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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
6419c6c2-627e-4cd8-b5fd-0ddc0933e103 | stylizing-3d-scene-via-implicit | 2105.13016 | null | https://arxiv.org/abs/2105.13016v3 | https://arxiv.org/pdf/2105.13016v3.pdf | Stylizing 3D Scene via Implicit Representation and HyperNetwork | In this work, we aim to address the 3D scene stylization problem - generating stylized images of the scene at arbitrary novel view angles. A straightforward solution is to combine existing novel view synthesis and image/video style transfer approaches, which often leads to blurry results or inconsistent appearance. Inspired by the high-quality results of the neural radiance fields (NeRF) method, we propose a joint framework to directly render novel views with the desired style. Our framework consists of two components: an implicit representation of the 3D scene with the neural radiance fields model, and a hypernetwork to transfer the style information into the scene representation. In particular, our implicit representation model disentangles the scene into the geometry and appearance branches, and the hypernetwork learns to predict the parameters of the appearance branch from the reference style image. To alleviate the training difficulties and memory burden, we propose a two-stage training procedure and a patch sub-sampling approach to optimize the style and content losses with the neural radiance fields model. After optimization, our model is able to render consistent novel views at arbitrary view angles with arbitrary style. Both quantitative evaluation and human subject study have demonstrated that the proposed method generates faithful stylization results with consistent appearance across different views. | ['Wei-Chen Chiu', 'Wei-Sheng Lai', 'Hung-Yu Tseng', 'Meng-Shiun Tsai', 'Pei-Ze Chiang'] | 2021-05-27 | null | null | null | null | ['video-style-transfer'] | ['computer-vision'] | [ 3.82326066e-01 6.02299683e-02 2.66994387e-01 -3.11589330e-01
-2.14028329e-01 -4.70571369e-01 4.84435797e-01 -7.58703172e-01
2.38007933e-01 7.24879861e-01 7.42475837e-02 -4.81202751e-02
2.32761249e-01 -9.69808519e-01 -8.57446432e-01 -7.42408395e-01
7.18410254e-01 1.05294384e-01 -1.97328150e-01 -1.01752758e-01
1.96754351e-01 7.50696182e-01 -1.40431106e+00 1.48877367e-01
1.01050735e+00 9.42846894e-01 2.74466783e-01 6.64834917e-01
-3.11804444e-01 6.53074980e-01 -6.79785192e-01 -5.60789347e-01
3.45347911e-01 -6.17038429e-01 -5.14461935e-01 5.33857107e-01
6.09814227e-01 -6.69353783e-01 -3.76309425e-01 1.03434658e+00
5.69798112e-01 1.93668798e-01 7.20640361e-01 -9.36327398e-01
-1.05990803e+00 -8.34575891e-02 -8.96463335e-01 -3.06348652e-01
3.32377344e-01 2.81181067e-01 5.64443171e-01 -8.24805021e-01
6.35127544e-01 1.39526796e+00 4.00477648e-01 7.53002703e-01
-1.48130000e+00 -6.09839022e-01 3.28496039e-01 -2.10601538e-01
-1.23592246e+00 -2.90530592e-01 1.15705085e+00 -3.33113104e-01
4.39779848e-01 1.63969442e-01 8.15743983e-01 1.24305594e+00
3.12558055e-01 7.23881125e-01 1.27295184e+00 -3.83599818e-01
1.46237448e-01 3.87351424e-01 -4.85987097e-01 7.28087544e-01
5.05977049e-02 2.98354328e-01 -3.58936131e-01 1.16881028e-01
1.50549912e+00 1.61350593e-01 -4.64732826e-01 -7.31396616e-01
-1.06071508e+00 5.76905608e-01 4.23534572e-01 -1.79825723e-02
-2.06642240e-01 6.91842660e-02 1.05339669e-01 1.13861568e-01
8.07350338e-01 5.63862383e-01 -2.75980622e-01 3.32590729e-01
-5.92536807e-01 3.06209177e-01 4.50181574e-01 1.15071404e+00
6.86721504e-01 5.29783130e-01 -2.99866438e-01 9.99766886e-01
1.98102251e-01 7.16266692e-01 1.67506441e-01 -1.16108024e+00
2.53084123e-01 3.81129831e-01 2.03925818e-01 -9.87999320e-01
-1.38445394e-02 -5.13495207e-01 -1.10264373e+00 6.44711137e-01
2.59628415e-01 -1.98067784e-01 -8.23768020e-01 1.71601295e+00
4.63591963e-01 1.45590574e-01 -9.97303054e-02 9.99396682e-01
8.29938233e-01 8.00581753e-01 -1.76818579e-01 -7.75563717e-02
1.19883847e+00 -9.36627388e-01 -7.43960023e-01 4.85706404e-02
-1.53241344e-02 -9.32958961e-01 1.26746297e+00 3.51285398e-01
-1.56069267e+00 -8.82721782e-01 -9.81948674e-01 -2.14829236e-01
-4.97419154e-03 1.68986455e-01 5.43324649e-01 5.00473440e-01
-1.12068689e+00 5.02679467e-01 -3.13508421e-01 -8.69332105e-02
3.72034967e-01 1.87989902e-02 -7.82421157e-02 2.15201732e-02
-9.14058924e-01 6.79351509e-01 1.76163703e-01 7.09127709e-02
-6.76603496e-01 -9.17663574e-01 -8.67063761e-01 1.09485902e-01
2.01232970e-01 -1.41641533e+00 1.00302589e+00 -1.35687876e+00
-2.13338423e+00 9.67992246e-01 -2.31540307e-01 1.43269375e-01
4.78088468e-01 -2.49993145e-01 -2.98541039e-01 1.24385834e-01
-1.09461479e-01 7.15868950e-01 1.26978970e+00 -1.85828424e+00
-4.28132176e-01 -2.08017528e-01 2.46619195e-01 5.11392534e-01
9.51366499e-03 -4.12548780e-01 -5.25058389e-01 -1.11517262e+00
1.31574005e-01 -6.39635921e-01 -1.30636856e-01 2.66908199e-01
-2.95650542e-01 3.96090627e-01 6.63673937e-01 -6.88256383e-01
7.07025528e-01 -2.13099909e+00 4.98840392e-01 5.15256263e-02
4.40983921e-01 6.80464953e-02 -2.00131819e-01 9.69061349e-03
-3.27354103e-01 -1.52271628e-01 -1.21472310e-02 -4.08266276e-01
-2.83395112e-01 4.43882272e-02 -6.05486095e-01 1.67624503e-01
1.04248606e-01 8.88190269e-01 -8.82382631e-01 -2.80584127e-01
4.90662575e-01 7.70511091e-01 -7.60734618e-01 6.25439107e-01
-3.66970897e-01 8.97770941e-01 -5.24340451e-01 4.44657564e-01
1.05864608e+00 -2.28951484e-01 -4.31268141e-02 -5.94876707e-01
8.29593539e-02 -1.40084237e-01 -1.00252426e+00 1.95454144e+00
-7.70064890e-01 3.24502021e-01 1.29126966e-01 -6.02398217e-01
9.84807968e-01 1.47589028e-01 2.94505477e-01 -8.15504551e-01
1.78704202e-01 -7.18117133e-02 -6.52624130e-01 -4.95774686e-01
5.21451712e-01 -5.03140092e-01 2.31607020e-01 4.09904301e-01
3.57682258e-02 -6.30122721e-01 -3.47714722e-01 -4.01783064e-02
3.05664390e-01 6.39796138e-01 2.76546255e-02 5.91523759e-02
6.17131948e-01 -6.05473816e-01 3.08938205e-01 3.51139873e-01
3.77570361e-01 1.03309166e+00 3.98274809e-01 -6.71647131e-01
-1.23601520e+00 -1.39721060e+00 1.04128741e-01 6.06770992e-01
3.60366642e-01 -6.34565158e-03 -8.57908487e-01 -5.34492016e-01
-3.26973706e-01 8.45666170e-01 -6.13558292e-01 -2.42557034e-01
-5.86044967e-01 -3.70456398e-01 8.19309577e-02 1.96388111e-01
7.47558355e-01 -1.07425082e+00 -3.59967560e-01 -8.65086541e-02
-2.26931766e-01 -1.06771469e+00 -7.69208014e-01 -2.88290411e-01
-9.65117216e-01 -8.37197900e-01 -1.10703051e+00 -6.21505260e-01
1.00266063e+00 4.57327068e-01 1.19915521e+00 -1.47650033e-01
-2.11199775e-01 3.49141061e-01 -1.22897467e-02 -1.31864384e-01
-4.95309204e-01 -2.15057954e-01 -1.98576286e-01 3.96317542e-01
-2.90307969e-01 -1.03252268e+00 -8.94223213e-01 2.34329343e-01
-1.05787694e+00 7.38227129e-01 5.28077483e-01 9.02144134e-01
5.90698481e-01 -2.67303437e-01 2.15247154e-01 -7.84960449e-01
4.32473451e-01 -9.52559859e-02 -7.57558882e-01 2.35092700e-01
-4.22437072e-01 1.64171010e-01 7.85948932e-01 -3.37377280e-01
-1.62775147e+00 -1.20612994e-01 -8.25245082e-02 -8.34154427e-01
-2.68498868e-01 -2.05964386e-01 -5.04253209e-01 -1.89891323e-01
5.21427512e-01 4.10013884e-01 -8.77215061e-03 -5.23478806e-01
7.03982353e-01 2.38271579e-01 4.13699806e-01 -7.47952580e-01
1.15369976e+00 6.04480922e-01 1.38515253e-02 -6.78412259e-01
-1.00068927e+00 1.40180200e-01 -6.43156528e-01 -3.43764126e-01
1.01552784e+00 -9.78273511e-01 -7.40990996e-01 5.93023539e-01
-1.42855227e+00 -3.71062458e-01 -5.84679484e-01 2.44745687e-01
-7.71010935e-01 4.25222188e-01 -5.29330015e-01 -5.03629982e-01
-3.35213155e-01 -1.13257504e+00 1.33586824e+00 3.47673386e-01
-6.57803416e-02 -8.96840274e-01 -4.78804410e-02 3.16858798e-01
4.68359768e-01 3.96108836e-01 1.25164330e+00 5.39182782e-01
-9.86788988e-01 2.94564545e-01 -5.73987305e-01 4.06514645e-01
4.02592927e-01 -1.45623058e-01 -1.22622883e+00 -2.34660596e-01
3.41916621e-01 -4.16654274e-02 4.80666041e-01 5.68789721e-01
1.52711403e+00 -2.83272356e-01 3.91062871e-02 1.27280843e+00
1.43998253e+00 3.88124168e-01 7.21946836e-01 2.47033089e-02
1.00463378e+00 6.50636435e-01 1.17751569e-01 3.73662233e-01
5.41596264e-02 7.75784492e-01 3.21189255e-01 -4.91225153e-01
-4.28955942e-01 -5.94624937e-01 1.54510409e-01 6.14076614e-01
-2.22736225e-01 -3.13658893e-01 -2.75394507e-02 1.07909329e-01
-1.38995981e+00 -9.94614899e-01 3.06436986e-01 2.13131499e+00
6.87137365e-01 -4.25488912e-02 -1.93966165e-01 -3.08017462e-01
5.48417747e-01 3.89349371e-01 -7.89029121e-01 -5.51922679e-01
-1.02099366e-01 1.81483701e-01 9.68962163e-02 5.84362149e-01
-6.66189134e-01 8.89725149e-01 6.34634399e+00 8.02082360e-01
-1.27669728e+00 -9.36486498e-02 9.55128968e-01 -2.31521279e-01
-8.29103708e-01 -2.17921719e-01 -4.42584097e-01 3.31542134e-01
2.64257222e-01 -9.73551795e-02 7.48486161e-01 6.27168000e-01
2.73938477e-01 1.22494534e-01 -1.01294267e+00 1.28512561e+00
2.98982143e-01 -1.42931068e+00 7.54685462e-01 -1.20643154e-01
9.55776870e-01 -5.72767794e-01 3.79664481e-01 -8.92011151e-02
1.17169522e-01 -9.80997562e-01 9.11916256e-01 8.46830606e-01
1.23181522e+00 -7.11437464e-01 1.16197631e-01 8.80355909e-02
-9.72676039e-01 2.93375701e-01 -2.69520164e-01 1.83049634e-01
2.56168395e-01 5.20234704e-01 -2.96835393e-01 7.23413944e-01
6.97817981e-01 7.83001065e-01 -3.97155702e-01 7.03665555e-01
-2.84656256e-01 -1.84985425e-03 2.24963918e-01 2.67888516e-01
-8.53670761e-02 -7.43562162e-01 6.61841750e-01 7.05120981e-01
4.32915002e-01 2.01241553e-01 -2.49767557e-01 1.52320123e+00
-1.32861689e-01 -1.64689943e-02 -9.15255547e-01 5.01652002e-01
-1.13167502e-01 1.20304894e+00 -4.57468778e-01 -2.35950187e-01
-2.75526643e-01 1.36700857e+00 2.72878736e-01 8.97542596e-01
-7.69557595e-01 -2.41960853e-01 5.45377851e-01 2.45633885e-01
2.53464520e-01 -1.36904940e-01 -5.68305731e-01 -1.48208857e+00
7.43548572e-02 -8.26126635e-01 -2.17149377e-01 -1.60010183e+00
-1.22895515e+00 8.11626673e-01 8.86908770e-02 -1.47082126e+00
-9.72013548e-02 -5.95543504e-01 -7.38982499e-01 1.25524402e+00
-1.52287436e+00 -1.15427458e+00 -5.40549815e-01 5.81253111e-01
7.16441214e-01 -1.21067837e-01 7.32532740e-01 1.68347806e-01
-3.62757921e-01 5.58847249e-01 5.54192327e-02 -1.93237096e-01
7.65383244e-01 -1.19113982e+00 4.62747902e-01 6.27360284e-01
-1.93274498e-01 4.85251993e-01 5.61040401e-01 -4.12787229e-01
-1.17950392e+00 -1.08513522e+00 2.91168422e-01 -4.10592109e-01
3.88042368e-02 -3.97387356e-01 -8.31341088e-01 4.75593597e-01
5.18457770e-01 -2.68141121e-01 4.31794703e-01 -2.54814297e-01
-2.33746991e-01 -2.55291939e-01 -1.11256397e+00 1.10142338e+00
1.12114990e+00 -4.35388535e-01 -3.48538220e-01 9.34403241e-02
8.35410655e-01 -5.30524552e-01 -6.07057631e-01 2.44218558e-01
6.24809086e-01 -1.15127790e+00 1.17662859e+00 -3.60550582e-01
8.21299553e-01 -4.12201047e-01 3.68973352e-02 -1.61493433e+00
-4.75356162e-01 -8.20221663e-01 -7.33630210e-02 1.18021548e+00
2.00816974e-01 -5.26516676e-01 6.31805241e-01 5.22201121e-01
-1.39569595e-01 -6.17737591e-01 -4.25235957e-01 -3.36919129e-01
-2.88578179e-02 -2.23121839e-03 7.87132680e-01 8.03486288e-01
-5.87483704e-01 5.70017695e-01 -6.47797287e-01 7.66709298e-02
8.45611811e-01 4.93105859e-01 1.01429272e+00 -8.45921457e-01
-5.57428241e-01 -4.22097236e-01 1.55255333e-01 -1.40831327e+00
9.69485864e-02 -6.55363977e-01 -1.93473294e-01 -1.26782286e+00
1.72920048e-01 -3.82738322e-01 1.32763371e-01 -1.57801926e-01
-1.97624907e-01 1.39963835e-01 1.51312381e-01 8.81021619e-02
-7.04984516e-02 9.44396615e-01 2.18328166e+00 -7.23024085e-02
-3.22565645e-01 1.20691873e-01 -8.41808021e-01 8.26818109e-01
5.38954914e-01 -1.20662274e-02 -8.23194087e-01 -8.70382249e-01
3.00157517e-01 3.83714676e-01 5.84028542e-01 -6.87799454e-01
-2.45705605e-01 -2.42132396e-01 8.29073429e-01 -5.60522676e-01
4.71008927e-01 -8.31451476e-01 3.87059480e-01 1.21492274e-01
-3.20077479e-01 -3.40167969e-03 2.78512150e-01 6.24195457e-01
-4.77017425e-02 -3.38972695e-02 1.07214665e+00 -3.04730982e-01
-2.52181143e-01 4.99177814e-01 7.44883576e-03 -6.24656305e-02
7.62120187e-01 -4.21472043e-01 -4.62528877e-02 -5.11825740e-01
-7.41290033e-01 -1.26482621e-01 7.72185445e-01 5.01366138e-01
8.13620865e-01 -1.59750509e+00 -5.47511935e-01 7.11089253e-01
-7.96231106e-02 3.10053259e-01 6.12715542e-01 3.74592751e-01
-7.57803321e-01 5.40076755e-02 -3.99482280e-01 -5.64645767e-01
-8.98500562e-01 7.50481904e-01 6.90325439e-01 -1.35579690e-01
-8.38835955e-01 6.02277100e-01 1.14470196e+00 -5.13686836e-01
1.22767366e-01 -2.43862137e-01 -8.84527490e-02 -3.64099681e-01
3.99632126e-01 1.02990702e-01 -3.37271601e-01 -4.39691603e-01
3.74554068e-01 1.07112539e+00 -6.37649521e-02 -1.20235719e-01
1.22804368e+00 -4.33329254e-01 1.68437827e-02 3.11824381e-01
1.15249908e+00 2.41105817e-02 -1.65821350e+00 -2.60784239e-01
-9.50595975e-01 -7.43621349e-01 -7.96740502e-02 -7.34143078e-01
-1.26704729e+00 9.94485199e-01 6.46119118e-01 -1.14264421e-01
1.28588820e+00 -2.70976514e-01 8.92708480e-01 1.66975288e-03
2.74396241e-01 -8.18406820e-01 3.00865918e-01 3.22601527e-01
1.23354852e+00 -9.25530136e-01 -1.52354687e-01 -5.71444333e-01
-5.13589680e-01 1.06884909e+00 8.40140879e-01 -2.17348307e-01
5.11869133e-01 -7.41126854e-03 1.38282716e-01 -2.57353365e-01
-3.43061835e-01 3.63971531e-01 5.38680792e-01 7.09603727e-01
3.16940695e-01 -1.86138391e-01 1.79947451e-01 2.48814628e-01
-1.27773806e-01 -3.61864418e-02 3.74330103e-01 2.69701719e-01
-8.55315756e-03 -8.83715332e-01 -3.90281916e-01 2.11304016e-02
-1.99909881e-01 -7.15878084e-02 -1.62888974e-01 5.46816885e-01
1.66541293e-01 4.26079839e-01 2.39831321e-02 -1.50931269e-01
6.82834804e-01 -1.24342062e-01 8.17178190e-01 -5.20378053e-01
-2.68525124e-01 3.57759178e-01 -3.22092652e-01 -6.39051795e-01
-4.45216358e-01 -1.95187971e-01 -7.22879112e-01 -4.27222669e-01
-1.49041221e-01 -1.54672071e-01 3.02868694e-01 6.53158844e-01
3.23116958e-01 9.13589239e-01 1.03513992e+00 -1.19455850e+00
-2.38071263e-01 -5.49193859e-01 -7.67308950e-01 5.03345788e-01
2.06949711e-01 -5.69298506e-01 -2.66274005e-01 3.02704096e-01] | [9.340160369873047, -3.165867805480957] |
1d671a23-23bd-40e1-85aa-75eced087138 | scalable-approximate-inference-for-state | 1910.00879 | null | https://arxiv.org/abs/1910.00879v2 | https://arxiv.org/pdf/1910.00879v2.pdf | The Neural Moving Average Model for Scalable Variational Inference of State Space Models | Variational inference has had great success in scaling approximate Bayesian inference to big data by exploiting mini-batch training. To date, however, this strategy has been most applicable to models of independent data. We propose an extension to state space models of time series data based on a novel generative model for latent temporal states: the neural moving average model. This permits a subsequence to be sampled without drawing from the entire distribution, enabling training iterations to use mini-batches of the time series at low computational cost. We illustrate our method on autoregressive, Lotka-Volterra, FitzHugh-Nagumo and stochastic volatility models, achieving accurate parameter estimation in a short time. | ['Dennis Prangle', 'Tom Ryder', 'Andrew Golightly', 'Isaac Matthews'] | 2019-10-02 | null | null | null | null | ['normalising-flows'] | ['methodology'] | [-3.46671462e-01 -7.64007643e-02 -8.65435004e-02 -3.90164554e-01
-6.56319439e-01 -3.12889308e-01 8.89692307e-01 -4.20606852e-01
-4.02875215e-01 9.55501318e-01 5.01730852e-02 -5.39788187e-01
-3.11259836e-01 -9.42991734e-01 -5.73455155e-01 -7.36262858e-01
-2.01731384e-01 8.96332741e-01 1.37571722e-01 1.37902483e-01
7.16826767e-02 4.42161351e-01 -1.07917094e+00 -2.80614913e-01
5.17386615e-01 8.02710354e-01 1.14579056e-03 7.99578071e-01
-2.30963185e-01 7.82495618e-01 -6.00387931e-01 -4.67473418e-01
1.45422742e-01 -3.45390588e-01 -1.43439919e-01 -2.25477099e-01
-1.54720187e-01 -6.57659054e-01 -3.05580378e-01 7.81386793e-01
1.71504632e-01 3.15818250e-01 1.16660261e+00 -9.20400381e-01
-3.45787615e-01 6.35955572e-01 -3.56339216e-01 3.88987690e-01
-2.30849609e-01 -5.65597899e-02 5.92800379e-01 -5.15023947e-01
2.55436659e-01 1.39145195e+00 7.18349516e-01 3.83076102e-01
-1.56319523e+00 -4.24641073e-01 2.01002344e-01 3.59923765e-02
-1.30193055e+00 -2.79317081e-01 6.22065425e-01 -2.60868520e-01
1.25473166e+00 -8.49705413e-02 8.02429438e-01 1.43271470e+00
5.41430295e-01 7.62541592e-01 1.07656479e+00 -2.42575034e-01
8.01189959e-01 -2.20571667e-01 7.35984892e-02 1.40204787e-01
1.21592507e-01 1.39813557e-01 -5.77469170e-02 -6.25785708e-01
9.55032885e-01 6.26259267e-01 3.85967463e-01 -1.24758050e-01
-5.10356784e-01 1.24536014e+00 -2.48557448e-01 1.41075388e-01
-7.87058592e-01 5.90827346e-01 2.54674345e-01 4.72356468e-01
8.94960403e-01 -3.86017025e-01 -4.87520099e-01 -4.23190296e-01
-1.32554269e+00 5.08353412e-01 1.14694548e+00 6.97232306e-01
5.57464838e-01 6.15501761e-01 -9.78516415e-02 5.13808727e-01
4.69980866e-01 1.02353024e+00 4.11977649e-01 -1.11666739e+00
8.22566152e-02 -2.08154872e-01 4.55455899e-01 -3.33830148e-01
-1.67156994e-01 -3.23640108e-01 -1.05502665e+00 8.67831632e-02
4.80464697e-01 -3.41099620e-01 -1.03914964e+00 1.75540817e+00
3.00861537e-01 6.91560745e-01 -5.76174147e-02 2.00048387e-01
-2.52919614e-01 1.05698824e+00 -2.67402828e-03 -7.17704654e-01
9.38598394e-01 -3.14375430e-01 -8.13551843e-01 -1.67434961e-02
1.48593396e-01 -4.67569530e-01 5.01268983e-01 6.90638065e-01
-1.28948343e+00 -4.03050512e-01 -5.02833545e-01 2.65773833e-01
-3.03889573e-01 -2.97632545e-01 6.80601120e-01 7.96036363e-01
-9.08513486e-01 9.00214612e-01 -1.50565767e+00 -1.03849970e-01
1.97029144e-01 1.96908101e-01 2.64960796e-01 1.17614791e-01
-1.05262470e+00 7.68436551e-01 4.10683125e-01 1.58242449e-01
-1.20884180e+00 -7.87081242e-01 -4.29790139e-01 3.04307789e-01
3.19682837e-01 -7.94579744e-01 1.45333076e+00 -6.36901855e-01
-1.93218279e+00 -7.41612911e-02 -5.02453446e-01 -1.06761265e+00
4.57450330e-01 -1.51257962e-01 -4.37190711e-01 -2.57792473e-02
-2.31481105e-01 -1.36398360e-01 1.34985328e+00 -5.54075837e-01
-1.66506901e-01 -3.97860467e-01 -4.29586112e-01 -4.09835458e-01
-1.21298674e-02 -9.39327553e-02 -6.36707991e-02 -7.19187856e-01
-2.61832960e-02 -9.11577880e-01 -3.95495683e-01 -5.87241590e-01
7.27385059e-02 -5.13918340e-01 5.71181476e-01 -7.21437216e-01
1.36096513e+00 -1.97055507e+00 2.67979577e-02 3.90889138e-01
-1.85741469e-01 4.07473706e-02 1.76527843e-01 8.32681656e-01
2.26652026e-01 -1.16046071e-01 -1.99801087e-01 -7.40375102e-01
1.35924533e-01 6.18358254e-01 -8.33137274e-01 4.16758358e-01
-1.13878645e-01 9.02852714e-01 -7.68002808e-01 -1.72674954e-01
3.35602522e-01 3.71358812e-01 -5.23687243e-01 1.48956165e-01
-4.30588067e-01 1.93203732e-01 -4.54789639e-01 2.22255230e-01
5.93266904e-01 -3.33432108e-01 1.18471652e-01 4.63277072e-01
-1.62637249e-01 1.19871169e-01 -1.39010394e+00 1.25336635e+00
-6.01873279e-01 3.26069117e-01 -1.11467786e-01 -1.16893876e+00
7.03653097e-01 5.94562888e-01 4.78415549e-01 -2.51450181e-01
4.92716557e-04 2.46452108e-01 -2.56302595e-01 -2.37181664e-01
3.01538140e-01 -5.95969319e-01 -7.47739002e-02 5.14081717e-01
2.37870947e-01 -3.91420692e-01 3.61353785e-01 8.19965079e-02
8.67669463e-01 3.41074049e-01 2.54867822e-01 6.49870234e-03
-1.05353273e-01 -3.33142221e-01 4.54172701e-01 1.27751577e+00
2.44194701e-01 2.67595291e-01 4.78006750e-01 -4.28719342e-01
-1.36739528e+00 -1.54739451e+00 -2.53846407e-01 7.69782364e-01
-6.88420355e-01 -3.88149738e-01 -4.77215320e-01 -1.11898431e-03
9.29959640e-02 1.08180666e+00 -6.43196106e-01 3.49784121e-02
-4.07572567e-01 -9.66621399e-01 2.98390150e-01 3.93107861e-01
2.73111891e-02 -9.05419230e-01 -5.51195741e-01 7.71812022e-01
3.09587598e-01 -5.99142194e-01 -4.50237712e-04 2.83948064e-01
-1.34269619e+00 -3.43974739e-01 -9.74762857e-01 1.50039151e-01
1.13347597e-01 -2.99681097e-01 1.07119858e+00 -7.59201705e-01
-1.27125010e-01 4.64270592e-01 1.50262535e-01 -6.35095596e-01
-5.26555955e-01 -2.65856922e-01 3.27520818e-01 5.00498898e-02
3.93356353e-01 -8.99858594e-01 -6.03199124e-01 -2.82216966e-02
-9.86204684e-01 -3.23158592e-01 3.35943043e-01 7.18948603e-01
5.12867749e-01 1.08929249e-02 4.60049957e-01 -6.63269997e-01
6.08265698e-01 -7.62435257e-01 -1.01501179e+00 1.35516271e-01
-6.97235286e-01 2.43794158e-01 6.56853199e-01 -8.28876317e-01
-1.37650228e+00 -2.69712180e-01 -6.26239702e-02 -9.54581201e-01
-1.22196022e-02 7.28597999e-01 5.73927045e-01 6.13539875e-01
1.51721537e-01 4.26145911e-01 1.28637940e-01 -7.80478835e-01
4.12945569e-01 5.60043275e-01 3.79548281e-01 -5.38871765e-01
4.35867876e-01 7.47708440e-01 2.74686038e-01 -8.88576508e-01
-5.17669201e-01 -4.15575743e-01 -4.26544696e-01 5.92321232e-02
6.89646125e-01 -9.42410529e-01 -7.09021688e-01 6.38740599e-01
-9.42565918e-01 -5.61613679e-01 -5.56138992e-01 8.66222799e-01
-8.38266611e-01 2.48381779e-01 -8.51523221e-01 -1.46999335e+00
-9.00180787e-02 -5.72843015e-01 9.51343298e-01 9.79127735e-02
-2.04070434e-01 -1.42441761e+00 5.97474813e-01 -3.12906742e-01
6.15240037e-01 4.91072424e-02 7.86362112e-01 -7.78108060e-01
-3.65911424e-01 -5.28737545e-01 3.47964346e-01 4.67264682e-01
-1.75743597e-03 4.13909912e-01 -6.60367846e-01 -2.47386426e-01
3.61712903e-01 -2.52786130e-02 8.77491474e-01 1.08663666e+00
8.88322055e-01 -3.09172839e-01 -3.35829914e-01 2.50450909e-01
1.29205358e+00 2.92612702e-01 4.82618570e-01 -1.65730730e-01
2.56199747e-01 1.47079125e-01 2.16602847e-01 8.90943766e-01
2.65954077e-01 3.59724283e-01 9.45688561e-02 4.85807300e-01
6.63839996e-01 -3.21416140e-01 4.70338583e-01 7.05629110e-01
-2.18098089e-01 -1.72406375e-01 -6.77498937e-01 6.87839925e-01
-1.98353910e+00 -1.56873214e+00 -2.30229259e-01 2.48721886e+00
6.40840590e-01 3.74340922e-01 3.77327651e-01 -2.86410391e-01
2.92995125e-01 6.53399974e-02 -6.20667636e-01 -3.91632855e-01
1.35154545e-01 4.69277650e-01 4.87587690e-01 5.88050306e-01
-8.66650164e-01 7.36033738e-01 7.43450117e+00 9.13038909e-01
-9.57326353e-01 3.95256460e-01 4.98125315e-01 -5.19185543e-01
-2.42653340e-01 2.41906613e-01 -9.56694424e-01 8.63770664e-01
1.92624819e+00 -2.77063251e-01 4.48479891e-01 7.72618771e-01
6.38372183e-01 -1.57883927e-01 -8.60871673e-01 6.75946414e-01
-4.65029985e-01 -1.03930330e+00 -1.15503356e-01 2.92544723e-01
8.42485964e-01 3.26073945e-01 1.29111037e-01 5.54740429e-01
7.75373459e-01 -6.77077711e-01 4.51402158e-01 1.08149183e+00
1.04259148e-01 -9.37403083e-01 5.06358445e-01 7.56218851e-01
-8.30919921e-01 -1.99568585e-01 -6.75409675e-01 -2.43155599e-01
7.92424619e-01 9.34984088e-01 -4.30998921e-01 2.28799939e-01
5.31557739e-01 3.35679054e-01 1.32336393e-01 8.56491983e-01
4.27592210e-02 1.13679075e+00 -9.75124836e-01 -1.21974148e-01
3.87010336e-01 -7.49872983e-01 5.63737273e-01 8.64287198e-01
7.50250518e-01 -1.29518718e-01 3.84883136e-02 9.43469346e-01
5.39753973e-01 -3.35124403e-01 -5.91965318e-01 -1.71766356e-01
3.75687003e-01 8.40813339e-01 -6.19689167e-01 -7.33600199e-01
-6.08092248e-01 8.20680916e-01 7.55502507e-02 6.08693540e-01
-7.79264867e-01 -1.41640306e-01 3.46912920e-01 5.97744063e-02
8.70561957e-01 -5.97434103e-01 2.28815719e-01 -1.30673373e+00
-3.28295469e-01 -4.79932904e-01 4.17563438e-01 -6.71996593e-01
-1.38707387e+00 1.76254764e-01 7.00452745e-01 -8.35768282e-01
-1.38528347e+00 -4.03168768e-01 -7.90376723e-01 1.20400643e+00
-9.68950152e-01 -9.27777052e-01 6.10678315e-01 6.31072402e-01
4.24450547e-01 -7.23791793e-02 6.95302606e-01 -5.42640202e-02
-3.89399081e-01 -5.73354699e-02 9.63236332e-01 -3.22312087e-01
2.53286362e-01 -1.33039999e+00 7.39153624e-01 6.51579678e-01
3.27859163e-01 9.85661685e-01 1.07742214e+00 -8.38041425e-01
-1.32321858e+00 -7.68619001e-01 6.69903755e-01 -4.47284937e-01
9.88333225e-01 -3.54537278e-01 -1.07645655e+00 8.70731115e-01
1.52588382e-01 -1.89852551e-01 5.20339727e-01 2.64836133e-01
-8.12194869e-02 6.63540512e-02 -9.05863166e-01 4.51441199e-01
3.86538237e-01 -5.30781567e-01 -5.82472205e-01 2.43714169e-01
2.90187299e-01 2.07418948e-02 -1.02065885e+00 -3.12249968e-03
5.33142030e-01 -7.51944065e-01 8.98208141e-01 -7.48003066e-01
6.54711127e-02 -1.18641350e-02 -4.37766463e-02 -1.26402390e+00
-3.79344821e-01 -1.12133300e+00 -8.75783861e-01 1.11619222e+00
1.65198803e-01 -8.28794897e-01 5.32770038e-01 7.09979117e-01
2.06842467e-01 -3.24700058e-01 -1.33262575e+00 -1.09999752e+00
3.45400631e-01 -7.43717909e-01 5.25647283e-01 3.19880307e-01
-3.14249396e-01 9.76126641e-02 -8.60473394e-01 5.97686619e-02
9.44366217e-01 1.34663865e-01 7.43585765e-01 -1.45635712e+00
-6.89341068e-01 -3.14057201e-01 1.59769524e-02 -1.18725538e+00
3.58765274e-01 -2.79904872e-01 -7.17170388e-02 -1.19264913e+00
2.16292247e-01 7.03226104e-02 -3.56785059e-01 3.79875787e-02
-1.12217501e-01 -3.11080813e-02 -1.06213681e-01 2.99429208e-01
-3.21632206e-01 6.15857065e-01 6.71222270e-01 2.57149339e-01
-1.54129595e-01 6.93256378e-01 2.84248926e-02 6.22202396e-01
6.84183240e-01 -6.32759213e-01 -6.28805399e-01 9.98174772e-03
2.57604748e-01 5.75247049e-01 4.44112152e-01 -7.56786883e-01
1.27394482e-01 -4.05213773e-01 3.42278242e-01 -9.65882242e-01
5.88273644e-01 -6.59570456e-01 6.90078974e-01 5.15255451e-01
-1.03900783e-01 1.25347704e-01 1.52125522e-01 1.05112517e+00
-4.78887707e-02 -5.61634541e-01 6.24658287e-01 -2.98213333e-01
-2.50923157e-01 4.13650274e-01 -8.66931617e-01 -2.15099767e-01
6.72256649e-01 1.02251753e-01 3.17026168e-01 -8.22260678e-01
-1.08551788e+00 1.26055107e-02 3.19602415e-02 -6.85362220e-02
3.06853533e-01 -1.05356777e+00 -5.69030106e-01 8.02143440e-02
-5.06480396e-01 -2.67399460e-01 5.75532436e-01 7.72298932e-01
-5.23991324e-02 5.34020841e-01 1.17824107e-01 -6.37843430e-01
-7.49980569e-01 8.22121382e-01 2.45053008e-01 -5.70728779e-01
-4.41715539e-01 3.47648859e-01 -2.31369779e-01 -1.51729643e-01
-1.13285579e-01 -4.02413398e-01 1.92038938e-01 1.83761090e-01
6.30738735e-01 7.38458991e-01 -2.46685162e-01 -6.31034076e-02
8.91193375e-02 1.96323067e-01 -1.63308427e-01 -7.52677917e-01
1.50524306e+00 -2.87379652e-01 -5.61722554e-02 1.32429051e+00
1.02311528e+00 -2.87775338e-01 -1.54287815e+00 -2.05485120e-01
-1.31432474e-01 -1.02215424e-01 1.91510975e-01 -2.81302840e-01
-5.44209778e-01 1.10542595e+00 5.55230081e-01 8.14002573e-01
6.66193485e-01 -1.02724977e-01 4.98322666e-01 4.10443366e-01
4.62376744e-01 -1.16655231e+00 -2.78509259e-01 5.99874616e-01
3.33126903e-01 -9.29260552e-01 1.87806189e-02 4.12197113e-01
-3.13315779e-01 1.03562450e+00 -2.63606578e-01 -4.55804884e-01
1.15758467e+00 3.74982148e-01 -3.26669455e-01 9.97876450e-02
-1.14717805e+00 4.00491208e-02 1.06100380e-01 3.41382146e-01
2.30358973e-01 1.68612272e-01 -1.06622100e-01 4.22720194e-01
1.08385300e-02 4.97182071e-01 3.99327219e-01 8.78609776e-01
-2.90471762e-01 -1.02955270e+00 -2.68821776e-01 6.94700778e-01
-9.18803036e-01 -2.31895864e-01 4.31162268e-01 6.43206298e-01
-6.14123940e-01 6.96985185e-01 5.19588470e-01 5.06667852e-01
-1.81359798e-02 5.11949062e-01 5.44407547e-01 -3.85922074e-01
-1.82192206e-01 7.92689383e-01 -2.13478237e-01 -2.96373576e-01
-1.78981915e-01 -1.01063955e+00 -7.75539637e-01 -5.61312735e-01
-5.89497723e-02 2.80829042e-01 6.79137707e-01 1.11260676e+00
2.46390700e-01 2.95714200e-01 6.00871503e-01 -1.04622364e+00
-1.37594128e+00 -1.28064525e+00 -9.51996744e-01 1.47047743e-01
3.86765182e-01 -4.43271726e-01 -5.41467190e-01 -1.07958121e-02] | [6.887795448303223, 3.730541467666626] |
b3af9ee5-973a-4fbd-b5df-12b2c1d304ee | tracking-instances-as-queries | 2106.11963 | null | https://arxiv.org/abs/2106.11963v2 | https://arxiv.org/pdf/2106.11963v2.pdf | Tracking Instances as Queries | Recently, query based deep networks catch lots of attention owing to their end-to-end pipeline and competitive results on several fundamental computer vision tasks, such as object detection, semantic segmentation, and instance segmentation. However, how to establish a query based video instance segmentation (VIS) framework with elegant architecture and strong performance remains to be settled. In this paper, we present \textbf{QueryTrack} (i.e., tracking instances as queries), a unified query based VIS framework fully leveraging the intrinsic one-to-one correspondence between instances and queries in QueryInst. The proposed method obtains 52.7 / 52.3 AP on YouTube-VIS-2019 / 2021 datasets, which wins the 2-nd place in the YouTube-VIS Challenge at CVPR 2021 \textbf{with a single online end-to-end model, single scale testing \& modest amount of training data}. We also provide QueryTrack-ResNet-50 baseline results on YouTube-VIS-2021 val set as references for the VIS community. | ['Wenyu Liu', 'Bin Feng', 'Ying Shan', 'Yu Li', 'Xinggang Wang', 'Yuxin Fang', 'Shusheng Yang'] | 2021-06-22 | null | null | null | null | ['video-instance-segmentation'] | ['computer-vision'] | [ 1.33172452e-01 -5.40540181e-02 -5.43575943e-01 -6.43609941e-01
-1.18036246e+00 -4.38105881e-01 2.94339567e-01 -2.31872216e-01
-8.35235298e-01 3.48803729e-01 -3.06605399e-01 1.34942206e-02
1.08823225e-01 -5.24564147e-01 -1.09498131e+00 -2.20032677e-01
1.31252974e-01 5.85886061e-01 9.23943639e-01 -1.19798765e-01
7.61979911e-03 -1.24524191e-01 -1.43439281e+00 3.67450029e-01
6.58422410e-01 1.48060298e+00 3.33628476e-01 8.97205532e-01
-6.28738478e-02 8.27278852e-01 -3.48830700e-01 -5.81977189e-01
4.09484297e-01 -2.84493715e-01 -9.59207356e-01 2.57305615e-02
1.28401303e+00 -7.47650266e-01 -7.53918767e-01 1.04698777e+00
5.47138155e-01 1.76634207e-01 2.62216330e-01 -1.35202301e+00
-7.87911057e-01 6.28505230e-01 -7.04925239e-01 7.57523835e-01
6.69628605e-02 5.41528404e-01 1.29343915e+00 -1.00638056e+00
9.04533327e-01 1.01403785e+00 4.37971771e-01 4.53356385e-01
-9.20881212e-01 -7.14168727e-01 3.33049297e-01 4.93815392e-01
-1.45951247e+00 -4.30816799e-01 3.07655394e-01 -2.90520370e-01
1.06815743e+00 1.78027287e-01 7.76686728e-01 1.08363891e+00
-5.77806354e-01 1.63485622e+00 5.62885106e-01 3.75917941e-01
-8.44962448e-02 -2.45298333e-02 2.31182218e-01 7.06913173e-01
-1.61856905e-01 -4.83588837e-02 -5.37970424e-01 4.68597203e-01
6.98411763e-01 2.26774886e-02 -3.24849635e-01 -1.84693918e-01
-1.23893392e+00 5.15219152e-01 8.48327100e-01 -1.06566995e-02
-3.27584803e-01 6.45658910e-01 5.76662898e-01 2.31252626e-01
