 -9.38729167e-01 1.28291100e-01 1.14984274e+00
3.39962691e-02 -1.07815573e-02 5.54136634e-01 -6.83598459e-01
-1.25782251e+00 -1.26051104e+00 1.36382997e-01 -1.75324515e-01
6.35866880e-01 -1.56092972e-01 -5.97169101e-01 8.68455231e-01
4.51761961e-01 2.59329677e-01 1.36805385e-01 -5.70986211e-01
-1.38801306e-01 -1.26833081e-01 -8.36999536e-01 7.21947312e-01
1.25884187e+00 -3.36706519e-01 -2.78763354e-01 2.01752797e-01
7.67131925e-01 -2.83015639e-01 -6.12520993e-01 3.55276704e-01
6.33684814e-01 -1.30524242e+00 1.07633924e+00 -3.91932249e-01
4.21378911e-01 -7.48836696e-01 -7.16949105e-02 -1.15109777e+00
-2.33034566e-01 -6.93621397e-01 -5.13241172e-01 9.70024586e-01
-1.76712200e-01 -6.13205194e-01 6.15482271e-01 7.70838782e-02
-2.00425178e-01 -7.05672383e-01 -8.79333317e-01 -7.14024782e-01
-7.17367291e-01 -4.77376103e-01 5.84176123e-01 7.83240199e-01
-1.64336637e-01 1.06964111e-01 -5.50294697e-01 3.75075132e-01
5.95389307e-01 -1.92093983e-01 8.95149589e-01 -5.87208629e-01
-3.89430910e-01 -3.96448046e-01 -5.77776849e-01 -1.05473053e+00
4.56783324e-02 -2.89838970e-01 2.39379965e-02 -1.69118607e+00
1.45368859e-01 7.21277893e-02 -4.67114896e-01 4.55082238e-01
1.15429364e-01 2.52623707e-01 3.77148539e-01 1.61936015e-01
-1.12936926e+00 7.48924911e-01 1.36438823e+00 -1.60492927e-01
-1.89932868e-01 -1.99614078e-01 9.72784963e-03 1.04953015e+00
7.11943090e-01 -3.03881615e-01 -7.37799346e-01 -4.22937572e-01
-3.73269543e-02 4.44413841e-01 8.61816227e-01 -1.30001748e+00
4.18528676e-01 -2.92368680e-01 2.23828554e-01 -6.99228764e-01
6.28867269e-01 -7.23026097e-01 2.27207795e-01 5.80899119e-01
-9.77322534e-02 1.99427247e-01 5.76899089e-02 9.02987719e-01
-3.00250590e-01 2.24517703e-01 3.43622506e-01 -3.24434489e-01
-1.18575823e+00 5.02544343e-01 -6.76041007e-01 -6.99740052e-02
1.38943517e+00 -3.07947904e-01 -7.64796019e-01 -7.61268377e-01
-6.63523376e-01 5.31366825e-01 2.51777232e-01 3.59256238e-01
8.36297750e-01 -1.42563689e+00 -7.33983874e-01 4.12445590e-02
-4.59696949e-02 -9.60372612e-02 6.75789833e-01 1.13184392e+00
-4.92749274e-01 2.18129680e-01 -3.51484269e-01 -6.33709908e-01
-1.23890471e+00 6.17715716e-01 4.65534657e-01 -4.37157989e-01
-6.05944872e-01 5.67358375e-01 6.39676809e-01 -4.63867709e-02
3.92645806e-01 -2.49852464e-01 -4.27586228e-01 8.37231949e-02
7.48946488e-01 6.64776206e-01 -5.27191222e-01 -8.36160004e-01
-3.87369335e-01 3.28049064e-01 -4.81471792e-02 -6.22908697e-02
1.21981275e+00 -1.32849962e-01 1.25218794e-01 1.33249849e-01
9.49215472e-01 -4.74013925e-01 -1.94614005e+00 -1.93195120e-01
-5.21133900e-01 -3.23028207e-01 2.30547506e-02 -7.32924581e-01
-1.22063720e+00 7.18832552e-01 5.40489137e-01 1.42230704e-01
1.28215384e+00 -1.00411713e-01 7.03607440e-01 1.16699606e-01
4.04468238e-01 -1.04884994e+00 6.31096363e-01 3.23309958e-01
1.07214916e+00 -1.15611959e+00 5.09970672e-02 -2.51225829e-01
-6.17196441e-01 9.93549287e-01 9.19127762e-01 -8.19784254e-02
4.39194053e-01 1.88170195e-01 4.62233648e-03 -3.62908930e-01
-1.19760954e+00 -2.67815500e-01 -1.47574227e-02 6.44699693e-01
-1.89306542e-01 -6.86503723e-02 2.07157463e-01 3.44671756e-01
4.95506108e-01 2.21328676e-01 5.85946918e-01 8.60872984e-01
-1.24121644e-01 -6.04260087e-01 -4.36067402e-01 4.97515425e-02
-9.59601067e-03 1.99388355e-01 -1.79930493e-01 9.83607471e-01
3.39426935e-01 1.05619669e+00 2.09740832e-01 -4.75999236e-01
1.95499256e-01 -4.23486233e-01 1.22743368e-01 -3.46475571e-01
-3.62675011e-01 2.75388807e-01 4.03149605e-01 -1.20954323e+00
-9.29880679e-01 -6.33835018e-01 -1.23755348e+00 -3.59619558e-01
8.90140310e-02 -1.13442659e-01 1.12378828e-01 8.54883194e-01
5.25804400e-01 9.26430106e-01 3.11498165e-01 -1.19873738e+00
-1.62283555e-01 -1.00192928e+00 -3.75648826e-01 3.78461540e-01
3.70448202e-01 -8.72922122e-01 -4.39391583e-01 6.07780032e-02] | [8.347282409667969, 0.5789557695388794] |
15964460-759f-4d70-b600-0fc74bc1196d | conditional-permutation-invariant-flows | 2206.09021 | null | https://arxiv.org/abs/2206.09021v1 | https://arxiv.org/pdf/2206.09021v1.pdf | Conditional Permutation Invariant Flows | We present a novel, conditional generative probabilistic model of set-valued data with a tractable log density. This model is a continuous normalizing flow governed by permutation equivariant dynamics. These dynamics are driven by a learnable per-set-element term and pairwise interactions, both parametrized by deep neural networks. We illustrate the utility of this model via applications including (1) complex traffic scene generation conditioned on visually specified map information, and (2) object bounding box generation conditioned directly on images. We train our model by maximizing the expected likelihood of labeled conditional data under our flow, with the aid of a penalty that ensures the dynamics are smooth and hence efficiently solvable. Our method significantly outperforms non-permutation invariant baselines in terms of log likelihood and domain-specific metrics (offroad, collision, and combined infractions), yielding realistic samples that are difficult to distinguish from real data. | ['Frank Wood', 'Trevor Campbell', 'Jonathan Wilder Lavington', 'Setareh Dabiri', 'Justice Sefas', 'Yunpeng Liu', 'Vasileios Lioutas', 'Matthew Niedoba', 'Adam Ścibior', 'Berend Zwartsenberg'] | 2022-06-17 | null | null | null | null | ['scene-generation'] | ['computer-vision'] | [ 4.45248514e-01 1.45157456e-01 9.44507271e-02 -4.41190660e-01
-7.45234549e-01 -8.21762800e-01 1.23149872e+00 -4.89386320e-01
-2.32343048e-01 9.09691453e-01 2.07313448e-01 -7.60865808e-02
-4.04738098e-01 -9.47056949e-01 -1.17081189e+00 -8.98603439e-01
-4.05314416e-01 1.02142143e+00 1.72530577e-01 5.92816398e-02