6.31685793e-01 1.90510795e-01 -7.64604151e-01 -2.64579415e-01
-1.29295170e+00 3.70548695e-01 5.58402061e-01 1.39911580e+00
6.51817441e-01 7.78693780e-02 -7.16096878e-01 9.02248323e-01
2.49681249e-01 8.53697896e-01 2.71574967e-02 -1.12999129e+00
6.83959842e-01 5.00936806e-01 1.02168284e-02 -6.99602842e-01
-2.17142701e-02 -5.00248909e-01 -5.08348107e-01 -3.99822056e-01
2.66765624e-01 4.41414164e-03 -1.30440521e+00 1.40054476e+00
3.97321254e-01 7.91818321e-01 -6.05373122e-02 1.05337918e+00
1.54608619e+00 7.28775263e-01 2.83614010e-01 4.69876220e-04
1.31917036e+00 -1.53827846e+00 -3.24879408e-01 -1.98735297e-01
5.30891716e-01 -4.81240362e-01 1.13514721e+00 2.12111250e-01
-1.26271248e+00 -7.20011234e-01 -5.79345405e-01 -4.99057442e-01
-2.70603776e-01 1.88974187e-01 4.95908409e-01 1.50117889e-01
-1.35074711e+00 2.22489700e-01 -8.42183232e-01 -3.71010989e-01
1.02671504e+00 2.88502902e-01 -1.26234815e-01 -2.17700034e-01
-1.02183151e+00 -1.03072040e-01 3.21062893e-01 1.18162811e-01
-1.38551128e+00 -1.03078115e+00 -7.70848870e-01 -1.82477236e-02
1.00644803e+00 -6.67424023e-01 1.42738152e+00 -1.02641797e+00
-9.47679400e-01 1.20451677e+00 -3.17344874e-01 -8.37054610e-01
9.55956161e-01 -4.52198476e-01 -9.09068808e-02 5.60854793e-01
5.52983046e-01 1.31169987e+00 7.20502436e-01 -9.65522110e-01
-1.02385962e+00 -3.46579343e-01 4.07595664e-01 3.44378538e-02
-4.08037975e-02 1.58011280e-02 -1.48882997e+00 -4.67944771e-01
2.48698369e-02 -8.20831180e-01 -7.85031244e-02 2.55130202e-01
-6.91273153e-01 -7.77006745e-01 7.91100264e-01 -4.72768337e-01
9.74709153e-01 -2.31795621e+00 -1.63884938e-01 -3.31113845e-01
4.63476241e-01 5.49200475e-01 -4.36973959e-01 9.70705971e-02
1.24004327e-01 3.03548515e-01 -2.35369161e-01 -5.22619247e-01
2.91014854e-02 -6.10232763e-02 -2.70125300e-01 3.44138712e-01
2.63652980e-01 1.59907293e+00 -9.45552409e-01 -7.22530484e-01
1.31154224e-01 3.56510520e-01 -6.08632147e-01 1.81046382e-01
-6.43721759e-01 2.10395246e-03 -7.84063041e-01 9.68101978e-01
6.07053101e-01 -6.02299511e-01 -4.55870986e-01 -4.03054565e-01
2.39108980e-01 -2.97856126e-02 -7.01215923e-01 1.92968309e+00
1.21986136e-01 9.19609308e-01 2.27482781e-01 -9.96657014e-01
4.49947417e-01 8.85679424e-02 7.49119163e-01 -1.25377703e+00
-6.91052303e-02 2.09492654e-01 -4.83004838e-01 -8.29411387e-01
7.27268040e-01 5.85920930e-01 1.55333042e-01 3.66716608e-02
1.37602448e-01 -1.70604710e-03 5.61333358e-01 5.89767396e-01
9.91097510e-01 3.47014099e-01 -4.71957415e-01 -6.84297904e-02
1.77603275e-01 3.51166964e-01 4.25417840e-01 9.61915016e-01
-5.98516464e-01 9.67728734e-01 3.98843229e-01 -2.28590816e-01
-7.39675581e-01 -8.94678175e-01 -1.71410576e-01 1.36039138e+00
7.24332273e-01 -3.47078264e-01 -6.96534455e-01 -8.04089785e-01
-2.07649264e-02 3.53334248e-01 -6.18524134e-01 1.43249840e-01
-6.38834238e-01 -3.96370709e-01 7.94835210e-01 5.75003386e-01
1.03039849e+00 -1.32238460e+00 -2.95341283e-01 2.84490824e-01
-3.29310864e-01 -1.67215717e+00 -6.58777118e-01 -2.68652529e-01
-7.18988538e-01 -1.37847733e+00 -1.08950686e+00 -8.61279786e-01
1.55647799e-01 7.63926625e-01 1.35248530e+00 6.68783784e-02
-3.72434318e-01 6.12989306e-01 -4.89344835e-01 -1.83647707e-01
3.32987338e-01 3.13781142e-01 -6.70244634e-01 -1.09688543e-01
5.65683663e-01 -8.21617842e-02 -1.25013614e+00 5.83585501e-01
-9.25850630e-01 -5.64430282e-02 3.15706193e-01 4.54481542e-01
1.14042079e+00 -5.26123405e-01 4.81183410e-01 -8.75980735e-01
1.31448045e-01 -5.29724181e-01 -5.88244975e-01 3.73625070e-01
-5.08988082e-01 -5.72800398e-01 5.25625162e-02 -1.13403492e-01
-5.30387342e-01 -1.07417628e-01 -4.59603846e-01 -8.14944744e-01
-1.57197654e-01 2.58789390e-01 1.10497981e-01 1.52674466e-01
5.73051751e-01 5.10661602e-01 -3.48897845e-01 -4.37927812e-01
5.41388154e-01 5.63564301e-01 8.19022000e-01 -3.24818671e-01
4.32098985e-01 4.87581968e-01 -5.79146266e-01 -7.79068470e-01
-1.19092762e+00 -9.58558679e-01 -2.07340509e-01 -7.61150569e-02
1.39387858e+00 -1.41010451e+00 -6.93585396e-01 4.96596605e-01
-1.02747178e+00 -7.36553252e-01 -3.88826102e-01 1.46301642e-01
-5.85088253e-01 1.48102880e-01 -6.07669890e-01 -4.35108036e-01
-5.69675803e-01 -1.48762608e+00 1.54724777e+00 3.91000152e-01
4.03762698e-01 -4.73923802e-01 -3.14264238e-01 8.26039732e-01
3.51272017e-01 8.64996836e-02 2.03086957e-01 -7.18318999e-01
-1.22742414e+00 -1.72343150e-01 -8.71491194e-01 4.87356037e-01
-4.32708502e-01 6.24800706e-03 -1.01388490e+00 -3.81085813e-01
-5.01275718e-01 -5.57480097e-01 1.44663799e+00 5.72687984e-01
1.52911162e+00 -7.69884214e-02 -3.52931738e-01 1.11819851e+00
1.49729037e+00 -1.74067579e-02 5.70812583e-01 1.64387599e-01
1.04704976e+00 2.21696019e-01 7.44730771e-01 9.56374183e-02
7.82882750e-01 6.79797053e-01 8.69161069e-01 -1.98648915e-01
-4.20263708e-01 -1.15845129e-01 1.29735723e-01 2.38366336e-01
9.45719332e-02 -5.33959568e-01 -6.67400539e-01 9.39292192e-01
-1.85360372e+00 -8.86908114e-01 -3.28430384e-01 1.73551917e+00
5.71256220e-01 2.70716935e-01 4.33651865e-01 -5.28750539e-01
6.05027854e-01 5.07058084e-01 -1.01347256e+00 3.74336898e-01
-1.82146460e-01 1.22937813e-01 6.55982792e-01 2.10263193e-01
-1.45347202e+00 1.63583839e+00 5.11428547e+00 1.12091470e+00
-1.25248051e+00 2.34758422e-01 1.01657093e+00 -3.22122186e-01
-3.17062736e-02 -3.65646005e-01 -1.00728667e+00 3.48204523e-01
6.48348987e-01 2.43830085e-02 2.33536422e-01 1.12187064e+00
2.33532563e-01 -1.05100833e-01 -9.68959033e-01 1.29806376e+00
-1.08270474e-01 -1.59067321e+00 8.18787962e-02 -1.88140243e-01
7.47935653e-01 1.01733005e+00 1.23712726e-01 6.35636687e-01
-3.99619192e-02 -9.37564850e-01 8.05628479e-01 3.08751315e-01
1.13035679e+00 -3.22834373e-01 6.07189357e-01 2.60174945e-02
-1.37667668e+00 2.27720320e-01 -3.62054288e-01 5.92892289e-01
2.78183162e-01 9.66651812e-02 -5.48311353e-01 4.26585197e-01
1.26695967e+00 1.28710914e+00 -6.93476975e-01 1.44335759e+00
6.84614554e-02 8.71569097e-01 -4.98607069e-01 6.62334114e-02
7.17607498e-01 3.72396596e-02 4.74708170e-01 1.13639343e+00
-7.80027881e-02 1.29545137e-01 4.45055127e-01 9.53667283e-01
-6.02622032e-01 4.18375625e-04 -1.32685646e-01 -3.08201492e-01
3.17463636e-01 1.18522835e+00 -8.49984825e-01 -6.71230793e-01
-3.60866100e-01 1.05322731e+00 2.07260344e-02 8.22719514e-01
-1.07081378e+00 -1.45722985e-01 6.87244415e-01 1.63000062e-01
8.20078731e-01 3.57040092e-02 1.04562365e-01 -1.07151246e+00
2.06489578e-01 -8.01779687e-01 2.98414826e-01 -8.35025847e-01
-1.19274080e+00 7.44837463e-01 9.98362079e-02 -8.98457170e-01
-2.56410264e-03 -4.04204130e-01 -2.51521319e-01 4.86540347e-01
-1.79387033e+00 -1.13805187e+00 -6.24354303e-01 8.99909079e-01
1.15257287e+00 1.74690157e-01 -1.71401016e-02 5.94516158e-01
-7.16239631e-01 8.93822372e-01 -9.77174044e-02 5.31570792e-01
5.61834872e-01 -1.26958227e+00 7.40447283e-01 6.99874520e-01
4.75490838e-01 1.10717617e-01 3.16408187e-01 -4.06628966e-01
-1.44031262e+00 -1.51817715e+00 4.94567871e-01 -6.16208851e-01
5.98756433e-01 -2.80851126e-01 -8.47938597e-01 7.51105487e-01
1.11723609e-01 6.85809970e-01 1.08615510e-01 -3.33024591e-01
-3.29446495e-01 -3.17741960e-01 -9.16549683e-01 3.60402852e-01
1.42024589e+00 -6.11655474e-01 -1.05666630e-01 8.25315475e-01
1.38906014e+00 -6.18713915e-01 -7.65065253e-01 5.08732915e-01
2.83721209e-01 -1.01307213e+00 1.32008076e+00 -7.73824036e-01
5.53786576e-01 -1.79622114e-01 -2.75281996e-01 -4.80447263e-01
1.77078858e-01 -6.94789946e-01 -3.15643623e-02 1.06663489e+00
5.05664289e-01 -2.12371126e-01 1.21680582e+00 4.56428975e-01
-2.41345882e-01 -1.24809384e+00 -8.55108976e-01 -6.48619354e-01
-1.92492276e-01 -8.25902820e-01 3.53312582e-01 6.94326937e-01
-1.00201511e+00 2.13047981e-01 -1.74057871e-01 -6.27282485e-02
5.98288059e-01 7.04538599e-02 1.04326928e+00 -7.77058840e-01
-3.90890449e-01 -5.32657206e-01 -4.22907233e-01 -1.96491563e+00
-1.59209922e-01 -9.50673044e-01 2.27562971e-02 -1.79099190e+00
2.19949320e-01 -4.31869090e-01 -3.50611746e-01 3.83530766e-01
-3.80358666e-01 7.26746976e-01 4.18010056e-01 4.92491215e-01
-1.54558992e+00 6.06388330e-01 1.44322896e+00 -4.61898386e-01
-1.81587830e-01 1.87648106e-02 -5.79513013e-01 3.59255821e-01
3.50979537e-01 -4.87343103e-01 -6.25063121e-01 -1.01658022e+00
1.77663833e-01 2.03627691e-01 7.60778069e-01 -7.82578290e-01
2.19789132e-01 -6.12733793e-03 -2.60198256e-03 -9.52770710e-01
4.97557223e-01 -5.59211731e-01 -3.02503049e-01 1.04585186e-01
-4.62719142e-01 2.27798130e-02 1.41225979e-01 1.04829729e+00
-4.23393190e-01 9.68310162e-02 5.37347019e-01 -2.42014229e-01
-1.41627657e+00 1.17427742e+00 5.52453101e-01 8.43098283e-01
1.19615555e+00 -3.82566065e-01 -5.05346358e-01 -1.84201106e-01
-7.95485258e-01 8.03948343e-01 7.97852129e-02 7.08929539e-01
6.48120642e-01 -8.33888710e-01 -9.81942236e-01 -2.23121017e-01
3.59763950e-01 5.71053922e-01 5.82254410e-01 1.07896829e+00
-8.10358346e-01 4.35508966e-01 4.05795038e-01 -1.27781498e+00
-1.02321351e+00 4.26703990e-01 4.77913737e-01 3.54342274e-02
-8.64835024e-01 1.65177608e+00 4.44099426e-01 -1.19705454e-01
5.32858849e-01 -3.20897460e-01 -4.61804233e-02 1.70806665e-02
4.89652097e-01 1.84875503e-01 -1.05906188e-01 -6.59666300e-01
-3.50168049e-01 5.76940954e-01 -3.38323563e-01 1.68866128e-01
1.21475828e+00 -1.14343867e-01 2.03306735e-01 1.34623885e-01
1.50848866e+00 -6.92351997e-01 -1.63939345e+00 -4.78129268e-01
-1.50573626e-01 -5.26460826e-01 8.60994030e-03 -7.24698961e-01
-1.78849614e+00 9.41816926e-01 7.03688323e-01 3.29220623e-01
9.79090571e-01 4.04210448e-01 1.30384588e+00 4.81575400e-01
2.12992385e-01 -9.28891659e-01 1.34940282e-01 2.57263154e-01
8.19199681e-01 -1.72482407e+00 -3.14206481e-01 -4.94132340e-01
-5.97221553e-01 6.05428219e-01 8.84025276e-01 -3.31816375e-01
5.77951491e-01 -3.21123332e-01 8.64088163e-02 -4.62821782e-01
-5.64392865e-01 -5.85660458e-01 4.73554730e-01 3.62632304e-01
1.04018696e-01 -1.52595416e-01 2.15862263e-02 1.92025483e-01
-7.97698740e-03 2.05935091e-01 1.89199615e-02 5.59209287e-01
-4.50777441e-01 -4.13633406e-01 4.33481991e-01 6.52327776e-01
-6.34864748e-01 -3.99450004e-01 -2.65068948e-01 9.76244926e-01
-1.62544906e-01 9.59657967e-01 1.04640901e-01 -1.96575224e-01
4.25442219e-01 -3.90570432e-01 1.06230296e-01 -4.51288998e-01
-6.98256314e-01 2.07399920e-01 1.68484205e-03 -1.21456254e+00
-4.99068111e-01 -3.72358412e-01 -1.33751798e+00 -6.14718385e-02
-1.93349481e-01 -1.78634048e-01 5.72911143e-01 8.81255686e-01
7.24460363e-01 5.23473203e-01 1.88864648e-01 -5.34505188e-01
-3.73300701e-01 -8.54695618e-01 -3.29496890e-01 4.74340439e-01
3.33175898e-01 -2.69085944e-01 -4.75757793e-02 -7.97095615e-03] | [9.30655574798584, 0.028429890051484108] |
3f52be41-d506-400a-a2a4-7e855bc72000 | l-scaled-attention-a-novel-fast-attention | 2201.02912 | null | https://arxiv.org/abs/2201.02912v1 | https://arxiv.org/pdf/2201.02912v1.pdf | λ-Scaled-Attention: A Novel Fast Attention Mechanism for Efficient Modeling of Protein Sequences | Attention-based deep networks have been successfully applied on textual data in the field of NLP. However, their application on protein sequences poses additional challenges due to the weak semantics of the protein words, unlike the plain text words. These unexplored challenges faced by the standard attention technique include (i) vanishing attention score problem and (ii) high variations in the attention distribution. In this regard, we introduce a novel {\lambda}-scaled attention technique for fast and efficient modeling of the protein sequences that addresses both the above problems. This is used to develop the {\lambda}-scaled attention network and is evaluated for the task of protein function prediction implemented at the protein sub-sequence level. Experiments on the datasets for biological process (BP) and molecular function (MF) showed significant improvements in the F1 score values for the proposed {\lambda}-scaled attention technique over its counterpart approach based on the standard attention technique (+2.01% for BP and +4.67% for MF) and state-of-the-art ProtVecGen-Plus approach (+2.61% for BP and +4.20% for MF). Further, fast convergence (converging in half the number of epochs) and efficient learning (in terms of very low difference between the training and validation losses) were also observed during the training process. | ['Akshay Deepak', 'Md Shah Fahad', 'Ashish Ranjan'] | 2022-01-09 | null | null | null | null | ['protein-function-prediction'] | ['medical'] | [ 3.95726502e-01 1.73032388e-01 2.86992759e-01 -1.63934752e-01
-6.01782978e-01 -2.80248910e-01 3.76629412e-01 5.81215978e-01
-8.14611077e-01 9.28562820e-01 -9.25029442e-02 -3.95719498e-01
-5.82679873e-03 -4.68020082e-01 -9.29857671e-01 -9.51736093e-01
4.60016802e-02 5.50699830e-01 2.11246237e-01 -4.41919893e-01
1.70759737e-01 4.55733925e-01 -1.41857922e+00 5.49913310e-02
8.76037955e-01 9.16324019e-01 5.40065885e-01 6.38178766e-01
-4.59852219e-01 7.22044706e-01 -6.29230499e-01 -4.41044837e-01
-1.09106407e-01 -2.27199942e-01 -7.59934723e-01 -3.23684305e-01
8.98948833e-02 2.19226941e-01 -1.64394751e-01 1.00407779e+00
7.12781250e-01 5.40473521e-01 6.88190758e-01 -7.17086256e-01
-1.05723143e+00 4.01312888e-01 -6.96252406e-01 4.31833982e-01
9.80133042e-02 6.28320500e-02 1.14218664e+00 -1.04867148e+00
5.72911739e-01 1.08399570e+00 5.15813053e-01 6.04978800e-01
-1.14768219e+00 -4.00853723e-01 1.10122845e-01 4.61995244e-01
-1.28458667e+00 2.17768580e-01 3.09788227e-01 -5.75589061e-01
1.63123202e+00 3.69382203e-02 1.28705487e-01 1.05838096e+00
4.38213110e-01 5.63556969e-01 5.37754834e-01 -4.19375181e-01
3.13151926e-02 1.28863469e-01 6.55344605e-01 4.18190569e-01
6.96161389e-02 -2.54542112e-01 -1.95304647e-01 -1.53248623e-01
3.83946627e-01 -6.51648715e-02 -4.13932323e-01 -1.86937153e-02
-9.25973296e-01 1.08008349e+00 4.39845204e-01 4.84756976e-01
-7.83234000e-01 1.20025858e-01 6.44442081e-01 8.32170174e-02
5.68085492e-01 6.63623571e-01 -8.89811456e-01 -9.50307623e-02
-4.90354776e-01 3.91617596e-01 5.69200635e-01 7.04886198e-01
5.98445714e-01 1.71987161e-01 -3.75964254e-01 9.83170569e-01
2.48017404e-02 1.16925500e-01 7.74838984e-01 -8.51560608e-02
3.51327747e-01 5.03297806e-01 2.10933462e-01 -1.00607765e+00
-4.58051205e-01 -6.77501798e-01 -7.62093425e-01 -5.52975088e-02
4.19471383e-01 -2.02210143e-01 -1.11992502e+00 2.00212550e+00
3.66341352e-01 1.73630103e-01 1.80194721e-01 8.61576915e-01
8.42130005e-01 9.75164056e-01 4.77916867e-01 -2.00045362e-01
1.58914626e+00 -9.76741374e-01 -8.35970402e-01 -4.60462905e-02
4.77339208e-01 -7.12036490e-01 1.12218821e+00 1.87835023e-01
-9.69330966e-01 -7.06906676e-01 -1.05586362e+00 -2.91235745e-01
-6.06070638e-01 -5.14093749e-02 5.01242936e-01 4.10103619e-01
-7.86429048e-01 7.02413857e-01 -4.26247507e-01 -4.45427269e-01
2.59821832e-01 7.31779993e-01 -3.17234904e-01 1.42741680e-01
-1.35742748e+00 1.14184701e+00 6.08443320e-01 -7.76856318e-02
-5.70426762e-01 -9.81167972e-01 -6.40808225e-01 6.11134350e-01
3.02466840e-01 -3.99695009e-01 9.97577667e-01 -9.90958095e-01
-1.39922500e+00 5.54504573e-01 -6.70030937e-02 -7.27404952e-01
2.19644651e-01 -4.90746439e-01 -8.84051025e-02 -4.35667150e-02
-1.69946596e-01 5.89255810e-01 3.07098508e-01 -8.01605284e-01
-4.51438963e-01 -3.22480530e-01 -5.87024614e-02 1.77303523e-01
-3.93370807e-01 3.66507359e-02 -3.61889035e-01 -6.79411829e-01
-5.36918044e-01 -6.34896696e-01 -3.06561768e-01 -3.51198912e-01
-6.00798391e-02 -4.76388752e-01 5.56526661e-01 -9.77764249e-01
1.05359936e+00 -2.00143909e+00 3.98558170e-01 -1.12754613e-01
9.92726758e-02 8.09655845e-01 -2.20073894e-01 5.11376679e-01
-4.04841751e-01 -2.94718333e-03 -2.72019774e-01 -1.43618569e-01
-1.71516329e-01 1.64483145e-01 5.04868990e-03 3.49735349e-01
5.33886611e-01 8.34312022e-01 -7.59014487e-01 -1.30850956e-01
2.40400612e-01 8.51182938e-01 -5.02627313e-01 2.35686094e-01
-2.96927124e-01 2.64759123e-01 -2.66373843e-01 3.71053308e-01
5.60184240e-01 -2.18736574e-01 2.36034974e-01 -2.90680438e-01
-5.71225248e-02 -8.31877887e-02 -5.95304549e-01 1.24800360e+00
-3.73133123e-01 4.13207233e-01 -1.44520581e-01 -1.14086223e+00
1.10458159e+00 5.47340095e-01 5.58292270e-01 -4.94468421e-01
4.42398787e-01 -2.17343587e-03 4.22130138e-01 -4.39119428e-01
3.72552454e-01 -2.48704314e-01 2.78727561e-01 2.05070719e-01
5.33376634e-01 5.66070378e-01 2.17479900e-01 -1.08483069e-01
9.83048022e-01 8.22717026e-02 4.86304790e-01 -5.62293291e-01
9.15790319e-01 -1.78642824e-01 5.62089443e-01 5.37432373e-01
-1.94008768e-01 3.44557524e-01 5.68909585e-01 -5.72074473e-01
-1.18946898e+00 -2.90777177e-01 2.52928510e-02 1.46857691e+00
-8.83926302e-02 -2.11698055e-01 -8.91548991e-01 -5.89002848e-01
4.16084751e-02 6.11068070e-01 -8.41421187e-01 -2.32925579e-01
-4.91233617e-01 -1.13887310e+00 4.04775262e-01 6.99016511e-01
6.82883114e-02 -1.38016093e+00 -5.43698251e-01 4.78152484e-01
3.40510607e-02 -9.56445396e-01 -3.71388644e-01 5.74213147e-01
-4.56843913e-01 -9.38123047e-01 -1.03128648e+00 -8.54542732e-01
3.95525575e-01 -1.43467754e-01 8.17508578e-01 5.81524409e-02
-3.04319620e-01 -1.95896566e-01 -4.45258200e-01 -7.27290571e-01
-8.45449120e-02 1.67068034e-01 -1.66672722e-01 1.63225055e-01
6.60395384e-01 -4.01534975e-01 -4.64764059e-01 -7.37746432e-02
-8.92247975e-01 -2.80905843e-01 7.13533640e-01 1.28020859e+00
6.47699535e-01 -2.99394369e-01 9.26809549e-01 -9.01522100e-01
8.35128367e-01 -6.91505015e-01 -5.30302703e-01 1.93517923e-01
-5.49860716e-01 2.35528201e-01 8.42439711e-01 -6.13651574e-01
-8.90745640e-01 4.37398907e-03 -5.35419464e-01 -3.76501143e-01
-2.23711450e-02 5.79187334e-01 4.11637016e-02 -1.78446978e-01
5.43952644e-01 2.64597803e-01 -1.77696848e-03 -6.55928016e-01
1.20629109e-01 4.83241051e-01 3.69253427e-01 -3.65389287e-01
2.55215049e-01 -1.08202897e-01 -8.68461952e-02 -8.72975647e-01
-5.00498056e-01 -5.04649818e-01 -3.32881302e-01 2.42163002e-01
9.98942375e-01 -5.18308818e-01 -1.04032707e+00 2.91655004e-01
-1.20542514e+00 -2.73304023e-02 4.46538348e-03 4.73964065e-01
-3.62548530e-01 6.63839102e-01 -6.81275368e-01 -7.88126647e-01
-9.47409987e-01 -1.29904068e+00 8.02901208e-01 1.63476408e-01
-3.08649629e-01 -1.02886760e+00 -9.94854197e-02 2.14542791e-01
4.96150553e-01 3.18389177e-01 1.43588758e+00 -1.19553840e+00
4.07998823e-02 -1.57764480e-01 -3.52786541e-01 4.36993569e-01
-1.23346381e-01 -1.20547138e-01 -8.29334676e-01 -3.11120480e-01
-5.16450070e-02 -2.29860485e-01 7.45226681e-01 3.76585066e-01
1.02092850e+00 -6.23391829e-02 -3.53838429e-02 3.68515253e-01
1.61020410e+00 5.50796032e-01 4.88366067e-01 3.42202455e-01
6.15723073e-01 4.85789269e-01 5.88175952e-01 4.22741652e-01
-3.32319945e-01 8.30648422e-01 4.22052145e-01 -2.28561878e-01
9.19313580e-02 1.37109235e-01 8.51534232e-02 5.36944330e-01
-8.81305784e-02 -7.00946212e-01 -8.77041519e-01 6.39943719e-01
-1.92452335e+00 -6.99842334e-01 -2.43086711e-01 2.13544750e+00
7.96767414e-01 -6.52422532e-02 -9.78210047e-02 -8.74833029e-04
7.97131956e-01 -9.83519703e-02 -5.91637790e-01 -8.93144846e-01
-5.47240376e-02 7.48633444e-01 4.84149396e-01 5.15212774e-01
-1.06254792e+00 9.57869232e-01 5.37548780e+00 1.05992460e+00
-1.13616991e+00 1.30581453e-01 5.84073246e-01 1.04274787e-01
8.25074017e-02 -5.12560666e-01 -8.20454538e-01 6.39286518e-01
1.18255031e+00 -1.83339924e-01 1.58895835e-01 8.72045696e-01
7.12729841e-02 2.61025429e-01 -8.27449739e-01 7.05893576e-01
3.00898310e-02 -1.16819334e+00 1.48631185e-01 3.87967676e-02
5.48837006e-01 1.60048142e-01 -5.93647510e-02 5.71866214e-01
6.20152429e-02 -1.05083776e+00 3.73460203e-01 3.03165406e-01
6.39740109e-01 -1.17515469e+00 1.35696006e+00 2.03449100e-01
-9.27419603e-01 -2.50727199e-02 -7.40386665e-01 8.44329968e-02
6.00989275e-02 5.06716192e-01 -8.69665861e-01 6.79042935e-01
5.62024951e-01 3.57772648e-01 -2.25030527e-01 7.02221751e-01
1.09589584e-01 4.99763668e-01 -1.32443950e-01 -4.17387187e-01
6.76178515e-01 -5.55932112e-02 3.43628764e-01 1.46379650e+00
2.08170414e-01 7.89146312e-03 5.88757806e-02 7.20373690e-01
-2.21917465e-01 5.90855479e-01 -1.20843679e-01 -1.35861218e-01
7.82295540e-02 1.14808643e+00 -4.77399886e-01 -3.95935297e-01
-3.29926193e-01 8.78170013e-01 6.27065480e-01 3.42409581e-01
-1.14577138e+00 -6.94345355e-01 8.74599755e-01 -1.95772275e-01
7.59633064e-01 1.03656910e-01 2.24484596e-02 -7.26935089e-01
-2.12431863e-01 -7.76496470e-01 2.82531112e-01 -5.11682093e-01
-1.29485059e+00 7.66813517e-01 -3.96373451e-01 -4.97429967e-01
6.45742044e-02 -9.12411034e-01 -4.13712651e-01 1.25173974e+00
-1.43161213e+00 -9.10238802e-01 -8.38037506e-02 3.44257891e-01
8.90178919e-01 -1.52000293e-01 9.96334314e-01 5.45713007e-01
-6.20855987e-01 7.93587327e-01 5.28410614e-01 -2.15487212e-01
4.08102214e-01 -1.26121998e+00 4.99647826e-01 5.19003451e-01
-2.15289310e-01 8.17888796e-01 8.63729537e-01 -4.65368390e-01
-1.23617423e+00 -1.14175439e+00 1.20750821e+00 -5.28796017e-02
6.51639462e-01 -2.85615772e-01 -1.25068748e+00 4.70126688e-01
2.02567801e-01 -3.42397213e-01 7.20763743e-01 -1.76968966e-02
-9.18025896e-02 2.48536959e-01 -1.29510736e+00 5.03962040e-01
4.96280432e-01 -1.76796272e-01 -4.38108563e-01 4.78735387e-01
9.87525880e-01 -1.25928670e-01 -9.72089350e-01 4.85511065e-01
5.38208783e-01 -6.75715208e-01 9.76941645e-01 -8.93260419e-01
4.14887518e-01 -3.27614844e-02 -1.01010151e-01 -9.18869436e-01
-7.30489552e-01 -4.06614035e-01 -9.73582715e-02 1.04355204e+00
4.94007587e-01 -6.68204963e-01 5.52966058e-01 2.03583598e-01
-2.80065656e-01 -1.26819074e+00 -8.45085323e-01 -6.11771166e-01
2.20881596e-01 1.75550245e-02 3.11925352e-01 7.78683841e-01
-1.55062512e-01 5.80483556e-01 -5.48651814e-01 -1.06052190e-01
1.85656786e-01 -2.82536179e-01 4.01302934e-01 -1.08513641e+00
-4.48834479e-01 -3.24171752e-01 -5.61116338e-01 -8.46316576e-01
6.32764474e-02 -7.12988555e-01 6.13318048e-02 -1.28246140e+00
2.86464244e-01 1.73337951e-01 -8.15879345e-01 5.14898837e-01
-6.00483596e-01 1.03922978e-01 1.36789009e-01 -1.37404740e-01
-1.82458282e-01 6.09815061e-01 8.31860185e-01 -1.91027336e-02
5.52996574e-03 -2.50728160e-01 -7.52389252e-01 2.56984413e-01
8.59625757e-01 -3.51359129e-01 -3.77577901e-01 -3.24329287e-01
-6.67135194e-02 -2.29920864e-01 -8.00012648e-02 -8.91869307e-01
-1.06847920e-01 6.12596003e-03 2.27734178e-01 -5.66474795e-01
4.29697126e-01 -6.47372663e-01 3.13043073e-02 5.50309479e-01
-4.03434187e-01 1.65832356e-01 3.94914478e-01 7.62371659e-01
-1.07187949e-01 -2.94634283e-01 9.60618079e-01 -2.05145925e-01
-5.03079176e-01 1.85782805e-01 -4.89431351e-01 -1.92251787e-01
1.12954414e+00 -1.53044552e-01 -1.44358665e-01 -1.94536895e-01
-8.13615561e-01 1.51676282e-01 -5.59154991e-03 4.40777272e-01
3.86078954e-01 -8.46924305e-01 -7.15647519e-01 1.51570678e-01
-3.23391333e-02 -3.31613362e-01 4.40289617e-01 8.03745389e-01
-6.71645045e-01 7.82361090e-01 -3.59795392e-01 -1.91941649e-01
-1.36694145e+00 7.89265215e-01 2.96482593e-01 -6.28918409e-01
-3.78306091e-01 1.16119790e+00 5.79156756e-01 -2.32677773e-01
2.72485942e-01 -2.75161117e-01 -3.77871841e-01 7.83525035e-02
5.14470816e-01 2.62168556e-01 3.41168821e-01 -6.37989044e-01
-3.78108770e-01 4.40727890e-01 -5.12293458e-01 3.62360984e-01
1.64936948e+00 2.83594161e-01 2.02607028e-02 1.41506091e-01
1.32982516e+00 -4.78376389e-01 -9.19341266e-01 -1.56290561e-01
2.36450881e-02 -1.49847999e-01 -8.89401510e-02 -1.01624501e+00
-8.44076216e-01 1.08680010e+00 6.98589265e-01 -1.11574017e-01
8.28790843e-01 -3.16752970e-01 9.13487792e-01 3.03752810e-01
-5.19839786e-02 -8.46053064e-01 -2.41391480e-01 7.66954422e-01
7.94302940e-01 -9.62490082e-01 -1.67944729e-01 -1.93673030e-01
-6.84874713e-01 9.44958568e-01 5.85036337e-01 -3.37152034e-02
2.55828798e-01 -5.80434464e-02 -1.32760897e-01 -3.09523731e-01
-8.53122711e-01 -1.27441198e-01 1.85562894e-01 5.89608192e-01
8.18949580e-01 -1.93139926e-01 -8.61345947e-01 6.82508349e-01
2.46763855e-01 -1.35167032e-01 2.17741951e-01 8.43442678e-01
-5.98152578e-01 -1.07617736e+00 -6.50539249e-02 3.20184022e-01
-1.01851475e+00 -3.30463946e-01 -3.87946218e-01 7.03528821e-01
2.18991131e-01 7.29884684e-01 -1.68875784e-01 -2.39249542e-01
2.91951299e-01 3.80845219e-01 1.56648651e-01 -5.38441956e-01
-1.14905477e+00 1.07689016e-01 -2.43251305e-02 -3.53547692e-01
-1.55532420e-01 -2.48338550e-01 -1.35211909e+00 -1.42160118e-01
-5.98089516e-01 3.46462607e-01 6.99523211e-01 7.85130441e-01
5.82794428e-01 8.20340335e-01 1.34479105e-01 -7.72351146e-01
-7.36227572e-01 -1.34272516e+00 -3.31835121e-01 6.81369066e-01
6.58159181e-02 -7.48865783e-01 -1.15256608e-01 -1.48748696e-01] | [4.776734828948975, 5.601308822631836] |
368d6faf-a44f-4a1a-97e9-40b017f9641a | pytrial-a-comprehensive-platform-for | 2306.04018 | null | https://arxiv.org/abs/2306.04018v1 | https://arxiv.org/pdf/2306.04018v1.pdf | PyTrial: A Comprehensive Platform for Artificial Intelligence for Drug Development | Drug development is a complex process that aims to test the efficacy and safety of candidate drugs in the human body for regulatory approval via clinical trials. Recently, machine learning has emerged as a vital tool for drug development, offering new opportunities to improve the efficiency and success rates of the process. To facilitate the research and development of artificial intelligence (AI) for drug development, we developed a Python package, namely PyTrial, that implements various clinical trial tasks supported by AI algorithms. To be specific, PyTrial implements 6 essential drug development tasks, including patient outcome prediction, trial site selection, trial outcome prediction, patient-trial matching, trial similarity search, and synthetic data generation. In PyTrial, all tasks are defined by four steps: load data, model definition, model training, and model evaluation, which can be done with a couple of lines of code. In addition, the modular API design allows practitioners to extend the framework to new algorithms and tasks easily. PyTrial is featured for a unified API, detailed documentation, and interactive examples with preprocessed benchmark data for all implemented algorithms. This package can be installed through Python Package Index (PyPI) and is publicly available at https://github.com/RyanWangZf/PyTrial. | ['Jimeng Sun', 'Cao Xiao', 'Tianfan Fu', 'Brandon Theodorou', 'Zifeng Wang'] | 2023-06-06 | null | null | null | null | ['synthetic-data-generation', 'synthetic-data-generation'] | ['medical', 'miscellaneous'] | [ 1.59859046e-01 -2.38894165e-01 -6.01466954e-01 -2.46485904e-01
-7.34903514e-01 -5.14707327e-01 4.48856175e-01 5.06936908e-01
-1.02588460e-01 6.91186428e-01 -3.41982156e-01 -7.75928795e-01
-2.68492043e-01 -7.24478662e-01 -2.19657898e-01 -6.33553445e-01
-3.12243737e-02 9.06633377e-01 -1.83218390e-01 1.97844714e-01
2.21862137e-01 6.68187737e-01 -1.13737345e+00 5.02891898e-01
1.07494926e+00 8.67729545e-01 1.61081359e-01 3.64670932e-01
2.33871520e-01 3.79125118e-01 -3.36586028e-01 -1.91837102e-01
8.10841173e-02 -4.91347194e-01 -6.60597742e-01 -4.26540554e-01
-2.56286979e-01 3.39430943e-02 2.37537682e-01 7.41120458e-01
1.02414548e+00 -4.50682670e-01 5.47927916e-01 -1.17345524e+00
-1.26055762e-01 9.30090249e-02 -1.27221271e-02 -2.00394869e-01
4.86088365e-01 5.97734272e-01 8.39974880e-01 -1.01678467e+00
5.41919410e-01 8.24282885e-01 4.83196855e-01 6.31267846e-01
-1.20575249e+00 -8.66933107e-01 -6.89874053e-01 2.04715028e-01
-1.42982900e+00 -4.39341217e-01 2.77345508e-01 -7.87107289e-01
8.58624279e-01 5.28981149e-01 6.55963600e-01 9.75485384e-01
4.07250136e-01 6.80998921e-01 8.92499387e-01 -9.63152722e-02
6.22457743e-01 2.00822532e-01 4.62173447e-02 4.51679856e-01
1.11078545e-01 3.88671756e-01 -1.04760461e-01 -6.98011458e-01
5.74695766e-01 2.17105165e-01 -1.63505390e-01 -2.75297165e-01
-1.08538747e+00 9.61800992e-01 6.90329596e-02 1.01657182e-01
-5.44858456e-01 -4.19236690e-01 4.22613740e-01 2.00316131e-01
1.86180994e-01 5.99394441e-01 -7.58224487e-01 -1.44105583e-01
-6.18238330e-01 5.15675545e-01 9.29302514e-01 5.48631728e-01
3.09498698e-01 -3.33872586e-01 -1.99713737e-01 9.06054914e-01
3.37047487e-01 2.93931544e-01 7.05505550e-01 -6.66357517e-01
-6.30458593e-02 7.01367199e-01 3.53518315e-02 -3.92555445e-01
-6.56408966e-01 -3.89654368e-01 -7.58123875e-01 1.89015001e-01
1.80239514e-01 -1.82789087e-01 -8.81390989e-01 1.22951949e+00
7.49135852e-01 4.98930924e-02 -1.28087029e-01 5.93531549e-01
1.05479193e+00 3.19977194e-01 6.18806124e-01 -4.40761894e-01
1.31874216e+00 -6.40148222e-01 -5.21696866e-01 1.24860816e-01
1.04968953e+00 -9.22968507e-01 7.87323236e-01 4.06463504e-01
-1.05074513e+00 -2.98054576e-01 -7.66082644e-01 2.17861101e-01
-5.46623886e-01 1.22980438e-02 8.41163397e-01 7.12375224e-01
-5.53586543e-01 8.45588863e-01 -9.03749347e-01 -3.04146379e-01
8.35489869e-01 5.96366227e-01 -4.45635736e-01 -1.64174870e-01
-1.24320495e+00 9.70580161e-01 4.11279649e-01 -3.22951823e-02
-8.06533575e-01 -1.24015343e+00 -9.29345250e-01 -2.47607991e-01
-1.61211230e-02 -1.13240421e+00 1.18160999e+00 -4.17333364e-01
-1.47638965e+00 7.86680341e-01 3.43817845e-02 -1.09830581e-01
3.44198167e-01 6.56801090e-02 -3.00005227e-01 -1.89010829e-01
7.36498237e-02 3.66712004e-01 1.04586035e-01 -3.91606361e-01
-3.34668100e-01 -6.41351581e-01 -6.82508528e-01 1.98962726e-02
1.76774710e-01 4.53208655e-01 -4.13244098e-01 -6.20584667e-01
-4.70410138e-01 -8.89615297e-01 -5.75038433e-01 2.85622454e-03
-2.59485662e-01 -4.22765464e-01 3.94356877e-01 -6.50825024e-01
1.39517617e+00 -2.02225113e+00 -1.88573316e-01 4.32253957e-01
5.52825145e-02 6.48680806e-01 -1.72709897e-01 6.32664680e-01
-6.03784442e-01 1.70549676e-01 -1.79574326e-01 1.97067708e-01
-2.36272901e-01 -3.38829011e-01 2.11518049e-01 4.29475099e-01
2.12786183e-01 1.11771524e+00 -9.08298016e-01 -1.72159210e-01
4.45998967e-01 4.49814558e-01 -4.83996004e-01 2.71531641e-01
-4.88739848e-01 8.66912961e-01 -8.78788531e-01 9.95105624e-01
7.53385723e-01 -6.52576864e-01 4.90981013e-01 2.21256495e-01
3.09597962e-02 3.75743538e-01 -9.68250036e-01 1.51111770e+00
-1.23342194e-01 -1.13812029e-01 -3.27048868e-01 -7.89691985e-01
7.65317738e-01 5.56886554e-01 8.81927371e-01 -6.03206396e-01
9.32205915e-02 4.21325266e-01 1.50120899e-01 -6.82776630e-01
-3.83980125e-01 1.14128571e-02 3.08569908e-01 5.55316269e-01
-3.06614697e-01 -5.36797829e-02 2.44711176e-01 -3.25753361e-01
1.12670279e+00 1.80898145e-01 6.64494455e-01 1.23160794e-01
3.96486938e-01 3.49631280e-01 5.12223005e-01 4.27554041e-01
1.33753391e-02 1.89210936e-01 2.78776586e-01 -7.19292223e-01
-1.12102818e+00 -9.36199069e-01 -9.87535179e-01 6.16622686e-01
-4.66401041e-01 -4.54457104e-01 -4.74120408e-01 -3.57224405e-01
1.17565364e-01 4.93729144e-01 -4.81730014e-01 8.27728882e-02
-1.40321895e-01 -1.23839211e+00 3.87281328e-01 7.09531605e-02
2.38458976e-01 -1.34351730e+00 -1.59014747e-01 5.05122781e-01
2.75587678e-01 -6.03004754e-01 -1.73462868e-01 -3.62059884e-02
-8.62183690e-01 -1.55416739e+00 -7.16710329e-01 -8.47895861e-01
4.68153179e-01 -3.48013550e-01 8.57029736e-01 1.92450806e-01
-8.21504235e-01 -1.57599717e-01 -1.57075003e-01 -7.90944993e-01
-5.53447545e-01 -9.95901600e-02 -3.66448127e-02 -4.51151758e-01
5.83335280e-01 -5.06438971e-01 -9.00621772e-01 5.22835851e-01
-9.26705897e-01 7.58723542e-02 5.99529922e-01 9.17089939e-01
8.58828068e-01 -2.92084992e-01 7.03203440e-01 -9.90545750e-01
7.61845112e-01 -6.97464526e-01 -7.14137316e-01 2.61930823e-01
-8.40837181e-01 -2.38832027e-01 3.92181695e-01 -1.84093401e-01
-4.53184247e-01 4.05606806e-01 -7.01690555e-01 1.05330780e-01
-4.20935690e-01 9.69486058e-01 -1.67179316e-01 9.72888991e-02
6.38882875e-01 8.21977258e-02 5.61651707e-01 -7.49791503e-01
-1.00839613e-02 9.73925233e-01 -2.89871395e-01 -1.51979059e-01
2.36195564e-01 -3.09392624e-02 8.49433765e-02 -5.38391531e-01
-3.90502572e-01 -3.79098266e-01 -1.33772850e-01 2.67474979e-01
6.34966135e-01 -9.16489840e-01 -1.01683021e+00 5.30855834e-01
-8.41167867e-01 -5.39857626e-01 1.56537756e-01 7.24120378e-01
-3.17574501e-01 1.39243394e-01 -4.76629168e-01 -2.49190152e-01
-8.36101234e-01 -1.54949832e+00 7.27522552e-01 1.50318384e-01
-6.03346229e-01 -1.05826271e+00 3.96833062e-01 2.86491424e-01
2.92443126e-01 4.26710486e-01 1.16442060e+00 -1.19404411e+00
-3.49584520e-01 -6.04638040e-01 1.81424215e-01 2.06376627e-01
1.54312968e-01 2.53891259e-01 -7.38165379e-01 -2.22049952e-01
-2.47243062e-01 -2.41483122e-01 2.65301108e-01 7.12075651e-01
1.39293480e+00 -2.43507877e-01 -8.01635623e-01 6.55636787e-01
1.10185897e+00 7.91661620e-01 9.07813668e-01 2.97356844e-01
1.53209299e-01 4.02017266e-01 7.16203630e-01 6.13931000e-01
1.58244699e-01 9.78246868e-01 2.36356065e-01 -5.21444261e-01
2.48063177e-01 1.76220521e-01 7.51987025e-02 9.48700830e-02
-3.27404924e-02 1.31431431e-01 -1.10099411e+00 5.52870668e-02
-1.67519629e+00 -9.25745130e-01 -3.68623167e-01 2.55188489e+00
1.21704972e+00 -3.32641244e-01 3.93324137e-01 -1.93481907e-01
4.71403301e-01 -6.62917912e-01 -6.88043356e-01 -4.83624011e-01
1.80197075e-01 7.66253054e-01 2.91233808e-02 4.30095106e-01
-1.05993557e+00 7.93300033e-01 6.69565201e+00 9.92259383e-01
-1.22323549e+00 -1.86354786e-01 1.15571141e+00 8.75790790e-02
-3.84943224e-02 -1.23538151e-01 -6.60246193e-01 6.60119534e-01
9.59477782e-01 -5.33820033e-01 1.59416020e-01 9.11070168e-01
8.28964889e-01 1.61657199e-01 -1.34934676e+00 6.21624291e-01
-6.20708942e-01 -1.77733564e+00 -3.25388052e-02 1.79971948e-01
2.33930618e-01 1.00781120e-01 6.96566254e-02 1.10658623e-01
2.80950725e-01 -1.31850469e+00 -2.29881749e-01 1.90857798e-01
9.61488128e-01 -5.37808299e-01 7.49029160e-01 1.21762961e-01
-6.92547739e-01 -1.57298986e-02 -1.91766303e-02 1.27662286e-01
7.75621235e-02 7.59591997e-01 -1.25742543e+00 6.64449096e-01
5.07206321e-01 9.41446185e-01 -5.84472001e-01 1.73058033e+00
-1.09446354e-01 4.22503948e-01 -6.89401403e-02 -1.17625771e-02
-2.87512928e-01 -2.97670275e-01 2.15828016e-01 9.59616601e-01
2.54005104e-01 -4.96447971e-03 1.62028626e-01 6.71389163e-01
3.66056919e-01 4.22999412e-01 -4.22826916e-01 -1.89586326e-01
5.31964958e-01 1.10813940e+00 -3.36080074e-01 -1.67307988e-01
-1.86292037e-01 5.34747422e-01 -3.65370035e-01 2.90085912e-01
-8.42847764e-01 -5.54385304e-01 8.56071830e-01 2.71378219e-01
-8.15325975e-02 3.75594378e-01 -4.74457294e-01 -6.60357475e-01
-2.46534824e-01 -1.37730646e+00 6.92396402e-01 -5.73564529e-01
-1.18524241e+00 5.72143435e-01 -1.52018189e-01 -1.39800680e+00
-1.90895230e-01 -8.12936246e-01 -7.24364877e-01 1.09352088e+00
-1.09468400e+00 -8.78838897e-01 -5.60003333e-02 4.09205377e-01
2.37007827e-01 -3.44587475e-01 1.35947359e+00 7.26008296e-01
-9.76709843e-01 5.40010571e-01 2.31696144e-01 -3.34019437e-02
8.35102439e-01 -7.49056339e-01 3.53655815e-01 4.34307307e-02
-4.68686521e-01 8.55096340e-01 4.78901297e-01 -7.73514688e-01
-1.18717515e+00 -1.11556339e+00 1.00601017e+00 -4.13126439e-01
7.48123944e-01 1.26635293e-02 -6.00417256e-01 4.30411816e-01
-6.30548820e-02 -2.47591838e-01 1.43070400e+00 -2.77554560e-02
5.84448874e-02 -2.93541644e-02 -1.05587041e+00 5.70489407e-01
2.64166981e-01 -4.05822247e-02 -1.01504773e-02 9.31679547e-01
3.58124495e-01 -4.31831568e-01 -1.44585788e+00 6.85436249e-01
7.11124837e-01 -3.92499834e-01 1.06123388e+00 -8.99268210e-01
3.23004007e-01 -5.50891578e-01 6.13421857e-01 -9.88248348e-01
-3.36169124e-01 -6.62545741e-01 1.54076099e-01 5.89075565e-01
1.11280298e+00 -8.16592097e-01 7.41310179e-01 8.19014013e-01
-3.77773717e-02 -1.29765368e+00 -8.30714226e-01 -4.92397785e-01
3.76734324e-02 -3.16711426e-01 6.76181495e-01 1.01818979e+00
5.04603624e-01 4.06827092e-01 -1.02338307e-01 -9.02353451e-02
3.39484334e-01 -6.09687716e-03 6.27547443e-01 -1.21759653e+00
-5.23998201e-01 -6.60399139e-01 -3.76651734e-01 -5.28310776e-01
-3.88930082e-01 -1.11075366e+00 -6.12267494e-01 -1.53160179e+00
3.84147167e-01 -8.38799119e-01 -2.27855727e-01 9.04647648e-01
-3.16556364e-01 1.24254175e-01 -2.89040476e-01 3.33724380e-01
-1.26412749e-01 1.15503035e-01 1.24812353e+00 -6.86393008e-02
-5.31916976e-01 5.94142795e-01 -6.87309980e-01 2.67510533e-01
1.02491546e+00 -6.26827359e-01 -1.11303933e-01 1.44933060e-01
2.26481277e-02 7.86631703e-02 2.57520616e-01 -5.93926191e-01
-3.41598950e-02 -4.45763409e-01 6.00346863e-01 -3.72802973e-01
-7.19237316e-04 -6.25860155e-01 9.12342846e-01 1.00415027e+00
-2.26123467e-01 1.23394758e-01 3.56606692e-01 6.50259778e-02
-9.21270028e-02 -1.31400913e-01 6.80478990e-01 -1.08659342e-01
-1.68596119e-01 7.66855538e-01 -3.18760186e-01 -4.38278496e-01
1.38531709e+00 -1.55187309e-01 -2.86348015e-01 7.57743120e-02
-1.09092855e+00 4.15951103e-01 3.93443227e-01 2.37539530e-01
4.44418877e-01 -1.23224032e+00 -9.03203547e-01 2.54947305e-01
4.21200871e-01 -9.01859179e-02 1.87205479e-01 1.21009612e+00
-8.35355878e-01 5.98331451e-01 -1.17863128e-02 -4.56130028e-01
-1.48663640e+00 6.64794385e-01 5.94067514e-01 -4.31287408e-01
-2.99295455e-01 4.47459728e-01 4.72430214e-02 -4.56919879e-01
1.36803567e-01 1.41303927e-01 -2.80021220e-01 -2.66979009e-01
9.14263189e-01 9.18039232e-02 3.97834688e-01 -8.18334799e-03
-5.46004832e-01 3.46582979e-01 -2.27532297e-01 4.18082803e-01
1.65896845e+00 6.49696231e-01 -2.68973202e-01 -1.51186988e-01
9.28037524e-01 -3.89715105e-01 -6.68661952e-01 1.85127944e-01
1.20890677e-01 -2.35705331e-01 -1.00053273e-01 -1.36189914e+00
-6.71934962e-01 5.61447322e-01 8.17802787e-01 -1.56078592e-01
1.00308657e+00 -3.59598882e-02 3.71379346e-01 3.48729603e-02
6.27215281e-02 -7.71210074e-01 -2.63042361e-01 1.68027356e-01
8.96819949e-01 -1.21393692e+00 2.71568716e-01 -4.19566542e-01
-6.52843177e-01 9.04678285e-01 1.96669921e-01 5.32647252e-01
8.84564400e-01 8.68220627e-02 1.46870792e-01 -3.12512994e-01
-8.24685037e-01 8.35448503e-02 2.69917458e-01 5.06114721e-01
9.19066787e-01 8.51477906e-02 -8.36048722e-01 6.73703194e-01
3.21713507e-01 6.54698372e-01 -1.50686800e-02 9.33492303e-01
-1.07902437e-01 -2.09870124e+00 -1.55594170e-01 5.05398333e-01
-6.56127453e-01 -5.56507051e-01 -2.39416137e-01 5.90503097e-01
1.63552128e-02 7.59656250e-01 -3.14032704e-01 -1.51397809e-01
2.98936188e-01 2.03091223e-02 2.18989581e-01 -8.15080583e-01
-7.47296631e-01 2.79152632e-01 1.44335717e-01 -5.00382245e-01
-1.94797646e-02 -5.06722748e-01 -1.02065539e+00 -1.52430832e-01
-2.18688250e-01 4.39944565e-01 8.30850363e-01 6.57260776e-01
9.58454311e-01 2.34739706e-01 7.63111532e-01 -3.61456484e-01
-3.50647539e-01 -9.46648836e-01 -4.07407224e-01 1.98928520e-01
-2.33128980e-01 -3.65202487e-01 -4.15135324e-02 5.96565567e-02] | [5.35670804977417, 5.739792823791504] |
37935041-ebf9-4219-b182-050c93effc47 | gazedirector-fully-articulated-eye-gaze | 1704.08763 | null | http://arxiv.org/abs/1704.08763v1 | http://arxiv.org/pdf/1704.08763v1.pdf | GazeDirector: Fully Articulated Eye Gaze Redirection in Video | We present GazeDirector, a new approach for eye gaze redirection that uses
model-fitting. Our method first tracks the eyes by fitting a multi-part eye
region model to video frames using analysis-by-synthesis, thereby recovering
eye region shape, texture, pose, and gaze simultaneously. It then redirects
gaze by 1) warping the eyelids from the original image using a model-derived
flow field, and 2) rendering and compositing synthesized 3D eyeballs onto the
output image in a photorealistic manner. GazeDirector allows us to change where
people are looking without person-specific training data, and with full
articulation, i.e. we can precisely specify new gaze directions in 3D.
Quantitatively, we evaluate both model-fitting and gaze synthesis, with
experiments for gaze estimation and redirection on the Columbia gaze dataset.
Qualitatively, we compare GazeDirector against recent work on gaze redirection,
showing better results especially for large redirection angles. Finally, we
demonstrate gaze redirection on YouTube videos by introducing new 3D gaze
targets and by manipulating visual behavior. | ['Andreas Bulling', 'Peter Robinson', 'Louis-Philippe Morency', 'Tadas Baltrusaitis', 'Erroll Wood'] | 2017-04-27 | null | null | null | null | ['gaze-redirection'] | ['computer-vision'] | [ 1.82713883e-03 1.29392713e-01 -6.45026118e-02 -2.96285570e-01
-1.59590870e-01 -8.86656880e-01 4.19042319e-01 -7.56592751e-01
-2.16606751e-01 3.88571978e-01 4.58567202e-01 -1.43190667e-01
2.58263409e-01 5.76288626e-02 -7.83940315e-01 -5.46221375e-01
3.26092392e-01 -1.04589825e-02 7.28421062e-02 4.57537211e-02
6.94232643e-01 5.46176136e-01 -2.15447760e+00 -3.83621529e-02
3.97064984e-01 6.04090273e-01 -2.31452018e-01 1.11720407e+00
3.43899190e-01 5.67405581e-01 -3.33889276e-01 -2.28801116e-01
3.82579088e-01 -6.28220439e-01 -7.40777016e-01 4.35418934e-01
1.16285014e+00 -5.06718278e-01 9.62042287e-02 7.67143667e-01
2.39945382e-01 1.28158405e-01 5.02489269e-01 -1.38443267e+00
-5.88626444e-01 -2.77023107e-01 -1.06938732e+00 3.10483128e-01
1.02430022e+00 3.84072721e-01 5.80922127e-01 -8.81790638e-01
1.03428972e+00 1.50391066e+00 3.68814856e-01 1.08856761e+00
-1.52042580e+00 -9.40156102e-01 2.18379647e-01 -6.92199171e-03
-1.27488780e+00 -9.12137866e-01 6.03935540e-01 -6.75925434e-01
5.49821734e-01 5.86570859e-01 7.84556270e-01 1.00235653e+00
1.51454553e-01 4.76793677e-01 1.29081893e+00 -6.02130830e-01
-2.14247838e-01 1.22802919e-02 -1.95755493e-02 6.34902358e-01
-2.01532274e-01 2.74656296e-01 -9.02067840e-01 3.50763947e-02
9.12884176e-01 -1.36209577e-01 -8.15753043e-01 -5.77731550e-01
-1.24776077e+00 2.91588455e-01 3.22635859e-01 -3.66316497e-01
-4.60907556e-02 6.48464262e-03 -2.34768972e-01 2.66168654e-01
7.09288716e-01 4.03827131e-01 -1.75198153e-01 -1.56772107e-01
-8.01303208e-01 4.86183017e-01 5.02480626e-01 1.00786316e+00
7.13781178e-01 -3.34998608e-01 -1.29294187e-01 3.01296294e-01
7.74351656e-01 7.17670739e-01 2.85331786e-01 -1.34510422e+00
3.70988585e-02 4.42732990e-01 3.46931934e-01 -6.96111262e-01
-5.07929802e-01 5.28659344e-01 5.35671301e-02 9.82142866e-01
7.35890031e-01 -2.90719837e-01 -9.71982419e-01 1.81451547e+00
7.29755878e-01 6.06897213e-02 -3.75494897e-01 1.24867046e+00
7.18018651e-01 1.96781799e-01 -8.01826864e-02 -2.40779296e-01
1.65647340e+00 -8.61172676e-01 -8.00602794e-01 2.20475215e-02
4.45276707e-01 -9.11397636e-01 1.14153326e+00 5.22462726e-01
-1.37907052e+00 -4.45349008e-01 -5.87990403e-01 -4.29912359e-01
1.76365659e-01 -1.88656487e-02 1.84781581e-01 8.25457156e-01
-1.46884537e+00 2.33819678e-01 -8.72526169e-01 -5.19997418e-01
2.53153831e-01 5.38978040e-01 -5.74761987e-01 4.38038886e-01
-5.07968545e-01 7.59221852e-01 -5.08990824e-01 -1.27917573e-01
-4.42650408e-01 -9.08973694e-01 -1.07442987e+00 -1.01886190e-01
1.63345769e-01 -7.75616109e-01 1.42989552e+00 -1.20897353e+00
-1.92016780e+00 1.27996373e+00 -1.06070411e+00 1.27727911e-01
4.62169141e-01 -2.54406989e-01 -3.24709415e-01 5.73548563e-02
-2.69032925e-01 1.06498981e+00 1.47210884e+00 -1.25106883e+00
-6.84453428e-01 -6.61519170e-01 2.66040802e-01 4.63637263e-01
7.85368308e-02 4.31503564e-01 -6.44384325e-01 -1.21638052e-01
-9.18550119e-02 -1.24760509e+00 5.08219540e-01 5.53959906e-01
-5.76033473e-01 -1.45374164e-01 1.04987717e+00 -6.03323936e-01
1.25768721e+00 -2.27070594e+00 4.10862416e-01 1.48140565e-01
9.22334552e-01 -8.27545524e-02 8.59498158e-02 -9.11350772e-02
-5.27399719e-01 2.46585254e-02 4.06511009e-01 -8.67153466e-01
-3.22683245e-01 -3.93576473e-01 -1.44266248e-01 6.25351131e-01
1.58858355e-02 8.89366806e-01 -6.48846626e-01 -3.58417779e-01
1.00691371e-01 6.46851122e-01 -8.99195611e-01 1.30034909e-01
4.52671945e-02 7.37978637e-01 -2.00062711e-02 5.55925429e-01
6.69946134e-01 -3.31011027e-01 2.75985483e-04 -1.69081882e-01
-5.66689432e-01 1.22637548e-01 -6.84920907e-01 1.41596615e+00
-2.39145413e-01 1.39512479e+00 1.98544888e-03 2.24554092e-01
6.34389579e-01 2.51214933e-02 1.86700925e-01 -5.49990296e-01
4.32555854e-01 -3.10105264e-01 -1.73939869e-01 -7.48321235e-01
4.59969342e-01 1.51289925e-01 3.30109745e-01 1.14191997e+00
-1.68545559e-01 1.35704011e-01 1.31841311e-02 -1.69044808e-02
3.99375618e-01 6.68377340e-01 9.85535458e-02 -2.56817430e-01
4.32867855e-01 -3.58560652e-01 -2.05714852e-01 -1.58572540e-01
-2.72384405e-01 1.06100452e+00 6.90986931e-01 -2.80531019e-01
-8.95745456e-01 -9.82495964e-01 -1.40755558e-02 1.19776976e+00
2.02124685e-01 -2.36603156e-01 -1.40076363e+00 -5.17807424e-01
-1.09293081e-01 6.40625000e-01 -1.23463380e+00 6.42145351e-02
-4.95027542e-01 -6.60791844e-02 1.13122568e-01 6.88826516e-02
-6.96130097e-02 -9.61461008e-01 -1.02207351e+00 -6.43309593e-01
-1.57372013e-01 -6.69700086e-01 -1.22354972e+00 -6.77564263e-01
-4.92715269e-01 -1.44930589e+00 -9.85828400e-01 -4.65892226e-01
1.08263457e+00 6.44219100e-01 8.37693095e-01 1.60818860e-01
-1.42879039e-01 4.40048516e-01 3.15373167e-02 -3.49958092e-01
-2.87840635e-01 -1.71296999e-01 3.81339848e-01 3.30743074e-01
5.00547886e-01 -2.45351389e-01 -7.19225824e-01 5.01051545e-01
-4.22671050e-01 1.83195308e-01 -2.00690299e-01 3.07521343e-01
9.39785317e-02 -8.64693105e-01 -2.80411065e-01 -7.42932677e-01
6.51712835e-01 -3.91128436e-02 -1.01741779e+00 1.66109681e-01
-5.26736081e-01 -1.55436248e-02 -1.17709838e-01 -5.22367895e-01
-1.04761708e+00 -1.77749157e-01 3.89335960e-01 -8.77533257e-01
-1.72617570e-01 -4.97626334e-01 2.02216074e-01 -3.03428978e-01
1.07729709e+00 -2.35793352e-01 5.64974785e-01 -4.28267747e-01
4.57355261e-01 6.97619021e-01 4.80611145e-01 -4.19194959e-02
7.42835343e-01 8.56458366e-01 -1.64827853e-01 -6.92851961e-01
-8.23888123e-01 -3.40137303e-01 -9.89147842e-01 -4.88387108e-01
1.12016881e+00 -6.39595628e-01 -1.68135679e+00 7.19818115e-01
-1.02397919e+00 -3.06482524e-01 -1.06760979e-01 4.32566881e-01
-5.66094458e-01 1.97370406e-02 -5.35612032e-02 -6.21119857e-01
-1.23910025e-01 -1.26545930e+00 1.38518035e+00 6.14340365e-01
-6.84470832e-01 -9.65626359e-01 2.92327970e-01 2.04257891e-01
1.42130122e-01 1.40831143e-01 4.47666585e-01 4.10799794e-02
-5.11825502e-01 2.04660729e-01 -2.23583862e-01 -3.28583494e-02
3.38023692e-01 6.05503142e-01 -1.32591701e+00 -1.60954013e-01
-3.99264544e-01 -7.53023997e-02 2.41489768e-01 6.03144288e-01
6.80859566e-01 -1.47778362e-01 -6.31357551e-01 9.69623148e-01
6.28162980e-01 -8.25417321e-03 5.79507172e-01 3.15809548e-01
8.62849355e-01 1.03864467e+00 3.79463613e-01 2.32096195e-01
5.85121751e-01 7.98279643e-01 3.19713414e-01 -1.27069458e-01
-4.22589362e-01 -2.15810031e-01 2.06940100e-01 -1.11787627e-02
-4.33696777e-01 -1.24785267e-01 -8.15661788e-01 1.21396288e-01
-1.33370662e+00 -8.65848660e-01 -3.14004958e-01 2.23768997e+00
8.56706262e-01 -2.77182728e-01 5.14805555e-01 -1.71571121e-01
7.60072291e-01 -1.89906880e-01 -7.27509797e-01 -2.82034427e-01
3.97188127e-01 2.82020457e-02 3.73921663e-01 8.95313561e-01
-6.30520821e-01 9.99560058e-01 7.32703304e+00 -2.29600042e-01
-1.60974586e+00 -1.21236287e-01 1.98879942e-01 -8.37110221e-01
-3.82999212e-01 -7.11733028e-02 -9.24468040e-01 3.81169885e-01
6.13155365e-01 -1.73023567e-01 6.92620635e-01 2.96845585e-01
4.79715168e-01 -2.76446581e-01 -1.35472870e+00 1.26414549e+00
5.45787156e-01 -1.17182505e+00 -3.45144093e-01 4.34687674e-01
3.78920555e-01 -1.98609784e-01 2.33429909e-01 -3.69685113e-01
-1.35069475e-01 -9.90621150e-01 9.34039593e-01 1.02777100e+00
1.29284394e+00 -4.18603897e-01 -2.31853649e-01 -8.56319815e-02
-6.03182137e-01 2.04241306e-01 1.86328173e-01 -7.81204700e-02
1.04959808e-01 -4.50391054e-01 -7.85684168e-01 -1.75652102e-01
1.00390363e+00 7.68644035e-01 -7.18894839e-01 9.38176751e-01
-2.37867922e-01 1.65355518e-01 -3.42411041e-01 1.01774134e-01
-5.07349789e-01 -2.19512150e-01 7.97023296e-01 5.73375702e-01
5.19903600e-02 4.11283188e-02 -7.28130877e-01 1.01331508e+00
-1.03209630e-01 -3.18765581e-01 -4.05392230e-01 4.16987568e-01
3.76568079e-01 1.07915366e+00 -3.85492235e-01 9.79524329e-02
-6.41879365e-02 9.25673485e-01 -5.69901615e-03 8.09842944e-01
-6.41386509e-01 -2.87297875e-01 1.26619232e+00 6.75270081e-01
7.51374438e-02 -4.39582281e-02 -4.59635258e-02 -1.20210779e+00
-2.32636213e-01 -6.71852052e-01 -2.15045676e-01 -1.61513650e+00
-4.69558030e-01 7.55976379e-01 2.32203290e-01 -1.34763169e+00
-6.30110621e-01 -5.29537439e-01 -4.06121314e-01 1.16669929e+00
-1.36944032e+00 -1.13541806e+00 -6.69337213e-01 9.27742124e-01
3.70656282e-01 1.13673754e-01 6.02818549e-01 -1.14311658e-01
-6.48597240e-01 8.27436090e-01 -3.39837760e-01 -2.11298734e-01
1.18054974e+00 -1.04096758e+00 8.01390946e-01 6.54255569e-01
-1.21224999e-01 8.66752803e-01 8.06898952e-01 -1.45922884e-01
-1.18785203e+00 -3.52978081e-01 9.54015315e-01 -1.19473970e+00
4.52199131e-01 -6.77959025e-01 -7.74961054e-01 1.02911353e+00
3.48421037e-01 2.78706457e-02 5.27565539e-01 6.58760294e-02
-3.75348032e-01 3.06623913e-02 -1.08107531e+00 1.08320534e+00
1.13366771e+00 -5.30678213e-01 -5.21078825e-01 -3.19599546e-02
4.45061266e-01 -8.02798212e-01 -3.09667259e-01 -2.15755865e-01
1.06493855e+00 -1.19943511e+00 7.95111477e-01 -7.16883957e-01
5.59833720e-02 -4.05861586e-01 5.37969053e-01 -1.11445093e+00
-1.67874068e-01 -1.19718170e+00 -1.25259057e-01 9.81151581e-01
1.56344414e-01 -6.68658078e-01 6.45568013e-01 8.11998248e-01
3.90204042e-01 -2.81288564e-01 -4.36424255e-01 -1.27347605e-02
-3.93241197e-01 -1.91505775e-01 7.05242813e-01 6.26763821e-01
1.24940448e-01 1.52137995e-01 -3.42245728e-01 2.30001900e-02
6.48305893e-01 -3.07870600e-02 1.46941209e+00 -1.33423316e+00
1.70626178e-01 -3.80002707e-01 -2.58495599e-01 -1.14025676e+00
1.95375606e-01 -2.11832836e-01 -1.98894426e-01 -8.14040601e-01
-1.50549889e-01 3.88800763e-02 4.55724865e-01 4.66277152e-01
-1.36815041e-01 6.79073215e-01 3.60074103e-01 6.26525700e-01
-1.84399679e-01 -4.27261218e-02 1.70572662e+00 4.78607118e-01
-5.06185412e-01 1.50668114e-01 -8.73639226e-01 7.75500655e-01
3.64114285e-01 -2.90249765e-01 -5.27055264e-01 -3.63562435e-01
4.98932600e-01 2.26609670e-02 6.84190333e-01 -5.92263341e-01
2.46359587e-01 -1.69493034e-01 4.01611149e-01 -3.75843823e-01
3.38302851e-01 -4.87634152e-01 5.64806499e-02 -2.23760100e-04
-2.65563965e-01 1.83996990e-01 3.45806748e-01 3.70231330e-01
2.39272028e-01 -1.36249112e-02 8.50692093e-01 3.69469434e-01
-4.84319121e-01 2.24238500e-01 -2.84379154e-01 1.11562029e-01
1.12181532e+00 -6.27689898e-01 -2.75323689e-01 -5.57610273e-01
-1.07715106e+00 1.03329597e-02 1.24191833e+00 8.21157873e-01
5.29361188e-01 -1.00873184e+00 -4.48412240e-01 1.00544369e+00
1.52322501e-01 -1.40833527e-01 1.54998943e-01 1.02502954e+00
-6.91534340e-01 2.58293420e-01 -3.19510728e-01 -1.11602020e+00
-1.81225717e+00 5.15335083e-01 6.44077003e-01 7.07348168e-01
-4.02606100e-01 9.26098287e-01 7.75549054e-01 6.27363706e-03
4.09362614e-02 -3.64131659e-01 -5.70861220e-01 1.12311974e-01
9.68732655e-01 3.25028807e-01 -1.51425973e-01 -9.36831772e-01
-4.09191072e-01 1.19910133e+00 -9.40102488e-02 -2.57772177e-01
8.34701180e-01 -8.78355980e-01 3.88073586e-02 2.79251695e-01
1.13579595e+00 5.88627994e-01 -1.76733577e+00 9.00333971e-02
-5.91226518e-01 -7.82128990e-01 -1.38556048e-01 -5.10703206e-01
-9.50829506e-01 8.45090449e-01 5.71672857e-01 9.35216174e-02
1.23832381e+00 5.34658618e-02 3.44807684e-01 -1.57337233e-01
2.45447662e-02 -4.65707302e-01 -4.97486852e-02 2.26583451e-01
8.70729327e-01 -1.09092367e+00 -8.48756209e-02 -2.33115703e-01
-6.43819511e-01 9.68207717e-01 5.65449893e-01 -5.43370247e-02
1.03523326e+00 4.79509309e-02 6.05712950e-01 -4.40892607e-01
-6.44312561e-01 -8.06264728e-02 9.26850855e-01 8.23537767e-01
2.91231662e-01 -3.70953768e-01 2.86807925e-01 -1.26108348e-01
-6.05892181e-01 1.72994778e-01 6.61554813e-01 5.30694127e-01
-1.41421303e-01 -8.06092322e-01 -7.77121425e-01 -3.08156218e-02
-2.86810011e-01 -2.43457556e-02 -4.18066561e-01 1.05016696e+00
-9.52331200e-02 8.10734451e-01 6.33155286e-01 -2.76048928e-01
3.83980006e-01 1.55106140e-02 8.43923271e-01 -6.50508463e-01
-3.55964184e-01 2.64785320e-01 -2.42863417e-01 -9.47171092e-01
-5.14095962e-01 -9.61573601e-01 -8.89453411e-01 -6.18266225e-01
-2.70365626e-01 -3.89594406e-01 6.81136250e-01 7.37626016e-01
7.06473053e-01 2.88467258e-01 6.09635890e-01 -1.40576255e+00
2.95402911e-02 -9.34714139e-01 -5.19231379e-01 3.51947308e-01
1.12686479e+00 -9.23845828e-01 -6.87439799e-01 5.88439465e-01] | [14.130725860595703, 0.064551942050457] |
8a3dc3f9-4695-42f0-a68a-b58d6a32e7ef | tetrahedral-diffusion-models-for-3d-shape | 2211.13220 | null | https://arxiv.org/abs/2211.13220v1 | https://arxiv.org/pdf/2211.13220v1.pdf | Tetrahedral Diffusion Models for 3D Shape Generation | Recently, probabilistic denoising diffusion models (DDMs) have greatly advanced the generative power of neural networks. DDMs, inspired by non-equilibrium thermodynamics, have not only been used for 2D image generation, but can also readily be applied to 3D point clouds. However, representing 3D shapes as point clouds has a number of drawbacks, most obvious perhaps that they have no notion of topology or connectivity. Here, we explore an alternative route and introduce tetrahedral diffusion models, an extension of DDMs to tetrahedral partitions of 3D space. The much more structured 3D representation with space-filling tetrahedra makes it possible to guide and regularize the diffusion process and to apply it to colorized assets. To manipulate the proposed representation, we develop tetrahedral convolutions, down- and up-sampling kernels. With those operators, 3D shape generation amounts to learning displacement vectors and signed distance values on the tetrahedral grid. Our experiments confirm that Tetrahedral Diffusion yields plausible, visually pleasing and diverse 3D shapes, is able to handle surface attributes like color, and can be guided at test time to manipulate the resulting shapes. | ['Konrad Schindler', 'Jan D. Wegner', 'Torben Peters', 'Nikolai Kalischek'] | 2022-11-23 | null | null | null | null | ['3d-shape-generation'] | ['computer-vision'] | [-1.38754979e-01 5.15077263e-02 6.31411612e-01 -1.15486778e-01
-1.84753478e-01 -9.06603336e-01 1.05762601e+00 -3.53210010e-02
-2.27467954e-01 4.78748918e-01 -2.05793723e-01 -2.15051576e-01
-2.33066410e-01 -1.29936659e+00 -7.00463057e-01 -1.10244751e+00
-1.62399590e-01 9.71551597e-01 3.38908434e-01 -1.76374510e-01
4.25778419e-01 1.17044318e+00 -1.68450928e+00 1.55282497e-01
9.00452971e-01 8.10420930e-01 1.35559887e-01 4.56187278e-01
-6.17995679e-01 8.42611045e-02 -3.16480309e-01 -4.57738966e-01
1.82076871e-01 -1.01982569e-02 -5.29193461e-01 1.26016885e-01
1.11981211e-02 1.40615374e-01 -3.59797105e-02 1.10749102e+00
3.63061637e-01 1.21633820e-01 1.39588416e+00 -1.16355479e+00
-1.22642338e+00 2.27845341e-01 -5.21969378e-01 -2.70429254e-01
1.80513546e-01 1.44139871e-01 5.05528331e-01 -1.08060408e+00
8.74703884e-01 1.71343720e+00 6.66047990e-01 7.72621930e-01
-1.60127854e+00 -2.09633261e-01 -5.09753041e-02 -2.28697628e-01
-1.21621668e+00 1.14887752e-01 1.04168332e+00 -6.65937483e-01
7.29967535e-01 4.63984579e-01 9.54152405e-01 1.21766794e+00
2.71893889e-01 6.53362036e-01 1.37771082e+00 -3.46701801e-01
7.53012180e-01 5.67933284e-02 -2.16332555e-01 6.51365697e-01
2.83826947e-01 -1.24862954e-01 -1.46325320e-01 -3.75624269e-01
1.19520247e+00 2.84318142e-02 -1.96854860e-01 -9.40995395e-01
-1.23529959e+00 9.13634300e-01 5.91075063e-01 4.25888032e-01
-5.76227427e-01 3.85991991e-01 -6.46049576e-03 -5.68359904e-03
8.30633581e-01 3.72006923e-01 -7.67932236e-02 3.23293321e-02
-6.83578134e-01 5.32854795e-01 8.11916828e-01 7.33130574e-01
8.77110422e-01 1.27420798e-01 -9.22502056e-02 6.16719186e-01
5.89304805e-01 7.95869231e-01 1.69005409e-01 -1.20946372e+00
-8.84395465e-02 6.84249401e-01 1.60122395e-01 -1.00418234e+00
-2.84129381e-01 -5.24906740e-02 -1.19518793e+00 8.47151041e-01
3.09158385e-01 7.31974617e-02 -1.16018569e+00 1.39417887e+00
4.73725885e-01 -1.80236120e-02 -5.41057885e-02 9.49244857e-01
5.44822931e-01 8.78208935e-01 -1.01290084e-01 2.19815314e-01
9.91031647e-01 -1.35636047e-01 -2.28013724e-01 3.87601376e-01
2.79539347e-01 -5.44444084e-01 9.14759636e-01 5.72555602e-01
-1.37841988e+00 -2.62764037e-01 -7.82900453e-01 -2.82947589e-02
-5.73318660e-01 -3.27293664e-01 7.21095145e-01 8.06431115e-01
-1.62394392e+00 1.03946650e+00 -1.05706549e+00 -2.61576235e-01
5.61339676e-01 2.88693458e-01 -1.52949184e-01 -1.68616287e-02
-9.63171422e-01 8.42622757e-01 1.97115578e-02 1.38820603e-01
-6.14245474e-01 -6.38629675e-01 -7.55338430e-01 -9.81956050e-02
-2.41882429e-01 -1.17112768e+00 7.01728106e-01 -6.08535767e-01
-1.75926292e+00 1.06772101e+00 -4.86541875e-02 -2.27649778e-01
5.65462708e-01 3.89213115e-01 2.65314132e-01 1.47489607e-02
-1.50760338e-01 1.00474012e+00 1.15245879e+00 -1.67463934e+00
2.39253014e-01 -4.28015023e-01 -5.51892072e-02 1.54056817e-01
-6.04351312e-02 -4.43848580e-01 -1.76400423e-01 -7.56952703e-01
5.02652287e-01 -1.05104721e+00 -5.01999140e-01 4.87044960e-01
-3.67622465e-01 -2.38889173e-01 7.95181751e-01 -1.74401984e-01
3.21912080e-01 -2.07829356e+00 5.84078729e-01 6.18336082e-01
3.22685301e-01 -8.18641186e-02 9.96139832e-03 3.34314406e-01
-3.28634232e-02 4.49730039e-01 -8.39155257e-01 -3.88189852e-01
3.22524965e-01 4.55545723e-01 -4.42022197e-02 3.06362212e-01
5.23681700e-01 8.47625256e-01 -7.83194602e-01 -1.31758571e-01
2.94793338e-01 1.04793370e+00 -7.32649028e-01 -1.97821915e-01
-5.20835698e-01 5.38518906e-01 -6.79520249e-01 2.78354198e-01
1.06272936e+00 -1.22699201e-01 -3.73759687e-01 3.48411016e-02
-2.38706931e-01 1.55010456e-02 -1.31024408e+00 1.69345284e+00
-4.07902867e-01 2.78467536e-01 1.62545934e-01 -6.75713778e-01
1.29182851e+00 1.35211453e-01 6.54610455e-01 -3.26946437e-01
3.18420026e-03 3.24281454e-01 -2.06014380e-01 -1.25778735e-01
3.44148189e-01 -3.38778436e-01 1.47184134e-01 7.04753876e-01
-1.94633707e-01 -9.98335242e-01 1.25591949e-01 7.95795321e-02
8.91720057e-01 3.43707502e-01 -4.84470159e-01 -4.24324721e-01
4.71013963e-01 5.87996617e-02 3.47493100e-03 4.34464514e-01
5.15884459e-01 1.04693067e+00 5.03706157e-01 -5.59237838e-01
-1.25443339e+00 -1.51734328e+00 -1.11696199e-01 3.93762112e-01
7.22158924e-02 1.53617695e-01 -9.20832276e-01 -2.40342349e-01
2.42427707e-01 8.53026390e-01 -7.83310533e-01 -8.00670460e-02
-4.07801300e-01 -1.07707548e+00 9.49716493e-02 1.99121028e-01
2.85490870e-01 -1.14633906e+00 -5.64823270e-01 2.20924214e-01
1.67750925e-01 -4.62068558e-01 -1.60487875e-01 3.43873873e-02
-1.29836404e+00 -8.01779687e-01 -1.25033796e+00 -5.04153550e-01
9.88183379e-01 2.60839403e-01 1.14039886e+00 5.18922843e-02
-2.50464946e-01 7.17270017e-01 -1.55138060e-01 -2.79215217e-01
-7.13161945e-01 -3.27152044e-01 1.38529897e-01 2.21555591e-01
1.80055708e-01 -1.02079499e+00 -5.75800717e-01 1.24955639e-01
-1.24858499e+00 6.31740093e-02 3.67951632e-01 4.22319382e-01
7.79525697e-01 -1.98382214e-02 1.27423093e-01 -7.37915456e-01
8.18835676e-01 -2.42247581e-01 -5.27010798e-01 -3.82448733e-02
-4.07227218e-01 4.08394367e-01 3.61967385e-01 -4.08050805e-01
-1.01504719e+00 6.80671306e-03 -3.30148309e-01 -6.27986848e-01
-3.30075800e-01 1.78492516e-01 -2.36381963e-02 -2.13572070e-01
6.56653106e-01 3.27977687e-01 1.96770445e-01 -5.99939346e-01
7.96010375e-01 2.39916399e-01 -8.76393169e-03 -9.38734949e-01
9.51964796e-01 9.16281104e-01 2.59367943e-01 -1.00381994e+00
1.64943814e-01 1.25727579e-01 -9.22346294e-01 -2.30815560e-01
1.00220227e+00 -3.37119222e-01 -7.68776357e-01 6.94127977e-01
-1.54861569e+00 -3.55013788e-01 -6.14517391e-01 2.14303005e-02
-7.68934548e-01 3.86559218e-01 -5.55335701e-01 -8.71425450e-01
-2.89132744e-01 -1.25451016e+00 1.06708562e+00 2.55013928e-02
-1.18296698e-01 -1.15948439e+00 2.89340299e-02 -2.79235959e-01
6.88308775e-01 3.84877920e-01 1.32974887e+00 -2.22615525e-02
-7.23563790e-01 1.80163719e-02 -1.34089679e-01 2.77616024e-01
-8.20730254e-02 3.82227778e-01 -9.40391898e-01 -1.92565974e-02
1.64708540e-01 -2.90905982e-02 8.06033313e-01 5.74308515e-01
1.25054002e+00 -2.50041373e-02 -3.90710741e-01 5.14534950e-01
1.12566614e+00 7.89572150e-02 7.91216612e-01 1.49238855e-01
6.05544984e-01 7.07804382e-01 -2.97613591e-01 2.97649235e-01
4.14160341e-01 4.00204360e-01 5.35376906e-01 -1.25667164e-02
-1.39367267e-01 -9.63201076e-02 1.50929660e-01 8.57108831e-01
-4.86303568e-01 -1.78585336e-01 -9.05938148e-01 1.87553927e-01
-1.38343751e+00 -7.30563879e-01 -5.30510664e-01 2.15896654e+00
6.83303237e-01 1.92124039e-01 -7.51215592e-02 1.81803763e-01
6.45126939e-01 7.72148222e-02 -5.45735538e-01 -4.63734329e-01