2.10468873e-01 7.74483681e-01 -1.68068099e+00 -1.62844911e-01
8.38574946e-01 5.15475333e-01 1.71537578e-01 9.14376915e-01
2.79998183e-01 6.08948290e-01 -2.70861298e-01 -1.96950629e-01
4.47565228e-01 -1.98602527e-01 -5.29012203e-01 4.27894294e-01
8.23949993e-01 -1.68661252e-01 -5.26172876e-01 9.11913157e-01
7.19658732e-02 5.29681861e-01 1.44657028e+00 -1.40556109e+00
-6.15482450e-01 1.30461484e-01 -5.77170908e-01 -2.11944748e-02
-1.64219424e-01 6.18163228e-01 1.06345570e+00 -6.22186899e-01
9.03021336e-01 1.60977173e+00 2.94235826e-01 1.60598963e-01
-1.86014378e+00 -4.69034314e-01 6.98787421e-02 -1.73865050e-01
-1.44649959e+00 -3.26390058e-01 6.36134028e-01 -8.79601240e-01
5.66840172e-01 1.11260898e-01 4.17129219e-01 1.12495232e+00
3.20362985e-01 3.71154279e-01 7.92428255e-01 -9.13690403e-03
2.44037926e-01 -1.91303730e-01 -1.21370211e-01 8.27941716e-01
2.80524850e-01 3.95599276e-01 -4.35951114e-01 -1.17058091e-01
1.05475163e+00 -2.93093115e-01 1.08698428e-01 -8.67974937e-01
-1.11581588e+00 9.04325843e-01 2.80131817e-01 -4.05187488e-01
-1.52217567e-01 5.90830982e-01 -1.63132817e-01 -3.69877905e-01
1.86779812e-01 2.35388592e-01 -1.59369797e-01 -1.95217997e-01
-7.10336566e-01 7.71438777e-01 7.56586492e-01 1.10565078e+00
9.45123196e-01 1.64894853e-02 -2.30820641e-01 4.49259073e-01
2.91434735e-01 9.68496680e-01 -4.71890539e-01 -1.41089725e+00
3.99094582e-01 1.81287467e-01 3.94615352e-01 -8.40249538e-01
-2.59394050e-01 -3.63732964e-01 -7.05638230e-01 4.83608723e-01
7.12663889e-01 -2.86073267e-01 -1.28508985e+00 2.11816692e+00
1.81435317e-01 4.22494620e-01 -2.51156062e-01 6.68803692e-01
1.94675624e-01 8.24621141e-01 2.72185326e-01 1.32131130e-01
9.09460187e-01 -5.68256021e-01 -2.62652725e-01 -8.00622031e-02
3.48094583e-01 -5.14262795e-01 1.02527511e+00 4.18344140e-02
-1.00067401e+00 -4.17719543e-01 -6.85324550e-01 -1.10972889e-01
-2.60247946e-01 3.52320112e-02 4.80001986e-01 5.40807962e-01
-1.05724168e+00 5.46796501e-01 -1.08863771e+00 -2.02654913e-01
6.58411980e-01 1.68185279e-01 -2.56517529e-01 -1.11597413e-02
-6.89720988e-01 5.94841957e-01 1.03257209e-01 -5.03174625e-02
-1.35743833e+00 -1.05281603e+00 -9.82857525e-01 1.71348155e-01
1.80563644e-01 -1.02438068e+00 9.08148944e-01 -2.58158743e-01
-1.44276106e+00 7.19945669e-01 -2.89495111e-01 -3.93372059e-01
8.55026186e-01 -1.60049379e-01 -5.46532311e-02 -1.93520874e-01
3.14331383e-01 1.14288580e+00 7.05182731e-01 -1.33222938e+00
-4.64903325e-01 -3.60849593e-03 1.75265514e-03 3.20381552e-01
2.27720350e-01 -3.53226274e-01 -3.78627747e-01 -4.41449314e-01
-4.10292208e-01 -1.20475924e+00 -4.43475097e-01 2.76087850e-01
-6.77450657e-01 -3.07967830e-02 9.18496192e-01 -4.61199820e-01
6.65517271e-01 -1.97409296e+00 1.93265051e-01 4.49170142e-01
5.44067249e-02 -3.80812228e-01 -3.39652002e-01 1.80337414e-01
1.38081327e-01 2.33834893e-01 -6.92037225e-01 -2.23999903e-01
2.52710879e-01 3.24129164e-01 -4.08254743e-01 5.07890046e-01
8.40877473e-01 8.75423133e-01 -9.87839580e-01 -3.10525030e-01
4.30512130e-01 6.77880347e-01 -8.19411516e-01 1.19456686e-02
-4.38554108e-01 6.13000274e-01 -3.02836984e-01 2.54489124e-01
7.90232122e-01 -2.46922523e-01 5.15913926e-02 -5.51821478e-02
-2.20040500e-01 1.73025042e-01 -1.23769522e+00 1.45170605e+00
-1.36982292e-01 7.39048362e-01 -7.89242238e-02 -6.78021908e-01
7.79661000e-01 -2.56395221e-01 4.81816143e-01 -9.41313431e-02
2.22831033e-02 -3.43074918e-01 3.50364558e-02 -2.08047181e-01
6.07242942e-01 -9.25679132e-02 -2.33721524e-01 4.88983274e-01
1.74906000e-01 -3.53731155e-01 4.51707393e-01 4.29863334e-01
1.06672835e+00 3.94960344e-01 -1.42175704e-01 -5.42940378e-01
-1.03370234e-01 -1.02169499e-01 4.50819731e-01 8.91550124e-01
-4.71257418e-02 7.50294864e-01 7.56830275e-01 -8.49130601e-02
-1.36924052e+00 -1.75840366e+00 -3.95721704e-01 6.54567838e-01
2.90040046e-01 -9.23380032e-02 -5.65020204e-01 -4.47010338e-01
2.93466389e-01 8.89828205e-01 -7.18195379e-01 -8.10609311e-02
-6.82057083e-01 -9.52128351e-01 2.66094774e-01 5.34248590e-01
2.44845986e-01 -7.82137156e-01 -2.40656182e-01 1.67815223e-01
-1.69764787e-01 -1.23666692e+00 -7.15979695e-01 -1.04045942e-01
-5.86537123e-01 -1.01136959e+00 -3.82750779e-01 -4.73725766e-01
8.28465044e-01 -1.58503633e-02 1.18826008e+00 -6.00378454e-01
-7.58957088e-01 4.20021921e-01 3.89522523e-01 -3.14848095e-01
-3.47758204e-01 -1.36796460e-01 6.31851703e-02 1.63329259e-01
1.94242112e-02 -7.29716957e-01 -5.12200892e-01 3.93617600e-01
-6.70965731e-01 1.50450751e-01 3.95298481e-01 6.35222852e-01
7.98716366e-01 2.95089800e-02 2.45667156e-02 -4.83415753e-01
3.96007329e-01 -5.40240347e-01 -1.05516088e+00 6.70296103e-02
-3.79912369e-02 2.92833149e-01 1.59604385e-01 -4.46435899e-01
-1.33256388e+00 3.42724502e-01 5.20414472e-01 -4.38309491e-01
-3.68221283e-01 6.63138330e-02 -2.81086326e-01 -2.90850662e-02
8.42668176e-01 4.95210327e-02 6.77682161e-02 -1.68918326e-01
8.88186693e-01 8.98036640e-03 9.25209284e-01 -1.03193283e+00
1.17223489e+00 9.10986245e-01 6.86087430e-01 -8.91626060e-01
-4.83881503e-01 6.07404800e-04 -7.75962770e-01 -3.61807585e-01
1.15402293e+00 -7.65085101e-01 -8.18251729e-01 3.28614146e-01
-1.07272112e+00 -5.73554933e-01 -5.28916597e-01 5.79056144e-01
-9.58571851e-01 8.93283188e-02 -3.75578195e-01 -9.64112043e-01
4.02410537e-01 -1.01078868e+00 1.21656036e+00 1.78317696e-01
-3.74435559e-02 -1.08450103e+00 2.21474633e-01 8.92477185e-02
2.65982568e-01 7.40018964e-01 9.21720326e-01 3.58192883e-02
-1.35431051e+00 6.86770072e-03 -4.29205656e-01 2.59624690e-01
1.33517608e-02 7.47384250e-01 -8.24647546e-01 1.24699853e-01
-6.10976338e-01 -1.68159269e-02 9.90291357e-01 9.28215981e-01
1.15277445e+00 -1.37385070e-01 -4.66956794e-01 6.68179035e-01
1.11976671e+00 -8.52977037e-02 6.47754014e-01 -1.66125789e-01
8.31648588e-01 7.60752738e-01 2.20148236e-01 4.03500170e-01
4.80638355e-01 4.58573669e-01 5.49645483e-01 1.16826221e-01
-2.21282169e-01 -4.14584070e-01 2.98112601e-01 2.00667307e-01
-2.33616218e-01 -5.93096554e-01 -9.70271707e-01 6.81546688e-01
-1.93222034e+00 -1.24856997e+00 -2.85764515e-01 2.12340975e+00
3.13754350e-01 3.39638948e-01 1.67759702e-01 -6.08147085e-01
8.08025479e-01 5.61941676e-02 -6.24587059e-01 5.69085479e-02
-3.07778239e-01 2.43886277e-01 7.68435895e-01 8.15304875e-01
-1.38665235e+00 9.12442327e-01 7.63602209e+00 6.38885260e-01
-6.05437338e-01 -2.69434512e-01 6.83861077e-01 -2.76184350e-01
-4.51054692e-01 2.10948274e-01 -7.79366970e-01 4.32946324e-01
7.84509182e-01 -2.78815836e-01 4.78893280e-01 6.56424701e-01
4.00788754e-01 -7.71818459e-02 -1.14383221e+00 6.55767143e-01
-1.89362437e-01 -1.71122146e+00 1.46244168e-01 4.71816510e-01
9.48315501e-01 1.62383303e-01 4.28501159e-01 1.86613828e-01
1.12813401e+00 -1.06801569e+00 6.77292705e-01 9.17308688e-01
8.12097549e-01 -7.37838030e-01 -1.32443517e-01 1.33844122e-01
-1.03134978e+00 2.36331001e-01 -1.60386100e-01 8.46416354e-02
7.93093503e-01 5.57247162e-01 -7.81138897e-01 1.93661720e-01
4.19890314e-01 7.94756413e-01 -9.47077498e-02 1.04705310e+00
-1.40849859e-01 7.66486704e-01 -6.80408180e-01 2.92033285e-01
2.00917304e-01 -7.06754863e-01 9.73682523e-01 1.24633443e+00
3.02263588e-01 -2.00814947e-01 1.87449425e-01 1.56234086e+00
-1.30512148e-01 -4.35621977e-01 -6.65903270e-01 3.16687860e-02
3.98063660e-01 1.25627863e+00 -7.32223690e-01 -1.42297462e-01
-8.99333227e-03 4.86321181e-01 3.52644235e-01 7.17485607e-01
-1.08333254e+00 -2.97932714e-01 1.22774434e+00 2.60640442e-01
4.99594927e-01 -5.77562571e-01 -3.24688375e-01 -1.04195523e+00
-1.08158275e-01 5.50477840e-02 1.54829603e-02 -7.92443752e-01
-1.41846550e+00 1.20683029e-01 4.27004367e-01 -1.04432321e+00
-3.64967763e-01 -6.90071166e-01 -7.47014582e-01 8.72302294e-01
-1.28554106e+00 -1.12910199e+00 -1.73433006e-01 2.48722866e-01
7.12129623e-02 7.34213367e-02 3.96738023e-01 1.87603086e-01
-6.37412310e-01 2.14175254e-01 3.70461755e-02 -7.67622069e-02
6.37396753e-01 -1.31107378e+00 7.65460610e-01 9.40175593e-01
4.92386669e-02 3.94158661e-01 7.66861498e-01 -7.63435781e-01
-8.94732535e-01 -1.52161777e+00 5.69195628e-01 -1.12883282e+00
7.55752921e-01 -5.84286988e-01 -5.73923290e-01 8.74576211e-01
-2.86261469e-01 1.13334857e-01 2.61079043e-01 -4.86702211e-02
-3.26872379e-01 5.76292910e-02 -1.06721449e+00 7.19648540e-01
1.37308657e+00 -4.15778756e-01 -1.16738312e-01 6.06227279e-01
5.15256107e-01 -3.42028975e-01 -5.73036909e-01 3.52124542e-01
5.07738769e-01 -5.40277779e-01 1.03834426e+00 -9.33712244e-01
3.93728882e-01 -5.19519866e-01 -1.30776957e-01 -1.11483979e+00
-6.36123478e-01 -7.75378704e-01 -3.01543390e-03 8.77230108e-01
4.51782197e-01 -3.82548869e-01 9.30590510e-01 7.92351782e-01
-1.21428296e-01 -2.93126494e-01 -9.32778120e-01 -1.03830266e+00
1.27214417e-01 -5.18038273e-01 5.50061345e-01 6.58642173e-01
-6.99943542e-01 2.83625185e-01 -5.24134159e-01 2.81038493e-01
1.10067511e+00 -1.81428894e-01 1.06432796e+00 -1.26467407e+00
-2.65598446e-01 -5.88152707e-01 -6.70371711e-01 -1.25543439e+00
2.85818040e-01 -8.73915792e-01 2.32716158e-01 -1.34701347e+00
4.06406194e-01 -5.25072098e-01 2.13922307e-01 1.28732130e-01
5.66722415e-02 1.54045895e-01 6.52156696e-02 2.08570287e-02
-4.81225938e-01 7.74262846e-01 1.35204732e+00 -4.19066176e-02
-5.66881783e-02 -1.22840077e-01 -4.96235877e-01 7.25429773e-01
5.94565928e-01 -4.82784241e-01 -5.37283480e-01 -3.10711265e-01
-5.41949011e-02 -2.69116700e-01 7.37037718e-01 -8.46502602e-01
-6.49490282e-02 -7.25072980e-01 2.38626719e-01 -5.76664031e-01
5.46231449e-01 -4.46945846e-01 3.74173373e-01 7.43935779e-02
-3.22866857e-01 -2.49835476e-01 8.46088901e-02 9.26844239e-01
2.10496888e-01 3.00439328e-01 7.78223693e-01 2.23437622e-01
-4.18026537e-01 6.70854926e-01 -5.51266611e-01 3.13961595e-01
1.13200235e+00 -1.91664487e-01 -4.23042804e-01 -6.66165471e-01
-8.49428773e-01 2.34431341e-01 3.72281015e-01 3.67326558e-01
1.62101641e-01 -1.49622965e+00 -8.84497046e-01 1.68688446e-01
2.26880573e-02 3.02667052e-01 2.58321524e-01 5.31175137e-01
-4.94943738e-01 2.75285751e-01 -2.36116901e-01 -9.13913190e-01
-7.77002573e-01 2.56998181e-01 4.10133541e-01 -2.75279611e-01
-3.91684592e-01 7.51030803e-01 8.13839972e-01 -6.59686327e-01
-4.59857136e-02 -3.44182849e-01 2.64833182e-01 -4.07876790e-01
2.14421496e-01 6.09481871e-01 -4.81378406e-01 -8.96656156e-01
-1.98894754e-01 5.40418565e-01 1.29893914e-01 -4.24905747e-01
1.11982429e+00 6.01806007e-02 2.36248225e-01 2.09680602e-01
1.04999352e+00 -1.93036824e-01 -2.10604095e+00 1.67630892e-02
-3.23004246e-01 -4.67078596e-01 -2.23560125e-01 -5.30180871e-01
-7.31527925e-01 7.91806102e-01 3.88269573e-01 -1.43646672e-01
3.55429828e-01 9.45866182e-02 3.79188448e-01 1.95333734e-01
1.08102709e-01 -8.88810575e-01 1.19321994e-01 7.19559133e-01
9.08181369e-01 -1.02951121e+00 -1.67992070e-01 -6.03131652e-01
-4.39578712e-01 8.45009804e-01 5.99208772e-01 -5.32688022e-01
1.05821383e+00 5.47693253e-01 -3.56517911e-01 -5.38288914e-02
-7.35719085e-01 -1.68227553e-01 4.68649030e-01 8.79075468e-01
-4.51560616e-02 3.15738946e-01 1.54130414e-01 -2.15315558e-02
-4.24531221e-01 -1.40695319e-01 3.99782032e-01 7.33342648e-01
-3.24930549e-01 -9.63309646e-01 -2.36967325e-01 5.49781561e-01
1.23162471e-01 8.32763389e-02 -7.04255998e-02 9.05893028e-01
2.73348726e-02 5.64701200e-01 7.12610424e-01 -1.04756698e-01
1.88466057e-01 -2.30624869e-01 7.20012426e-01 -4.48991060e-01
3.03074449e-01 -2.38865949e-02 5.14180325e-02 -6.05839133e-01
-2.52506733e-01 -1.10151911e+00 -1.01417446e+00 -4.68396395e-01
-5.32278307e-02 -3.50286663e-01 5.49846590e-01 8.14380288e-01