-2.63795465e-01 3.48328501e-01 2.36009523e-01 4.95855331e-01
-6.88379109e-01 7.13820815e-01 6.09705830e+00 6.70534015e-01
-1.08400655e+00 6.30253926e-02 6.86205089e-01 1.32679269e-01
-1.08798397e+00 -1.93807825e-01 -3.38552684e-01 4.54203516e-01
3.91851157e-01 1.40733914e-02 3.96546006e-01 5.97715557e-01
2.61182189e-01 9.70337689e-02 -9.50322688e-01 1.12233889e+00
-2.06194356e-01 -1.67799509e+00 5.38685322e-01 7.75540620e-02
7.87785828e-01 1.63240079e-02 3.12140048e-01 -1.31618619e-01
5.47822952e-01 -9.94864047e-01 8.49718928e-01 8.16808879e-01
6.83191895e-01 -6.69605851e-01 1.64725378e-01 3.79155010e-01
-8.13698828e-01 4.17312682e-01 -7.03291833e-01 1.93905026e-01
1.84040412e-01 8.81067812e-01 -4.33131099e-01 2.74051338e-01
7.61550188e-01 4.71140087e-01 -2.71428436e-01 1.19823158e+00
-1.08214960e-01 1.69511348e-01 -6.36369646e-01 -3.94687682e-01
3.37221712e-01 -9.09804225e-01 7.07595408e-01 1.01065278e+00
7.14878798e-01 4.13594255e-03 -2.69938648e-01 1.80084252e+00
7.04562292e-02 -8.24065283e-02 -8.05867195e-01 7.89939165e-02
2.86864042e-01 1.13936484e+00 -1.09614658e+00 -1.14752539e-01
-8.31368789e-02 1.16184235e+00 8.25715065e-02 5.41652620e-01
-4.97378856e-01 -2.19798431e-01 7.38276541e-01 2.94236898e-01
2.74860889e-01 -7.38638639e-01 -4.65261996e-01 -8.89391780e-01
-1.77701890e-01 -4.10049975e-01 -3.23950768e-01 -1.18015122e+00
-1.52250791e+00 6.84252679e-01 1.06283501e-02 -9.54477787e-01
7.39529729e-02 -9.48527515e-01 -7.37554014e-01 1.15099323e+00
-1.25149858e+00 -8.40223074e-01 -1.02795213e-01 4.84735727e-01
1.62664220e-01 3.32696944e-01 8.63399565e-01 -6.22645877e-02
-3.55449468e-02 -2.32502714e-01 2.22000495e-01 -2.48708174e-01
8.62884372e-02 -1.60614097e+00 9.29383874e-01 3.58605742e-01
1.98769242e-01 5.10685444e-01 6.15051746e-01 -6.31752253e-01
-1.65348625e+00 -7.71686137e-01 4.06095624e-01 -7.37285137e-01
4.51811671e-01 -6.05342925e-01 -1.24156845e+00 1.57564685e-01
4.41085026e-02 -2.16220319e-01 1.91307366e-01 -2.99749613e-01
-1.88201874e-01 2.78600872e-01 -1.26809239e+00 7.54388809e-01
1.12448525e+00 -2.32474521e-01 -4.62347567e-01 4.40203339e-01
4.52579796e-01 -3.42332006e-01 -8.63044143e-01 1.61413342e-01
1.13355264e-01 -1.15303993e+00 1.19026434e+00 -2.15498775e-01
2.20924720e-01 -2.61070997e-01 2.22336411e-01 -1.67019737e+00
-3.29194695e-01 -6.84031308e-01 1.27390653e-01 8.97869706e-01
4.55898136e-01 -6.80511892e-01 1.04199684e+00 7.60916412e-01
-1.69135511e-01 -6.01968110e-01 -1.06863511e+00 -6.49122596e-01
5.95729172e-01 -6.19198024e-01 6.83231175e-01 7.22572803e-01
-4.84587371e-01 -2.90605873e-01 2.48554483e-01 4.50944528e-02
7.56040931e-01 1.43867806e-01 3.97604197e-01 -1.84031689e+00
5.55971302e-02 -8.73144269e-01 -2.48488352e-01 -1.20245123e+00
6.98717916e-03 -1.07591915e+00 -8.89770985e-02 -1.80760622e+00
-2.51341134e-01 -1.06874228e+00 2.77330130e-01 6.34577498e-02
2.58242637e-01 3.57248217e-01 8.08796436e-02 1.98591292e-01
6.14476725e-02 8.87548447e-01 1.65679133e+00 -2.27816343e-01
-1.78879172e-01 2.73661930e-02 -3.81835550e-01 7.32802391e-01
6.15546584e-01 -5.29743373e-01 -3.93088341e-01 -6.32796764e-01
5.30321658e-01 -2.53767908e-01 6.42529607e-01 -8.66347134e-01
3.49000245e-02 2.73777143e-04 5.59075356e-01 -6.37372255e-01
5.83797097e-01 -7.70743608e-01 3.79070908e-01 3.06564867e-01
3.50194499e-02 2.06254780e-01 1.53680086e-01 5.50364673e-01
8.45312625e-02 -4.17815775e-01 8.87581289e-01 -6.17580354e-01
-1.98927030e-01 4.23920453e-01 -6.26924872e-01 -1.00500099e-01
8.93668771e-01 -5.87787271e-01 1.13063797e-01 -1.90526575e-01
-9.16869462e-01 -2.18948588e-01 1.00717986e+00 2.88702428e-01
8.66269290e-01 -1.42581403e+00 -5.62029421e-01 4.20540899e-01
-3.41996968e-01 4.33703303e-01 3.05507600e-01 4.70387876e-01
-9.54382002e-01 1.15630910e-01 -1.83090314e-01 -9.44589257e-01
-4.82267797e-01 5.55462360e-01 4.73938107e-01 2.98917800e-01
-7.87855804e-01 1.00038159e+00 2.19779179e-01 -7.00927973e-01
-1.15394946e-02 -6.71880484e-01 -1.47077003e-02 3.44952173e-03
1.96448773e-01 1.97184861e-01 8.28133151e-02 -4.61298257e-01
-2.16053188e-01 9.95115042e-01 3.53951097e-01 -2.82442540e-01
1.60890698e+00 6.90567493e-02 -5.03337383e-01 4.47348982e-01
8.51116717e-01 -1.89265609e-01 -1.49788988e+00 1.76025629e-01
-8.29363838e-02 -3.03268373e-01 -7.27417320e-02 -4.13898796e-01
-1.13700879e+00 1.19386077e+00 3.68104428e-01 6.79051101e-01
6.07916236e-01 5.64548410e-02 6.36592209e-01 2.57717818e-01
5.01557767e-01 -6.07841730e-01 1.69399172e-01 5.78958452e-01
1.14540088e+00 -7.21264422e-01 -3.56716394e-01 -4.78928387e-01
-4.77414936e-01 1.33027256e+00 1.36637494e-01 -4.75436807e-01
8.99407864e-01 3.47261131e-01 -4.57668416e-02 -4.77033794e-01
-5.06923497e-01 6.71768114e-02 3.20291638e-01 1.00090647e+00
3.39734852e-01 5.86452335e-03 9.65753943e-02 -6.91843405e-02
-6.88800439e-02 -2.36592278e-01 6.00673854e-01 6.84919715e-01
-4.39713359e-01 -1.19539297e+00 -7.24353790e-01 2.60977030e-01
1.36411428e-01 -4.28058691e-02 -4.21777278e-01 6.63366735e-01
1.30236417e-01 3.60378355e-01 2.92988241e-01 -6.04757108e-02
3.77646446e-01 2.37222463e-01 7.85495460e-01 -5.56547821e-01
-2.34318659e-01 -1.00892581e-01 -4.59710032e-01 -2.39308551e-01
-4.77826506e-01 -7.34809697e-01 -1.27416337e+00 -4.29438621e-01
7.56685659e-02 2.54316419e-01 9.70261097e-01 6.72585726e-01
4.92096812e-01 2.77873516e-01 4.23655242e-01 -1.68778813e+00
-3.79630864e-01 -7.32343674e-01 -7.15604544e-01 4.62701529e-01
3.29692848e-02 -9.74776685e-01 -4.57672805e-01 -5.37970960e-02] | [8.940215110778809, -3.616727352142334] |
ffd237f5-9f4f-40cc-9a1c-43fd64f57335 | drug-repurposing-for-sars-cov-2-using | 2206.01047 | null | https://arxiv.org/abs/2206.01047v2 | https://arxiv.org/pdf/2206.01047v2.pdf | Drug Repurposing For SARS-COV-2 Using Molecular Docking | Drug repurposing is an unconventional approach that is used to investigate new therapeutic aids of existing and shelved drugs. Recent advancement in technologies and the availability of the data of genomics, proteomics, transcriptomics, etc., and with the accessibility of large and reliable database resources, there are abundantly of opportunities to discover drugs by drug repurposing in an efficient manner. The recent pandemic of SARS-COV-2, that caused the death of 6,245,750 human beings to date, has tremendously increase the exceptional usage of bioinformatics tools in interpreting the molecular characterizations of viral infections. In this paper, we have employed various bioinformatics tools such as AutoDock-Vina, PyMol etc. We have found a leading drug candidate Cepharanthine that has shown better results and effectiveness than recently used antiviral drug candidates such as Favipiravir, IDX184, Remedesivir, Ribavirin and etc. This paper has analyzed Cepharanthine potential therapeutic importance as a drug of choice in managing COVID-19 cases. It is anticipated that proposed study would be beneficial for researchers and medical practitioners in handling SARS-CoV-2 and its variant related diseases. | ['Hina Bashir', 'Muhammad Ismail', 'Abdul Majid', 'Imra Aqeel'] | 2022-06-02 | null | null | null | null | ['molecular-docking'] | ['medical'] | [ 1.71200424e-01 -6.21650100e-01 -4.55948822e-02 4.69324477e-02
6.97564930e-02 -7.64053881e-01 3.11448127e-01 4.35064971e-01
-2.54626691e-01 1.57945454e+00 7.70660415e-02 -5.59704065e-01
-1.81687936e-01 -3.64602774e-01 -4.52384427e-02 -7.52073050e-01
-2.09162191e-01 8.53644848e-01 -5.77884093e-02 -3.27401042e-01
4.04534280e-01 9.05149996e-01 -9.33658540e-01 -8.66127759e-02
1.22048843e+00 1.60261355e-02 2.96651244e-01 7.26499081e-01
7.25802481e-02 9.13278386e-03 -2.76414394e-01 -9.57202241e-02
3.05796266e-01 -5.40914118e-01 -5.96507847e-01 -2.87555903e-01
-2.73262888e-01 -3.71948957e-01 2.32139125e-01 5.57106912e-01
5.46919048e-01 1.29775726e-03 9.39779639e-01 -9.59834099e-01
-4.71214801e-01 -3.05393070e-01 -9.07697737e-01 5.27427316e-01
3.75943273e-01 4.04427230e-01 3.78546029e-01 -8.07639182e-01
9.11671817e-01 1.20032191e+00 4.61073488e-01 4.81276572e-01
-8.10768604e-01 -3.94895107e-01 -7.95020700e-01 3.89049768e-01
-1.27010751e+00 -1.53530002e-01 1.71423152e-01 -6.11204028e-01
1.50103641e+00 1.51129365e-01 4.76989895e-01 4.02778447e-01
7.53884852e-01 1.68205518e-02 1.17309058e+00 -1.29264235e-01
-3.82131189e-02 3.04841101e-01 1.87517554e-01 6.96537197e-01
5.97219288e-01 7.38242790e-02 -1.43042639e-01 -7.77243495e-01
4.42732841e-01 5.30355096e-01 -2.74070263e-01 5.33037558e-02
-9.33777213e-01 7.99993396e-01 -2.39488482e-01 4.04615462e-01
-6.66221499e-01 -4.06590939e-01 6.54802561e-01 -6.69378936e-02
2.95126624e-02 1.62480906e-01 -1.00458097e+00 -1.18880600e-01
-4.20454800e-01 2.26256967e-01 3.86207342e-01 2.93247521e-01
3.66495669e-01 7.42549673e-02 4.50846136e-01 6.07499897e-01
3.49834949e-01 5.82370996e-01 1.29395694e-01 -1.34326339e-01
-1.32719889e-01 5.35219133e-01 3.70400250e-01 -7.43443727e-01
-5.63872159e-01 -6.63601607e-02 -6.97958887e-01 -1.10259742e-01
2.14934066e-01 -3.25070143e-01 -8.79204810e-01 1.43485200e+00
5.24890721e-01 3.53226155e-01 2.28217840e-01 6.99451268e-01
2.96319216e-01 7.78671622e-01 5.33196688e-01 -9.03177857e-01
1.72683942e+00 -4.15977180e-01 -6.68540716e-01 6.49454236e-01
1.56742454e-01 -1.08684027e+00 3.50777030e-01 6.64820194e-01
-6.19730353e-01 -2.13510543e-01 -1.00147653e+00 4.85968113e-01
-5.58574200e-01 -1.22115523e-01 8.39861751e-01 8.54288340e-01
-7.98088372e-01 7.60830641e-01 -8.66250813e-01 -1.10142398e+00
5.22270739e-01 7.29106724e-01 -5.39009035e-01 -3.05347834e-02
-8.49464417e-01 1.23008049e+00 4.79627609e-01 -1.47442296e-01
-7.73418903e-01 -5.05900919e-01 -2.26797704e-02 -3.67790192e-01
5.49040958e-02 -1.08436012e+00 3.12380940e-01 -1.13157317e-01
-1.23309541e+00 6.99255347e-01 -2.77543962e-01 -3.80743772e-01
-1.43199340e-01 7.79221281e-02 -4.09707874e-01 1.30141884e-01
-8.28776136e-02 1.83587223e-01 2.99950063e-01 -6.80737495e-01
-4.34219331e-01 -1.09160376e+00 -4.08767104e-01 1.56128287e-01
3.59203279e-01 4.84692067e-01 2.89442241e-01 -3.98026794e-01
-3.57857734e-01 -9.10815477e-01 -3.90481353e-01 -6.49967134e-01
-3.07548702e-01 -2.81060547e-01 1.00167704e+00 -6.01262629e-01
7.42629647e-01 -1.65915239e+00 1.55107930e-01 1.98856965e-01
1.48924902e-01 7.13501573e-01 2.59350479e-01 1.17076671e+00
-2.00483687e-02 5.17505109e-01 -1.52239978e-01 7.31060207e-01
-6.33664966e-01 7.51766935e-02 -1.80255920e-01 7.97125340e-01
1.17665336e-01 4.62512851e-01 -6.87979043e-01 -1.82898462e-01
1.50418043e-01 8.31436336e-01 -3.04719567e-01 -2.91153044e-02
-3.37382346e-01 8.24040771e-01 -9.52638388e-01 9.89357352e-01
1.07767153e+00 -1.23745702e-01 4.91163731e-01 -1.30291358e-01
-1.98998854e-01 -1.25313818e-01 -5.51448762e-01 9.78196859e-01
2.73597121e-01 1.40044063e-01 -3.46509725e-01 -7.44005442e-01
7.07816124e-01 7.77886271e-01 3.26246172e-01 -2.78316766e-01
3.20596814e-01 2.68022656e-01 2.12564498e-01 -6.61061108e-01
2.68088914e-02 -5.04211128e-01 8.24580073e-01 7.76591673e-02
-3.64130795e-01 4.66635168e-01 3.44476998e-01 -5.39275929e-02
9.16564763e-01 3.07683349e-01 9.55407917e-01 -4.46089536e-01
8.77131283e-01 4.68154043e-01 6.56130791e-01 4.88692001e-02
-4.83818054e-01 -1.85579270e-01 1.56577870e-01 -4.52890307e-01
-1.32055187e+00 -9.31302905e-01 -6.02834404e-01 5.64519167e-01
-2.05872476e-01 3.62352245e-02 -4.42767918e-01 -1.28419772e-01
-1.18882187e-01 2.64350742e-01 -2.43974432e-01 4.73301351e-01
-5.67269504e-01 -9.43034112e-01 5.72499156e-01 -3.63714993e-01
4.14794236e-01 -9.40207720e-01 -4.51572448e-01 5.60238838e-01
4.66023117e-01 -5.74879467e-01 1.07229427e-01 2.24215332e-02
-1.42747748e+00 -1.24175727e+00 -5.17314315e-01 -6.42503262e-01
4.35805470e-01 2.66220868e-01 3.80190104e-01 3.36717814e-01
-7.41078675e-01 -4.44232345e-01 -4.62571561e-01 -6.54004753e-01
-3.66283804e-01 -5.04621923e-01 6.26214564e-01 -6.45713389e-01
7.26486802e-01 -7.19375551e-01 -8.04052532e-01 -1.23521812e-01
-5.47479153e-01 -1.67229235e-01 3.89594048e-01 5.95976055e-01
4.58004057e-01 1.34424329e-01 1.15797091e+00 -1.28028202e+00
8.01915646e-01 -9.51727331e-01 -4.62034434e-01 9.71302576e-03
-8.11115980e-01 -1.33206710e-01 8.01818252e-01 1.67442784e-01
-9.69043970e-01 -2.21473560e-01 -2.33538061e-01 3.13141853e-01
-2.45455623e-01 5.27180433e-01 9.64080468e-02 5.78989759e-02
4.92813826e-01 2.09726796e-01 1.87611997e-01 -6.99446619e-01
-1.02836661e-01 8.48126173e-01 1.17619015e-01 -3.35117042e-01
4.98106986e-01 4.25693035e-01 3.60389203e-01 -1.12542522e+00
1.48022726e-01 -6.69392765e-01 -2.17683278e-02 1.01172984e-01
1.30298829e+00 -5.52796781e-01 -1.23764944e+00 4.07748759e-01
-1.16874242e+00 7.00636685e-01 8.52679670e-01 4.99901801e-01
-1.09668829e-01 5.83338618e-01 -5.76560199e-01 -7.53568828e-01
-8.50962222e-01 -1.18215072e+00 2.50565231e-01 5.17159462e-01
-3.85470316e-02 -7.70757496e-01 7.44672894e-01 6.36707008e-01
5.06382465e-01 7.11370885e-01 1.52668142e+00 -9.56702054e-01
-6.48846090e-01 9.67479348e-02 -1.53807282e-01 -1.19346820e-01
6.03607595e-01 4.99283016e-01 -5.30454814e-01 -1.90781981e-01
-4.52958867e-02 -1.44484490e-01 3.24687093e-01 4.42019284e-01
1.93758860e-01 -1.13496467e-01 -6.59364641e-01 3.49838376e-01
1.84970462e+00 1.35792220e+00 8.64210904e-01 2.64499307e-01
4.14386630e-01 2.26626590e-01 6.63794696e-01 7.70372450e-01
-6.43974394e-02 5.19527137e-01 2.48000681e-01 2.91458905e-01
5.60333669e-01 1.69017017e-01 -1.88883856e-01 4.03402299e-01
-8.27359080e-01 -3.25561136e-01 -9.40944970e-01 1.79445550e-01
-1.43812442e+00 -1.12518263e+00 -4.36965764e-01 2.10890079e+00
9.32338774e-01 -4.45684314e-01 1.68707535e-01 -2.47760817e-01
5.17049015e-01 -3.93910706e-01 -5.68849027e-01 -8.95643651e-01
-6.98112324e-02 5.94310284e-01 2.72758603e-01 3.86053503e-01
-7.50710547e-01 8.21020484e-01 5.80299473e+00 5.70448458e-01
-9.16862488e-01 -8.73879939e-02 3.07121307e-01 4.55178440e-01
-2.55651195e-02 3.48458946e-01 -5.67683876e-01 4.24881488e-01
1.10670400e+00 -4.89059746e-01 5.40762246e-01 5.16742885e-01
8.62528801e-01 -3.61650318e-01 -6.68962598e-01 6.56485796e-01
-2.72048324e-01 -1.63372624e+00 -2.06567105e-02 3.65612566e-01
5.87875485e-01 3.19841132e-02 -2.38095865e-01 -2.92122513e-01
-1.16108537e-01 -1.10539770e+00 -6.64223194e-01 3.43519986e-01
4.15158480e-01 -1.23521173e+00 1.11345208e+00 2.23231509e-01
-7.44021595e-01 3.83349180e-01 -5.03998041e-01 1.17444254e-01
2.42619514e-01 2.08846107e-01 -1.60511541e+00 4.15256321e-01
3.49033847e-02 4.79413778e-01 -7.60392100e-02 1.16768122e+00
2.45112836e-01 2.28970528e-01 -1.32671773e-01 -4.52607274e-02
8.73888955e-02 -7.02955782e-01 4.59015697e-01 9.18782592e-01
-3.21718939e-02 7.06160069e-01 -5.69679588e-03 2.76646048e-01
3.08138698e-01 6.23612940e-01 -8.14270914e-01 -6.84193671e-01
3.73191774e-01 8.43085647e-01 -7.53617704e-01 -5.03371775e-01
-2.51134098e-01 6.68506086e-01 -2.87676036e-01 2.52603233e-01
-4.44272012e-01 -2.01857388e-01 8.99633706e-01 2.66720086e-01
1.05634250e-01 -1.44688055e-01 2.11012155e-01 -5.50852895e-01
-9.49811995e-01 -1.35020792e+00 3.69537979e-01 -5.49823344e-01
-9.64458406e-01 6.84625745e-01 6.23810552e-02 -9.82339203e-01
-1.26968175e-01 -5.71914792e-01 -6.07804716e-01 1.15224266e+00
-1.04785454e+00 -1.06655669e+00 3.70693833e-01 4.54312503e-01
6.33638263e-01 -5.72721541e-01 1.09879017e+00 1.25789240e-01
-5.08529842e-01 -1.96103454e-01 5.63991010e-01 -6.64297581e-01
7.16229796e-01 -6.31942630e-01 -4.08186674e-01 4.37374592e-01
-7.10542560e-01 1.23932350e+00 1.21680748e+00 -1.06849849e+00
-1.64839232e+00 -5.44017971e-01 6.45392656e-01 -1.77856639e-01
5.88389337e-01 2.82552809e-01 -6.63971364e-01 4.40820277e-01
8.04795384e-01 -7.62743652e-01 1.17215061e+00 -3.84419382e-01
-2.55785678e-02 3.77202272e-01 -1.53067410e+00 6.61176920e-01
2.49125525e-01 -1.63814679e-01 -5.94897509e-01 7.39698291e-01
4.49027240e-01 8.70453119e-02 -8.72097909e-01 4.48630720e-01
6.16658807e-01 -1.08836257e+00 8.72844517e-01 -1.20302820e+00
-3.31577733e-02 -8.51217806e-01 3.04740928e-02 -8.91563416e-01
-1.94535702e-01 -6.42722487e-01 4.96152222e-01 8.06187987e-01
3.09065074e-01 -6.14552379e-01 6.73473179e-01 1.89343005e-01
2.53255546e-01 -8.94563496e-01 -7.73811936e-01 -2.55486131e-01
-2.17207357e-01 4.96290565e-01 1.84779719e-01 1.01716888e+00
1.15436114e-01 6.20804429e-01 -7.32416451e-01 -8.04616585e-02
5.31323671e-01 -5.34178317e-02 4.80013251e-01 -1.36743832e+00
-2.45580167e-01 1.55597791e-01 -5.79212964e-01 -1.08008400e-01
-4.78974789e-01 -5.01138210e-01 -9.24288213e-01 -1.51087153e+00
7.07744420e-01 -2.45405331e-01 -5.30416146e-03 1.38732359e-01
-4.36739856e-03 6.72400417e-03 -9.97297913e-02 2.88138688e-01
1.31362408e-01 3.01276017e-02 1.18109369e+00 2.76413709e-01
-5.06063879e-01 -1.53156951e-01 -5.73420942e-01 6.39254153e-01
1.01110280e+00 -5.81363857e-01 -6.53885722e-01 4.34036255e-01
3.31754148e-01 2.36959025e-01 -2.64689744e-01 -1.07100944e-03
-1.93451762e-01 -6.98151588e-01 3.31432104e-01 -1.08896625e+00
1.64793164e-01 -7.78289437e-01 1.02123165e+00 9.02067065e-01
7.17826307e-01 3.06448102e-01 3.04495513e-01 5.36779463e-01
2.21505716e-01 -2.37227559e-01 8.51204932e-01 -3.84816974e-01
-5.55255294e-01 1.97497070e-01 -7.07991183e-01 -1.43392265e-01
1.30019963e+00 -6.55022681e-01 -5.91493964e-01 2.77537376e-01
-5.84136605e-01 -1.02485299e-01 5.89251876e-01 1.09410673e-01
6.31240010e-01 -6.82742476e-01 -8.46284688e-01 -3.95217584e-03
-3.76036502e-02 -6.31556630e-01 4.33054596e-01 1.10206199e+00
-1.38211989e+00 9.83072639e-01 -9.56997514e-01 -2.82614321e-01
-1.75272059e+00 9.51304436e-01 -9.86537412e-02 -1.05382003e-01
-2.50340462e-01 3.78974676e-01 6.90626577e-02 2.48871028e-01
-3.61595571e-01 3.35051715e-01 -5.69850385e-01 -1.45909697e-01
5.70007801e-01 6.17652535e-01 4.09075953e-02 -7.64942348e-01
-9.55289125e-01 7.04524100e-01 -5.41086137e-01 5.03526390e-01
1.58102202e+00 3.88332978e-02 -6.06355906e-01 -1.22964174e-01
9.19742525e-01 2.51246393e-01 -2.78440863e-01 2.90756792e-01
-1.23761103e-01 -6.20496631e-01 -3.40839386e-01 -1.07150018e+00
-3.08926374e-01 1.88730001e-01 8.54097307e-01 -1.53882578e-01
9.78523612e-01 -2.47121423e-01 6.49154365e-01 1.89275995e-01
6.91473901e-01 -6.42388880e-01 -4.47143167e-01 3.35959196e-01
6.95676148e-01 -1.16855335e+00 2.02197328e-01 -2.30292380e-01
-6.81203008e-01 8.98487985e-01 2.11368248e-01 7.71816894e-02
5.29108405e-01 2.55060941e-01 -4.45280634e-02 -3.91048491e-01
-1.11950123e+00 8.12103301e-02 -1.92845479e-01 8.32568705e-01
9.08707678e-01 1.90754220e-01 -1.32507253e+00 1.32957205e-01
4.39172983e-01 4.79195803e-01 5.73383152e-01 1.02400684e+00
-8.19525003e-01 -1.57524478e+00 -4.82678294e-01 3.89368951e-01
-8.08185399e-01 -3.76171649e-01 -4.11267519e-01 9.44068670e-01
2.45743722e-01 7.78374076e-01 -4.07637417e-01 -2.69582309e-02
2.18578577e-02 2.46752217e-01 5.17457485e-01 -4.21021670e-01
-4.54545945e-01 5.12307286e-01 2.54468858e-01 1.52416632e-01
-6.19792998e-01 -4.70598191e-01 -1.72147310e+00 -8.38744164e-01
-2.07648665e-01 4.46760327e-01 9.18836892e-01 9.74891663e-01
6.09319508e-01 -1.13933114e-02 5.99901438e-01 -1.24638185e-01
-2.04357088e-01 -9.41332221e-01 -7.28531837e-01 -8.28996301e-02
7.57999122e-02 -5.66743910e-01 3.36932344e-03 2.58517146e-01] | [4.6529130935668945, 5.0839738845825195] |
8659100a-8e7a-4a4b-a64f-17060f798897 | unsupervised-domain-adaptive-person-re | 1908.10359 | null | https://arxiv.org/abs/1908.10359v1 | https://arxiv.org/pdf/1908.10359v1.pdf | Unsupervised Domain-Adaptive Person Re-identification Based on Attributes | Pedestrian attributes, e.g., hair length, clothes type and color, locally describe the semantic appearance of a person. Training person re-identification (ReID) algorithms under the supervision of such attributes have proven to be effective in extracting local features which are important for ReID. Unlike person identity, attributes are consistent across different domains (or datasets). However, most of ReID datasets lack attribute annotations. On the other hand, there are several datasets labeled with sufficient attributes for the case of pedestrian attribute recognition. Exploiting such data for ReID purpose can be a way to alleviate the shortage of attribute annotations in ReID case. In this work, an unsupervised domain adaptive ReID feature learning framework is proposed to make full use of attribute annotations. We propose to transfer attribute-related features from their original domain to the ReID one: to this end, we introduce an adversarial discriminative domain adaptation method in order to learn domain invariant features for encoding semantic attributes. Experiments on three large-scale datasets validate the effectiveness of the proposed ReID framework. | ['Vittorio Murino', 'Xiangping Zhu', 'Pietro Morerio'] | 2019-08-27 | null | null | null | null | ['pedestrian-attribute-recognition'] | ['computer-vision'] | [ 1.11835070e-01 -2.99753040e-01 -3.02599426e-02 -9.26168323e-01
-3.13204587e-01 -6.54203236e-01 7.58009970e-01 3.23708445e-01
-4.28446114e-01 9.61051583e-01 2.39987850e-01 3.38661969e-01
-3.95704135e-02 -9.19222295e-01 -6.33030951e-01 -8.92611146e-01
3.30594659e-01 4.08511728e-01 -1.41135141e-01 -3.39160077e-02
2.10239261e-01 5.43554604e-01 -1.70154607e+00 -8.82393792e-02
8.79927754e-01 7.87024796e-01 -1.02669373e-01 7.28176609e-02
-4.10471261e-01 2.43401095e-01 -5.22566199e-01 -5.94739377e-01
3.81204635e-01 -4.05711740e-01 -8.70358825e-01 2.41300762e-01
6.49842560e-01 -3.93181324e-01 -2.90762007e-01 1.11854839e+00
4.87341315e-01 2.56897330e-01 8.93931150e-01 -1.52784383e+00
-8.34071815e-01 1.44642532e-01 -4.95888114e-01 -5.29217049e-02
5.38254738e-01 -1.13825038e-01 6.71855807e-01 -6.81463480e-01
6.85385048e-01 1.24390912e+00 6.29832566e-01 7.17310071e-01
-1.18179405e+00 -8.68390203e-01 3.57693493e-01 4.09846604e-01
-1.70424557e+00 -4.29407507e-01 1.17776287e+00 -4.02157933e-01
5.50226532e-02 5.56516200e-02 6.43319786e-01 1.30941510e+00
-2.68717855e-01 7.19765842e-01 1.37135279e+00 -4.68792647e-01
2.17735842e-01 5.23751616e-01 2.06613421e-01 3.54744673e-01
4.18370277e-01 -1.58141330e-01 -4.51658159e-01 -2.20537364e-01
5.18099368e-01 1.23512045e-01 1.20920882e-01 -6.32303357e-01
-1.03546572e+00 6.42345011e-01 3.66461545e-01 -3.13618146e-02
-1.68741703e-01 -2.54855782e-01 7.07399845e-01 2.35093787e-01
5.67644536e-01 3.94606330e-02 -2.50196517e-01 2.43348092e-01
-2.09089532e-01 3.89851868e-01 5.09732127e-01 1.38079298e+00
1.14894187e+00 -1.16755590e-01 -1.62863418e-01 9.72703457e-01
1.10482238e-01 6.84573591e-01 3.56626928e-01 -5.90039074e-01
1.56040296e-01 7.30417907e-01 3.10311377e-01 -8.93090248e-01
-2.53152847e-01 -2.28495121e-01 -8.87101173e-01 -3.22424471e-02
4.87665355e-01 8.99758935e-02 -7.56945372e-01 1.83215570e+00
7.82075286e-01 1.69524103e-01 1.26995355e-01 8.80900085e-01
7.39854515e-01 2.17233196e-01 6.75563991e-01 2.72532880e-01
1.50293386e+00 -4.26941425e-01 -6.28245890e-01 -9.81737077e-02
4.33568597e-01 -6.66289389e-01 8.23073208e-01 -2.96240896e-02
-5.14630973e-01 -8.07610631e-01 -9.08150256e-01 5.02408743e-02
-7.49478996e-01 2.29473993e-01 5.15074849e-01 9.45451319e-01
-6.17305458e-01 2.81413794e-01 -2.70025790e-01 -8.82912695e-01
4.48697090e-01 4.62409556e-01 -7.22508729e-01 -2.37015694e-01
-1.21614277e+00 7.25727439e-01 5.17723441e-01 -4.64597195e-02
-6.86975777e-01 -3.67364645e-01 -1.02050400e+00 -2.91064709e-01
1.10554449e-01 -9.29824114e-01 7.83615947e-01 -1.17966831e+00
-1.18325067e+00 1.29880941e+00 -3.96743178e-01 -3.02393913e-01
5.54511487e-01 -1.77671179e-01 -5.02444386e-01 -1.08732440e-01
3.89351159e-01 5.12425661e-01 9.72115695e-01 -1.46058190e+00
-7.41544485e-01 -7.79824138e-01 -1.58212706e-02 9.21617448e-02
-7.68841982e-01 2.81079896e-02 -1.55847147e-01 -6.18279099e-01
-1.43217236e-01 -9.16937470e-01 1.57028228e-01 6.69429451e-02
-4.47104245e-01 -4.40349102e-01 7.66521037e-01 -7.83134401e-01
5.68856061e-01 -2.09489894e+00 -8.18663314e-02 3.67886305e-01
-1.27531216e-01 7.16914758e-02 8.11677054e-03 1.07967287e-01
-5.32435067e-02 -8.19434747e-02 -1.28743351e-01 -3.84317994e-01
4.53019999e-02 4.12387818e-01 9.69237238e-02 7.23160625e-01
3.27221543e-01 5.62869549e-01 -8.03859293e-01 -7.93541193e-01
3.38903189e-01 4.97612000e-01 -2.09846184e-01 3.02331209e-01
2.64626056e-01 9.67824221e-01 -8.63733888e-01 7.49318361e-01
1.00793743e+00 3.18119913e-01 -4.68603261e-02 -2.40202159e-01
6.80617839e-02 -2.18204349e-01 -1.26167381e+00 1.67489851e+00
-3.21818501e-01 2.14908272e-01 -4.01926739e-03 -1.07363462e+00
1.31156635e+00 1.36235133e-01 4.45582688e-01 -6.06995523e-01
1.95700169e-01 4.94540185e-02 -4.03493971e-01 -3.69805753e-01
4.32507634e-01 -3.92646790e-02 -2.19811231e-01 2.29351103e-01
6.60586432e-02 4.69981432e-01 6.89244270e-02 -4.79115471e-02
4.18499172e-01 3.42079937e-01 5.70684135e-01 -2.55569011e-01
1.00753760e+00 -2.47573890e-02 8.11595201e-01 5.95620811e-01
-4.04018641e-01 3.31572741e-01 -1.77748740e-01 -7.46314049e-01
-1.53696704e+00 -1.10801029e+00 -3.40933561e-01 1.20861363e+00
4.93222982e-01 3.49458419e-02 -8.59826386e-01 -7.09037960e-01
1.63175330e-01 4.71352577e-01 -7.60410964e-01 -2.89128870e-01
-4.73981529e-01 -5.69055676e-01 5.01971185e-01 7.53383875e-01
8.99588108e-01 -7.49613225e-01 -5.77644669e-02 1.56634912e-01
-2.61829883e-01 -1.38372540e+00 -3.71192008e-01 -2.52489775e-01
-4.15170968e-01 -9.99303162e-01 -8.60181689e-01 -8.47988367e-01
1.06023228e+00 2.46810660e-01 7.58318782e-01 -5.67743443e-02
-3.47427875e-01 6.10353112e-01 -5.94202936e-01 -4.90564644e-01
-3.05926114e-01 5.25074750e-02 4.81411606e-01 6.11783147e-01
7.47572005e-01 -4.75586712e-01 -6.90763295e-01 5.29703736e-01
-5.87450862e-01 -4.92009148e-02 5.54193318e-01 8.53515923e-01
5.51794112e-01 -6.72274306e-02 7.98032284e-01 -9.75715697e-01
3.41864020e-01 -4.57699150e-01 -2.33277246e-01 2.92690992e-01
-4.02980417e-01 2.12238543e-02 7.98414111e-01 -4.23628539e-01
-1.39977610e+00 2.48718649e-01 -3.34288552e-03 -1.18122391e-01
-8.91580820e-01 -1.79023772e-01 -7.87284732e-01 -2.35265449e-01
4.89000112e-01 4.36214924e-01 -1.28922388e-01 -5.41898191e-01
2.58194506e-01 8.31949472e-01 8.37331116e-01 -1.07864857e+00
1.04334271e+00 6.74419343e-01 2.10999146e-01 -5.41881740e-01
-7.33549774e-01 -7.35442877e-01 -1.21870279e+00 -8.75367150e-02
9.28831875e-01 -1.08833814e+00 -4.90838408e-01 5.46976149e-01
-9.10973251e-01 4.39770132e-01 -1.90352842e-01 2.93523759e-01
-5.11511445e-01 5.86930931e-01 -1.16087750e-01 -8.59851241e-01
-2.71439701e-01 -7.09899008e-01 1.02091110e+00 4.95287120e-01
-1.97144166e-01 -8.68484735e-01 -2.06332847e-01 3.86764228e-01
2.54206836e-01 4.02753890e-01 7.28453100e-01 -1.00552809e+00
-1.85681418e-01 -3.51502866e-01 -3.79479319e-01 2.72396296e-01
5.38288057e-01 -4.37371314e-01 -1.25966907e+00 -4.37425345e-01
-4.12326336e-01 -1.44488707e-01 4.71063077e-01 6.84555154e-03
1.16212070e+00 -1.61385268e-01 -5.52500904e-01 5.21336138e-01
1.26450646e+00 4.80542704e-02 3.80437195e-01 6.72527611e-01
7.79893398e-01 7.63870358e-01 8.76197457e-01 6.87930524e-01
4.23112661e-01 8.35109949e-01 1.02053858e-01 -1.04449816e-01
-1.70374870e-01 -3.84128541e-01 1.28754139e-01 1.55176133e-01
-2.86214441e-01 7.29072094e-02 -7.70673037e-01 6.41588330e-01
-1.65664172e+00 -7.98807859e-01 -9.08012688e-02 2.41555810e+00
6.40492260e-01 -1.21305302e-01 3.86691064e-01 7.95559958e-02
1.20407128e+00 -3.76904905e-01 -5.60176432e-01 -2.39022106e-01
-2.12366864e-01 -1.05419956e-01 6.02417886e-01 2.81547960e-02
-1.51176715e+00 7.76865125e-01 4.79863119e+00 7.41782904e-01
-4.82641667e-01 6.23048563e-03 3.81473809e-01 6.79411650e-01
-1.12728529e-01 3.16518322e-02 -9.39136267e-01 5.35949290e-01
5.41676939e-01 -3.44799906e-01 1.59338787e-01 1.09281182e+00
5.17319934e-03 1.43711314e-01 -1.20951843e+00 9.74434316e-01
2.03674838e-01 -6.12286031e-01 2.47610852e-01 -8.38194713e-02
5.12937248e-01 -9.43501711e-01 4.61867126e-03 1.79348275e-01
1.96049348e-01 -7.93416262e-01 4.64652449e-01 4.98941034e-01
9.34609294e-01 -1.09204388e+00 9.76109624e-01 1.11215413e-01
-1.36919010e+00 7.37692863e-02 -7.19312549e-01 1.08348958e-01
-1.37341544e-01 1.61154732e-01 -9.30800438e-01 8.66765738e-01
8.14681172e-01 6.97201073e-01 -7.16610968e-01 1.04214919e+00
-1.79322630e-01 1.28848687e-01 -9.88292769e-02 2.49001518e-01
-1.73338875e-01 -1.69613957e-01 4.27739590e-01 1.21948409e+00
2.34543949e-01 1.53834328e-01 3.61075014e-01 7.48659909e-01
-1.22148171e-01 3.40550333e-01 -7.66042411e-01 3.15425754e-01
6.60540938e-01 1.10509229e+00 -3.70955646e-01 -2.64530629e-01
-5.15634775e-01 1.37867129e+00 1.62159815e-01 3.30478936e-01
-6.15418494e-01 -2.66018957e-01 9.11756814e-01 1.05153523e-01
5.78838065e-02 -3.95122804e-02 -1.65565640e-01 -9.13128376e-01
7.66757056e-02 -4.74888355e-01 4.76571947e-01 -4.42736149e-01
-1.83707821e+00 2.36531258e-01 1.64430104e-02 -1.46253467e+00
-2.00548902e-01 -4.82844085e-01 -3.94148558e-01 1.05196059e+00
-1.72132421e+00 -1.94038904e+00 -8.51111948e-01 1.04465628e+00
3.81927997e-01 -3.79063189e-01 1.00897801e+00 4.34191734e-01
-3.92657280e-01 9.68002975e-01 2.23907799e-01 6.22297883e-01
1.22717977e+00 -1.17142463e+00 4.11756366e-01 6.48055971e-01
-2.94001818e-01 7.45160341e-01 8.56369674e-01 -6.56352699e-01
-1.35806465e+00 -1.33851600e+00 5.48137665e-01 -3.94383043e-01
3.05056423e-01 -3.27473253e-01 -9.38885331e-01 7.75457323e-01
-9.76223573e-02 1.02126747e-01 9.19041336e-01 2.18898237e-01
-5.99187374e-01 -3.29950333e-01 -1.53099668e+00 2.31356427e-01
1.15022433e+00 -4.94991034e-01 -5.05407214e-01 1.90432504e-01
2.13060290e-01 -1.02097526e-01 -1.05624807e+00 2.66526341e-01
6.16033912e-01 -6.24642432e-01 1.33485615e+00 -6.25736594e-01
-5.02515323e-02 -3.91772389e-01 -1.04221530e-01 -1.32594204e+00
-3.61351907e-01 -9.09824073e-02 5.41861713e-01 1.95901120e+00
-3.08038622e-01 -8.00524533e-01 7.54663408e-01 8.09040844e-01
1.03077061e-01 1.15170754e-01 -8.98059666e-01 -1.01309943e+00
8.67960826e-02 1.55935124e-01 9.96991575e-01 1.09715962e+00
-4.28145587e-01 5.14142998e-02 -5.10632634e-01 3.81124020e-01
1.09339523e+00 1.83178186e-01 1.22088432e+00 -1.63743508e+00
2.46604696e-01 -1.12364717e-01 -1.03764975e+00 -5.38640380e-01
6.63597286e-01 -7.88497925e-01 -2.56075203e-01 -1.19872165e+00
6.76687717e-01 -8.43535542e-01 -2.69101202e-01 4.90271300e-01
-3.72929603e-01 1.21068165e-01 -6.18274277e-03 2.44299516e-01
-3.90206307e-01 5.79582751e-01 1.00777769e+00 -4.76865381e-01
1.96380541e-01 2.17675767e-03 -6.61739767e-01 6.02892935e-01
9.41472769e-01 -3.85722697e-01 -2.12244198e-01 -1.95737943e-01
-3.56495559e-01 -4.45909888e-01 7.57804155e-01 -1.07305539e+00
2.98205465e-01 -1.84789255e-01 9.14340258e-01 -2.18088031e-01
3.47572148e-01 -9.87506628e-01 1.19179608e-02 -8.44317377e-02
-1.98872298e-01 -1.50889367e-01 1.28316134e-01 7.29784429e-01
-2.78348684e-01 -9.17365104e-02 8.17990363e-01 -1.46898434e-01
-1.32850432e+00 4.03723598e-01 1.38331637e-01 -7.75790811e-02
1.18587255e+00 -5.13871551e-01 -1.59275055e-01 -1.94107205e-01
-6.90484941e-01 2.20933348e-01 8.56221557e-01 3.78645658e-01
5.43589473e-01 -1.55187690e+00 -9.03904021e-01 2.97143191e-01
8.06080818e-01 -1.97053358e-01 3.29213500e-01 2.78945416e-01
1.71434355e-03 1.12270616e-01 -7.89235950e-01 -4.67578053e-01
-1.50266767e+00 8.39952886e-01 1.44036323e-01 2.67534912e-01
-6.74619198e-01 5.47954857e-01 3.42743874e-01 -5.45341909e-01
-7.06488453e-03 4.01272178e-01 -3.68469656e-01 -6.81966469e-02
5.26175141e-01 2.89661527e-01 -6.75545558e-02 -1.05130172e+00
-3.82596016e-01 8.46195877e-01 -1.67986616e-01 2.88469464e-01
1.01475060e+00 -5.08245349e-01 9.56493244e-03 8.10569227e-02
1.08015501e+00 -1.03095658e-01 -1.10574257e+00 -4.71007854e-01
1.80668049e-02 -8.63433301e-01 -4.92900938e-01 -5.53696871e-01
-7.16450930e-01 7.13594317e-01 9.49521244e-01 -2.14886963e-01
1.13901186e+00 -2.21354645e-02 7.83615768e-01 3.08433414e-01
7.42867827e-01 -1.23785257e+00 -4.13337857e-01 4.42220159e-02
4.46181059e-01 -1.48541975e+00 -3.42703387e-02 -6.03728592e-01
-6.89005554e-01 1.07638443e+00 7.02126384e-01 2.26281118e-02
2.01138303e-01 -9.27183032e-02 -4.01315354e-02 1.70456409e-01
1.20684259e-01 -3.42785984e-01 1.65302590e-01 1.08375704e+00
2.65770048e-01 2.36183926e-01 -3.03053975e-01 5.06285489e-01
-1.84062794e-01 -1.07234329e-01 2.81541824e-01 8.88912439e-01
-3.19397330e-01 -1.42464352e+00 -7.49228358e-01 8.63955691e-02
-2.03628480e-01 4.18406457e-01 -4.88105208e-01 7.81939864e-01
3.34113121e-01 9.66399133e-01 -2.15523411e-02 -2.23653704e-01
3.34471911e-01 2.67603576e-01 5.94559133e-01 -2.20508873e-01
-2.56147623e-01 -5.45271397e-01 1.07812963e-01 -1.42511234e-01
-6.51284277e-01 -7.76252985e-01 -1.04421282e+00 -4.74790841e-01
1.08875960e-01 4.26779054e-02 5.26931167e-01 1.00946128e+00
8.27286765e-02 6.10277280e-02 7.10989773e-01 -5.87439716e-01
-4.75191265e-01 -7.03335881e-01 -7.22048104e-01 1.28927791e+00
1.76483527e-01 -1.04160178e+00 1.30508235e-02 4.63577747e-01] | [14.720922470092773, 1.0430488586425781] |
a9aa16b4-d755-406e-996b-f6744e04fa24 | low-level-online-control-of-the-formula-1 | 2303.00372 | null | https://arxiv.org/abs/2303.00372v1 | https://arxiv.org/pdf/2303.00372v1.pdf | Low-level Online Control of the Formula 1 Power Unit with Feedforward Cylinder Deactivation | Since 2014, the F\'ed\'eration Internationale de l'Automobile has prescribed a parallel hybrid powertrain for the Formula 1 race cars. The complex low-level interactions between the thermal and the electrical part represent a non-trivial and challenging system to be controlled online. We present a novel controller architecture composed of a supervisory controller for the energy management, a feedforward cylinder deactivation controller, and a track region-dependent low-level nonlinear model predictive controller to optimize the engine actuators. Except for the nonlinear model predictive controller, the proposed controller subsystems are computationally inexpensive and are real time capable. The framework is tested and validated in a simulation environment for several realistic scenarios disturbed by driver actions or grip conditions on the track. In particular, we analyze how the control architecture deals with an unexpected gearshift trajectory during an acceleration phase. Further, we demonstrate how an increased maximum velocity trajectory impacts the online low-level controller. Our results show a suboptimality over an entire lap with respect to the benchmark solution of 49 ms and 64 ms, respectively, which we deem acceptable. Compared to the same control architecture with full knowledge of the disturbances, the suboptimality amounted to only 2 ms and 17 ms. For all case studies we show that the cylinder deactivation capability decreases the suboptimality by 7 to 8 ms. | ['Christopher H. Onder', 'Alberto Cerofolini', 'Pol Duhr', 'Camillo Balerna', 'Giona Fieni', 'Marc-Philippe Neumann'] | 2023-03-01 | null | null | null | null | ['energy-management'] | ['time-series'] | [ 9.24395584e-03 7.43475318e-01 -1.20487101e-01 3.31851333e-01
-6.48035930e-05 -6.70305073e-01 6.87446237e-01 7.58207142e-02
-1.87705621e-01 6.02971971e-01 -7.77097762e-01 -4.60967004e-01
-4.84809011e-01 -4.30401474e-01 -9.34245884e-01 -8.04996908e-01
4.12336364e-02 6.11534715e-01 3.55190039e-01 -2.88647324e-01
8.68546497e-03 8.15507889e-01 -2.05448699e+00 -4.58220452e-01
7.00256348e-01 9.32517111e-01 3.73754352e-01 5.84945440e-01
4.79475111e-01 4.46180016e-01 -4.23382699e-01 2.65633345e-01
2.59746432e-01 -1.48698777e-01 -2.21346915e-01 2.86894619e-01
-2.50587314e-01 -1.71889827e-01 -2.79422849e-02 8.69000137e-01
3.28094035e-01 5.78559101e-01 6.15133464e-01 -1.65219808e+00
5.54558933e-01 1.35152817e-01 -6.90393895e-02 -3.88716459e-02
-2.36260086e-01 5.88152409e-01 1.34478733e-01 -4.14275378e-01
5.90628862e-01 9.33655679e-01 3.49577546e-01 3.05029273e-01
-1.00727022e+00 -5.03712475e-01 1.66871041e-01 3.86769176e-01
-1.55574727e+00 -3.71094078e-01 7.46527374e-01 -5.21402717e-01
1.09823346e+00 3.61754775e-01 6.34312332e-01 5.84892750e-01
6.18717730e-01 2.38643944e-01 9.05536473e-01 -1.83585271e-01
3.11896354e-01 4.26801592e-01 -1.26158923e-01 3.33465517e-01
4.21187609e-01 6.27045453e-01 3.15107644e-01 4.09538805e-01
2.83540845e-01 -2.11787060e-01 1.15263045e-01 -3.92817020e-01
-7.32650757e-01 3.37978512e-01 5.93975261e-02 7.49990568e-02
-4.30051476e-01 1.55804157e-01 3.75222385e-01 -1.12858303e-02
-1.71889812e-01 6.47564590e-01 -5.68311810e-01 -4.41295773e-01
-5.78305185e-01 3.52258056e-01 9.24216390e-01 1.35031796e+00
3.37757468e-01 3.78311843e-01 5.29668853e-02 2.45445937e-01
1.37084335e-01 5.88024437e-01 -6.08217977e-02 -9.68001187e-01
1.71512946e-01 6.01530671e-01 6.08031452e-01 -5.85752487e-01
-8.12566757e-01 -5.13674736e-01 -5.19091010e-01 4.81917083e-01
1.97543666e-01 -4.42700952e-01 -4.92935300e-01 1.26645029e+00
5.04501343e-01 -1.68609217e-01 -6.45874739e-02 9.39909756e-01
-1.71152353e-01 5.88315964e-01 -1.02305405e-01 -5.05792499e-01
1.35241938e+00 -7.47353971e-01 -1.09796917e+00 -6.19916059e-02
5.24555862e-01 -7.47943699e-01 7.27669418e-01 6.66761398e-01
-1.25935757e+00 -6.51165664e-01 -1.42560089e+00 3.84052753e-01
-5.85684538e-01 6.73100770e-01 -2.68122286e-01 4.17967290e-01
-6.19168282e-01 7.45835423e-01 -8.13963711e-01 -1.54692546e-01
-4.45659369e-01 4.33520377e-01 2.03977466e-01 6.36734605e-01
-1.16776693e+00 1.54836798e+00 5.47018528e-01 4.48912829e-01
-9.00182366e-01 -7.83975720e-01 -4.90400672e-01 1.41868621e-01
9.16726291e-01 -2.93652207e-01 1.37363148e+00 -3.53114516e-01
-2.16160846e+00 -4.66474146e-02 1.25384703e-01 -3.86496067e-01
5.59267044e-01 -2.53994942e-01 -5.72440624e-01 1.12343222e-01
-5.05790710e-01 6.16546236e-02 6.91261172e-01 -1.03568280e+00
-7.06818998e-01 2.64202744e-01 -3.01429834e-02 9.34582353e-02
4.70678471e-02 -3.86506408e-01 -4.41275090e-02 -2.68709332e-01
-6.50612891e-01 -1.45384514e+00 -1.62928328e-01 -2.95597523e-01
-2.84021318e-01 -1.63247898e-01 1.18258834e+00 -2.34108761e-01
1.28546023e+00 -2.04206133e+00 2.82977045e-01 3.06287736e-01
-5.04516602e-01 3.68020415e-01 4.09954250e-01 5.20331085e-01
-3.07747930e-01 -6.18201718e-02 -3.73719633e-02 -1.17529398e-02
2.13730097e-01 2.39123166e-01 -1.40022844e-01 4.63842124e-01
2.91282147e-01 2.36482143e-01 -5.69223821e-01 1.13200828e-01
5.50266981e-01 3.66052032e-01 -4.83527958e-01 2.81930327e-01
-3.41298878e-01 2.09525138e-01 -3.83092672e-01 1.92188531e-01
4.84069705e-01 4.41006720e-01 2.65748650e-01 -2.87075788e-01
-7.37144351e-01 -1.69030145e-01 -1.45249367e+00 9.32399511e-01
-5.58756173e-01 2.90930212e-01 8.88370812e-01 -8.55895758e-01
1.07872009e+00 3.09383154e-01 4.54176962e-01 -6.30671680e-01
6.74163222e-01 3.10786217e-01 8.18971097e-02 -4.33109045e-01
5.05850375e-01 -1.82771057e-01 4.18260284e-02 -9.08274949e-02
-2.17503384e-01 -7.37239182e-01 3.77355486e-01 -2.96868891e-01
8.85821223e-01 2.50192493e-01 -5.53718023e-02 -7.72779465e-01
1.10524321e+00 3.79458964e-01 5.47696114e-01 -2.70190805e-01
-2.19153473e-03 -4.21089172e-01 6.34403765e-01 1.47488669e-01
-1.11673975e+00 -5.16975999e-01 3.13229486e-02 4.58359241e-01
6.18345916e-01 -2.63583779e-01 -8.16343844e-01 8.67170282e-03
3.18350233e-02 1.37468398e+00 -1.92366615e-02 -7.78589010e-01
-7.62073994e-01 -1.51429132e-01 -9.17316005e-02 5.65337241e-01
1.14583597e-01 -1.85784981e-01 -1.21228492e+00 2.60959625e-01
3.91457975e-01 -1.39795899e+00 -3.93985927e-01 2.96327263e-01
-5.80216050e-01 -1.36545205e+00 8.90877023e-02 -3.34825754e-01
6.45774603e-01 -1.70145288e-01 6.54697418e-01 -3.84902209e-02
-4.09899086e-01 2.83958554e-01 1.51795859e-03 -7.98768997e-01
-6.24658942e-01 -5.57249375e-02 4.16597813e-01 -1.83519483e-01
-1.81008056e-01 1.09188065e-01 -4.13866848e-01 8.91125262e-01
-6.19689643e-01 8.67232233e-02 4.23262477e-01 5.70867419e-01
7.69324064e-01 8.36557984e-01 3.54933947e-01 -3.06603163e-01
3.89801472e-01 -2.79984385e-01 -1.45184708e+00 -1.11432947e-01
-8.55762780e-01 7.95670301e-02 1.17368698e+00 -4.21861798e-01
-1.33136308e+00 5.11170566e-01 2.16722772e-01 -6.90921366e-01
-7.23016635e-02 -1.01428881e-01 -3.01219553e-01 9.29698348e-02
-2.26339083e-02 -4.59146500e-02 5.34620523e-01 -3.62046599e-01
2.23604783e-01 3.64040375e-01 5.64828932e-01 -5.07862449e-01
1.22399569e+00 4.25188616e-02 7.60867596e-01 -7.03683734e-01
1.05588242e-01 -5.46630956e-02 -5.64415455e-01 -7.09217012e-01
8.85910928e-01 -7.24050343e-01 -1.47593927e+00 1.86221778e-01
-9.83890951e-01 -6.03802025e-01 -3.92142475e-01 6.32282436e-01
-1.05196953e+00 -1.88651159e-01 -2.44476885e-01 -1.12633061e+00
-2.12815162e-02 -1.46580851e+00 8.59125912e-01 2.96188205e-01
-2.73961693e-01 -6.68578863e-01 -2.14569792e-01 7.57148787e-02
7.01632321e-01 4.32594806e-01 9.28683102e-01 -8.01126361e-02
-6.98989213e-01 -1.72866762e-01 5.45161903e-01 2.53383338e-01
-2.07772151e-01 2.60691673e-01 -6.35735929e-01 -6.30744040e-01
8.71708170e-02 3.40129048e-01 -3.06935385e-02 1.77861322e-02
8.77963543e-01 -2.34542161e-01 -6.44697547e-01 7.46912286e-02
1.57504070e+00 7.77477205e-01 1.41901165e-01 2.26013854e-01
1.52341053e-01 8.97421539e-01 1.43736494e+00 4.97623384e-01
6.01123422e-02 1.16447973e+00 9.51007664e-01 2.16972344e-02
2.01680243e-01 9.43737030e-02 5.42664826e-01 5.37406266e-01
-1.50540620e-01 -1.02794943e-02 -7.28345454e-01 4.01767582e-01
-1.61652911e+00 -5.57186186e-01 -3.94494742e-01 2.37689233e+00
4.29076195e-01 2.21539527e-01 8.42392743e-02 3.10443252e-01
7.79899597e-01 -5.18442988e-01 -5.37858069e-01 -1.01521790e+00
4.79683995e-01 -5.97494543e-02 1.08422720e+00 5.16925514e-01
-7.26850033e-01 4.28547919e-01 5.75312805e+00 7.57848859e-01
-1.23476529e+00 -1.42590404e-01 3.48322764e-02 -6.78454697e-01
2.71507919e-01 -2.71428917e-02 -9.90308821e-01 6.67872429e-01
1.72374403e+00 -7.26774335e-01 5.38545430e-01 1.08508074e+00
6.97398782e-01 -3.77072185e-01 -1.02858889e+00 3.98152918e-01
-3.34196240e-01 -8.77611935e-01 -7.23279834e-01 6.47969022e-02
3.01688194e-01 -3.71926904e-01 -1.18868083e-01 5.22907078e-01
-9.07774419e-02 -5.88109434e-01 1.00171781e+00 7.04661369e-01
5.39031386e-01 -1.34437001e+00 5.61834097e-01 7.64132082e-01
-1.36226714e+00 -5.23158312e-01 1.21523015e-01 -1.01958930e-01
6.92585289e-01 2.44365185e-01 -6.77505553e-01 7.00668454e-01
3.52290809e-01 1.52049825e-01 -3.48932058e-01 7.40229905e-01
-3.77800576e-02 2.00320199e-01 -5.52077174e-01 -4.42425102e-01
2.38863021e-01 -4.34478432e-01 8.49629521e-01 6.40003502e-01
4.66137916e-01 8.33094418e-02 1.09517850e-01 8.47222447e-01
7.64102697e-01 -2.63315618e-01 -5.68096101e-01 9.27360058e-02
2.17143536e-01 1.37335002e+00 -4.44262683e-01 -2.95324355e-01
4.67823707e-02 3.66765141e-01 -4.74634796e-01 3.08549166e-01
-1.58599102e+00 -7.82011926e-01 1.06341684e+00 4.94043022e-01
3.03553045e-01 -4.93764102e-01 2.86634956e-02 -2.70514786e-01
5.89002036e-02 -3.55838448e-01 -3.25784355e-01 -7.96877027e-01
-4.49613541e-01 3.06406051e-01 6.21314943e-01 -1.57961428e+00
-7.54903197e-01 -9.52140987e-01 -4.87077624e-01 7.92061090e-01
-1.54725957e+00 -1.83351606e-01 -8.63947272e-02 1.97401941e-01
5.33062398e-01 -1.36349732e-02 4.52627897e-01 4.48492259e-01
-1.05553687e+00 2.65950933e-02 2.99592465e-01 -8.44864786e-01
4.73676264e-01 -8.27991247e-01 -2.78373390e-01 8.57103825e-01
-1.41320968e+00 2.62401134e-01 1.28087485e+00 -5.01736760e-01
-2.18309140e+00 -1.27626944e+00 5.30753434e-01 8.67999624e-03
8.30043495e-01 -3.16147417e-01 -7.57727325e-01 4.00269747e-01
4.57212716e-01 -1.86850637e-01 -2.37351865e-01 -8.11868608e-01
7.49435782e-01 -3.18177491e-01 -1.02540183e+00 5.73319077e-01
5.75912952e-01 -1.01324953e-01 -2.73995876e-01 3.15802395e-01
6.39114082e-01 -6.13775134e-01 -1.09993804e+00 5.74372411e-01
2.08067283e-01 -3.19535673e-01 5.61979771e-01 -8.46355185e-02
-9.74326804e-02 -7.69373894e-01 4.39927846e-01 -1.35520816e+00
-2.44521108e-02 -1.03609681e+00 -4.19246644e-01 9.87838507e-01
1.59508884e-01 -6.50246620e-01 4.24098909e-01 5.97495675e-01
-5.40077209e-01 -8.65659118e-01 -1.23405647e+00 -1.23062408e+00
-2.93240603e-02 -3.98384839e-01 5.55408895e-01 2.12749660e-01
1.95779234e-01 1.88759699e-01 -6.28611147e-02 6.13689125e-01
3.07219386e-01 -9.97008663e-03 8.12883794e-01 -8.53913128e-01
-2.69599289e-01 -6.13097072e-01 -1.43197477e-01 -4.65911090e-01
2.31624976e-01 -4.37063932e-01 4.73767936e-01 -1.20541453e+00
-5.50645769e-01 -1.85978618e-02 5.00138104e-03 9.27545205e-02
2.37362176e-01 -6.51349127e-01 2.31729418e-01 -2.92965621e-01
-5.19973114e-02 7.29364514e-01 1.10623515e+00 -2.89536953e-01
-1.58525661e-01 3.31613839e-01 2.21159354e-01 6.21547937e-01
7.52861261e-01 -2.40206882e-01 -9.00481939e-01 8.34542662e-02
-5.45537807e-02 1.78969905e-01 3.15722406e-01 -1.37583351e+00
4.05940652e-01 -2.95714259e-01 -3.50909412e-01 -8.22490990e-01
5.14559984e-01 -1.66737878e+00 8.82559717e-01 1.06517291e+00
1.17315546e-01 1.52170405e-01 6.81220114e-01 4.68233585e-01
-1.90977320e-01 4.62515764e-02 8.21868241e-01 5.59246600e-01
-6.17635131e-01 -1.72036007e-01 -9.56255436e-01 -6.86409056e-01
1.94558764e+00 -2.23831743e-01 -3.39735240e-01 -1.41634513e-03
-7.58943975e-01 8.05617332e-01 3.03506494e-01 6.15494072e-01
-5.26915155e-02 -1.01114094e+00 -6.21786825e-02 1.97481811e-01
-2.53563911e-01 -2.21176565e-01 5.14675081e-01 1.15308368e+00
-3.41943830e-01 9.56578493e-01 -4.00568277e-01 -5.54866791e-01
-1.26543307e+00 9.36084330e-01 6.55085623e-01 2.24294383e-02
-3.82873118e-01 8.61002877e-03 -1.84460700e-01 -1.35452285e-01
-1.07522652e-01 -7.28556395e-01 -1.84741825e-01 3.30855176e-02
3.61259021e-02 1.03892016e+00 3.56168598e-01 -5.73992014e-01
-4.18694913e-01 7.80683815e-01 5.83814919e-01 1.57076359e-01
1.00143611e+00 -2.06672344e-02 3.28628153e-01 4.35322195e-01
8.28246236e-01 -4.80522275e-01 -1.44983053e+00 5.98659217e-01
-1.53118834e-01 8.66761133e-02 1.16965532e-01 -7.02200174e-01
-1.05023217e+00 5.04150391e-01 5.68354011e-01 1.95301488e-01
1.25682068e+00 -5.64839005e-01 2.57484794e-01 3.26138020e-01
4.06211823e-01 -1.68422186e+00 -4.82738316e-01 7.78640211e-01
1.10519707e+00 -3.19498092e-01 1.37602150e-01 -7.01097310e-01
-5.40161788e-01 1.12091815e+00 8.66616905e-01 -1.53336957e-01
8.51380765e-01 7.42252946e-01 -4.39713359e-01 2.79509395e-01
-1.15637398e+00 5.95926940e-02 1.33259445e-01 2.43467242e-02
-6.18552007e-02 -1.43363168e-02 -6.24941289e-01 6.51298106e-01
1.08583935e-01 1.25732303e-01 6.95789337e-01 9.63627160e-01
-4.12250042e-01 -8.26603591e-01 -6.31145895e-01 -1.35870308e-01
9.22788382e-02 9.07824159e-01 1.57751158e-01 1.47979009e+00
2.95596272e-01 8.89751554e-01 4.04352188e-01 -3.76856893e-01
1.26896238e+00 1.22252278e-01 9.17710438e-02 -4.18535262e-01
-7.27883697e-01 2.09729880e-01 3.87396872e-01 -1.03444183e+00
-8.99627209e-02 -4.59636152e-01 -1.65590477e+00 -1.69004664e-01
-4.59569037e-01 1.86280116e-01 1.00025165e+00 8.22920680e-01
6.04871988e-01 1.11570561e+00 6.20235085e-01 -1.02778494e+00
-7.77202249e-01 -7.99437821e-01 -6.38147593e-01 -2.44253010e-01
9.22044888e-02 -1.29525304e+00 -5.71845531e-01 -1.86175093e-01] | [5.434308052062988, 2.2043256759643555] |
3b5c49aa-6a72-4140-b8b4-efb16073b6c8 | question-answering-over-knowledge-base-using | null | null | https://aclanthology.org/N16-2016 | https://aclanthology.org/N16-2016.pdf | Question Answering over Knowledge Base using Factual Memory Networks | null | ['Sarthak Jain'] | 2016-06-01 | null | null | null | naacl-2016-6 | ['knowledge-base-question-answering'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.288315773010254, 3.8666677474975586] |
e24f7a23-faae-46d4-9c75-9f67e7aedbbc | designing-novel-cognitive-diagnosis-models | 2307.04429 | null | https://arxiv.org/abs/2307.04429v1 | https://arxiv.org/pdf/2307.04429v1.pdf | Designing Novel Cognitive Diagnosis Models via Evolutionary Multi-Objective Neural Architecture Search | Cognitive diagnosis plays a vital role in modern intelligent education platforms to reveal students' proficiency in knowledge concepts for subsequent adaptive tasks. However, due to the requirement of high model interpretability, existing manually designed cognitive diagnosis models hold too simple architectures to meet the demand of current intelligent education systems, where the bias of human design also limits the emergence of effective cognitive diagnosis models. In this paper, we propose to automatically design novel cognitive diagnosis models by evolutionary multi-objective neural architecture search (NAS). Specifically, we observe existing models can be represented by a general model handling three given types of inputs and thus first design an expressive search space for the NAS task in cognitive diagnosis. Then, we propose multi-objective genetic programming (MOGP) to explore the NAS task's search space by maximizing model performance and interpretability. In the MOGP design, each architecture is transformed into a tree architecture and encoded by a tree for easy optimization, and a tailored genetic operation based on four sub-genetic operations is devised to generate offspring effectively. Besides, an initialization strategy is also suggested to accelerate the convergence by evolving half of the population from existing models' variants. Experiments on two real-world datasets demonstrate that the cognitive diagnosis models searched by the proposed approach exhibit significantly better performance than existing models and also hold as good interpretability as human-designed models. | ['Xingyi Zhang', 'Yaochu Jin', 'Limiao Zhang', 'Ye Tian', 'Cheng Zhen', 'Haiping Ma', 'Shangshang Yang'] | 2023-07-10 | null | null | null | null | ['architecture-search'] | ['methodology'] | [ 2.54815280e-01 2.98312247e-01 -1.21776145e-02 -2.95016319e-01
-8.98252055e-02 -3.06214303e-01 1.32112861e-01 -1.08556405e-01
1.71130095e-02 4.96854067e-01 -2.64960229e-01 -4.79494810e-01
-9.01744068e-01 -1.00080228e+00 -3.00913632e-01 -7.13103592e-01
3.82164121e-01 6.96918070e-01 2.19553351e-01 -3.60433340e-01
3.39360416e-01 2.02741712e-01 -2.05060768e+00 8.74668956e-02
1.95131719e+00 9.18828249e-01 6.22897446e-01 3.44954044e-01
-4.39972907e-01 5.66848874e-01 -7.92134106e-01 -3.90778095e-01
-2.56210089e-01 -6.47726238e-01 -7.86538422e-01 1.33793429e-01
-2.85890907e-01 1.08126275e-01 1.32385492e-01 1.25046396e+00
7.08127499e-01 3.04601341e-01 6.08598709e-01 -1.10782957e+00
-1.17075992e+00 6.72858834e-01 -2.15293288e-01 1.06026880e-01
2.91726023e-01 -4.62308247e-03 6.16580009e-01 -4.70882446e-01
1.18722782e-01 1.37845862e+00 3.82034630e-01 7.83391178e-01
-9.55885231e-01 -5.26773274e-01 3.25957060e-01 6.11605823e-01
-1.48258126e+00 1.77686866e-02 9.79995728e-01 -4.13577706e-01
5.14049232e-01 5.64014971e-01 1.10785007e+00 8.37233126e-01
6.18957058e-02 8.20228338e-01 8.63297462e-01 -5.78787506e-01
5.00840425e-01 3.09386909e-01 2.11580247e-01 1.10831463e+00
2.29313448e-01 8.86245817e-02 -2.23939419e-01 4.90011275e-03
4.42385882e-01 -2.51882941e-01 -2.93631375e-01 -1.91608101e-01
-6.40743911e-01 6.83265448e-01 3.68299693e-01 3.35506767e-01
-3.80634665e-01 -3.78857464e-01 -1.11614786e-01 1.64959311e-01
-2.86809728e-02 7.65431523e-01 -2.58742839e-01 8.52907151e-02
-3.71403396e-01 1.73418164e-01 4.12365079e-01 8.68718624e-01
3.98181647e-01 2.62142956e-01 -1.26098096e-01 9.64898288e-01
3.78012031e-01 3.68018478e-01 1.15017271e+00 -6.34696722e-01
1.47320017e-01 1.39263415e+00 -3.90290171e-01 -9.63226318e-01
-3.06399614e-01 -1.03765094e+00 -9.27066743e-01 5.52059151e-02
-2.42185056e-01 -3.91401090e-02 -7.43731916e-01 1.52405751e+00
4.49687809e-01 1.81574285e-01 2.85957277e-01 8.49682450e-01
9.76712704e-01 6.15708172e-01 3.01901978e-02 -3.47736418e-01
1.54781508e+00 -9.19043779e-01 -7.50122726e-01 -3.09740156e-01
3.64728034e-01 -2.40225628e-01 1.07118165e+00 6.10453725e-01
-1.17564845e+00 -6.51547670e-01 -1.14856660e+00 3.19314480e-01
-3.76771510e-01 2.04634011e-01 5.97277761e-01 1.00701773e+00
-1.05056310e+00 -6.89060614e-02 -3.07731956e-01 6.28358796e-02
3.17105770e-01 4.81591076e-01 3.84617090e-01 -9.89327654e-02
-1.24649668e+00 7.69852102e-01 7.91610837e-01 2.40030169e-01
-6.01227522e-01 -5.82746625e-01 -6.28847480e-01 3.86053681e-01
4.37589496e-01 -1.03449023e+00 8.90734196e-01 -1.18223190e+00
-1.80269694e+00 4.97728437e-01 1.90021452e-02 1.64762267e-03
3.83208543e-01 4.89281207e-01 -5.89353859e-01 -3.89251374e-02
-2.93160111e-01 4.86294717e-01 7.02982783e-01 -1.23558271e+00
-7.91976213e-01 -2.36385718e-01 2.68864125e-01 4.62094843e-01
-9.99840379e-01 -1.14612296e-01 -4.48082685e-01 -6.01668298e-01
3.45461607e-01 -8.25124800e-01 -3.51651430e-01 -3.39889109e-01
-1.30440369e-01 -3.91523689e-01 3.70876700e-01 -5.74998677e-01
1.84943163e+00 -1.72807336e+00 6.71019018e-01 5.35332501e-01
2.87704289e-01 4.19717133e-01 -1.03815235e-01 -4.02971894e-01
6.59307912e-02 1.93376422e-01 -2.51271576e-01 -1.71886869e-02
1.90476403e-01 3.78599674e-01 2.47665033e-01 -3.74049902e-01
4.27043252e-02 9.53782976e-01 -8.23160648e-01 -7.25384474e-01
-1.30715206e-01 9.85913947e-02 -8.42803776e-01 2.71206945e-01
-3.00065935e-01 2.06521019e-01 -1.09191024e+00 9.14662004e-01
4.04639691e-01 -2.91638643e-01 3.37361753e-01 2.31783599e-01
4.23540100e-02 -2.06071690e-01 -1.02344751e+00 1.38713670e+00
-5.54063499e-01 4.57844734e-02 -6.05055466e-02 -1.10329568e+00
1.17079961e+00 2.74614930e-01 7.38072470e-02 -7.86342263e-01
3.81769925e-01 1.41158178e-01 4.86201942e-01 -7.66611040e-01
1.42761126e-01 3.77031676e-02 8.72612186e-03 3.88791114e-01
-2.88772702e-01 1.88466236e-02 -1.90192508e-03 -4.73170280e-01
8.31480563e-01 -8.38017836e-02 1.89691916e-01 -4.78469342e-01
1.05443537e+00 5.64481802e-02 8.38672817e-01 5.31124294e-01
-1.96577869e-02 8.29305053e-02 5.28289229e-02 -6.45694554e-01
-5.03238618e-01 -9.19671714e-01 2.40654349e-02 1.04426372e+00
2.12286562e-01 -3.65068465e-02 -1.02285624e+00 -5.87744772e-01
-3.59644860e-01 8.84692371e-01 -4.10353780e-01 -7.69400060e-01
-6.24860525e-01 -1.10860312e+00 5.85592210e-01 3.44933063e-01
7.03068197e-01 -1.17820835e+00 -8.15211952e-01 2.81450987e-01
-2.11784840e-01 -4.36754942e-01 -1.26719639e-01 -7.37939849e-02
-7.71140933e-01 -8.66079867e-01 -3.20688426e-01 -1.17846191e+00
9.87257659e-01 7.76736811e-02 9.35633898e-01 7.19207168e-01
-2.88167775e-01 3.52458470e-02 -2.55156666e-01 -5.51878035e-01
-5.66345334e-01 2.09375381e-01 7.99903274e-02 -5.73723428e-02
2.65445083e-01 -7.10926294e-01 -3.48111868e-01 2.59618312e-01
-1.00471878e+00 3.42211872e-01 7.71132052e-01 9.00886416e-01
5.55998504e-01 8.66939843e-01 9.55258489e-01 -2.51992851e-01
1.18745327e+00 -4.07731056e-01 -6.25644624e-01 9.17189002e-01
-1.05033422e+00 3.44721466e-01 8.20221782e-01 -6.93858802e-01
-1.35793269e+00 -2.45409489e-01 3.75000462e-02 -6.06016695e-01
-2.56577134e-02 8.22070479e-01 -5.92932642e-01 -2.23673984e-01
5.37070751e-01 5.79320014e-01 2.01080162e-02 -3.74278784e-01
-1.95589408e-01 6.64875329e-01 3.28940332e-01 -1.06339490e+00
5.72515488e-01 -3.19031447e-01 4.33406234e-02 -4.36101079e-01
-6.38614178e-01 3.58725756e-01 -3.44410688e-01 -4.21506584e-01
9.33402300e-01 -4.58578438e-01 -1.11463404e+00 2.89778471e-01
-8.62657011e-01 1.40004739e-01 1.13732971e-01 2.91249007e-01
-2.40337595e-01 -3.13036032e-02 -2.45460629e-01 -9.69056070e-01
-4.47575182e-01 -1.61562562e+00 4.56799418e-01 5.66920042e-01
-1.20423034e-01 -1.07869744e+00 -3.28176707e-01 7.29581118e-01
3.80188972e-01 -1.26308173e-01 1.69351923e+00 -5.35776079e-01
-5.61459303e-01 5.72982095e-02 2.71815926e-01 -5.79900034e-02
-3.06726918e-02 -9.45450813e-02 -7.63774753e-01 -4.15098704e-02
1.80461898e-01 -8.93312842e-02 3.42569262e-01 2.05065176e-01
1.58768451e+00 -4.99512643e-01 -3.11487675e-01 5.88985145e-01
1.13757849e+00 9.83495891e-01 4.75625873e-01 6.52671456e-01
5.45865357e-01 7.92372525e-01 4.36856002e-01 2.05798358e-01
5.99566519e-01 4.82228816e-01 3.47737461e-01 4.20093328e-01
2.55791426e-01 -1.34028435e-01 3.85445207e-02 1.23062634e+00
-2.23570243e-01 -4.10591841e-01 -1.20202506e+00 2.60976762e-01
-1.79843795e+00 -7.37983048e-01 2.53949076e-01 1.82770228e+00
7.79214859e-01 4.37232992e-03 -1.57557815e-01 2.25439161e-01
8.57400239e-01 -4.68503922e-01 -6.07110560e-01 -4.76874381e-01
-1.67637214e-01 1.56510338e-01 -2.51655281e-01 2.59330124e-01
-5.94494164e-01 6.41033590e-01 5.65180969e+00 1.11785650e+00
-9.99150753e-01 -1.24697417e-01 6.48837090e-01 -1.70607418e-01
-7.25059628e-01 -3.47479820e-01 -5.83303213e-01 6.72601104e-01
6.70636833e-01 -5.25493979e-01 7.44576752e-01 6.86951816e-01
1.45498648e-01 3.84350777e-01 -5.83440363e-01 9.33744133e-01
4.03605141e-02 -1.14533353e+00 4.90037829e-01 1.53251395e-01
8.28509748e-01 -1.07258451e+00 3.14326704e-01 5.12999952e-01
1.67182013e-01 -1.04872870e+00 8.04580688e-01 6.55740559e-01
3.61898959e-01 -1.06033850e+00 5.45193255e-01 7.50578105e-01
-1.12224853e+00 -7.39537239e-01 -4.51437473e-01 1.44091368e-01
-1.91329733e-01 1.78905711e-01 -5.96265495e-01 7.16694951e-01
5.25944829e-01 1.24405593e-01 -9.61890519e-01 9.14682686e-01
-2.14118943e-01 3.33368361e-01 4.83631752e-02 -4.98152316e-01
2.47154355e-01 -3.86255354e-01 5.16292155e-01 4.82869953e-01
8.19917083e-01 5.22014797e-01 2.26786762e-01 1.07852650e+00
2.45333090e-01 1.06085218e-01 -1.68491587e-01 5.28850257e-02
7.90678561e-01 8.81591558e-01 -6.35319471e-01 -2.65778840e-01
6.20425195e-02 5.82666636e-01 4.63294923e-01 1.17407054e-01
-9.25511837e-01 -1.33973643e-01 5.01930237e-01 -2.44136348e-01
-1.32938847e-01 4.20781761e-01 -5.53709090e-01 -9.09938514e-01
4.39915843e-02 -1.15105116e+00 4.96182352e-01 -9.12150264e-01
-9.12210286e-01 9.35508728e-01 1.02789970e-02 -1.10920811e+00
-2.82195419e-01 -3.65441740e-01 -7.74802506e-01 9.46741521e-01
-1.07259226e+00 -8.77308309e-01 -4.44715291e-01 4.63172704e-01
6.80588067e-01 -6.63641036e-01 8.23540390e-01 1.96202680e-01
-1.10353744e+00 7.82391429e-01 -1.57454297e-01 -4.70566124e-01
-3.58820744e-02 -1.07806408e+00 -1.26940846e-01 6.22591555e-01
-3.56980294e-01 6.66225374e-01 5.76207280e-01 -6.61509812e-01
-1.47985673e+00 -9.18073356e-01 6.73231304e-01 1.99856963e-02
2.57836819e-01 2.00372964e-01 -1.10969293e+00 1.74967587e-01
4.78462391e-02 -6.54852986e-01 7.38889217e-01 -1.15431301e-01
3.04312468e-01 1.91542190e-02 -1.27279067e+00 8.87201965e-01
1.34471130e+00 2.46973895e-02 -7.84560800e-01 -7.19536794e-03
8.50786030e-01 -3.23486835e-01 -7.72057593e-01 5.73981345e-01
4.83046800e-01 -7.45883405e-01 1.03762567e+00 -5.60157597e-01
4.76006776e-01 -3.39376062e-01 2.61921495e-01 -1.49294603e+00
-7.42718935e-01 -2.18069673e-01 -2.66238034e-01 1.09735703e+00
2.72487581e-01 -8.25764120e-01 8.02958786e-01 7.22059429e-01
-5.30232251e-01 -1.24615574e+00 -7.01302052e-01 -6.27922595e-01
-1.20355524e-01 -4.29618694e-02 1.35429764e+00 1.12979412e+00
6.15954027e-03 2.12435335e-01 2.72429168e-01 3.39577913e-01
5.67318797e-01 3.44642818e-01 7.32753277e-02 -1.62381613e+00
-4.57427025e-01 -1.05362618e+00 -3.00482422e-01 -8.05268705e-01
4.61815178e-01 -8.72258186e-01 -1.33353397e-01 -1.30760574e+00
8.68999958e-02 -7.88211823e-01 -3.32739621e-01 4.81153905e-01
-4.98281270e-01 -4.26444829e-01 -1.32878609e-02 -2.80652032e-03
-3.49696100e-01 7.54464447e-01 1.56235158e+00 -3.06704372e-01
-3.31260264e-01 7.26738572e-02 -9.64455664e-01 7.79290617e-01
8.20694983e-01 -1.92036599e-01 -1.09063423e+00 -5.74116945e-01
2.39941150e-01 3.15437049e-01 5.68918809e-02 -1.00688791e+00
4.55597073e-01 -4.59038287e-01 3.99691790e-01 -2.78523624e-01
1.54383108e-01 -8.37613344e-01 4.76553828e-01 7.00479686e-01
-3.38294804e-01 3.99588525e-01 1.36305198e-01 3.31866056e-01
-1.16716020e-01 -6.76590323e-01 4.84806567e-01 -1.02281980e-01
-6.99386299e-01 3.33489850e-02 -4.89087909e-01 -2.03974336e-01
1.16382980e+00 -6.62490726e-01 -1.77736297e-01 1.10088445e-01
-6.78752244e-01 6.40201509e-01 1.75726950e-01 5.97334564e-01
8.61232221e-01 -1.37557840e+00 -4.11654145e-01 4.33609337e-01
-5.27447425e-02 2.53192902e-01 5.53375840e-01 4.02904183e-01
-5.60603678e-01 5.00386834e-01 -3.50066423e-01 -4.67540145e-01
-1.29144907e+00 6.38740659e-01 5.26328444e-01 -2.85543978e-01
-2.62033999e-01 7.38903284e-01 3.16251665e-01 -6.27446830e-01
2.42489710e-01 -1.39619291e-01 -7.28938937e-01 -5.43410629e-02
4.72541034e-01 4.81827021e-01 8.01374838e-02 -2.52303004e-01
-1.51319429e-01 4.85939026e-01 1.85433581e-01 2.23784864e-01
1.19900239e+00 -1.05158307e-01 -8.74500871e-02 -1.23852260e-01