6.14203572e-01 4.98520076e-01 4.40697104e-01 -1.22068834e+00
-4.12237883e-01 -7.57508039e-01 -4.34609413e-01 5.80776095e-01
2.55790442e-01 -9.49589670e-01 -5.60973704e-01 4.92566496e-01] | [9.043949127197266, -3.2916462421417236] |
bd19fb7c-8040-4e31-8a5b-08969f96b00f | howkgpt-investigating-the-detection-of | 2305.18226 | null | https://arxiv.org/abs/2305.18226v2 | https://arxiv.org/pdf/2305.18226v2.pdf | HowkGPT: Investigating the Detection of ChatGPT-generated University Student Homework through Context-Aware Perplexity Analysis | As the use of Large Language Models (LLMs) in text generation tasks proliferates, concerns arise over their potential to compromise academic integrity. The education sector currently tussles with distinguishing student-authored homework assignments from AI-generated ones. This paper addresses the challenge by introducing HowkGPT, designed to identify homework assignments generated by AI. HowkGPT is built upon a dataset of academic assignments and accompanying metadata [17] and employs a pretrained LLM to compute perplexity scores for student-authored and ChatGPT-generated responses. These scores then assist in establishing a threshold for discerning the origin of a submitted assignment. Given the specificity and contextual nature of academic work, HowkGPT further refines its analysis by defining category-specific thresholds derived from the metadata, enhancing the precision of the detection. This study emphasizes the critical need for effective strategies to uphold academic integrity amidst the growing influence of LLMs and provides an approach to ensuring fair and accurate grading in educational institutions. | ['Michail Maniatakos', 'Yasir Zaki', 'Talal Rahwan', 'Manaar Alam', 'Christoforos Vasilatos'] | 2023-05-26 | null | null | null | null | ['specificity'] | ['natural-language-processing'] | [ 1.50866911e-01 3.19523245e-01 -1.70867741e-01 -1.69828370e-01
-9.44606662e-01 -7.86930919e-01 8.48369241e-01 6.92470253e-01
-4.11622792e-01 6.40217304e-01 5.44780195e-01 -7.21844375e-01
-4.89812762e-01 -7.17641950e-01 -2.91344762e-01 -4.32324320e-01
5.25159895e-01 3.73363703e-01 -1.50658116e-01 1.66525040e-02
8.17609489e-01 4.81918782e-01 -1.64762163e+00 3.44320118e-01
1.35565388e+00 2.85668224e-01 2.10330963e-01 9.28933382e-01
-6.18608057e-01 1.19072199e+00 -1.38060629e+00 -7.20310450e-01
-2.54323512e-01 -6.89154267e-01 -1.04184198e+00 2.08684802e-01
7.95284688e-01 -1.01911739e-01 -1.48210734e-01 7.34075546e-01
4.68647659e-01 4.25371110e-01 8.32308173e-01 -1.35163593e+00
-8.37345302e-01 9.28322315e-01 7.63724074e-02 3.81137758e-01
5.76739013e-01 4.00791168e-02 9.85571086e-01 -6.88429534e-01
5.63349485e-01 9.29665983e-01 4.76321608e-01 3.45088899e-01
-1.06095111e+00 -6.09252274e-01 -9.37825441e-02 1.29247159e-01
-1.19676185e+00 -3.13681275e-01 4.14491594e-01 -8.99228454e-01
6.67874396e-01 2.84198403e-01 6.47312939e-01 1.02891624e+00
1.30831629e-01 6.43820882e-01 1.13880193e+00 -7.99272895e-01
9.23917592e-02 3.85683000e-01 1.75855771e-01 4.11236227e-01
3.20422620e-01 -6.72014594e-01 -7.81933308e-01 -3.72695744e-01
3.39988023e-01 -3.12165707e-01 -1.53000169e-02 2.63561219e-01
-1.18054128e+00 6.59497976e-01 -1.42630115e-01 4.94661421e-01
-1.64376318e-01 -2.97816724e-01 1.76678330e-01 2.96582073e-01
1.69235572e-01 9.09801960e-01 -3.60516906e-02 -7.15868294e-01
-1.33447599e+00 4.22992796e-01 9.26051915e-01 8.10959995e-01
2.97153890e-01 -3.81671079e-02 -4.87990081e-01 9.28435326e-01
1.80473581e-01 2.06726760e-01 8.16806793e-01 -8.53389740e-01
5.66928983e-01 1.05563772e+00 -8.28710571e-03 -8.46448302e-01
1.21603087e-01 -4.44757521e-01 -1.65739715e-01 -1.18059650e-01
7.34559059e-01 -2.26674825e-01 -6.94891810e-01 1.30984104e+00
6.44872710e-02 -1.29337028e-01 -8.13792832e-03 4.88045812e-01
8.53663504e-01 4.68260437e-01 4.89027113e-01 1.64154217e-01
1.18554640e+00 -5.58752596e-01 -6.76635385e-01 1.53582208e-02
1.01418972e+00 -1.12694025e+00 1.06542516e+00 4.09119755e-01
-1.07962096e+00 -3.73514116e-01 -7.22564459e-01 -3.52776833e-02
-4.58670884e-01 1.89176977e-01 2.40975827e-01 8.69223475e-01
-9.82830107e-01 3.57400864e-01 -3.15646857e-01 -3.00656646e-01
2.33624548e-01 2.72836804e-01 -7.60695264e-02 1.31686674e-02
-1.00262821e+00 8.48798990e-01 -8.14152956e-02 -2.93452829e-01
-5.17535925e-01 -9.28524315e-01 -6.75020695e-01 1.64582759e-01
-9.54972059e-02 -1.77470371e-01 1.50725377e+00 -5.96936464e-01
-1.21132791e+00 9.35823977e-01 6.31323755e-02 -4.26076241e-02
5.71475565e-01 2.15266541e-01 -3.02288741e-01 5.31952716e-02
3.94315541e-01 2.89061844e-01 4.37426805e-01 -7.46952951e-01
-7.38814712e-01 -1.79014787e-01 -9.69354361e-02 2.33750612e-01
-9.06816721e-01 2.42858917e-01 2.28523180e-01 -5.24027348e-01
1.36017799e-02 -7.21949160e-01 8.64883810e-02 -9.01981235e-01
-2.60017246e-01 -9.29894030e-01 5.03491998e-01 -5.95578313e-01
1.64402688e+00 -1.59746528e+00 -2.96717167e-01 3.00982147e-01
4.18554217e-01 2.74650455e-01 1.28925610e-02 7.89571404e-01
1.95883051e-01 3.89067829e-01 4.64594513e-01 -1.56134158e-01
4.69021559e-01 -5.79665452e-02 -3.38082731e-01 3.02277990e-02
5.64784221e-02 5.76827884e-01 -1.14140987e+00 -5.62519729e-01
6.18373677e-02 7.87272081e-02 -2.87135303e-01 4.88571972e-01
1.02349490e-01 4.88706194e-02 -4.40204948e-01 6.21371627e-01
2.23812144e-02 -1.56226948e-01 1.13666072e-01 8.08049142e-01
-5.64420223e-01 7.79564619e-01 -8.63441825e-01 1.12424719e+00
-4.27758694e-01 9.98085678e-01 -4.63433713e-02 -6.53666854e-01
1.30564713e+00 4.47231978e-01 3.31535399e-01 -3.13045561e-01
9.02485996e-02 2.91695923e-01 2.08706036e-01 -5.61585784e-01
9.84813452e-01 2.02012524e-01 -1.68610424e-01 8.68571997e-01