5.96479058e-01 -3.47887576e-01 -7.36567318e-01 -1.66600849e-02
-1.60820633e-02 -2.00490832e-01 1.32428393e-01 -1.06278241e+00
-1.11549962e+00 5.89301407e-01 6.60254598e-01 1.94908410e-01
1.75362384e+00 -3.34871143e-01 4.22249347e-01 5.14749467e-01
4.41663891e-01 -8.24751198e-01 1.51177853e-01 3.31298828e-01
7.36968696e-01 -6.65448606e-01 -3.33245635e-01 -4.24743146e-01
-4.86608595e-01 1.04935527e+00 1.12676454e+00 4.89257872e-01
2.33903795e-01 1.45211080e-02 -2.08334967e-01 -2.50015229e-01
-8.80374074e-01 6.82349177e-03 5.83198726e-01 5.23182273e-01
-3.56405973e-03 8.07605162e-02 -4.15739477e-01 1.35200202e+00
-5.02435148e-01 -1.19113408e-01 2.35416889e-01 9.60723400e-01
-7.55178630e-01 -1.01349688e+00 -6.42531753e-01 2.66836077e-01
2.52890587e-02 -1.15068732e-02 -4.09588277e-01 4.07470316e-01
5.38226008e-01 8.99396658e-01 -8.99653956e-02 -5.90015471e-01
3.60844463e-01 2.49960259e-01 5.41797340e-01 -3.62091660e-01
-8.52329671e-01 -2.41861060e-01 -1.97032899e-01 -1.01762488e-02
-1.33525312e-01 -3.10346544e-01 -1.35431814e+00 -3.04860979e-01
-3.20390105e-01 4.41615671e-01 3.85415226e-01 1.05779135e+00
2.99353600e-01 1.15339386e+00 4.18197811e-01 -3.30863267e-01
-4.55481887e-01 -6.18911386e-01 -1.42284587e-01 -1.64601877e-01
-3.08357567e-01 -9.51235771e-01 -4.83449996e-02 -9.91038233e-03] | [8.191364288330078, 3.4047327041625977] |
d208aec8-d6e9-48f5-b2d0-1f0e4b9e76c8 | robots-that-ask-for-help-uncertainty | 2307.01928 | null | https://arxiv.org/abs/2307.01928v1 | https://arxiv.org/pdf/2307.01928v1.pdf | Robots That Ask For Help: Uncertainty Alignment for Large Language Model Planners | Large language models (LLMs) exhibit a wide range of promising capabilities -- from step-by-step planning to commonsense reasoning -- that may provide utility for robots, but remain prone to confidently hallucinated predictions. In this work, we present KnowNo, which is a framework for measuring and aligning the uncertainty of LLM-based planners such that they know when they don't know and ask for help when needed. KnowNo builds on the theory of conformal prediction to provide statistical guarantees on task completion while minimizing human help in complex multi-step planning settings. Experiments across a variety of simulated and real robot setups that involve tasks with different modes of ambiguity (e.g., from spatial to numeric uncertainties, from human preferences to Winograd schemas) show that KnowNo performs favorably over modern baselines (which may involve ensembles or extensive prompt tuning) in terms of improving efficiency and autonomy, while providing formal assurances. KnowNo can be used with LLMs out of the box without model-finetuning, and suggests a promising lightweight approach to modeling uncertainty that can complement and scale with the growing capabilities of foundation models. Website: https://robot-help.github.io | ['Anirudha Majumdar', 'Andy Zeng', 'Dorsa Sadigh', 'Zhenjia Xu', 'Jake Varley', 'Fei Xia', 'Leila Takayama', 'Peng Xu', 'Noah Brown', 'Stephen Tu', 'Sumeet Singh', 'Alexandra Bodrova', 'Anushri Dixit', 'Allen Z. Ren'] | 2023-07-04 | null | null | null | null | ['conformal-prediction', 'conformal-prediction'] | ['computer-vision', 'reasoning'] | [-1.09033950e-01 9.51790750e-01 -1.55575484e-01 -6.10253513e-01
-1.00735116e+00 -5.72645605e-01 9.09368157e-01 1.37213439e-01
-4.15270180e-01 7.39261866e-01 4.31422263e-01 -3.52304220e-01
-3.75660777e-01 -3.99709761e-01 -8.29520881e-01 -1.83739305e-01
-3.91236365e-01 1.28648627e+00 1.27483830e-01 -5.24712145e-01
3.76889855e-01 2.17478484e-01 -1.27307642e+00 2.05926508e-01
7.77986825e-01 8.31641555e-01 5.97078979e-01 3.27498347e-01
2.80207515e-01 1.14539135e+00 -2.56807834e-01 -2.94091087e-02
4.62315232e-01 2.71139860e-01 -1.02111328e+00 -1.97070956e-01
-6.87795579e-02 -4.32756662e-01 -3.17622811e-01 9.30359244e-01
2.59380460e-01 2.75252402e-01 6.59193516e-01 -1.62807608e+00
-5.10651231e-01 9.92794454e-01 7.49337748e-02 -3.25181633e-01
7.46550262e-01 6.30472481e-01 8.01450789e-01 -7.21856892e-01
5.59871078e-01 1.82981384e+00 8.08443069e-01 5.24132013e-01
-1.31315386e+00 -4.73178029e-01 4.14495617e-02 1.98834717e-01
-1.35743415e+00 -7.75721669e-01 1.20990276e-02 -5.28240204e-01
1.38071227e+00 -7.29584694e-02 1.75015330e-01 1.18303967e+00
6.94507539e-01 6.57816470e-01 1.07771778e+00 -1.37540132e-01
5.79503119e-01 1.38353273e-01 -2.39483014e-01 6.07382238e-01
1.61740214e-01 2.94391453e-01 -8.65867376e-01 -4.27198112e-01
5.27868271e-01 -4.22555596e-01 -4.96203499e-03 -5.66031873e-01
-1.66555285e+00 6.40727699e-01 3.06373179e-01 -2.68355429e-01
-4.85496998e-01 4.00282472e-01 1.02899954e-01 2.13395700e-01
-5.92789836e-02 1.07777262e+00 -5.96114695e-01 -3.63435417e-01
-4.34265316e-01 7.03314960e-01 9.80120122e-01 1.53771126e+00
6.74606383e-01 -1.88211456e-01 -1.04199044e-01 6.11755610e-01
4.53686655e-01 4.13454473e-01 5.84155202e-01 -1.87834525e+00
6.68543220e-01 2.40848750e-01 7.82182157e-01 -8.44899118e-01
-9.81769502e-01 1.23225622e-01 -3.73242766e-01 5.53654253e-01
3.05872142e-01 -1.50412813e-01 -8.83075356e-01 1.92497230e+00
9.50445607e-02 -2.67888188e-01 4.02773499e-01 9.02157485e-01
1.43435106e-01 5.36471546e-01 1.43123940e-01 1.21697508e-01
1.15558755e+00 -9.98806775e-01 -4.40262496e-01 -9.74136531e-01
5.96919060e-01 -3.70712608e-01 1.39383340e+00 6.36152029e-01
-1.22708309e+00 -2.94962764e-01 -9.68761086e-01 -2.17172876e-01
-1.11804813e-01 -3.42555016e-01 8.06064546e-01 1.52407989e-01
-1.26400197e+00 8.91198754e-01 -1.12304604e+00 -4.90358561e-01
2.13635772e-01 2.82764047e-01 -5.92878819e-01 -1.81052878e-01
-1.05731630e+00 1.58445644e+00 8.67178261e-01 -8.20905343e-02
-1.29848361e+00 -2.70592868e-01 -1.11948538e+00 -1.73968207e-02
7.25869477e-01 -6.89529419e-01 1.94101548e+00 -2.73894042e-01
-1.61183012e+00 3.15393806e-01 8.84454548e-02 -8.30600917e-01
6.57440722e-01 -4.73073989e-01 1.03534393e-01 -1.26274422e-01
6.01616025e-01 1.34159470e+00 4.48788822e-01 -1.07348990e+00
-5.15959263e-01 -1.55799881e-01 2.86418378e-01 6.44275188e-01
2.28086561e-01 -3.12324494e-01 7.22151101e-02 -1.66572593e-02
3.85136455e-01 -1.32767200e+00 -6.08949065e-01 -1.44462019e-01
-4.68900830e-01 -2.46232018e-01 1.35774538e-02 -4.03557569e-01
4.87244397e-01 -1.77729774e+00 2.49129996e-01 -6.88286796e-02
-9.48134512e-02 -4.39736396e-01 -1.71836689e-01 6.51586413e-01
3.97848010e-01 1.47217527e-01 -2.55886793e-01 -5.21506250e-01
5.51078618e-01 4.35388595e-01 -4.18402255e-01 3.25299442e-01
1.74745262e-01 8.64688456e-01 -1.06280196e+00 -6.21861458e-01
2.76883602e-01 -5.17312773e-02 -5.53530514e-01 2.06162900e-01
-8.49907756e-01 4.21629637e-01 -2.77334094e-01 5.83767951e-01
3.12300414e-01 -1.27747938e-01 2.93067306e-01 4.68667418e-01
4.17450294e-02 6.35012209e-01 -1.15164554e+00 2.07153225e+00
-4.64073688e-01 3.58494788e-01 2.01698661e-01 -2.89810181e-01
6.36785865e-01 1.74172714e-01 3.02767078e-03 -3.47393692e-01
1.26600079e-02 2.54714817e-01 -2.54698303e-02 -2.90371060e-01
7.83253610e-01 -1.19498260e-01 -5.59371769e-01 2.50714004e-01
2.16916606e-01 -8.88316870e-01 4.53073182e-04 2.31268957e-01
1.28943062e+00 4.90951926e-01 5.35819650e-01 -4.86724675e-01
-1.66900098e-01 5.12481153e-01 5.33245206e-01 1.09711361e+00
-3.02791327e-01 3.51339221e-01 4.36935842e-01 -2.18806207e-01
-9.94155228e-01 -1.08485186e+00 5.35751358e-02 1.11299992e+00
3.52255642e-01 -4.55722898e-01 -6.72073007e-01 -2.51167238e-01
3.26265320e-02 1.70128787e+00 -4.95210856e-01 -1.77349851e-01
-1.79024220e-01 -2.22948954e-01 6.04487956e-01 5.33900023e-01
2.81656086e-01 -1.24535036e+00 -1.09180272e+00 2.07686678e-01
-3.86360794e-01 -1.18699288e+00 -2.66745836e-01 6.54979348e-01
-6.76072001e-01 -8.36669803e-01 -9.77010354e-02 -4.45667922e-01
3.90172511e-01 1.39866367e-01 1.13262749e+00 -5.15006542e-01
2.24265695e-01 6.56359196e-01 -2.21743360e-01 -6.78097904e-01
-6.42821074e-01 -2.03392968e-01 5.85657716e-01 -8.92142236e-01
1.37761896e-02 -6.71369731e-01 -1.81695282e-01 4.13094759e-01
-4.88212287e-01 3.65222782e-01 7.43388295e-01 7.95359373e-01
4.13191587e-01 -4.30968590e-02 3.60528648e-01 -6.25003994e-01
9.73601162e-01 -7.63299704e-01 -5.50064564e-01 2.10793003e-01
-9.54335392e-01 4.47740048e-01 4.65572536e-01 -3.57565433e-01
-1.15412521e+00 1.84202433e-01 4.00271147e-01 -2.47909695e-01
-3.49799991e-01 5.12018502e-01 -1.54868558e-01 1.96970627e-01
1.04865324e+00 4.03927192e-02 1.00876074e-02 -2.24884614e-01
6.52317286e-01 4.68854249e-01 7.91745603e-01 -9.51450527e-01
5.64499855e-01 2.55789489e-01 1.91916607e-03 -2.28436306e-01
-5.18976092e-01 -8.86753723e-02 -5.19275963e-01 5.90260997e-02
6.60813749e-01 -1.03085220e+00 -7.66172886e-01 1.42903417e-01
-1.22151577e+00 -9.00795519e-01 -3.25296044e-01 4.77146029e-01
-1.32639587e+00 2.42991716e-01 -6.65600538e-01 -1.04963410e+00
1.33470982e-01 -1.26697361e+00 1.07408822e+00 1.81659535e-01
-5.83506823e-01 -5.65859437e-01 -2.00919494e-01 2.27774501e-01
4.43035036e-01 7.35391751e-02 8.57406497e-01 -9.30056512e-01
-5.21094918e-01 -7.24333972e-02 1.13837346e-01 -1.43497661e-01
-3.39186549e-01 -7.77066529e-01 -8.17472577e-01 -8.98018777e-02
3.70836034e-02 -7.89639771e-01 3.03857028e-01 1.20699108e-01
7.44696856e-01 -6.33573949e-01 -4.32663441e-01 2.44196907e-01
8.87488246e-01 1.10908248e-01 5.45205235e-01 7.48652220e-01
1.03314005e-01 6.73013508e-01 1.24216938e+00 6.57500148e-01
1.08534098e+00 6.08025670e-01 7.82765150e-01 7.49145925e-01
2.66987145e-01 -6.93608582e-01 5.36176920e-01 9.99530405e-02
2.49284297e-01 -1.18162431e-01 -1.33851504e+00 4.30633545e-01
-2.23038554e+00 -7.49819994e-01 3.72953147e-01 2.07656980e+00
8.63026261e-01 3.79591346e-01 -1.49073631e-01 -5.36918044e-01
3.76975030e-01 -1.35005713e-01 -7.68533826e-01 -3.82003367e-01
2.36081466e-01 -4.85071540e-01 5.60791850e-01 8.42818439e-01
-7.13758707e-01 1.25930274e+00 6.64722633e+00 6.44266903e-01
-4.34799373e-01 1.47784144e-01 5.07553279e-01 -1.06380507e-01
-2.48809159e-01 2.53902644e-01 -6.32253349e-01 1.68848246e-01
1.11608613e+00 -2.51447499e-01 1.05539083e+00 1.40725613e+00
4.57766876e-02 -5.76216877e-01 -1.71807790e+00 8.25378895e-01
-1.14337750e-01 -1.02969134e+00 -4.13731188e-01 -1.12052761e-01
4.10417318e-01 4.32912767e-01 7.01140333e-03 7.37798095e-01
1.05082822e+00 -1.37529266e+00 1.51694953e+00 5.85341394e-01
4.76179808e-01 -5.05815983e-01 6.78081512e-01 1.15127993e+00
-6.54402912e-01 -3.85089219e-01 -5.85577011e-01 -3.18873465e-01
3.91019791e-01 1.27640888e-02 -1.50966549e+00 2.53038436e-01
6.70368314e-01 2.27990702e-01 -3.99643809e-01 7.17552006e-01
-3.59945089e-01 -5.45824207e-02 -5.38843989e-01 -9.35283229e-02
2.10838467e-01 1.22817434e-01 5.99789321e-01 1.02465308e+00
4.50758576e-01 2.44917482e-01 4.56529140e-01 1.03298426e+00
3.96458298e-01 -4.68909830e-01 -7.75165319e-01 1.05494387e-01
1.08309734e+00 8.27533603e-01 -3.34653646e-01 -1.59451768e-01
1.69403598e-01 8.67643237e-01 5.35477340e-01 4.19177189e-02
-8.37929070e-01 1.34800076e-02 5.97829163e-01 -5.74773885e-02
-2.69065261e-01 -6.32900596e-01 -4.42604125e-01 -7.70611942e-01
-5.74490763e-02 -1.06726217e+00 1.66276783e-01 -1.36693311e+00
-9.97891963e-01 5.48811853e-01 4.63790536e-01 -1.08582056e+00
-7.66682386e-01 -5.56119621e-01 -1.92758694e-01 6.76156342e-01
-1.13125384e+00 -1.16177738e+00 -1.28245443e-01 2.77672499e-01
7.41026998e-01 -1.38347253e-01 1.09501219e+00 -6.05695724e-01
1.42768800e-01 1.42483547e-01 -2.76920766e-01 -6.26614749e-01
8.17172527e-01 -1.22933412e+00 7.31455147e-01 4.87305164e-01
-1.27791524e-01 7.30157375e-01 1.14781010e+00 -8.49667072e-01
-1.45783544e+00 -7.77152538e-01 6.21943831e-01 -9.51379120e-01
6.76599622e-01 -2.93527186e-01 -5.37661135e-01 1.25074291e+00
3.75764109e-02 -3.16590518e-01 2.19398990e-01 3.37689400e-01
-3.25186342e-01 4.07522559e-01 -1.45119333e+00 8.56368303e-01
1.18019116e+00 -3.19106668e-01 -1.08646035e+00 7.30363607e-01
1.12454629e+00 -8.54312718e-01 -8.18281829e-01 3.38049352e-01
4.94894266e-01 -1.11274588e+00 7.77597964e-01 -5.25415659e-01
2.65964895e-01 -3.12462747e-01 -6.75985515e-01 -1.59636164e+00
-6.00542605e-01 -8.62718105e-01 1.34856347e-02 8.27247977e-01
5.94933510e-01 -7.62702465e-01 3.94239992e-01 1.48688304e+00
-5.67275047e-01 -5.27897000e-01 -1.01447940e+00 -9.12828267e-01
1.55561892e-02 -9.08912420e-01 8.05957079e-01 5.80220759e-01
6.64378405e-01 2.21387759e-01 -3.04978341e-01 5.33010423e-01
6.31426215e-01 -4.80851352e-01 7.51572192e-01 -1.12281263e+00
-4.50465173e-01 -2.67251045e-01 -2.46538937e-01 -9.24573898e-01
2.89726913e-01 -6.87615454e-01 7.89257705e-01 -1.69456112e+00
4.64123823e-02 -7.42830873e-01 6.14077635e-02 8.15863132e-01
2.77780801e-01 -5.56142986e-01 5.45753896e-01 4.68700528e-01
-9.34926689e-01 5.57492554e-01 7.87888944e-01 -5.98339625e-02
-2.37588465e-01 -2.32068077e-01 -9.28632796e-01 1.14877498e+00
1.00479591e+00 -4.56009537e-01 -5.41055262e-01 -6.73131883e-01
3.73671204e-01 5.22699237e-01 3.18498790e-01 -1.31697190e+00
5.04372537e-01 -7.00203478e-01 1.40398636e-01 -1.93191126e-01
7.54982293e-01 -7.59498894e-01 1.90772340e-01 2.86476225e-01
-8.25199187e-01 1.77723333e-01 4.03982192e-01 7.65024245e-01
3.35926801e-01 -4.11761284e-01 4.73542958e-01 -4.55070019e-01
-1.02822912e+00 -9.82221663e-02 -5.01227677e-01 3.46143283e-02
9.49246585e-01 7.94737637e-02 -4.84844983e-01 -7.22553492e-01
-6.72542632e-01 7.38241732e-01 8.27549398e-01 5.73308110e-01
5.62572360e-01 -1.09185195e+00 -4.39723283e-01 -3.31597120e-01
3.10794175e-01 3.03610921e-01 -5.33377863e-02 5.97174525e-01
-1.96065798e-01 7.34097123e-01 -2.92540550e-01 -5.10518670e-01
-5.49313903e-01 5.37620783e-01 2.87472099e-01 -8.10982734e-02
-6.69302166e-01 8.87851894e-01 1.17931537e-01 -1.07347631e+00
4.00130242e-01 -5.86453021e-01 1.32521793e-01 -5.73013961e-01
4.12475795e-01 2.75616050e-01 -2.27136686e-01 -3.76840383e-01
-4.85750824e-01 -8.08068588e-02 3.46107215e-01 -7.25500703e-01
1.20326793e+00 -3.59745473e-01 6.16052859e-02 7.26904452e-01
1.97989106e-01 -3.01086247e-01 -1.74097729e+00 2.42290571e-02
4.63548064e-01 -3.57794762e-01 -2.58456826e-01 -1.31848812e+00
1.14688963e-01 6.66630745e-01 2.61925876e-01 -9.78012905e-02
4.18951303e-01 2.37207755e-01 3.72665852e-01 1.12342453e+00
1.36188912e+00 -1.32958603e+00 4.29335348e-02 7.81212687e-01
1.24981165e+00 -1.43026745e+00 -7.77314743e-03 -8.28718301e-04
-1.27491832e+00 9.15276587e-01 8.68308604e-01 1.00255467e-01
1.50134742e-01 3.65161330e-01 -1.01057351e-01 -4.47888859e-02
-1.22686183e+00 9.33129415e-02 -9.42336023e-02 1.09945774e+00
-1.08597763e-01 4.64844942e-01 3.38586211e-01 9.11249042e-01
-6.70665026e-01 -1.77514344e-01 7.31944144e-01 8.70868087e-01
-8.43364000e-01 -6.92959130e-01 -6.19260967e-01 2.83687323e-01
2.16687530e-01 -3.04985605e-02 -3.48898292e-01 7.75702000e-01
-2.50753760e-02 1.35090768e+00 -3.27296138e-01 -4.67664242e-01
2.11782590e-01 2.42610767e-01 5.26385725e-01 -8.71202230e-01
-1.79240927e-01 -3.27174455e-01 5.26289105e-01 -1.13341737e+00
1.81141227e-01 -8.98445547e-01 -1.71947849e+00 -9.74995270e-02
-4.97732870e-02 1.99498218e-02 7.08353579e-01 9.22604024e-01
4.66278136e-01 1.05839998e-01 6.13751039e-02 -1.48767924e+00
-1.25916576e+00 -1.14697707e+00 -3.45591277e-01 -7.28569168e-04
1.51904359e-01 -8.35809350e-01 -2.18596891e-01 -3.17256182e-01] | [4.3663010597229, 0.9578589200973511] |
e6aed30c-eb6f-4771-9863-30c5f957e747 | kidney-segmentation-in-neck-to-knee-body-mri | 2006.06996 | null | https://arxiv.org/abs/2006.06996v1 | https://arxiv.org/pdf/2006.06996v1.pdf | Kidney segmentation in neck-to-knee body MRI of 40,000 UK Biobank participants | The UK Biobank is collecting extensive data on health-related characteristics of over half a million volunteers. The biological samples of blood and urine can provide valuable insight on kidney function, with important links to cardiovascular and metabolic health. Further information on kidney anatomy could be obtained by medical imaging. In contrast to the brain, heart, liver, and pancreas, no dedicated Magnetic Resonance Imaging (MRI) is planned for the kidneys. An image-based assessment is nonetheless feasible in the neck-to-knee body MRI intended for abdominal body composition analysis, which also covers the kidneys. In this work, a pipeline for automated segmentation of parenchymal kidney volume in UK Biobank neck-to-knee body MRI is proposed. The underlying neural network reaches a relative error of 3.8%, with Dice score 0.956 in validation on 64 subjects, close to the 2.6% and Dice score 0.962 for repeated segmentation by one human operator. The released MRI of about 40,000 subjects can be processed within two days, yielding volume measurements of left and right kidney. Algorithmic quality ratings enabled the exclusion of outliers and potential failure cases. The resulting measurements can be studied and shared for large-scale investigation of associations and longitudinal changes in parenchymal kidney volume. | ['Joel Kullberg', 'Håkan Ahlström', 'Lowe Lundin', 'Andreas Wallin', 'Andreas Östling', 'Taro Langner', 'Robin Strand', 'Lukas Maldonis', 'Dag Lindgren', 'Albin Karlsson', 'Daniel Olmo'] | 2020-06-12 | null | null | null | null | ['kidney-function'] | ['medical'] | [ 1.86298974e-02 3.58359426e-01 -7.77188092e-02 -6.09012365e-01
-6.36274636e-01 -5.23641527e-01 2.20624983e-01 8.33761513e-01
-5.66153586e-01 6.19850159e-01 2.72705674e-01 -2.27287859e-01
-1.04663335e-01 -7.07413971e-01 -4.76840049e-01 -6.08960092e-01
-5.87482333e-01 8.97569060e-01 -1.89028364e-02 7.02290356e-01
-5.37806749e-02 6.21590495e-01 -5.61707795e-01 -5.38334288e-02
7.91303992e-01 8.02966714e-01 9.00387466e-02 9.43731129e-01
-2.12698311e-01 5.66984951e-01 -9.54036266e-02 -5.89483321e-01
3.81843746e-01 -6.12079442e-01 -4.50927734e-01 4.41361405e-03
8.99705231e-01 -6.54330432e-01 -7.74776042e-02 8.42711926e-01
8.97048235e-01 -2.32568339e-01 4.62610871e-01 -5.78786731e-01
-5.86982548e-01 9.84957814e-01 -4.35751140e-01 3.35019261e-01
8.13027087e-04 4.68157053e-01 2.15835243e-01 -6.11287236e-01
5.93846619e-01 7.75267005e-01 8.85915101e-01 3.42420697e-01
-1.23390782e+00 -6.24780536e-01 -4.62153137e-01 -3.05830449e-01
-9.69777405e-01 -5.32828748e-01 -5.74460439e-02 -7.62229383e-01
5.05086362e-01 3.55290800e-01 1.01778913e+00 2.65036315e-01
4.70292747e-01 2.08299771e-01 1.32878649e+00 -1.00731224e-01
2.37933964e-01 -5.37040979e-02 2.49388199e-02 5.70227861e-01
5.83604872e-01 -1.85902715e-01 5.13075758e-03 -3.38260174e-01
1.00386477e+00 -2.32703295e-02 -3.12506482e-02 -4.76971686e-01
-1.51939952e+00 5.08553267e-01 3.99069875e-01 1.26449168e-01
-6.42056465e-01 1.91101924e-01 8.13603103e-01 1.93088502e-01
2.42949978e-01 6.86817020e-02 -4.84705031e-01 1.84996381e-01
-1.22887015e+00 9.49687734e-02 7.88515806e-01 6.00662827e-01
6.48954585e-02 -4.81365900e-03 -2.53803402e-01 8.22818696e-01
6.08469188e-01 7.82901645e-01 5.45851409e-01 -1.27270699e+00
2.37867519e-01 4.33911175e-01 -2.28839442e-01 -4.43490803e-01
-8.00568044e-01 -4.09940392e-01 -1.23613393e+00 2.92972550e-02
9.10473824e-01 -1.32242188e-01 -1.22815883e+00 1.31673479e+00
7.30013788e-01 -2.48694718e-01 -3.14292908e-01 1.23088706e+00
8.99635851e-01 -2.97885805e-01 5.51692784e-01 -2.08709955e-01
1.81007564e+00 -4.40469056e-01 -2.88519263e-01 -3.89264487e-02
3.10573399e-01 -5.84140539e-01 4.61828440e-01 9.58205760e-02
-1.38731754e+00 -1.86855257e-01 -5.97332835e-01 1.02621451e-01
-4.42803726e-02 -6.47577271e-02 6.49848402e-01 1.07751703e+00
-8.41600776e-01 7.50559330e-01 -1.10314989e+00 -4.93669242e-01
8.19957316e-01 2.28116050e-01 -5.95385373e-01 -1.70461372e-01
-9.33118105e-01 1.28707194e+00 5.44238463e-03 4.31136757e-01
-6.12132728e-01 -1.42888260e+00 -7.57684469e-01 -1.58161804e-01
-1.39689744e-01 -8.75276566e-01 1.00584149e+00 -2.37790152e-01
-1.37169480e+00 1.24334645e+00 1.63985521e-01 -6.45902991e-01
1.04017532e+00 2.17354745e-01 -2.10437745e-01 4.77980584e-01
3.56505275e-01 4.64449316e-01 1.79853514e-01 -5.86725295e-01
5.12178838e-02 -9.72683668e-01 -9.68597174e-01 2.26257928e-03
6.04845583e-01 3.34573597e-01 -2.16716781e-01 -4.59561020e-01
4.90320712e-01 -8.66486430e-01 -5.71652293e-01 4.28342015e-01
-2.85970643e-02 5.49600661e-01 -2.63887763e-01 -1.50827003e+00
7.14819908e-01 -1.56273770e+00 -5.01718998e-01 4.51695681e-01
5.75069547e-01 -4.74412888e-02 2.87544668e-01 -4.30433638e-02
-1.55043721e-01 1.70244262e-01 -5.23563027e-01 1.55584797e-01
-2.03915779e-02 -1.44333541e-01 5.70599198e-01 1.28990734e+00
-2.37738043e-01 1.25850916e+00 -7.89776742e-01 -4.96951550e-01
3.51683915e-01 3.29602271e-01 -7.09521919e-02 -1.24646515e-01
5.12900412e-01 6.96843505e-01 -1.42062634e-01 7.77586937e-01
8.31163049e-01 -6.82466328e-02 5.83434343e-01 -1.28018335e-02
-1.23002511e-02 1.33806840e-01 -8.36882412e-01 1.73013759e+00
-8.23658928e-02 2.28908807e-01 6.34379745e-01 -7.20680892e-01
8.99338424e-01 3.68578315e-01 7.15921938e-01 -1.11401939e+00
-5.39713725e-02 3.26205343e-01 6.12097681e-01 -4.76475209e-01
-3.03440034e-01 -8.13405573e-01 1.67633906e-01 3.68224502e-01
-1.80450976e-01 1.68818504e-01 2.28910223e-01 1.77063495e-01
1.12137759e+00 -2.97340266e-02 1.03624202e-01 -6.57307327e-01
4.86392647e-01 -5.95413297e-02 5.76280117e-01 5.74963748e-01
-8.21656823e-01 7.21136868e-01 4.90691096e-01 -5.03917933e-01
-1.50951505e+00 -1.38770783e+00 -8.88829947e-01 5.05217731e-01
-3.47799957e-01 8.34825486e-02 -6.31006181e-01 -2.56415874e-01
6.69873953e-01 1.93553180e-01 -6.29422426e-01 3.51327360e-01
-6.46580160e-01 -1.06243348e+00 8.70593190e-01 5.13617396e-01
1.77126646e-01 -8.57407153e-01 -7.48472691e-01 4.23258603e-01
1.08492993e-01 -6.58012092e-01 -2.63598859e-01 3.05943750e-02
-1.38248122e+00 -1.38895845e+00 -1.45804822e+00 -3.09686035e-01
5.30122042e-01 -5.48347414e-01 1.29199994e+00 -2.56889090e-02
-7.93156564e-01 3.02739710e-01 2.57080644e-01 -3.43162298e-01
-5.63023090e-01 -3.26238722e-01 1.08500376e-01 -5.49153864e-01
1.81113720e-01 -6.52155459e-01 -1.17781639e+00 9.98496488e-02
-4.53149050e-01 -2.99724102e-01 7.95453191e-01 3.27543199e-01
6.68857872e-01 -7.21565008e-01 8.18647027e-01 -9.65099692e-01
3.08920115e-01 -5.83876848e-01 -5.66995621e-01 2.65603304e-01
-8.64701748e-01 -2.90588379e-01 2.41842732e-01 -1.14110120e-01
-7.12621510e-01 -5.56473397e-02 1.43629506e-01 -1.27655998e-01
-3.07392746e-01 4.89684761e-01 3.61383766e-01 -6.02189600e-02
4.75291461e-01 -1.74602866e-02 5.71224153e-01 -5.89905739e-01
3.37477744e-01 4.02066588e-01 7.63389885e-01 -4.51254129e-01
1.18831806e-01 3.19518805e-01 3.93191487e-01 -8.25487137e-01
-9.39094350e-02 -5.70060372e-01 -8.84916008e-01 -1.90722272e-01
8.55433583e-01 -9.58819151e-01 -7.88992226e-01 2.46400759e-01
-3.35966498e-01 -5.33490777e-01 -3.92307311e-01 9.17870820e-01
-4.27234024e-01 7.18917012e-01 -1.04381478e+00 -6.91665530e-01
-9.77261722e-01 -9.18179989e-01 6.26702726e-01 6.71067908e-02
-3.33994597e-01 -9.41208541e-01 2.81074047e-01 5.63979030e-01
6.59273028e-01 6.03063166e-01 7.96918690e-01 -7.20232129e-01
-2.97372013e-01 -3.97139400e-01 -3.36140543e-01 1.54161766e-01
-2.55800873e-01 -2.38832995e-01 -3.49510461e-01 -7.94431791e-02
6.57430738e-02 -2.93542556e-02 7.38632798e-01 1.03279555e+00
5.44279575e-01 -5.46549782e-02 8.08683708e-02 4.24749732e-01
1.35015738e+00 3.47499214e-02 7.20054209e-01 2.07915694e-01
4.86359835e-01 5.94571829e-01 -9.18046236e-02 2.84119099e-01
4.69248861e-01 3.07433447e-03 1.61368623e-01 -7.42985234e-02
-2.49992892e-01 3.58069777e-01 -1.66550115e-01 6.13920450e-01
-7.53090754e-02 5.95924318e-01 -1.29055285e+00 7.34762251e-01
-1.23094928e+00 -8.29370320e-01 -9.10731375e-01 2.44786382e+00
7.72307515e-01 -1.10453293e-01 3.67912114e-01 -5.64334452e-01
8.21369946e-01 -2.61171073e-01 -6.67548001e-01 -4.32058543e-01
2.56788582e-01 2.27556422e-01 1.08343041e+00 3.18621725e-01
-8.54722977e-01 6.62203878e-02 7.04243994e+00 -3.03671420e-01
-7.42542148e-01 1.71679765e-01 8.18274975e-01 -2.29899943e-01
2.51227105e-03 -2.15312626e-04 -3.46555263e-01 6.22968435e-01
1.29842901e+00 -1.00598447e-01 2.44755268e-01 4.92572814e-01
4.62601423e-01 -6.80176735e-01 -7.31437981e-01 5.41412115e-01
-2.50482559e-01 -1.11334872e+00 -6.61551893e-01 3.31825137e-01
3.94139200e-01 6.61646903e-01 -4.12719667e-01 6.03934750e-02
3.94379087e-02 -1.15196919e+00 3.36500347e-01 1.04007912e+00
1.10470986e+00 -3.90787780e-01 7.93082118e-01 2.31421113e-01
-7.82642007e-01 3.79494607e-01 -4.12073344e-01 5.60525395e-02
2.15769932e-01 1.29572499e+00 -1.00042510e+00 3.31518948e-01
5.88735163e-01 1.41493857e-01 -6.26664877e-01 1.44049120e+00
1.85985282e-01 5.61202466e-01 -5.12186825e-01 5.14518499e-01
-6.18659146e-02 -6.14516437e-01 2.04950839e-01 1.26667607e+00
1.01163261e-01 1.52305454e-01 -7.81404786e-03 9.75165308e-01
-1.45004332e-01 4.28341657e-01 -1.56875908e-01 1.12160191e-01
1.33605346e-01 1.56454265e+00 -1.04943943e+00 -6.47984087e-01
-4.30724740e-01 6.22515559e-01 4.09753770e-02 1.23323884e-03
-5.19638181e-01 4.50617112e-02 2.39437431e-01 4.60115016e-01
-7.65258446e-02 8.46985504e-02 -4.67122436e-01 -1.01587737e+00
6.27163425e-02 -8.04324508e-01 4.84635651e-01 -2.94591188e-01
-1.53265536e+00 -5.69783412e-02 -2.42621690e-01 -3.99587750e-01
5.53776808e-02 -2.89989859e-01 -4.81406838e-01 1.64249134e+00
-1.07097471e+00 -8.42004955e-01 -2.08595216e-01 -1.42785981e-01
-6.83265626e-02 2.98777282e-01 8.16586375e-01 5.48202753e-01
-5.31544745e-01 3.28793675e-01 2.11452395e-01 6.11329496e-01
7.70391107e-01 -1.69427991e+00 3.27886939e-01 4.63855743e-01
-5.94447434e-01 9.00073528e-01 3.89295757e-01 -1.22464395e+00
-1.37798619e+00 -1.02040660e+00 9.38281775e-01 -4.87812072e-01
4.10289466e-01 8.80287066e-02 -8.52790415e-01 6.73767567e-01
-8.63419920e-02 5.59937239e-01 1.14183700e+00 -5.47118299e-02
-7.83869997e-02 5.39321229e-02 -1.56499064e+00 1.19196244e-01
5.55660129e-01 -2.15937972e-01 -4.99568254e-01 2.78921694e-01
-5.78197949e-02 -6.82225168e-01 -1.86947024e+00 3.65736365e-01
1.30097675e+00 -9.00877953e-01 1.11689901e+00 -6.79573178e-01
3.49152535e-01 -2.94542462e-01 9.03622359e-02 -6.19246006e-01
-3.20522279e-01 -9.89976451e-02 -1.46717787e-01 1.10651159e+00
2.23646060e-01 -4.55964386e-01 1.00338519e+00 1.33018005e+00
7.82694593e-02 -6.75681055e-01 -8.72674525e-01 -6.03571117e-01
5.39365172e-01 -2.37690002e-01 1.80947036e-01 1.01869333e+00
2.54449621e-02 -2.44652703e-01 1.40670255e-01 -4.46756668e-02
1.19270444e+00 1.31645948e-01 5.74882448e-01 -1.31717122e+00
-3.66259832e-03 -6.01108670e-01 -4.54264402e-01 -1.76470906e-01
-3.39523911e-01 -1.37856865e+00 -2.19001204e-01 -1.86340094e+00
6.18308365e-01 -5.01630485e-01 -2.22351193e-01 1.66108757e-01
-5.93073107e-02 4.23250437e-01 2.43500233e-01 1.49179801e-01
-1.09246410e-01 3.89267574e-03 1.33643711e+00 1.34723336e-01
-5.22735454e-02 -1.79987237e-01 -5.70658743e-01 4.68543023e-01
9.30393577e-01 -5.74738145e-01 1.24413200e-01 1.28480187e-02
3.26126888e-02 3.93769771e-01 5.09365201e-01 -8.07287872e-01
1.25617042e-01 -3.81210707e-02 1.21438646e+00 -6.65511489e-01
-3.89893562e-01 -6.90210938e-01 6.97556555e-01 1.05631018e+00
-3.90169740e-01 -1.27150789e-01 -2.82100774e-02 4.05728072e-01
3.10327739e-01 -2.24868804e-01 9.91248012e-01 -9.83131826e-01
6.06481135e-02 3.13546151e-01 -4.40499276e-01 7.42425099e-02
7.88473785e-01 -1.86806753e-01 -6.51247948e-02 -1.15503073e-01
-1.18236876e+00 6.59227252e-01 4.19572115e-01 -2.70153612e-01
1.15805589e-01 -1.17762756e+00 -1.20413303e+00 5.45183022e-04
-3.16574514e-01 1.21764377e-01 6.42237067e-01 1.73476565e+00
-1.08892536e+00 3.79195571e-01 -2.20508516e-01 -7.71375418e-01
-9.28824842e-01 3.60797346e-01 7.48460054e-01 -2.93724239e-01
-1.20829606e+00 4.43582684e-01 -9.41668302e-02 -6.71056747e-01
-1.30586088e-01 -3.73957604e-01 1.27204865e-01 1.59334779e-01
5.34143388e-01 6.83535457e-01 8.47267807e-02 -6.44308925e-01
-5.88727534e-01 2.01893687e-01 1.99684530e-01 4.71015610e-02
1.42061985e+00 -4.37727153e-01 -5.44819534e-01 3.84424776e-01
1.05244720e+00 -4.33464088e-02 -7.98577666e-01 -1.57758538e-02
1.27469763e-01 -2.49502152e-01 1.20267630e-01 -1.21440232e+00
-1.12842536e+00 6.73030674e-01 9.04164851e-01 8.82574320e-02
7.19866812e-01 -1.11052282e-01 7.46764243e-01 -2.98870981e-01
-1.61960702e-02 -8.93991053e-01 -9.97300506e-01 -8.06930847e-03
7.47789502e-01 -1.08276093e+00 5.30281007e-01 -5.96115254e-02
-5.92275023e-01 1.12989509e+00 1.04606420e-01 -1.73507690e-01
4.39392745e-01 2.89282262e-01 3.58284414e-01 -4.73258436e-01
-3.41032445e-01 2.12389186e-01 2.48119801e-01 5.27660489e-01
8.24853122e-01 3.87130260e-01 -7.52083123e-01 4.07297254e-01
-4.71529067e-02 6.69522360e-02 4.18135822e-01 7.37003922e-01
-3.96663964e-01 -7.68557429e-01 -5.07620335e-01 1.14911687e+00
-1.02511430e+00 -1.30286694e-01 -1.42581820e-01 7.13739574e-01
-6.67784512e-02 2.65491456e-01 -8.15060921e-03 7.60275662e-01
2.60800600e-01 6.57685637e-01 6.71522260e-01 -3.90032530e-01
-8.52136672e-01 3.25240791e-01 8.31494182e-02 -5.46780586e-01
-5.18144250e-01 -9.73034024e-01 -1.38983512e+00 -3.17329288e-01
-1.98758304e-01 -1.67533979e-01 9.60881054e-01 7.28247702e-01
3.59449871e-02 3.53786409e-01 1.11736722e-01 -6.02467835e-01
-4.65189844e-01 -1.04153609e+00 -1.02591395e+00 2.81436890e-01
1.64163098e-01 -5.89073449e-02 -1.65625647e-01 7.52801523e-02] | [14.131939888000488, -2.4211668968200684] |
2b6cced5-5e3a-46c0-b6d5-c95263ce952e | multi-layer-kernel-ridge-regression-for-one | 1805.07808 | null | http://arxiv.org/abs/1805.07808v2 | http://arxiv.org/pdf/1805.07808v2.pdf | Multi-layer Kernel Ridge Regression for One-class Classification | In this paper, a multi-layer architecture (in a hierarchical fashion) by
stacking various Kernel Ridge Regression (KRR) based Auto-Encoder for one-class
classification is proposed and is referred as MKOC. MKOC has many layers of
Auto-Encoders to project the input features into new feature space and the last
layer was regression based one class classifier. The Auto-Encoders use an
unsupervised approach of learning and the final layer uses semi-supervised
(trained by only positive samples) approach of learning. The proposed MKOC is
experimentally evaluated on 15 publicly available benchmark datasets.