2.07115516e-01 -3.63909513e-01 4.35751885e-01 6.57553613e-01
1.14234555e+00 -6.29683435e-02 -1.91721454e-01 -3.94003242e-01
5.38036346e-01 1.81764108e-03 6.29203841e-02 9.27708268e-01
-2.77262241e-01 2.33961910e-01 4.93283480e-01 -1.12965360e-01
-8.28137219e-01 -7.18281627e-01 -1.14171945e-01 1.48519671e+00
-6.61818326e-01 -6.98257387e-01 -8.93302143e-01 -5.38029492e-01
9.97910053e-02 9.73992229e-01 -1.95112109e-01 -2.64433682e-01
-2.32672945e-01 -6.39157474e-01 8.65509331e-01 1.05956964e-01
1.51979268e-01 -9.28067267e-01 -5.11433303e-01 3.28432232e-01
-4.63011444e-01 -8.44711423e-01 -3.71118516e-01 -8.27918649e-02
-4.30596381e-01 -9.28172767e-01 -6.13813400e-01 -7.74507523e-01
8.83631885e-01 6.45036027e-02 1.13592911e+00 2.72593856e-01
-2.73300976e-01 8.01728785e-01 -3.09141338e-01 -6.69072986e-01
-9.30380285e-01 4.93153036e-01 6.76178467e-03 -3.82369161e-01
9.37110543e-01 -3.33414555e-01 -7.56308883e-02 1.54589890e-02
-8.90690982e-01 -1.21921942e-01 4.68408555e-01 4.70501333e-01
-1.63622901e-01 -9.42987278e-02 9.50037718e-01 -6.17514074e-01
1.27403367e+00 -5.71668208e-01 -4.55489606e-01 3.67205620e-01
-1.01392508e+00 -1.21855341e-01 6.33145988e-01 -5.92272639e-01
-1.02748656e+00 -4.55924422e-01 7.95438439e-02 2.57223006e-03
-4.25308496e-01 5.19363105e-01 1.94962159e-01 -2.21877202e-01
6.64218307e-01 1.88194960e-01 -1.88112289e-01 -3.88311416e-01
-1.23488829e-01 1.17738175e+00 4.69939411e-01 -9.22946930e-01
6.77004397e-01 -4.14654464e-01 -1.70208231e-01 -9.06084299e-01
-8.79796147e-01 -5.64738393e-01 -6.38953686e-01 -7.84965992e-01
5.08627057e-01 -7.70745039e-01 -9.60475743e-01 2.86922455e-01
-8.93870175e-01 -3.35953653e-01 -7.91738089e-03 5.54206371e-01
-1.40210390e-01 3.07345837e-01 -7.75832593e-01 -9.53339458e-01
-1.49179175e-01 -1.09213674e+00 5.24971664e-01 6.03994131e-01
-8.01170886e-01 -1.09315395e+00 1.46160021e-01 1.17944479e+00
1.23543710e-01 -8.88201147e-02 1.03959000e+00 -1.09187734e+00
-3.90852422e-01 -3.68449420e-01 1.81956917e-01 1.81392342e-01
-1.39033571e-01 3.37461621e-01 -9.14038599e-01 2.45731343e-02
-3.72366786e-01 -5.21248043e-01 2.15137333e-01 -7.07214847e-02
8.72428119e-01 -4.98669475e-01 1.39337063e-01 4.55966145e-02
9.71770525e-01 6.18337840e-02 2.24816665e-01 8.04299295e-01
5.90938210e-01 7.95428753e-01 4.08985168e-01 5.38148463e-01
6.25976026e-01 2.13483363e-01 -1.33369993e-02 4.32635874e-01
2.68948041e-02 -5.15712321e-01 5.62589049e-01 1.16179311e+00
6.45041689e-02 -2.66271859e-01 -1.37190759e+00 8.78919423e-01
-1.44361603e+00 -1.04884803e+00 -5.86578906e-01 2.10187984e+00
1.11855900e+00 5.50537445e-02 -4.36517298e-02 1.01843566e-01
5.71492910e-01 -5.44091463e-02 -4.01739515e-02 -9.45589542e-01
1.63081124e-01 3.37290138e-01 3.61380905e-01 5.41719019e-01
-4.84076977e-01 6.94545090e-01 6.64708471e+00 5.95725358e-01
-6.98458612e-01 -2.31990471e-01 5.26325405e-01 9.11171585e-02
-3.86402875e-01 -1.84459373e-01 -1.20984209e+00 6.58009171e-01
1.53893614e+00 -7.04923272e-01 7.65746236e-02 6.74595833e-01
4.33540970e-01 -1.58990443e-01 -7.73912966e-01 3.99272889e-01
3.58255237e-01 -1.08673966e+00 1.69490576e-02 2.66179413e-01
9.72843528e-01 -3.75950605e-01 5.96110597e-02 4.57399487e-01
8.22195590e-01 -9.51733649e-01 6.26345873e-01 4.54378366e-01
2.89323956e-01 -7.88684368e-01 6.93256736e-01 4.95842904e-01
-5.12908041e-01 -1.32298648e-01 -2.09470689e-01 -5.20297885e-01
-4.66678560e-01 3.06357741e-01 -1.70556760e+00 5.39142974e-02
2.98394769e-01 9.49171707e-02 -8.01670969e-01 9.49742734e-01
-4.94313866e-01 1.16572559e+00 1.36986420e-01 -3.26757044e-01
3.62239271e-01 -2.66810775e-01 3.42199355e-01 1.36434197e+00
5.31533718e-01 5.24725392e-02 2.85565734e-01 8.35520267e-01
-2.33118281e-01 1.93374723e-01 -3.99804145e-01 -5.54687738e-01
9.71695662e-01 1.46867514e+00 -7.15986192e-01 -3.70753497e-01
-1.85400113e-01 6.21444106e-01 2.09510729e-01 2.35127881e-01
-2.89356411e-01 -7.57736623e-01 6.29865468e-01 3.18560094e-01
-2.37285897e-01 -2.05109954e-01 -5.06473184e-01 -9.04382825e-01
-2.92491168e-01 -9.90400255e-01 3.28283519e-01 -7.28153527e-01
-9.84680474e-01 2.68177956e-01 -1.51832581e-01 -8.02332044e-01
-5.10897398e-01 -4.36907113e-01 -8.13587010e-01 1.15329611e+00
-1.07290888e+00 -9.12674904e-01 -4.45167571e-01 2.82489866e-01
4.40294266e-01 -1.73399776e-01 7.75000751e-01 1.29793137e-01
-7.55317152e-01 8.73249888e-01 1.50544375e-01 3.53256106e-01
1.13938367e+00 -1.48485148e+00 -1.08501539e-01 9.67128932e-01
-6.01020083e-02 8.08807611e-01 6.86052561e-01 -6.88766241e-01
-1.10113585e+00 -9.24359918e-01 1.75432563e+00 -9.69062388e-01
1.01472545e+00 -2.82683283e-01 -8.45222890e-01 5.30530632e-01
2.35730261e-01 -8.61163139e-01 1.23929095e+00 5.55983111e-02
1.61362477e-02 2.29247198e-01 -8.99597108e-01 5.83580434e-01
5.52359164e-01 -6.78099275e-01 -9.64575589e-01 5.22680819e-01
3.45527172e-01 -3.07145745e-01 -1.36902308e+00 -3.70300293e-01
2.32002407e-01 -6.56052351e-01 5.91683805e-01 -7.23687649e-01
7.44933367e-01 2.05548704e-01 4.32866633e-01 -1.27620316e+00
-2.84080237e-01 -7.28839397e-01 1.15005970e-01 1.76338208e+00
2.04395130e-01 -2.83238411e-01 8.83387327e-01 1.31823874e+00
-3.97087455e-01 -4.07589048e-01 -5.88965535e-01 -4.82314646e-01
2.63166517e-01 -1.95102483e-01 6.59985363e-01 1.44884014e+00
4.68329281e-01 1.25316679e-01 1.85787976e-01 -1.20589778e-01
5.18395424e-01 -8.98531675e-02 7.31173098e-01 -1.55590117e+00
3.09094191e-01 -7.13273585e-01 -4.81282771e-01 -3.30224514e-01
3.35079223e-01 -9.56405044e-01 -1.60652306e-02 -1.47826695e+00
-2.00408953e-03 -3.87179315e-01 -7.13155046e-02 6.19780362e-01
-2.51727521e-01 2.01580465e-01 1.57866791e-01 1.20305553e-01
-5.60407937e-01 -1.74914245e-02 1.05265570e+00 2.86082923e-01
-3.55483323e-01 -2.54949033e-02 -8.86192858e-01 4.77197945e-01