Experimental results verify the effectiveness of the proposed approach over 11
existing state-of-the-art kernel-based one-class classifiers. Friedman test is
also performed to verify the statistical significance of the claim of the
superiority of the proposed one-class classifiers over the existing
state-of-the-art methods. | ['Chandan Gautam', 'Alexandros Iosifidis', 'Aruna Tiwari', 'Sundaram Suresh'] | 2018-05-20 | null | null | null | null | ['one-class-classifier'] | ['methodology'] | [-1.02381609e-01 -9.77681726e-02 -3.56308728e-01 -5.30251265e-01
-4.73578274e-01 1.85132176e-02 6.31324291e-01 2.38104284e-01
-4.99554873e-01 9.26366985e-01 4.64089550e-02 -2.99359828e-01
-2.91516721e-01 -7.47277021e-01 -7.82631874e-01 -6.58517063e-01
-3.35980445e-01 -3.12598012e-02 2.52271801e-01 -9.54225883e-02
3.23550373e-01 5.32846689e-01 -1.69681966e+00 6.65588796e-01
7.45746672e-01 1.16984499e+00 -1.68719873e-01 9.76951838e-01
8.11210573e-02 1.19616103e+00 -3.88919294e-01 -9.34482813e-02
2.01148286e-01 -7.99321011e-02 -7.24003911e-01 -2.17651919e-01
3.07241470e-01 -1.35151327e-01 -1.95018962e-01 5.42604506e-01
1.09307030e-02 1.39889374e-01 1.07618248e+00 -1.29545820e+00
-1.18614447e+00 3.50808680e-01 -4.20632720e-01 1.18387572e-01
4.19786349e-02 -4.10248309e-01 8.70225251e-01 -1.12121797e+00
1.04599044e-01 8.58289421e-01 8.27201188e-01 4.49731201e-03
-1.20093107e+00 -7.96483815e-01 -4.70457137e-01 3.69480461e-01
-1.56856406e+00 3.45883854e-02 6.22938216e-01 -8.00514698e-01
1.15959418e+00 -1.17480069e-01 2.39734098e-01 8.00378442e-01
6.09065831e-01 6.99406624e-01 1.73806632e+00 -5.75579405e-01
6.93686083e-02 7.78987229e-01 6.70569122e-01 9.37152088e-01
2.66467750e-01 4.00992125e-01 -5.74127853e-01 -1.99519664e-01
4.60742503e-01 1.24781743e-01 3.15680355e-01 -3.65914613e-01
-9.52542126e-01 1.09993947e+00 6.45017445e-01 4.46768075e-01
-4.37378943e-01 -6.60193637e-02 5.84051251e-01 5.16526103e-01
4.74220902e-01 1.38181642e-01 -6.70838177e-01 1.88101918e-01
-1.03912830e+00 -4.78123501e-02 5.73864698e-01 6.86600506e-01
8.26666296e-01 4.64417189e-01 -4.19820193e-03 7.38163650e-01
3.05756241e-01 4.92272824e-02 1.01144660e+00 1.10436954e-01
-1.34640690e-02 1.00977802e+00 -2.09896356e-01 -4.87860441e-01
-3.77623826e-01 -6.94802046e-01 -7.81208158e-01 5.80153465e-01
-2.26577446e-02 -3.08492810e-01 -1.04491723e+00 9.96068299e-01
1.00084051e-01 5.16017437e-01 7.60444105e-01 4.47124660e-01
7.75311828e-01 8.75860155e-01 7.42517188e-02 1.13281757e-01
1.10022795e+00 -1.04416978e+00 -4.06434327e-01 3.64101559e-01
5.56819022e-01 -8.06029797e-01 6.16640151e-01 5.35112441e-01
-3.19199890e-01 -1.09454834e+00 -1.58412790e+00 2.37818137e-01
-1.16097462e+00 8.41550946e-01 8.85603905e-01 6.58342779e-01
-7.96464980e-01 5.13574362e-01 -4.33521152e-01 -1.24439009e-01
3.77839178e-01 7.04025865e-01 -8.11516702e-01 2.69927949e-01
-1.05850816e+00 9.42460179e-01 9.20229256e-01 5.19612879e-02
-8.46670091e-01 -5.22465765e-01 -7.82987058e-01 -2.89359614e-02
-1.65789530e-01 8.39067996e-03 7.20935822e-01 -1.13844681e+00
-1.59018219e+00 5.03286839e-01 1.92608252e-01 -6.70487821e-01
1.59714133e-01 -5.43128014e-01 -6.99587822e-01 -1.76732153e-01
-1.26393795e-01 4.36560750e-01 1.06414175e+00 -1.05772614e+00
-9.16445851e-01 -3.41729134e-01 -3.16409707e-01 1.68819819e-02
-4.22466546e-01 -8.75915866e-03 4.49086636e-01 -6.30917788e-01
-2.82608569e-01 -8.70974183e-01 2.11303458e-01 -6.03091538e-01
-2.82911330e-01 -5.03155589e-01 1.17245686e+00 -5.16926646e-01
1.35950470e+00 -2.20622897e+00 -1.53723404e-01 3.11589181e-01
-6.53211400e-02 3.83175015e-01 2.26775527e-01 6.38194621e-01
-5.59035003e-01 -1.33498237e-01 -8.02742317e-02 -2.82104667e-02
-3.29266697e-01 -2.68015917e-02 -1.29090786e-01 4.49085474e-01
4.42583203e-01 5.85062087e-01 -3.39331895e-01 -4.88175541e-01
3.79207134e-01 7.20920444e-01 -3.57084274e-02 5.52376568e-01
3.90048712e-01 5.91095462e-02 -1.57497466e-01 6.84002459e-01
5.08872867e-01 4.33394925e-05 -3.62985104e-01 -3.10891062e-01
-2.56867707e-01 -4.01082277e-01 -1.14366055e+00 1.13796508e+00
-6.23954237e-01 7.58155525e-01 -7.72486985e-01 -1.18013906e+00
1.28865719e+00 3.98306668e-01 9.83204842e-02 4.84426506e-02
1.69248611e-01 3.21716189e-01 2.73173928e-01 -3.17513406e-01
4.33719367e-01 -7.40247667e-02 4.79156747e-02 1.48245841e-01
7.24549532e-01 6.45104825e-01 -6.48022890e-02 -2.04798236e-01
7.25332379e-01 2.99981773e-01 7.27883697e-01 -2.37314850e-01
1.07097781e+00 5.16332500e-02 2.49223635e-01 4.78776455e-01
-4.86094877e-02 7.67620802e-02 2.71098197e-01 -7.48653710e-01
-1.05095792e+00 -1.07228577e+00 -4.69975173e-01 1.19741690e+00
-6.18359089e-01 -4.13684919e-02 -6.23381197e-01 -7.35595524e-01
3.26298952e-01 7.19514966e-01 -1.19496477e+00 -1.32645264e-01
-6.12493791e-02 -1.05099297e+00 8.14045250e-01 6.91649616e-01
6.26655638e-01 -1.03385913e+00 -6.32490754e-01 2.70517498e-01
7.17596591e-01 -9.98876810e-01 2.85976946e-01 8.25386822e-01
-8.62138450e-01 -1.14573729e+00 -4.29846108e-01 -9.46093678e-01
6.71803653e-01 -1.73547834e-01 4.90170866e-01 -2.91497201e-01
-5.14335513e-01 2.58385744e-02 -7.34088123e-01 -6.46655321e-01
-3.59745055e-01 2.62341321e-01 1.51106060e-01 7.46403158e-01
5.81988275e-01 -3.11932504e-01 -2.89468080e-01 -3.25527489e-02
-7.83523619e-01 -1.87590167e-01 1.16106737e+00 1.08874786e+00
3.71478915e-01 3.25307637e-01 8.36528718e-01 -1.24829960e+00
8.51355195e-01 -7.14479744e-01 -7.41864860e-01 3.96312237e-01
-9.35766459e-01 4.13934648e-01 9.31731761e-01 -3.68394047e-01
-9.04182494e-01 3.64576638e-01 3.26633930e-01 -4.64695334e-01
-2.18299240e-01 3.55637997e-01 5.76767623e-01 -3.37545395e-01
6.65197968e-01 6.15325868e-01 -5.63247390e-02 -4.31959033e-01
2.73578405e-01 1.14612412e+00 4.93296415e-01 -3.92321467e-01
7.92967141e-01 1.89210922e-02 -6.49265721e-02 -8.91179860e-01
-5.76360762e-01 -6.19796038e-01 -1.11419344e+00 -1.47319198e-01
1.10780144e+00 -9.82477963e-01 -4.54959750e-01 6.31371915e-01
-7.97892451e-01 6.96159841e-04 1.02960110e-01 8.10302794e-01
-3.81069958e-01 -2.65784085e-01 -5.78461647e-01 -1.18055665e+00
-6.15814507e-01 -1.06820774e+00 6.21252060e-01 4.84100670e-01
5.45323938e-02 -9.25519407e-01 1.74264029e-01 3.04303259e-01
4.85976249e-01 2.79258519e-01 9.64374840e-01 -1.19788420e+00
1.55442934e-02 -5.87958217e-01 -2.49097750e-01 8.28897476e-01
2.09722146e-01 1.12446465e-01 -1.24262047e+00 -1.25530183e-01
-3.99399042e-01 -6.91921830e-01 8.40061605e-01 1.12249784e-01
1.06193030e+00 -1.28410146e-01 -6.99976161e-02 4.08644527e-01
2.10788274e+00 2.73473889e-01 3.91569436e-01 5.76146781e-01
5.09242177e-01 2.16954753e-01 5.37563920e-01 3.39603394e-01
1.85032383e-01 3.33420038e-01 1.08836144e-01 -3.69906753e-01
2.40521386e-01 -5.06085828e-02 4.32246923e-01 9.09386933e-01
-2.71210819e-01 2.41115987e-01 -6.80788636e-01 3.53206486e-01
-1.69652629e+00 -7.22971022e-01 -1.69274330e-01 2.10915232e+00
6.74333096e-01 2.89415717e-01 1.19007900e-01 5.47355711e-01
4.27128732e-01 -1.28547817e-01 -1.73356682e-01 -1.10404325e+00
5.62354252e-02 7.50230730e-01 8.15266490e-01 3.55708182e-01
-1.56280649e+00 9.70108032e-01 5.90747023e+00 6.65620863e-01
-1.26746356e+00 3.45736332e-02 6.27745211e-01 4.73392665e-01
7.54038155e-01 -7.21572265e-02 -9.40162301e-01 2.87042350e-01
1.28036010e+00 3.31332326e-01 2.56137878e-01 1.27512813e+00
-2.28735030e-01 -2.97340661e-01 -1.02720833e+00 7.46495903e-01
1.87257946e-01 -1.18473864e+00 -1.60659805e-01 -3.96955982e-02
7.14489818e-01 9.07582045e-02 -2.09375266e-02 7.01923847e-01
3.10358405e-01 -1.20665228e+00 3.37496698e-01 8.44883084e-01
6.68258667e-01 -1.13545036e+00 1.26604724e+00 2.32870623e-01
-1.30289423e+00 -2.96447575e-01 -4.79017049e-01 8.81099701e-03
-6.76853716e-01 2.51081258e-01 -1.07367873e+00 8.30336571e-01
6.75114393e-01 4.07178879e-01 -1.09879100e+00 7.59186745e-01
8.88707936e-02 9.07198846e-01 4.80954759e-02 -2.89821863e-01
4.82247949e-01 3.55042145e-02 -9.86161008e-02 1.52613854e+00
-3.09736701e-03 -1.52322590e-01 1.06109574e-01 3.81227195e-01
2.67454267e-01 4.54285413e-01 -7.66613841e-01 3.27272490e-02
1.87732294e-01 1.35216427e+00 -3.90206695e-01 -6.35712445e-01
-7.60742366e-01 1.00908422e+00 5.89032054e-01 2.24457636e-01
-7.00776398e-01 -9.50974464e-01 1.92397729e-01 -8.67610127e-02
4.60922718e-01 -7.38908648e-02 -1.23222560e-01 -8.34953785e-01
-1.97386429e-01 -5.61252713e-01 3.98194075e-01 -6.32684588e-01
-1.34175730e+00 8.45372081e-01 1.11204118e-01 -1.19109738e+00
-1.66398212e-01 -1.02115071e+00 -5.52003920e-01 7.29009032e-01
-1.47436428e+00 -1.37741303e+00 -1.21847481e-01 6.51699483e-01
4.74678367e-01 -1.00773573e+00 1.05269289e+00 -2.80709919e-02
-4.81150627e-01 7.35657573e-01 4.46320087e-01 4.56243604e-01
6.41601920e-01 -1.44394815e+00 -3.71641099e-01 5.46481431e-01
1.93681911e-01 5.94995201e-01 1.05102621e-01 -4.96051252e-01
-1.24594474e+00 -1.15907705e+00 7.34888911e-01 -1.75132617e-01
6.43015325e-01 -2.65560031e-01 -8.64573419e-01 6.64414644e-01
4.26873684e-01 4.62367386e-01 1.30664515e+00 6.94965245e-03
-6.09279215e-01 -5.13504267e-01 -9.98135209e-01 -8.89204219e-02
-2.24560857e-01 -4.25048083e-01 -8.48560154e-01 -1.02184236e-01
3.35361332e-01 2.09426507e-02 -1.33343673e+00 4.91541445e-01
1.00947416e+00 -1.04220855e+00 7.25699723e-01 -8.71225774e-01
5.14183998e-01 -4.28385913e-01 -3.76293868e-01 -1.21666372e+00
-3.52159292e-01 2.73721784e-01 -1.08671926e-01 9.85788107e-01
6.27059162e-01 -7.51733243e-01 6.01935685e-01 -1.35003790e-01
2.66320221e-02 -1.30001593e+00 -5.34731090e-01 -7.39599049e-01
4.77350578e-02 -1.44051194e-01 1.82215005e-01 1.15999579e+00
-6.04914688e-02 4.77886885e-01 -4.89215970e-01 1.30676463e-01
6.35962486e-01 -2.21344754e-01 6.88530028e-01 -1.43374801e+00
-2.62328178e-01 -5.04672974e-02 -1.05053937e+00 1.82231203e-01
1.04243018e-01 -9.36756611e-01 -3.15747470e-01 -1.11437547e+00
1.58050328e-01 -2.20407069e-01 -9.42587674e-01 6.91233575e-01
-8.83544385e-02 1.18806839e-01 -1.22561246e-01 8.43718648e-02
-2.01686591e-01 2.94405788e-01 5.14853477e-01 2.49661192e-01
-2.51531154e-01 -1.16077904e-03 -3.61203104e-01 4.92863595e-01
1.00406599e+00 -5.80384970e-01 -5.48302233e-01 3.19680333e-01
-2.96371639e-01 -1.72419012e-01 2.71959186e-01 -1.45589519e+00
1.69723064e-01 7.68635124e-02 1.18017983e+00 -6.15577936e-01
6.66766539e-02 -8.45357656e-01 -8.51772800e-02 5.83163202e-01
-5.52288353e-01 3.62858266e-01 1.85391068e-01 5.97490788e-01
-4.64306831e-01 -3.07237208e-01 1.01173365e+00 1.61849663e-01
-7.71579802e-01 -1.53046520e-02 -4.14190710e-01 -3.77017409e-01
1.41791916e+00 -3.24138939e-01 -1.59881897e-02 1.67908788e-01
-7.15981901e-01 -2.06459925e-01 -1.03690587e-01 4.78433847e-01
7.23425329e-01 -1.39941859e+00 -8.34748685e-01 4.18842554e-01
3.93569857e-01 -3.91221315e-01 -1.87887266e-01 5.96608281e-01
-7.18042374e-01 7.51548290e-01 -7.68483877e-01 -3.63808662e-01
-1.31920993e+00 3.84718418e-01 1.51906624e-01 -6.00920141e-01
-3.38317484e-01 5.32162845e-01 -1.90958396e-01 -4.99242872e-01
1.22348256e-01 -2.03908607e-01 -5.00982225e-01 -1.32661864e-01
2.62064725e-01 4.44358528e-01 2.46855021e-01 -6.64909422e-01
-2.72111773e-01 4.87915277e-01 -4.89684045e-01 -3.20193022e-02
1.49776936e+00 7.88650870e-01 6.35746047e-02 8.57315302e-01
1.56626844e+00 -3.39502990e-01 -8.68752778e-01 -1.09390534e-01
2.76521921e-01 5.81060648e-02 5.44148386e-02 -8.37576509e-01
-7.15523899e-01 8.62988710e-01 1.08456981e+00 -2.91806608e-01
1.00466526e+00 -6.28851056e-01 2.82446414e-01 6.36345208e-01
-6.26368746e-02 -1.36898935e+00 -6.57746419e-02 2.98641175e-01
5.82851291e-01 -1.37859893e+00 2.88829982e-01 -8.66485387e-02
-7.67468870e-01 1.63003755e+00 7.82391310e-01 -7.34702408e-01
1.16558909e+00 2.70345867e-01 -7.63717666e-02 4.89910170e-02
-1.00263345e+00 -7.81303793e-02 5.91766238e-01 4.34917510e-01
8.38492215e-01 3.47642124e-01 -4.08690393e-01 6.50508285e-01
-2.71349579e-01 1.01774566e-01 3.60942662e-01 9.49523151e-01
-3.59974802e-01 -1.04768956e+00 -3.64175588e-01 7.42621064e-01
-5.32732069e-01 -4.16877912e-03 -3.90667439e-01 1.17576945e+00
3.14108819e-01 6.46306872e-01 -1.74549967e-01 -8.71774077e-01
-1.11537859e-01 5.57196319e-01 2.02540442e-01 -3.43742192e-01
-9.75025833e-01 -2.68291086e-01 4.84961458e-03 -1.24189153e-01
-3.05964321e-01 -3.59041750e-01 -1.09111130e+00 3.37108642e-01
-4.11322504e-01 2.22360805e-01 9.88717437e-01 8.17017078e-01
5.98184317e-02 6.16954863e-01 9.14959252e-01 -2.92725414e-01
-5.53827941e-01 -1.25787854e+00 -4.63763148e-01 1.21974304e-01
3.56954068e-01 -9.07360435e-01 -2.51095802e-01 1.54505298e-01] | [8.234586715698242, 3.848489999771118] |
e4888a35-466c-4cfe-9e86-c0ec51548fa9 | tnpar-topological-neural-poisson-auto | 2306.14114 | null | https://arxiv.org/abs/2306.14114v1 | https://arxiv.org/pdf/2306.14114v1.pdf | TNPAR: Topological Neural Poisson Auto-Regressive Model for Learning Granger Causal Structure from Event Sequences | Learning Granger causality from event sequences is a challenging but essential task across various applications. Most existing methods rely on the assumption that event sequences are independent and identically distributed (i.i.d.). However, this i.i.d. assumption is often violated due to the inherent dependencies among the event sequences. Fortunately, in practice, we find these dependencies can be modeled by a topological network, suggesting a potential solution to the non-i.i.d. problem by introducing the prior topological network into Granger causal discovery. This observation prompts us to tackle two ensuing challenges: 1) how to model the event sequences while incorporating both the prior topological network and the latent Granger causal structure, and 2) how to learn the Granger causal structure. To this end, we devise a two-stage unified topological neural Poisson auto-regressive model. During the generation stage, we employ a variant of the neural Poisson process to model the event sequences, considering influences from both the topological network and the Granger causal structure. In the inference stage, we formulate an amortized inference algorithm to infer the latent Granger causal structure. We encapsulate these two stages within a unified likelihood function, providing an end-to-end framework for this task. | ['Zhifeng Hao', 'Keli Zhang', 'Zijian Li', 'Yuguang Yan', 'Jie Qiao', 'Wei Chen', 'Yuequn Liu', 'Ruichu Cai'] | 2023-06-25 | null | null | null | null | ['causal-discovery'] | ['knowledge-base'] | [ 1.32353351e-01 -2.13690996e-01 1.61066260e-02 -1.77042767e-01
-4.83189106e-01 -5.73956549e-01 9.85469520e-01 6.73913285e-02
-1.50565326e-01 8.76324236e-01 2.79878616e-01 -3.32068086e-01
-4.43197489e-01 -1.17495680e+00 -1.04248917e+00 -7.59753942e-01
-3.53955805e-01 7.18164861e-01 1.69795066e-01 2.36820400e-01
1.98038235e-01 2.96383888e-01 -1.16004348e+00 4.39728871e-02
7.21049964e-01 4.99416798e-01 2.24651545e-01 7.00575471e-01
-1.28322199e-01 1.06432021e+00 -3.20700049e-01 -2.51706708e-02
-1.28830433e-01 -3.58488977e-01 -5.30373275e-01 -3.87887567e-01
-1.14394255e-01 -2.79280752e-01 -4.53314245e-01 7.23784804e-01
8.98059756e-02 1.80248022e-01 1.25584483e+00 -1.25703251e+00
-4.20763284e-01 8.33204985e-01 -8.38132143e-01 5.15735686e-01
1.17858313e-01 -6.50734007e-02 1.16792428e+00 -8.84209454e-01
4.20144737e-01 1.43012989e+00 5.75130165e-01 -2.62076318e-01
-1.32099164e+00 -6.99607134e-01 3.34899813e-01 1.23399794e-01
-1.39052188e+00 4.45275800e-03 7.91162431e-01 -6.70580029e-01
6.74462318e-01 5.49935736e-03 5.41080534e-01 1.52780199e+00
6.63211703e-01 6.77685142e-01 8.16797674e-01 -1.00203849e-01
4.81230140e-01 -5.38839042e-01 1.17090285e-01 2.84411818e-01
7.79422671e-02 2.21315175e-01 -8.59072447e-01 -4.81504261e-01
1.29412377e+00 1.46033570e-01 1.49242535e-01 1.00807413e-01
-9.07154739e-01 8.63017857e-01 2.63785899e-01 1.64109901e-01
-6.06504738e-01 7.08341599e-01 2.71081656e-01 -7.28524178e-02
5.99980652e-01 -1.67919040e-01 -3.78143907e-01 -3.10092382e-02
-7.66166449e-01 5.68787217e-01 7.24470019e-01 7.49227464e-01
5.13657928e-01 -1.85064003e-01 -1.76291466e-01 5.25912762e-01
7.38948166e-01 4.24559325e-01 -1.97277188e-01 -3.27210397e-01
3.24612886e-01 1.79531559e-01 7.66959190e-02 -1.21339095e+00
-3.50257337e-01 -4.65702504e-01 -1.28903913e+00 -3.59526336e-01
5.11890471e-01 -8.18956047e-02 -8.38027120e-01 2.03445411e+00
4.32976305e-01 1.07145643e+00 -5.52930534e-01 5.46198547e-01
4.31098819e-01 9.63224649e-01 3.25036645e-01 -2.41804302e-01
1.49279857e+00 -2.55698562e-01 -7.71527767e-01 -3.71485716e-03
1.37173787e-01 -6.94090486e-01 6.24578655e-01 3.58902425e-01
-1.08928192e+00 -2.93592721e-01 -6.77052915e-01 5.98292761e-02
-1.05176158e-01 -1.91772491e-01 7.56139457e-01 5.28948456e-02
-1.02329838e+00 4.02350187e-01 -1.18836188e+00 -1.25534713e-01
1.72458783e-01 1.16271026e-01 8.35537612e-02 -8.32912512e-04
-1.51653397e+00 5.31265259e-01 5.22954464e-01 2.52973050e-01
-1.32189441e+00 -9.13368762e-01 -5.05146563e-01 1.42026380e-01
4.78142083e-01 -1.04313421e+00 1.12706482e+00 -3.38163495e-01
-9.75447059e-01 3.26671541e-01 -4.87773925e-01 -3.54217172e-01
4.51821893e-01 -2.04639390e-01 -1.58717275e-01 -1.25602752e-01
2.90065855e-01 2.72380896e-02 9.05135512e-01 -1.21919835e+00
-6.38846695e-01 -1.00248814e-01 -9.58734676e-02 9.28661227e-03
3.08925640e-02 -2.68464163e-02 -4.62290972e-01 -1.08562732e+00
1.83016419e-01 -7.90505171e-01 -2.07076639e-01 -3.28616351e-01
-7.61102259e-01 -4.59263951e-01 4.41693962e-01 -5.94930232e-01
1.31359684e+00 -1.91475761e+00 2.94051796e-01 4.14909959e-01
4.73453403e-01 -4.65890527e-01 6.78463578e-02 6.20286703e-01
-3.05780619e-01 4.59180810e-02 -4.07604069e-01 -4.43678200e-01
-1.43253744e-01 2.16235936e-01 -6.88033462e-01 4.45235431e-01
3.63058746e-01 8.97386611e-01 -1.20634854e+00 -4.03321266e-01
2.99293697e-02 6.21477664e-01 -3.30804557e-01 2.93141276e-01
-3.69410038e-01 8.21349144e-01 -6.26841128e-01 2.93844223e-01
6.18422449e-01 -5.12990534e-01 1.53026029e-01 -4.96097468e-02
-2.73794174e-01 5.03026366e-01 -1.48749697e+00 1.27160072e+00
-3.61539423e-01 4.09851968e-01 -1.58312991e-01 -1.14353454e+00
7.80595720e-01 5.73699176e-01 6.02432013e-01 -4.02186722e-01
6.49932353e-03 1.47014976e-01 -1.46665737e-01 -3.09673399e-01
2.96839625e-01 -5.58305740e-01 -1.66549355e-01 6.84885144e-01
4.57264222e-02 3.66094619e-01 7.29543865e-02 4.75736827e-01
1.49730682e+00 9.02221054e-02 3.83696377e-01 -1.61680758e-01
-6.68498054e-02 -3.01981211e-01 7.45133638e-01 1.05616450e+00
3.37985486e-01 5.15034795e-01 8.63340139e-01 -5.49003780e-01
-9.94256914e-01 -1.76218867e+00 -1.72248095e-01 6.61662877e-01
-1.88426182e-01 -2.56847918e-01 -3.67276192e-01 -3.35402548e-01
-5.81360236e-02 7.20360339e-01 -9.16848779e-01 -5.99399619e-02
-7.76691973e-01 -1.13549256e+00 6.31175816e-01 5.97658098e-01
1.27534106e-01 -9.03999627e-01 -2.56955057e-01 5.02386332e-01
-3.66541654e-01 -8.80401552e-01 -1.90624818e-01 2.84393340e-01
-7.74402022e-01 -9.64388251e-01 -3.94039899e-01 -3.41770321e-01
4.22121197e-01 -7.71444440e-02 1.35349369e+00 -2.10648865e-01
-2.19720706e-01 2.61968613e-01 -1.36959344e-01 -3.26897144e-01
-1.96870044e-01 -8.51693824e-02 2.73834243e-02 3.45087610e-02
1.82620719e-01 -1.04237378e+00 -6.54699624e-01 1.47775725e-01
-1.11514127e+00 2.00291798e-01 6.98892236e-01 5.26041329e-01
6.57320440e-01 6.15451396e-01 7.73674428e-01 -6.13924921e-01
3.94169956e-01 -1.14861190e+00 -6.81009591e-01 1.05237484e-01
-4.51360941e-02 1.36802822e-01 6.04680538e-01 -5.15881419e-01
-1.19311607e+00 -1.00052118e-01 1.12001449e-01 -3.65295738e-01
-2.82295644e-01 9.09796655e-01 -2.00147361e-01 7.01000571e-01
1.69738308e-01 1.07699491e-01 -3.68638456e-01 -2.84063429e-01
4.13249463e-01 4.95216884e-02 5.35442710e-01 -8.33781123e-01
8.59080434e-01 7.20604181e-01 3.04672182e-01 -7.02296793e-01
-7.19841242e-01 -4.65670615e-01 -6.40465558e-01 -3.10646832e-01
9.46139574e-01 -9.06928957e-01 -7.81612277e-01 5.12525439e-01
-1.74675632e+00 -4.00896281e-01 -6.56623319e-02 4.91606534e-01
-5.41544259e-01 1.31866187e-01 -6.94661140e-01 -1.02652538e+00
7.88506791e-02 -8.57212782e-01 1.03403890e+00 -1.08908363e-01
-5.49462616e-01 -1.56190717e+00 4.31455851e-01 -1.78179383e-01
-4.89240997e-02 4.18209642e-01 1.03472650e+00 -4.46578413e-01
-1.02610302e+00 -1.17607355e-01 -4.27513599e-01 -3.46011668e-01
2.85839766e-01 3.10287662e-02 -7.07503378e-01 2.36253127e-01
2.02771589e-01 3.64577770e-01 9.39376593e-01 8.44313383e-01
9.06886339e-01 -2.65019685e-01 -5.37444115e-01 3.61666709e-01
1.25464487e+00 1.30228639e-01 5.53740561e-01 -1.88220739e-02
1.05313683e+00 7.58942127e-01 2.90605068e-01 7.21498787e-01
6.70734584e-01 4.50051963e-01 4.29780394e-01 1.42461702e-01
3.24607849e-01 -6.59555674e-01 2.10174203e-01 8.59650016e-01
3.11615746e-02 -4.37108129e-01 -1.31824529e+00 6.81108236e-01
-2.14377236e+00 -1.08841491e+00 -8.01636875e-01 2.10598612e+00
1.04359162e+00 1.40187636e-01 3.40905786e-02 -1.61044940e-01
7.20444083e-01 1.05689086e-01 -4.54894394e-01 8.95380229e-02
-1.20801054e-01 7.23952502e-02 2.87594348e-01 5.72535098e-01
-1.10851610e+00 7.13252544e-01 6.08593273e+00 9.01348054e-01
-7.58418262e-01 1.06041685e-01 6.34922326e-01 1.23105627e-02
-6.04204834e-01 3.12441736e-01 -8.17981482e-01 6.84892893e-01
9.66940880e-01 6.80240663e-03 2.90943176e-01 1.47860125e-01
7.57003725e-01 -2.51347095e-01 -1.27501464e+00 6.52513504e-01
-3.78733635e-01 -1.22805274e+00 2.02385679e-01 2.04012126e-01
5.88112891e-01 -3.23496610e-01 -1.12859793e-02 -2.11460888e-02
7.76827574e-01 -1.17008746e+00 6.51187479e-01 9.82245803e-01
5.52128255e-01 -7.27491140e-01 3.22707921e-01 5.05327940e-01
-1.50801480e+00 2.77892262e-01 -2.26650201e-02 -2.34153807e-01
8.62076640e-01 1.46295428e+00 -7.10066140e-01 7.64706910e-01
6.93438709e-01 9.50230718e-01 2.02894304e-02 9.36227918e-01
-3.66128355e-01 9.54217732e-01 -5.82814872e-01 1.87165499e-01
1.53321877e-01 -2.81947613e-01 7.65541077e-01 1.02178907e+00
3.90758008e-01 4.08437476e-02 -4.44546193e-02 1.42090786e+00
1.15717612e-01 -3.38911533e-01 -6.44958258e-01 2.03187438e-03
4.64054078e-01 9.00687993e-01 -1.10466707e+00 -1.65819332e-01
-2.99567938e-01 7.20181227e-01 2.92153597e-01 5.67010105e-01
-1.12065792e+00 1.91035181e-01 3.75298619e-01 2.10108608e-01
2.16957822e-01 -4.39540833e-01 -7.59054840e-01 -9.68818605e-01
-6.98042139e-02 -4.05950129e-01 4.70338225e-01 -5.43855965e-01
-1.66170144e+00 8.16000029e-02 4.14657563e-01 -7.95860827e-01
-1.82179153e-01 -3.56237262e-01 -9.62207556e-01 1.08671308e+00
-1.31680512e+00 -1.15374303e+00 5.88299260e-02 4.06042814e-01
2.69781351e-01 4.10422951e-01 3.49691331e-01 3.65864813e-01
-5.83109975e-01 3.45348530e-02 -1.59345253e-03 -4.52102013e-02
6.22410119e-01 -1.35229135e+00 4.65040565e-01 8.16356659e-01
-4.19723056e-02 9.09598351e-01 8.00633490e-01 -1.29998159e+00
-1.20645046e+00 -1.22062647e+00 9.40903842e-01 -7.27958798e-01
1.09626162e+00 -5.30509591e-01 -1.16162086e+00 9.61310148e-01
-1.42679498e-01 -2.97140360e-01 5.82992077e-01 3.85344625e-01
-1.89108849e-01 3.12912196e-01 -5.22089422e-01 7.63226151e-01
9.48531985e-01 -3.42615932e-01 -4.58206862e-01 2.37825677e-01
6.13772035e-01 -2.35833172e-02 -6.65017664e-01 3.96609366e-01
3.13861102e-01 -6.66358769e-01 1.19374764e+00 -3.63104671e-01
8.33401442e-01 -5.12043774e-01 1.09896325e-01 -1.11455262e+00
-4.96201515e-01 -5.83057940e-01 -3.99366766e-01 1.49082983e+00
2.22736150e-01 -6.36827707e-01 4.22522187e-01 3.15747052e-01
2.35896945e-01 -5.92384636e-01 -1.04644406e+00 -5.59152007e-01
2.18883574e-01 -8.82854342e-01 3.94373447e-01 9.50788498e-01
-4.41778243e-01 6.66436970e-01 -6.21585786e-01 4.98625278e-01
1.00155926e+00 -2.10208640e-01 4.34851915e-01 -1.48924482e+00
-4.35207188e-01 -3.41878593e-01 7.97966644e-02 -1.12737691e+00
1.92322731e-01 -6.93130195e-01 3.31629664e-01 -1.43727124e+00
5.38611114e-01 -6.39954686e-01 -1.89025551e-01 1.63294196e-01
-6.20554864e-01 -2.95110315e-01 -4.06022847e-01 3.51733714e-01
-3.20383996e-01 6.49282634e-01 8.63363981e-01 2.14069381e-01
-1.65737629e-01 1.59927867e-02 -3.07134062e-01 8.69914114e-01
8.28915954e-01 -6.82787538e-01 -5.10123432e-01 -4.00748730e-01
9.61219788e-01 3.55361015e-01 9.48465586e-01 -5.22297084e-01
5.15508115e-01 -3.69332016e-01 1.36942625e-01 -1.00447595e+00
1.35887235e-01 -5.24215460e-01 4.07919377e-01 7.88354501e-02
-8.87646526e-02 1.62263319e-01 1.71117947e-01 1.10981500e+00
-1.63200080e-01 6.78461939e-02 3.41100007e-01 -1.17516086e-01
-2.00130776e-01 4.52941149e-01 -6.97913110e-01 9.58120450e-02
7.79965758e-01 2.06225425e-01 -4.62192781e-02 -3.47699404e-01
-6.37784481e-01 2.84968048e-01 -7.47005567e-02 2.22971097e-01
5.61764836e-01 -1.36177969e+00 -8.42404187e-01 -3.70472930e-02
-3.03982347e-01 5.03395915e-01 4.83750492e-01 1.15550327e+00
-1.72518909e-01 3.21191996e-01 3.28142196e-01 -6.94526732e-01
-7.49969721e-01 5.24546802e-01 -1.04266033e-03 -7.44259238e-01
-5.27553916e-01 8.48958194e-01 8.12626362e-01 -3.07209790e-01
-2.69483197e-02 -3.16767246e-01 -8.71753618e-02 2.96165824e-01
4.57797766e-01 5.62027872e-01 -3.75978917e-01 -2.86465317e-01
-1.36724979e-01 3.08053613e-01 -2.77257897e-02 -4.94581133e-01
1.33091652e+00 -1.21471718e-01 -5.07375896e-01 1.10498500e+00
8.98217499e-01 -2.27688372e-01 -1.19620514e+00 -2.46808842e-01
1.29072651e-01 -8.90630111e-03 -1.48985133e-01 -4.20079052e-01
-6.13118470e-01 8.39600325e-01 -5.75435720e-02 4.42904502e-01
8.94205391e-01 1.80846438e-01 7.03468859e-01 -3.24141949e-01
2.74189174e-01 -7.49590814e-01 2.07209200e-01 8.53556216e-01
7.84484625e-01 -7.49447525e-01 -1.46161035e-01 -6.56604886e-01
9.49721318e-03 9.63199615e-01 1.95723221e-01 -3.50787163e-01
1.02561677e+00 5.03142238e-01 -3.38487506e-01 -4.09235865e-01
-9.52424407e-01 1.56646535e-01 3.25562984e-01 1.97150663e-01
5.11306047e-01 1.66237831e-01 -2.61185348e-01 5.18726230e-01
-1.25185013e-01 2.30313409e-02 2.60996819e-01 6.95240617e-01
-9.37852710e-02 -9.49285328e-01 -6.57750845e-01 4.13928390e-01
-4.62206036e-01 -2.76889116e-01 -1.50740966e-01 4.45686758e-01
5.62613597e-03 8.20238352e-01 4.19808656e-01 6.97551519e-02
8.18842202e-02 -8.54993984e-03 1.66806921e-01 -5.47965884e-01
-1.96535215e-01 3.60868514e-01 -2.72065759e-01 -3.01500827e-01
-1.22721002e-01 -8.80044580e-01 -1.19592810e+00 -4.39906657e-01
-7.38206729e-02 -1.23272285e-01 3.39618236e-01 1.18397772e+00
1.64973617e-01 1.02038801e+00 3.58601838e-01 -8.05820346e-01
-1.57157540e-01 -9.27290797e-01 -5.08784235e-01 2.28073582e-01
3.62866431e-01 -9.56929982e-01 -3.89869809e-01 1.71146095e-01] | [7.002326011657715, 3.634523391723633] |
b1730035-9be8-468a-b9c1-15e4ac399bde | digeo-discriminative-geometry-aware-learning | 2303.09674 | null | https://arxiv.org/abs/2303.09674v1 | https://arxiv.org/pdf/2303.09674v1.pdf | DiGeo: Discriminative Geometry-Aware Learning for Generalized Few-Shot Object Detection | Generalized few-shot object detection aims to achieve precise detection on both base classes with abundant annotations and novel classes with limited training data. Existing approaches enhance few-shot generalization with the sacrifice of base-class performance, or maintain high precision in base-class detection with limited improvement in novel-class adaptation. In this paper, we point out the reason is insufficient Discriminative feature learning for all of the classes. As such, we propose a new training framework, DiGeo, to learn Geometry-aware features of inter-class separation and intra-class compactness. To guide the separation of feature clusters, we derive an offline simplex equiangular tight frame (ETF) classifier whose weights serve as class centers and are maximally and equally separated. To tighten the cluster for each class, we include adaptive class-specific margins into the classification loss and encourage the features close to the class centers. Experimental studies on two few-shot benchmark datasets (VOC, COCO) and one long-tail dataset (LVIS) demonstrate that, with a single model, our method can effectively improve generalization on novel classes without hurting the detection of base classes. | ['Shih-Fu Chang', 'Guangxing Han', 'Shiyuan Huang', 'Jincheng Xu', 'Yulei Niu', 'Jiawei Ma'] | 2023-03-16 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Ma_DiGeo_Discriminative_Geometry-Aware_Learning_for_Generalized_Few-Shot_Object_Detection_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Ma_DiGeo_Discriminative_Geometry-Aware_Learning_for_Generalized_Few-Shot_Object_Detection_CVPR_2023_paper.pdf | cvpr-2023-1 | ['few-shot-object-detection'] | ['computer-vision'] | [ 1.80333138e-01 -1.15185589e-01 -4.40702766e-01 -3.96508276e-01