8.79518449e-01 -3.87020171e-01 -3.58898580e-01 -3.75463963e-02
2.34470055e-01 -2.84257978e-01 1.56662896e-01 -9.35850561e-01
4.46908414e-01 -5.71453273e-01 5.37765265e-01 -3.25269163e-01
-3.11472714e-01 -2.31012359e-01 -2.95846254e-01 1.78271681e-01
-8.18004251e-01 2.82935500e-01 6.64373785e-02 -1.52473077e-01
-2.17448488e-01 -6.51141465e-01 5.06619871e-01 -1.63593218e-01
-4.07996863e-01 -1.40356272e-01 -1.11852551e+00 1.63940221e-01
8.19001615e-01 -2.97371596e-01 -5.45535386e-01 -4.87886757e-01
-4.61648315e-01 2.08060101e-01 3.64608943e-01 4.98498231e-01
3.73419493e-01 -1.11171687e+00 -6.67982161e-01 2.80096591e-01
-3.11480742e-03 -2.13053048e-01 8.48510042e-02 7.61941373e-01
-4.62331951e-01 7.80761838e-01 -2.12073952e-01 -2.49263644e-02
-1.51930571e+00 -1.87750742e-01 -7.73196220e-02 -3.21240008e-01
-1.85252294e-01 9.15181935e-01 -4.25131947e-01 -4.71681565e-01
3.44103754e-01 -1.30237252e-01 -3.89859438e-01 3.65975648e-01
6.40743434e-01 7.44465709e-01 2.86459655e-01 -6.03897572e-01
2.10861698e-01 -3.80132824e-01 -2.40248054e-01 -2.01965645e-01
1.12820947e+00 -1.27526782e-02 -1.75389946e-01 4.60367918e-01
6.76703990e-01 3.99445981e-01 -7.40738451e-01 1.01383276e-01
4.69543785e-01 -4.70799536e-01 8.32478628e-02 -1.07841361e+00
-2.25207746e-01 6.91548347e-01 -6.09863438e-02 3.71887833e-01
6.31881356e-01 -2.54717678e-01 5.80241740e-01 2.48533025e-01
-1.19372323e-01 -1.43193698e+00 1.70991138e-01 6.85426772e-01
3.38667065e-01 -9.68141735e-01 -2.00059921e-01 1.57383367e-01
-3.99888426e-01 1.34461653e+00 1.00013244e+00 5.10653317e-01
-1.26897514e-01 2.17260271e-02 1.96479172e-01 3.53035852e-02
-8.12286496e-01 2.19621867e-01 4.07925695e-01 3.34285140e-01
1.04759514e+00 2.44751126e-01 -5.80517709e-01 6.35009766e-01
-8.15770686e-01 -1.36987388e-01 1.26182532e+00 9.59472001e-01
-7.04472780e-01 -1.35068429e+00 -8.78524542e-01 7.34627008e-01
-7.77348518e-01 -1.29044220e-01 -1.07342994e+00 5.51051021e-01
1.95806861e-01 1.16337073e+00 7.38820732e-02 -3.56841058e-01
1.79306027e-02 7.46318877e-01 1.34447426e-01 -7.21422553e-01
-1.11557937e+00 -3.82837385e-01 9.90515426e-02 2.31786281e-01
-1.56670377e-01 -7.94079244e-01 -1.03588140e+00 -5.44507444e-01
-2.69170612e-01 7.65509725e-01 7.46458650e-01 8.38786542e-01
3.66117209e-01 4.94796306e-01 3.49208891e-01 -2.79330701e-01
-8.52734685e-01 -1.21167493e+00 -3.83870751e-01 4.51494634e-01
1.62713498e-01 -2.30600700e-01 -5.28713882e-01 9.66964886e-02] | [11.135890007019043, 9.179248809814453] |
fbd044ca-e4c3-431d-a929-95bf2aa54a7e | unified-keypoint-based-action-recognition | 2303.15270 | null | https://arxiv.org/abs/2303.15270v1 | https://arxiv.org/pdf/2303.15270v1.pdf | Unified Keypoint-based Action Recognition Framework via Structured Keypoint Pooling | This paper simultaneously addresses three limitations associated with conventional skeleton-based action recognition; skeleton detection and tracking errors, poor variety of the targeted actions, as well as person-wise and frame-wise action recognition. A point cloud deep-learning paradigm is introduced to the action recognition, and a unified framework along with a novel deep neural network architecture called Structured Keypoint Pooling is proposed. The proposed method sparsely aggregates keypoint features in a cascaded manner based on prior knowledge of the data structure (which is inherent in skeletons), such as the instances and frames to which each keypoint belongs, and achieves robustness against input errors. Its less constrained and tracking-free architecture enables time-series keypoints consisting of human skeletons and nonhuman object contours to be efficiently treated as an input 3D point cloud and extends the variety of the targeted action. Furthermore, we propose a Pooling-Switching Trick inspired by Structured Keypoint Pooling. This trick switches the pooling kernels between the training and inference phases to detect person-wise and frame-wise actions in a weakly supervised manner using only video-level action labels. This trick enables our training scheme to naturally introduce novel data augmentation, which mixes multiple point clouds extracted from different videos. In the experiments, we comprehensively verify the effectiveness of the proposed method against the limitations, and the method outperforms state-of-the-art skeleton-based action recognition and spatio-temporal action localization methods. | ['Taiki Sekii', 'Fumiaki Sato', 'Ryo Hachiuma'] | 2023-03-27 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Hachiuma_Unified_Keypoint-Based_Action_Recognition_Framework_via_Structured_Keypoint_Pooling_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Hachiuma_Unified_Keypoint-Based_Action_Recognition_Framework_via_Structured_Keypoint_Pooling_CVPR_2023_paper.pdf | cvpr-2023-1 | ['video-classification', 'weakly-supervised-temporal-action', 'action-localization', 'spatio-temporal-action-localization'] | ['computer-vision', 'computer-vision', 'computer-vision', 'computer-vision'] | [ 3.72187763e-01 -2.35619202e-01 -4.49510396e-01 -1.12513162e-01
-5.46868742e-01 -1.68606251e-01 5.95816016e-01 -2.88554102e-01
-5.86121559e-01 3.35581332e-01 2.21238136e-01 4.09052819e-01
-2.19565090e-02 -6.45734310e-01 -8.14824939e-01 -8.68294477e-01
-3.45285162e-02 1.43340036e-01 7.50598073e-01 3.93975899e-02
2.34889358e-01 9.21925485e-01 -1.63191426e+00 3.40835422e-01
5.58016121e-01 1.28134012e+00 -7.51300380e-02 4.63846922e-01
-8.69212896e-02 8.27315390e-01 -5.60256183e-01 -1.87276542e-01
4.27998483e-01 -2.94161052e-01 -4.60869491e-01 4.75240111e-01
7.88910151e-01 -5.27138174e-01 -6.21951044e-01 8.11144471e-01
3.28683853e-01 2.89576173e-01 4.42079365e-01 -1.48366356e+00
-4.65882182e-01 -9.08832774e-02 -7.72796869e-01 2.38994077e-01
3.32556486e-01 5.24253130e-01 5.47644496e-01 -9.10596609e-01
4.34991211e-01 1.34087694e+00 8.32572937e-01 5.09766459e-01
-6.66071057e-01 -7.13535666e-01 4.27450150e-01 3.20657700e-01
-1.24440718e+00 -3.84528190e-01 1.04157495e+00 -3.94756079e-01
9.94601011e-01 8.37218612e-02 1.05842268e+00 1.24433267e+00
1.92756355e-01 1.25510979e+00 7.38837719e-01 -1.75976560e-01
1.19410440e-01 -6.36408508e-01 -1.77791089e-01 9.49211359e-01
-3.35685909e-02 3.12386062e-02 -7.52701402e-01 -1.11841112e-02
1.53795838e+00 5.47419965e-01 -2.19227970e-01 -6.83846772e-01
-1.65192497e+00 4.05241698e-01 3.99834275e-01 2.51817048e-01
-5.72115660e-01 4.80794609e-01 4.96518463e-01 -2.92483091e-01
3.19586068e-01 -1.69575363e-01 -6.47771418e-01 -1.76683292e-01
-9.02210474e-01 2.29765236e-01 1.80087060e-01 8.73945057e-01
6.11528575e-01 1.25169948e-01 -4.30321097e-01 4.00986940e-01
3.14941615e-01 6.86303496e-01 8.10815811e-01 -1.07879126e+00
5.52730083e-01 1.02738667e+00 1.61145002e-01 -1.20194995e+00
-4.24794823e-01 -1.47430956e-01 -7.12021112e-01 3.99878800e-01
7.56654084e-01 1.67712629e-01 -1.13675618e+00 1.53684068e+00
6.67048156e-01 6.34783864e-01 -2.70982027e-01 9.15601075e-01
6.35732532e-01 2.43851662e-01 2.49752343e-01 -2.07707450e-01
1.34742117e+00 -1.09765959e+00 -6.40040755e-01 -2.14960426e-02
4.45024222e-01 -2.68171310e-01 9.42068160e-01 2.42811769e-01
-1.08545971e+00 -8.78436685e-01 -9.03326571e-01 -4.99577560e-02
-3.03483427e-01 3.92700404e-01 7.08288491e-01 3.10930550e-01
-7.00637996e-01 6.73636556e-01 -1.31756485e+00 -2.97571987e-01
8.19852650e-01 4.08180952e-01 -7.26493299e-01 7.52101913e-02
-8.41404378e-01 6.03228211e-01 4.46556747e-01 2.09437102e-01
-8.30270946e-01 -5.53502262e-01 -1.08335733e+00 -6.84487000e-02
7.64066160e-01 -7.80087233e-01 1.02854991e+00 -1.13498580e+00
-1.73478234e+00 7.11176753e-01 -6.39301911e-02 -3.61971825e-01
7.44834602e-01 -5.09575784e-01 -2.72025198e-01 6.55480802e-01
1.95477098e-01 6.27543032e-01 1.18213272e+00 -7.37793803e-01
-8.33935678e-01 -6.08337879e-01 4.00783960e-03 2.08333269e-01
-2.71989882e-01 2.07608759e-01 -6.89616561e-01 -9.49366927e-01
3.21906269e-01 -5.39330542e-01 -1.53738394e-01 6.21037483e-01
-1.65172756e-01 -4.90070492e-01 1.04860592e+00 -5.22134960e-01
8.61541212e-01 -2.03106046e+00 1.84619308e-01 -1.43224567e-01
3.88999842e-02 4.28074181e-01 -1.21329524e-01 1.67511746e-01
-1.33112893e-01 -2.28930309e-01 -2.53107369e-01 -3.63950461e-01
-3.90784107e-02 3.85180563e-01 4.70802709e-02 7.92928159e-01
5.10586917e-01 1.14454126e+00 -9.51136053e-01 -8.71693015e-01
6.21309936e-01 5.75310528e-01 -1.85223639e-01 6.47274628e-02
-2.09609568e-01 6.45539939e-01 -7.24719882e-01 1.10565734e+00
4.63234812e-01 -1.67573735e-01 -4.88933742e-01 -4.47891295e-01
-2.79775579e-02 -2.34277263e-01 -1.45260966e+00 2.10659146e+00
-5.33124898e-03 -3.36161372e-03 -5.28744943e-02 -1.17373264e+00
8.07228088e-01 3.36601287e-01 1.03654110e+00 -4.69926566e-01
1.14737459e-01 -2.25863177e-02 -4.75999624e-01 -6.33703113e-01
1.05531976e-01 -3.94302746e-03 6.43718988e-02 2.25905314e-01
3.67517531e-01 3.31050158e-01 7.69772679e-02 -2.30832905e-01
1.09454429e+00 8.81153584e-01 3.36912096e-01 1.74968436e-01
8.38540673e-01 -2.55481362e-01 9.62928832e-01 5.14275432e-01
-6.46715045e-01 7.10004985e-01 1.93365395e-01 -7.32110620e-01
-7.43238926e-01 -9.23627198e-01 3.27207968e-02 1.02901065e+00
3.34171534e-01 -2.56445706e-01 -5.75770259e-01 -1.23939109e+00
7.34632239e-02 -1.36385374e-02 -8.11182141e-01 -1.70383036e-01
-1.00747454e+00 -2.84431100e-01 5.51495552e-01 1.02673697e+00
1.05260873e+00 -1.25770748e+00 -9.81039405e-01 1.46876514e-01
-6.77775592e-02 -1.22176898e+00 -5.61890483e-01 -1.04972325e-01
-1.05810523e+00 -1.36479867e+00 -9.92360175e-01 -5.61634064e-01
5.42303622e-01 6.53925315e-02 6.58052146e-01 1.06811328e-02
-2.77061552e-01 7.60145187e-01 -3.89038622e-01 -9.24315378e-02
7.75398985e-02 -4.65846717e-01 2.03585282e-01 5.74654341e-01
5.22104859e-01 -6.99298918e-01 -8.95174742e-01 4.65152204e-01
-8.63019288e-01 -1.07613072e-01 8.94232213e-01 6.07150078e-01
8.97436440e-01 3.47381681e-02 1.40812650e-01 -1.95563287e-01
-6.31559640e-02 -9.00320858e-02 -3.03180337e-01 4.10923481e-01
7.22613782e-02 -2.53940582e-01 3.87290627e-01 -6.41515136e-01
-9.46646810e-01 6.23064220e-01 1.08349361e-01 -9.50201869e-01
-6.59566998e-01 1.26707302e-02 -4.53722477e-01 -3.39782894e-01
4.59044516e-01 4.89307761e-01 1.76913932e-01 -6.36654735e-01
5.46992421e-01 3.02530587e-01 8.49342823e-01 -5.58583379e-01
7.31747031e-01 9.34736967e-01 1.14464760e-01 -6.60876751e-01
-7.05678999e-01 -6.87395155e-01 -1.32228887e+00 -5.75263500e-01
1.16065884e+00 -8.40400875e-01 -6.77732050e-01 9.62376773e-01
-1.27018130e+00 -1.86859816e-01 -6.08852386e-01 6.32540405e-01
-1.04192698e+00 7.83183813e-01 -5.82589149e-01 -7.75143802e-01
-1.77895650e-01 -9.22924876e-01 1.57418537e+00 2.20241189e-01
1.29783973e-01 -6.81384683e-01 -1.15045622e-01 4.29300666e-01
-1.63576022e-01 7.90178239e-01 3.15610021e-01 -5.78524232e-01
-6.94563091e-01 -3.74756157e-01 -1.31994560e-01 4.93093312e-01
2.19971105e-01 -5.60261831e-02 -7.89670527e-01 -1.14563823e-01
1.03065632e-01 -2.41488636e-01 8.80973995e-01 4.87035036e-01
1.14951181e+00 -1.65208727e-01 -4.95608211e-01 6.29625857e-01
1.08632207e+00 1.47903025e-01 6.59040272e-01 3.46006185e-01
1.06713974e+00 4.41973835e-01 7.41719306e-01 4.73945439e-01
2.14728080e-02 8.03843975e-01 5.58754981e-01 -1.74833938e-01
-2.18290567e-01 -2.43380427e-01 4.55347925e-01 3.96644682e-01
-6.00902855e-01 3.19649041e-01 -6.28581047e-01 3.95419061e-01
-2.15350270e+00 -1.20247006e+00 7.02993795e-02 2.07422566e+00
5.92831194e-01 1.39706701e-01 4.78628099e-01 3.15812677e-01
7.49298215e-01 2.71907926e-01 -8.40415955e-01 4.09157783e-01
-6.90589994e-02 2.40972787e-01 4.86866206e-01 -1.40086427e-01
-1.61085248e+00 8.97501051e-01 5.63385963e+00 9.78741705e-01
-9.48732138e-01 1.31428167e-01 1.19107783e-01 2.56708041e-02
4.50669557e-01 -3.42362255e-01 -5.96217394e-01 4.44333404e-01
3.29522222e-01 2.83823818e-01 -1.47271186e-01 9.33937550e-01
2.86563694e-01 6.57846630e-02 -1.08284616e+00 1.20741713e+00
2.42543548e-01 -1.31970143e+00 8.08529332e-02 -6.13276474e-02
4.05591398e-01 -2.88226336e-01 -1.89592719e-01 2.26606429e-01
-1.16212495e-01 -7.88838446e-01 8.12724411e-01 8.77706110e-01
5.51035583e-01 -6.00491166e-01 5.58964789e-01 3.57255757e-01
-1.54309249e+00 -4.42397028e-01 -1.81932420e-01 -1.53915092e-01
2.70564049e-01 1.52581602e-01 -1.81782573e-01 7.39671886e-01
9.47720528e-01 1.26540554e+00 -5.58456540e-01 1.09106886e+00
-3.78825068e-01 2.70469248e-01 -3.11317295e-01 2.63810426e-01
2.53656685e-01 -1.88114792e-02 5.24949253e-01 1.00418687e+00
1.45609602e-01 3.02103907e-01 5.42042792e-01 7.38128245e-01
2.81370491e-01 -6.97566941e-02 -3.82299364e-01 4.47516441e-02
7.90818259e-02 1.14910817e+00 -8.78408372e-01 -3.93206805e-01
-5.70503652e-01 1.17416811e+00 1.01619117e-01 4.77726281e-01
-8.38459313e-01 -2.11490989e-01 7.12159991e-01 2.49495879e-01
6.53361320e-01 -4.66377497e-01 3.16797756e-02 -1.29629385e+00
2.81046748e-01 -6.09165072e-01 5.67522526e-01 -7.84109771e-01
-1.24116147e+00 3.28793935e-02 1.68227434e-01 -1.57198858e+00
-1.84628237e-02 -7.72281289e-01 -7.32433915e-01 5.30028164e-01
-1.31370425e+00 -1.76525116e+00 -4.10296172e-01 1.13999951e+00
6.84392989e-01 -1.97018847e-01 5.60402930e-01 1.97223261e-01
-5.92274725e-01 5.25647521e-01 -3.79445016e-01 5.31357169e-01
3.67643744e-01 -1.02656484e+00 1.88874766e-01 9.93627727e-01
2.21251681e-01 3.74624491e-01 8.22076350e-02 -7.25439608e-01
-1.27416492e+00 -1.14539361e+00 3.85719717e-01 -6.23978674e-01
5.74131846e-01 -9.24949273e-02 -7.91861653e-01 5.98935366e-01
-3.67524683e-01 4.69967544e-01 3.84710610e-01 -4.00110275e-01
-1.50846511e-01 -1.38634890e-01 -1.04844475e+00 5.30109942e-01
1.34528577e+00 -3.24411094e-01 -8.23144197e-01 4.76369053e-01
7.43736565e-01 -4.67412293e-01 -8.65735769e-01 7.30762124e-01
5.12217522e-01 -1.02932525e+00 1.26481509e+00 -7.83634484e-01
1.71068922e-01 -5.81559956e-01 -1.15325162e-02 -4.85819459e-01
-4.23783630e-01 -5.49872398e-01 -7.06904471e-01 9.18623030e-01
-4.49392349e-01 -2.78925925e-01 1.17593157e+00 4.26784277e-01
-3.60936671e-01 -9.70719159e-01 -1.33054996e+00 -8.95979226e-01
-1.22604474e-01 -4.62132215e-01 5.59755504e-01 7.93341637e-01
-3.02186579e-01 -2.11443439e-01 -3.61669213e-01 2.84689307e-01
6.50767148e-01 -3.14770592e-03 9.21960533e-01 -1.12325227e+00
-1.84628323e-01 -5.23764968e-01 -1.15523565e+00 -1.29096556e+00
1.30937248e-01 -4.88739967e-01 -1.21458545e-01 -1.37500095e+00
-5.97782917e-02 -6.28829002e-02 -5.03968358e-01 8.11860025e-01
-9.56675261e-02 3.35865378e-01 1.72941908e-01 3.05063695e-01
-7.28138566e-01 7.45462954e-01 1.61890805e+00 -2.75604635e-01
-1.62330732e-01 2.10338414e-01 -6.86552376e-02 1.04347372e+00
4.77884620e-01 -1.63324431e-01 -2.84167260e-01 -2.53227741e-01
-3.75272959e-01 3.12166903e-02 9.73117590e-01 -1.37729931e+00
3.11560601e-01 -3.24383944e-01 7.42200673e-01 -7.41181731e-01
5.80730319e-01 -1.04053831e+00 -1.17788203e-01 4.29517686e-01
-6.14979565e-02 -2.25275531e-01 2.86225248e-02 9.95021224e-01
-1.33982971e-01 1.40838832e-01 6.65848494e-01 -3.91930401e-01
-9.78413284e-01 8.97361696e-01 4.70247380e-02 -1.37838274e-01
1.32636619e+00 -8.77202570e-01 1.52506873e-01 6.39330968e-02
-9.00264323e-01 3.01839411e-03 3.52591932e-01 4.29652750e-01
7.82674670e-01 -1.60056508e+00 -4.46536690e-01 3.76645863e-01
1.32219896e-01 1.62379280e-01 4.73853260e-01 1.25374377e+00
-3.15692127e-01 1.71654940e-01 -5.04017711e-01 -9.45601821e-01
-1.13635671e+00 6.54244840e-01 5.31818509e-01 -6.85880408e-02
-1.12475753e+00 8.07962298e-01 2.02449098e-01 -1.58958659e-01
5.19202948e-01 -6.45958841e-01 -2.59719014e-01 -1.24663949e-01
6.34286046e-01 4.86769676e-01 -1.49388209e-01 -9.93859529e-01
-4.56788540e-01 9.70027208e-01 2.80459553e-01 1.44187093e-01
1.12842560e+00 1.81715652e-01 1.41752675e-01 4.40533519e-01
1.05434585e+00 -4.04201031e-01 -1.85571539e+00 -3.57511908e-01
-2.79225588e-01 -7.75664032e-01 -2.13271558e-01 -5.89986265e-01
-1.18750632e+00 9.66602147e-01 6.53134644e-01 -1.94451064e-01
1.14944768e+00 -9.34938192e-02 1.06725860e+00 1.59150586e-01
5.06804466e-01 -1.24767470e+00 5.36260903e-01 2.08992526e-01
8.40767682e-01 -1.25761759e+00 8.19859430e-02 -2.53246486e-01
-4.44089264e-01 1.34263885e+00 7.94161379e-01 -3.02643180e-01
5.61531782e-01 -9.92245525e-02 -9.35785323e-02 -3.04451376e-01
-1.31111279e-01 -2.73762643e-01 3.96821946e-01 8.99440885e-01
-1.76282495e-01 -1.96778283e-01 -1.17147930e-01 7.01975048e-01
4.07456487e-01 2.24499419e-01 -9.26732421e-02 1.22088766e+00
-5.22120178e-01 -8.49364996e-01 -4.61663783e-01 1.78923845e-01
-2.40728721e-01 4.05450404e-01 -2.93403298e-01 1.03338814e+00
5.99416435e-01 5.45127809e-01 1.64743975e-01 -3.44820321e-01
4.82904017e-01 1.02854431e-01 6.09933972e-01 -4.83560890e-01
-4.03050005e-01 1.81229040e-01 -3.63929838e-01 -1.25289416e+00
-9.86028850e-01 -7.57810712e-01 -1.41917849e+00 1.35933653e-01
-2.51517385e-01 -3.46990228e-01 2.25563124e-01 1.19167876e+00
3.57020676e-01 4.17524308e-01 2.79188335e-01 -1.35607612e+00
-5.50145030e-01 -8.26925695e-01 -5.98296762e-01 6.32637203e-01
2.84885526e-01 -1.18335295e+00 -2.30729446e-01 3.58462363e-01] | [7.901422500610352, 0.40543219447135925] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.