-9.84201074e-01 -4.34452891e-01 3.16791564e-01 2.37733468e-01
-3.75234634e-01 5.96320570e-01 -2.03536168e-01 2.09458858e-01
-2.97579944e-01 -6.95920944e-01 -6.43953383e-01 -8.77201259e-01
1.18068524e-01 2.51076192e-01 8.09551001e-01 -2.89281029e-02
2.15974778e-01 4.61626768e-01 -1.86942089e+00 4.43628639e-01
1.03309500e+00 1.26264143e+00 2.16426432e-01 5.92087865e-01
9.61081088e-02 5.69009662e-01 -4.87931639e-01 -2.56199569e-01
3.96841168e-01 -3.17171037e-01 -2.48034731e-01 2.95334280e-01
5.82319021e-01 -3.06076348e-01 -2.64233202e-01 1.07191539e+00
6.03187621e-01 5.01206100e-01 9.74273920e-01 -1.20282054e+00
-6.30087733e-01 3.18275660e-01 -8.05936694e-01 5.57469964e-01
-6.63301870e-02 2.75622278e-01 1.02938557e+00 -1.35084307e+00
5.10799527e-01 8.95597637e-01 8.14996004e-01 6.85929775e-01
-1.03310382e+00 -6.89302266e-01 4.18534845e-01 3.73925567e-01
-1.76767468e+00 -5.16380966e-01 6.98042512e-01 -4.03180152e-01
6.52790189e-01 4.60208386e-01 5.58420777e-01 7.95681059e-01
-1.49194330e-01 8.86113107e-01 6.32624805e-01 -3.93788636e-01
4.44633871e-01 2.97529310e-01 3.74164015e-01 7.00548410e-01
3.08714598e-01 3.79297249e-02 -4.43940789e-01 -1.54194105e-02
4.25091982e-01 2.99639702e-01 -2.53889978e-01 -8.07073772e-01
-6.91985965e-01 1.00563693e+00 5.89675546e-01 1.07162148e-01
1.82986036e-01 -1.18509114e-01 4.24625188e-01 -3.39339599e-02
6.31235719e-01 4.65744846e-02 -3.19461644e-01 -3.11669111e-02
-8.39834571e-01 2.39216443e-02 4.79075938e-01 1.31918108e+00
9.08422887e-01 -1.85246959e-01 -5.32638252e-01 9.95618343e-01
-1.87133163e-01 2.42868140e-01 6.48227155e-01 -5.63904583e-01
3.26857477e-01 7.61073828e-01 -2.85263453e-02 -8.13660085e-01
-3.26551199e-01 -8.32360566e-01 -4.98600543e-01 2.53639258e-02
3.97235096e-01 -2.92728785e-02 -1.04324687e+00 1.44010913e+00
6.85862064e-01 3.86167973e-01 -9.48275030e-02 7.93741405e-01
8.21865499e-01 4.67106372e-01 -1.55699700e-01 -3.76248717e-01
1.15469408e+00 -1.00770843e+00 -2.91044205e-01 -2.69345284e-01
1.05682850e+00 -5.27686834e-01 1.24238932e+00 6.52559847e-02
-6.35166824e-01 -5.44457972e-01 -1.20145178e+00 8.00485760e-02
-5.07212937e-01 1.32913798e-01 5.28986454e-01 7.67417908e-01
-1.65155679e-01 5.16183317e-01 -7.66147852e-01 -2.69908816e-01
9.78385746e-01 1.63696185e-01 4.91608456e-02 -2.09509626e-01
-7.40239739e-01 5.21546721e-01 6.86945140e-01 -4.03756797e-01
-8.37880373e-01 -1.02145803e+00 -8.03127408e-01 1.16714887e-01
7.29241073e-01 -3.95016253e-01 9.60920274e-01 -9.55236673e-01
-1.09508812e+00 8.43542337e-01 1.46839797e-01 -4.50907707e-01
5.93841791e-01 -5.27557172e-02 -2.69652605e-01 1.48108795e-01
1.86419070e-01 7.05034018e-01 9.23556089e-01 -1.14159179e+00
-1.09671378e+00 -4.84887153e-01 -4.63289395e-02 1.83666512e-01
-7.74432600e-01 -2.27860510e-01 -4.12817776e-01 -7.29387403e-01
7.87963569e-02 -7.62233615e-01 -3.82618443e-03 3.63974988e-01
-1.56457677e-01 -4.35483247e-01 1.21724987e+00 -2.82467226e-03
1.36548519e+00 -2.50863600e+00 -1.21914618e-01 -5.67415953e-02
2.62158334e-01 3.41328949e-01 -3.98704968e-02 -2.73937464e-01
2.14256093e-01 -3.41283143e-01 -2.05474973e-01 -1.49849936e-01
-1.18495308e-01 3.29126626e-01 -2.86141098e-01 6.78784847e-01
2.49120414e-01 8.37313294e-01 -9.34694231e-01 -6.43484592e-01
3.24288428e-01 2.42127329e-01 -6.81139231e-01 9.72859412e-02
-8.73014480e-02 -2.12467685e-02 -4.65855956e-01 8.67884517e-01
7.37116754e-01 -2.73658693e-01 -3.17376822e-01 -1.53570563e-01
1.09037235e-01 -4.97990400e-01 -1.38821805e+00 1.53059411e+00
-2.62228012e-01 3.41192365e-01 -1.65124536e-01 -1.08805633e+00
9.68463242e-01 -2.37809077e-01 4.04854327e-01 -4.36762214e-01
2.91250557e-01 2.39579558e-01 -2.27861293e-02 -4.02262181e-01
2.51117736e-01 -2.06325978e-01 8.79810899e-02 -3.13602127e-02
3.18431288e-01 -2.43524723e-02 3.47149909e-01 1.40820473e-01
7.32902110e-01 -8.35601538e-02 5.40550888e-01 -3.18485916e-01
9.60708186e-02 -2.72838175e-01 8.36070657e-01 8.50990355e-01
-6.66414320e-01 9.42485571e-01 1.86446682e-01 -3.31756383e-01
-9.02418792e-01 -1.00422585e+00 -4.85928118e-01 1.54041111e+00
5.51856995e-01 -4.14890289e-01 -6.75668359e-01 -9.86477435e-01
1.17209464e-01 8.81569088e-01 -8.92214656e-01 -6.29584610e-01
-3.39153796e-01 -9.79703963e-01 2.11611420e-01 8.46809924e-01
9.47878510e-02 -4.95781809e-01 -8.37880015e-01 7.80972745e-03
1.26191854e-01 -1.05937350e+00 -8.09194088e-01 4.59832937e-01
-6.09932959e-01 -1.21583688e+00 -9.46424484e-01 -9.01111364e-01
6.54317498e-01 6.75221980e-01 6.79339349e-01 4.66715395e-02
-8.51536691e-01 2.25105643e-01 -5.94828844e-01 -5.07485271e-01
9.29255188e-02 8.50039199e-02 1.67572230e-01 2.38512427e-01
4.81440634e-01 -3.95536363e-01 -6.26592636e-01 7.01949954e-01
-6.27797186e-01 -2.28076249e-01 2.57369399e-01 9.91637647e-01
5.86222351e-01 -1.09534606e-01 7.38771439e-01 -6.25025153e-01
-5.96507341e-02 -5.42183697e-01 -4.02435958e-01 3.52592349e-01
-6.31974339e-01 -4.00731981e-01 5.17627954e-01 -7.56449461e-01
-9.25066888e-01 1.41876861e-01 3.09563398e-01 -7.99731553e-01
-2.46719718e-02 8.19383003e-03 -3.18433374e-01 -2.79133499e-01
1.15818167e+00 2.12690949e-01 -3.28870267e-01 -2.79124767e-01
4.40293610e-01 6.66636348e-01 4.36218917e-01 -6.04961276e-01
7.12539613e-01 7.58261204e-01 -2.13737652e-01 -9.20444548e-01
-1.39540517e+00 -1.01828647e+00 -9.65042770e-01 -2.25322723e-01
5.78192234e-01 -1.00968707e+00 -1.44138128e-01 1.67933553e-01
-6.67538583e-01 -1.67459846e-01 -8.27746630e-01 3.93355995e-01
-6.48612976e-01 3.46872628e-01 -1.72357991e-01 -9.10681129e-01
-2.50343204e-01 -8.55688155e-01 1.16169477e+00 5.33812881e-01
1.86502934e-02 -7.13173687e-01 -2.75769442e-01 1.61616221e-01
-4.57336754e-02 1.39875636e-01 5.37167192e-01 -7.60852993e-01
-3.90512079e-01 -5.01743674e-01 -4.59316343e-01 3.15066904e-01
9.53358933e-02 5.63477241e-02 -1.10780156e+00 -5.89818120e-01
-1.99423730e-01 -5.71258664e-01 1.25590050e+00 4.21445817e-01
1.43141651e+00 -1.18936133e-02 -5.86957216e-01 1.04488206e+00
1.32369947e+00 1.74618542e-01 3.33608747e-01 9.97961834e-02
6.81519210e-01 2.69718528e-01 9.68748987e-01 8.33839834e-01
-1.50359396e-04 7.62548923e-01 3.14054191e-01 2.09110886e-01
-3.29958975e-01 -1.34085327e-01 1.99216112e-01 4.01760936e-01
6.77180439e-02 -8.75448361e-02 -6.13753080e-01 6.42293572e-01
-1.96856546e+00 -9.14120317e-01 2.06589311e-01 2.20857549e+00
6.38389587e-01 3.39742303e-01 3.45239103e-01 1.30298510e-01
9.02657807e-01 -6.78627491e-02 -8.52360189e-01 2.96932340e-01
-2.98650980e-01 -2.33678684e-01 4.81286675e-01 1.86385527e-01
-1.34658098e+00 9.27936912e-01 5.38462353e+00 1.49890983e+00
-9.38231111e-01 2.38437802e-01 6.72542989e-01 -6.11000836e-01
2.43522078e-01 -5.93243539e-02 -1.44985092e+00 4.11715537e-01
3.43129337e-01 -2.58223176e-01 1.08442254e-01 1.17591047e+00
-1.55961767e-01 9.31021273e-02 -1.14517641e+00 1.10628164e+00
3.30477715e-01 -1.32654119e+00 -4.51736376e-02 -2.10054129e-01
9.70739007e-01 -1.31373346e-01 8.45710486e-02 5.86152554e-01
-1.51936248e-01 -6.25504315e-01 8.89583826e-01 3.81458670e-01
9.77183104e-01 -9.91776586e-01 3.09652656e-01 5.44650614e-01
-1.32739413e+00 -6.51973844e-01 -8.24353576e-01 5.18659083e-03
-1.78679198e-01 5.49848616e-01 -6.39204800e-01 3.25567454e-01
7.03788638e-01 8.50856960e-01 -7.48075843e-01 1.40763271e+00
1.74810737e-01 4.75076616e-01 -3.57535869e-01 -1.36056751e-01
2.15808883e-01 2.28454340e-02 7.28446603e-01 1.11597288e+00
3.19881439e-01 3.24496448e-01 6.74275756e-01 6.57183290e-01
4.83092740e-02 2.11747512e-01 -3.09815496e-01 2.76389718e-01
5.70729792e-01 1.34191859e+00 -9.93870199e-01 -5.39319217e-01
-5.16636074e-01 7.54884899e-01 4.64695275e-01 1.20813116e-01
-1.07112920e+00 -6.88535869e-01 4.37022597e-01 2.84576207e-01
6.60559535e-01 9.57449898e-02 -5.01536608e-01 -1.32331896e+00
6.40116408e-02 -4.32986200e-01 8.80741656e-01 -3.39413524e-01
-1.34020185e+00 1.63658708e-01 7.02912956e-02 -1.45440900e+00
2.39857942e-01 -6.45168066e-01 -7.60913968e-01 2.96046168e-01
-1.31109750e+00 -1.36120725e+00 -4.72367913e-01 4.99122798e-01
8.39195669e-01 -3.46952647e-01 4.75744605e-01 5.11270702e-01
-8.14532220e-01 1.11318445e+00 4.25909281e-01 8.11635032e-02
7.50715613e-01 -9.85278368e-01 -2.54988402e-01 7.01131940e-01
8.93098265e-02 2.98002511e-01 5.75291753e-01 -4.94068265e-01
-9.72424686e-01 -1.52989864e+00 2.01107621e-01 -4.03085202e-01
5.68142593e-01 -5.00653148e-01 -9.98301566e-01 5.17255187e-01
-6.25367999e-01 7.51548529e-01 6.85288966e-01 1.25979617e-01
-4.69234735e-01 -2.96400428e-01 -1.17278397e+00 3.24896604e-01
1.26617730e+00 -2.22045690e-01 -3.73343974e-01 6.67489290e-01
8.22844148e-01 -2.80760735e-01 -5.87036073e-01 5.24449110e-01
4.77649987e-01 -7.97264874e-01 1.04414237e+00 -6.91289544e-01
1.32726714e-01 -3.88189614e-01 -4.84447747e-01 -1.20369267e+00
-4.96013761e-01 -3.08654606e-01 -5.74063241e-01 1.20803022e+00
1.92972571e-01 -2.85046399e-01 9.66010451e-01 3.10771346e-01
-4.61471438e-01 -9.31198895e-01 -1.19796360e+00 -1.17449808e+00
6.19477406e-02 -3.47420275e-01 2.97745377e-01 9.25868571e-01
-1.39616266e-01 3.91919464e-01 -2.25726545e-01 1.75059974e-01
7.26698041e-01 3.44839483e-01 8.37465346e-01 -1.22606444e+00
-3.79758567e-01 -5.31737626e-01 -7.33881354e-01 -9.27512288e-01
-6.44469559e-02 -9.04446483e-01 1.75321326e-01 -1.00009787e+00
7.74002790e-01 -5.13784647e-01 -4.37272042e-01 4.55544442e-01
-4.52644914e-01 3.47018182e-01 1.78862214e-01 1.21821702e-01
-1.06414628e+00 9.43262935e-01 1.22903562e+00 -3.21512431e-01
-2.03746855e-01 6.82282671e-02 -6.25595808e-01 7.89691269e-01
4.01308805e-01 -4.28504884e-01 -3.13658535e-01 4.97428812e-02
-2.09634095e-01 -5.77500463e-01 3.71835589e-01 -1.34031856e+00
1.89133152e-01 -3.29301924e-01 5.74938536e-01 -6.77854419e-01
3.21215004e-01 -5.80760241e-01 -5.52770078e-01 4.07169878e-01
-2.78058589e-01 -7.97842801e-01 1.01686411e-01 9.48416233e-01
1.96681410e-01 -4.78099376e-01 1.58047199e+00 1.65672302e-01
-9.67910290e-01 5.87431133e-01 1.90685704e-01 4.00773615e-01
1.66196859e+00 -5.31517208e-01 -2.60107338e-01 7.88757354e-02
-7.77265966e-01 3.07665437e-01 4.50179845e-01 4.61632282e-01
5.89309812e-01 -1.39445031e+00 -6.89734519e-01 1.30697787e-01
7.66815960e-01 2.24446416e-01 3.68352830e-01 7.69247532e-01
-8.37665647e-02 1.17616966e-01 -1.46464147e-02 -8.75015914e-01
-1.32699466e+00 9.75856900e-01 5.09054601e-01 2.43143573e-01
-7.78740346e-01 1.29299498e+00 7.52007008e-01 -2.23262757e-01
3.36099684e-01 -1.49081483e-01 -9.96119007e-02 3.65408152e-01
8.46589684e-01 6.42583013e-01 -1.61103532e-02 -5.34644961e-01
-3.08973849e-01 5.87951422e-01 -2.82131255e-01 6.18739963e-01
1.25920355e+00 -2.17584342e-01 4.71446484e-01 6.62256062e-01
1.33244145e+00 -2.93406934e-01 -1.79558420e+00 -3.20036381e-01
-1.36979997e-01 -7.02569127e-01 4.57083471e-02 -4.95915323e-01
-8.58594954e-01 8.94600689e-01 1.00719190e+00 -1.35434903e-02
8.83875966e-01 2.13757053e-01 7.55779922e-01 3.14172208e-01
1.93828687e-01 -1.59068668e+00 4.27302450e-01 4.92213607e-01
6.05515838e-01 -1.40278602e+00 -5.82476938e-03 -6.72362506e-01
-4.03833032e-01 1.06942260e+00 9.15812731e-01 -5.26500419e-02
7.37843394e-01 1.41537473e-01 -3.87082607e-01 -1.54975235e-01
-6.20377362e-01 -3.46879184e-01 4.81046289e-01 6.74082935e-01
2.74952613e-02 1.82658449e-01 -5.35218231e-02 8.87866080e-01
1.31946146e-01 -1.20566502e-01 2.92361170e-01 1.00333869e+00
-1.14768100e+00 -2.26802230e-01 -3.37391019e-01 8.23747993e-01
-2.68614851e-02 6.18690699e-02 -1.05459362e-01 6.76879346e-01
4.58381981e-01 6.80145264e-01 2.86893964e-01 -3.38040173e-01
3.20724159e-01 -1.03297897e-01 5.71941972e-01 -8.46955895e-01
6.84286654e-02 7.00913966e-02 -2.44602770e-01 -2.06688225e-01
4.89099547e-02 -5.70879340e-01 -1.27252018e+00 1.42886519e-01
-9.47445095e-01 1.74178556e-02 1.39474884e-01 9.73431945e-01
2.27530047e-01 4.81065124e-01 8.11192989e-01 -7.94239461e-01
-1.01647139e+00 -8.40918899e-01 -8.94835532e-01 3.82390767e-01
1.62232235e-01 -1.08040166e+00 -5.40782750e-01 -6.42636865e-02] | [9.619672775268555, 2.142378568649292] |
711447a5-c0e2-4ff4-adf6-f5883bb1adb0 | learning-ensembles-of-potential-functions-for | null | null | http://openaccess.thecvf.com/content_iccv_2015/html/Hajimirsadeghi_Learning_Ensembles_of_ICCV_2015_paper.html | http://openaccess.thecvf.com/content_iccv_2015/papers/Hajimirsadeghi_Learning_Ensembles_of_ICCV_2015_paper.pdf | Learning Ensembles of Potential Functions for Structured Prediction With Latent Variables | Many visual recognition tasks involve modeling variables which are structurally related. Hidden conditional random fields (HCRFs) are a powerful class of models for encoding structure in weakly supervised training examples. This paper presents HCRF-Boost, a novel and general framework for learning HCRFs in functional space. An algorithm is proposed to learn the potential functions of an HCRF as a combination of abstract nonlinear feature functions, expressed by regression models. Consequently, the resulting latent structured model is not restricted to traditional log-linear potential functions or any explicit parameterization. Further, functional optimization helps to avoid direct interactions with the possibly large parameter space of nonlinear models and improves efficiency. As a result, a complex and flexible ensemble method is achieved for structured prediction which can be successfully used in a variety of applications. We validate the effectiveness of this method on tasks such as group activity recognition, human action recognition, and multi-instance learning of video events. | ['Hossein Hajimirsadeghi', 'Greg Mori'] | 2015-12-01 | null | null | null | iccv-2015-12 | ['group-activity-recognition'] | ['computer-vision'] | [ 5.15048683e-01 -8.35623592e-02 -4.82929438e-01 -7.12297916e-01
-6.40346646e-01 -2.10385025e-01 6.29335284e-01 -1.56739488e-01
-1.11770004e-01 9.55378234e-01 3.30220640e-01 4.08465527e-02
-1.62418202e-01 -5.28953791e-01 -7.75402248e-01 -1.18015277e+00
-1.88552693e-01 3.51537794e-01 6.72103837e-02 1.88629165e-01
1.48030892e-01 4.46241081e-01 -1.85103631e+00 6.17113888e-01
7.27028310e-01 9.46050406e-01 3.03099900e-01 4.84765261e-01
9.29797590e-02 1.24332273e+00 -4.45448935e-01 -1.90692306e-01
-1.33147046e-01 -3.97279888e-01 -7.66640365e-01 4.32076484e-01
2.40472972e-01 1.63606927e-01 -6.72122091e-02 6.09195590e-01
1.76260218e-01 4.90266502e-01 1.10263717e+00 -1.19250464e+00
-3.78964126e-01 3.15228134e-01 -1.17998309e-01 1.57600529e-02
3.79397392e-01 -7.49356151e-02 1.28039455e+00 -9.18245971e-01
4.02229875e-01 1.21713459e+00 5.47308862e-01 3.92297655e-01
-1.55848348e+00 -4.00910258e-01 4.26195323e-01 4.64082897e-01
-1.33538699e+00 -4.43948805e-01 6.48773909e-01 -8.53263259e-01
1.22962701e+00 4.35931057e-01 4.97839540e-01 1.24406505e+00
2.76821017e-01 9.20587361e-01 9.17403340e-01 -4.68637496e-01
1.96070433e-01 2.87249424e-02 2.66234279e-01 9.09175396e-01
-1.01403244e-01 2.72799823e-02 -8.88749182e-01 -4.80857521e-01
5.61627567e-01 2.32540876e-01 -4.24652845e-01 -5.21343946e-01
-1.22446203e+00 9.87495363e-01 1.15532048e-01 1.72756746e-01
-2.62023747e-01 1.56194851e-01 3.60805243e-02 1.44766346e-01
6.50359750e-01 4.70907897e-01 -3.97053003e-01 3.96637581e-02
-6.94425762e-01 2.45424658e-02 6.31260216e-01 5.79448044e-01
9.70317245e-01 6.30459413e-02 -2.86155939e-01 7.16199160e-01
4.91636872e-01 1.90095395e-01 4.04939950e-01 -9.95066464e-01
2.41486862e-01 6.43874466e-01 -8.78617447e-03 -7.25449860e-01
-2.48592332e-01 -4.59293872e-01 -7.92705595e-01 2.22473308e-01
2.59027064e-01 3.05895925e-01 -7.73857594e-01 1.57673895e+00
3.04027230e-01 3.60582054e-01 5.75674092e-03 4.88952339e-01
5.58701396e-01 8.00960422e-01 1.73189208e-01 -5.30674279e-01
8.59098017e-01 -1.04537284e+00 -6.75569415e-01 3.60851437e-02
6.41125560e-01 -4.06089574e-01 8.01612914e-01 4.76176113e-01
-9.71293807e-01 -7.13266373e-01 -8.54356885e-01 1.32345306e-02
-1.17985398e-01 3.88524532e-01 9.41765547e-01 2.98052728e-01
-7.49475360e-01 8.89550269e-01 -1.22320783e+00 1.24628849e-01
4.80733246e-01 5.98924875e-01 -3.62032235e-01 -3.04122455e-03
-8.33986282e-01 7.92929947e-01 2.86018252e-01 1.41085431e-01
-9.48268414e-01 -3.69179755e-01 -9.67285514e-01 1.57131240e-01
3.88011962e-01 -5.10503113e-01 9.60697114e-01 -1.12515461e+00
-1.46323025e+00 6.83959842e-01 -5.88313460e-01 -4.19298977e-01
1.96812972e-01 -1.09742291e-01 -3.80553365e-01 -1.55612193e-02
-9.44041088e-02 2.59553879e-01 1.21994770e+00 -8.60942483e-01
-1.78422421e-01 -3.56449455e-01 -1.79420754e-01 3.57455671e-01
-2.88896441e-01 2.87271515e-02 -2.52116527e-02 -7.03180373e-01
1.15261652e-01 -8.58249784e-01 -3.85621339e-01 -1.42057031e-01
-2.67158717e-01 -4.76794511e-01 4.90693986e-01 -4.61590588e-01
1.33226585e+00 -2.06381655e+00 7.00797319e-01 4.15642440e-01
2.36059070e-01 -8.05071294e-02 1.35970369e-01 2.35341221e-01
-2.45080203e-01 6.44381642e-02 -3.29167098e-01 -4.26258713e-01
-8.20662305e-02 4.92306083e-01 -1.87491104e-01 6.11468494e-01
2.47838825e-01 9.67307210e-01 -4.90721881e-01 -4.88461107e-01
2.66024679e-01 5.71212649e-01 -3.71840239e-01 3.76599640e-01
-4.21567023e-01 7.43599117e-01 -5.98775744e-01 5.53096950e-01
1.47963002e-01 -6.54271960e-01 3.28599125e-01 -1.47330269e-01
-7.89703429e-02 1.32959902e-01 -1.01104915e+00 1.52885258e+00
-1.30106717e-01 4.29615438e-01 -4.55559850e-01 -1.54565442e+00
8.84807050e-01 3.97893429e-01 7.72044361e-01 -2.89457858e-01
-2.02635065e-01 -1.31556690e-01 -3.09529036e-01 -5.44054270e-01
-4.06015106e-02 -8.55222940e-02 1.24383131e-02 2.96421498e-01
5.05442619e-01 5.71910143e-01 5.64274751e-02 -1.34342104e-01
1.02692461e+00 4.36058819e-01 4.75547761e-01 -3.98184657e-02
8.38869512e-01 -5.09005010e-01 6.79511905e-01 6.53789639e-01
2.30353728e-01 3.47563475e-01 3.21052760e-01 -6.56125724e-01
-6.89915955e-01 -8.17024648e-01 -2.96758235e-01 1.27924025e+00
-1.69686317e-01 -7.07603931e-01 -3.17801356e-01 -6.89760685e-01
-1.54263124e-01 4.43324417e-01 -6.57494962e-01 -1.78979531e-01
-5.28236330e-01 -9.21271861e-01 1.91042647e-01 7.03321874e-01
1.09731451e-01 -1.24241841e+00 -2.55427033e-01 2.20944002e-01
-3.11499715e-01 -9.58978117e-01 -3.87874573e-01 4.94642347e-01
-9.99603450e-01 -1.12156880e+00 -5.05409658e-01 -6.43946767e-01
6.99698746e-01 4.15231846e-02 1.14614308e+00 1.00032479e-01
-2.76622951e-01 5.16233444e-01 -2.09552184e-01 -1.33779254e-02
-3.54903936e-01 -2.92720884e-01 1.07451668e-03 5.80370963e-01
2.50476837e-01 -6.11539602e-01 -2.68316805e-01 4.65201169e-01
-7.06214845e-01 1.56969085e-01 5.27662933e-01 1.20304799e+00
9.81424093e-01 1.00696340e-01 4.08849090e-01 -8.25012863e-01
1.69793636e-01 -4.18959737e-01 -6.11099064e-01 7.67708719e-01
-5.38293600e-01 6.00444973e-01 4.80057746e-01 -6.74932599e-01
-1.25640392e+00 5.22311687e-01 2.21396193e-01 -4.08583164e-01
-2.55329430e-01 7.56554306e-01 -3.68579865e-01 -1.23551212e-01
7.76485324e-01 3.27580690e-01 -1.70158863e-01 -6.11781120e-01
-4.51201759e-02 3.45798790e-01 2.40076765e-01 -5.50689757e-01
3.11988831e-01 2.99200118e-01 3.09998721e-01 -8.28868747e-01
-9.65865731e-01 -6.26513124e-01 -6.54056191e-01 -3.10252458e-01
9.58269000e-01 -1.01488853e+00 -6.86493635e-01 3.69024307e-01
-8.70544195e-01 -3.60571206e-01 -1.99411288e-01 7.51170158e-01
-8.11316967e-01 4.67641801e-01 -4.13599491e-01 -8.68810534e-01
7.59466663e-02 -9.08197820e-01 1.04250836e+00 -1.59229413e-02
-1.14767753e-01 -1.17003739e+00 2.53577024e-01 6.15007341e-01
-5.43409772e-02 3.38974506e-01 8.28219533e-01 -4.18751329e-01
-7.66389072e-01 -2.22166125e-02 4.94348258e-01 5.79043210e-01
2.88389884e-02 2.54683904e-02 -9.07806396e-01 -3.13982368e-01
-1.15136309e-02 -5.20644963e-01 1.13551080e+00 5.39048731e-01
1.53845346e+00 -4.02310312e-01 -5.03067851e-01 6.99997306e-01
1.11267674e+00 1.05416052e-01 5.84192216e-01 -1.78012684e-01
7.48670936e-01 4.45612848e-01 4.05052692e-01 4.41842616e-01
1.90653652e-01 9.63357568e-01 1.40069917e-01 1.64011773e-02
2.47552156e-01 -2.08385572e-01 7.11138964e-01 4.29999024e-01
-7.40723670e-01 -1.79082602e-01 -8.23488772e-01 3.85965109e-02
-2.24653554e+00 -1.33368123e+00 -2.06243739e-01 2.19924498e+00
7.33557999e-01 -7.88393840e-02 4.28801961e-02 4.74593453e-02
7.52865374e-01 1.33932173e-01 -5.79698801e-01 2.01405250e-02
-2.05181897e-01 4.04685110e-01 1.71288237e-01 3.89354080e-01
-1.28438306e+00 6.32553458e-01 6.51276827e+00 8.29699516e-01
-8.10633183e-01 5.18099554e-02 3.95238400e-01 4.61625420e-02
-1.00036286e-01 1.03796966e-01 -1.00845873e+00 4.04991806e-01
9.78407443e-01 1.32022440e-01 4.86698389e-01 8.52995932e-01
-1.07597616e-02 1.34716518e-02 -1.23582315e+00 1.04765272e+00
2.29359679e-02 -1.44232213e+00 3.79290693e-02 2.77717352e-01
7.46180892e-01 -1.61989376e-01 -1.57687124e-02 3.32731932e-01
1.28118202e-01 -1.31648731e+00 4.25712764e-01 9.51146543e-01
6.91238463e-01 -6.46250844e-01 4.02492821e-01 6.20055914e-01
-1.08745587e+00 -3.02846313e-01 -4.28622901e-01 -6.56560957e-02
8.84404778e-02 2.86815763e-01 -5.46134710e-01 3.95905375e-01
6.02977931e-01 1.08895469e+00 -5.48356354e-01 9.78932023e-01
-1.94433257e-01 8.28895390e-01 2.72660982e-02 1.00288749e-01
-1.47247806e-01 -9.73433405e-02 3.45328927e-01 1.00877976e+00
2.16409579e-01 1.07420728e-01 4.84031916e-01 6.14983976e-01
1.90629244e-01 1.56609252e-01 -5.60876071e-01 -1.99862346e-01
-2.09719464e-01 1.08686614e+00 -5.83329797e-01 -2.58929580e-01
-5.77639043e-01 7.47364163e-01 5.80102623e-01 5.71714938e-01
-8.63394141e-01 2.85217732e-01 4.94986504e-01 8.82291719e-02
5.33669353e-01 -2.07881719e-01 7.84386545e-02 -1.76754272e+00
-3.84725071e-02 -7.98490286e-01 5.98000586e-01 -5.67908645e-01
-1.34499216e+00 6.65072918e-01 1.69721588e-01 -1.42165709e+00
-4.54345614e-01 -6.19542778e-01 -4.74790514e-01 6.97589934e-01
-1.27824593e+00 -1.15544820e+00 -1.97666049e-01 1.23873770e+00
3.86238068e-01 -4.09943938e-01 1.29425764e+00 -2.93450486e-02
-4.77885336e-01 3.99690568e-01 1.48713946e-01 -1.83881149e-01
4.65115935e-01 -1.28325462e+00 -2.38120869e-01 6.22450292e-01
5.22048235e-01 4.68190342e-01 4.49766099e-01 -5.85884452e-01
-9.53072667e-01 -1.03068113e+00 9.12666500e-01 -5.56562126e-01
2.93866813e-01 -5.06077349e-01 -1.06507730e+00 7.74546325e-01
-2.35182315e-01 2.19097018e-01 1.23024321e+00 3.93011928e-01
-2.67850071e-01 -4.48865034e-02 -7.15719402e-01 2.03556642e-01
1.07128048e+00 -6.19721293e-01 -3.93103898e-01 5.65713406e-01
3.76869857e-01 -1.31631181e-01 -9.55011666e-01 6.05352461e-01
6.16649032e-01 -8.08523118e-01 1.10718763e+00 -1.04904354e+00
1.40322894e-01 -2.69730359e-01 -2.85209358e-01 -9.86633778e-01
-6.24734640e-01 -6.29586399e-01 -8.71951699e-01 6.98607326e-01
3.15794259e-01 -4.02974188e-01 7.29661167e-01 7.04338968e-01
-1.57547906e-01 -7.33554304e-01 -8.87769699e-01 -5.47433257e-01
-4.41626757e-01 -3.56561065e-01 2.07449451e-01 8.26614738e-01
-5.77712469e-02 5.44061124e-01 -8.28387678e-01 4.21276316e-02
7.31155872e-01 3.61319005e-01 5.11264265e-01 -1.57326758e+00
-9.89161491e-01 -3.03114913e-02 -5.96236467e-01 -1.03800082e+00
6.06381476e-01 -1.05980456e+00 -4.98191305e-02 -1.12380254e+00
3.52856576e-01 -2.08045736e-01 -6.27097964e-01 6.24729872e-01
-8.93166661e-02 -1.54944491e-02 -2.93496661e-02 4.48004931e-01
-9.41109121e-01 7.35056341e-01 1.10071111e+00 -1.95580766e-01
-3.64046931e-01 6.99012935e-01 -3.74790162e-01 7.70957291e-01
8.29495311e-01 -5.57085633e-01 -5.42402148e-01 1.43390909e-01
1.59570947e-01 3.83457065e-01 5.87627232e-01 -7.54771173e-01
8.25188830e-02 -3.65417808e-01 6.34450376e-01 -5.26598394e-01
5.95746458e-01 -6.89870119e-01 4.90093350e-01 5.78333959e-02
-5.19083261e-01 -2.64538348e-01 -3.84081602e-01 9.00832653e-01
-4.54597980e-01 -2.46780559e-01 5.38786769e-01 -2.71591097e-01
-6.10513747e-01 4.11973387e-01 -5.03940761e-01 -3.79826158e-01
9.28850174e-01 -3.00511777e-01 1.11652218e-01 -4.07121241e-01
-1.39879620e+00 8.06191713e-02 8.51712450e-02 2.39092052e-01
6.52646244e-01 -1.38231409e+00 -4.38885182e-01 4.18580323e-01
1.42447978e-01 -7.54334256e-02 3.50945264e-01 9.26872730e-01
9.90299582e-02 4.53748614e-01 -1.53873920e-01 -7.70417154e-01
-1.49287140e+00 5.28505862e-01 3.40042502e-01 -7.22807825e-01
-5.74905336e-01 6.94476366e-01 4.23310399e-01 -2.11335495e-01
4.78886604e-01 -1.53019711e-01 -5.12563407e-01 2.12285873e-02
4.98501182e-01 3.10005218e-01 -8.68111476e-02 -8.94204557e-01
-2.85353303e-01 4.05827016e-01 7.80110434e-02 1.17978230e-01
1.66058552e+00 3.81649695e-02 1.19951926e-02 5.79476058e-01
1.29592264e+00 -3.29145014e-01 -1.69271505e+00 -4.16936904e-01
1.15796670e-01 -2.41781339e-01 4.08921093e-02 -5.29541373e-01
-7.45385766e-01 8.90485227e-01 4.35061723e-01 -6.84046373e-03
1.10654354e+00 2.53956258e-01 3.29651795e-02 7.10576952e-01
3.28143001e-01 -1.02673280e+00 4.20646906e-01 3.73982579e-01
9.34642971e-01 -1.18248415e+00 1.34930894e-01 -2.81452000e-01
-8.06623042e-01 1.33095706e+00 4.25378948e-01 5.38755991e-02
7.62452006e-01 3.21262814e-02 -4.79899526e-01 -1.41466856e-01
-1.04656756e+00 -2.20592082e-01 7.56864011e-01 4.74945962e-01
4.62791234e-01 2.07333956e-02 -1.99743822e-01 6.32051289e-01
2.77827531e-01 1.72144920e-02 1.18471712e-01 8.69760275e-01
-3.52436811e-01 -1.30749297e+00 -8.74784142e-02 4.20170128e-01
-5.58736920e-01 1.68448873e-02 -6.75582234e-03 2.69048721e-01
2.13342443e-01 7.34148800e-01 -2.48417169e-01 -3.11660469e-01
-5.88746220e-02 4.86081719e-01 6.45370662e-01 -7.35763967e-01
-3.24061453e-01 2.78851360e-01 4.57844930e-03 -6.99317694e-01
-1.01251447e+00 -9.57240105e-01 -1.12936962e+00 3.03305924e-01
-6.20425165e-01 2.01909482e-01 3.67142797e-01 1.01660931e+00
1.83534309e-01 2.81938285e-01 7.66251087e-01 -9.01252031e-01
-6.17511153e-01 -7.62167752e-01 -5.61979473e-01 3.26051027e-01
3.04149896e-01 -1.03979766e+00 -3.89824778e-01 4.36505586e-01] | [8.52204704284668, 0.8417503833770752] |
b84e81eb-8991-4e7e-ae2d-44588d2ce1d9 | what-can-a-cook-in-italy-teach-a-mechanic-in | 2306.08713 | null | https://arxiv.org/abs/2306.08713v1 | https://arxiv.org/pdf/2306.08713v1.pdf | What can a cook in Italy teach a mechanic in India? Action Recognition Generalisation Over Scenarios and Locations | We propose and address a new generalisation problem: can a model trained for action recognition successfully classify actions when they are performed within a previously unseen scenario and in a previously unseen location? To answer this question, we introduce the Action Recognition Generalisation Over scenarios and locations dataset (ARGO1M), which contains 1.1M video clips from the large-scale Ego4D dataset, across 10 scenarios and 13 locations. We demonstrate recognition models struggle to generalise over 10 proposed test splits, each of an unseen scenario in an unseen location. We thus propose CIR, a method to represent each video as a Cross-Instance Reconstruction of videos from other domains. Reconstructions are paired with text narrations to guide the learning of a domain generalisable representation. We provide extensive analysis and ablations on ARGO1M that show CIR outperforms prior domain generalisation works on all test splits. Code and data: https://chiaraplizz.github.io/what-can-a-cook/. | ['Dima Damen', 'Barbara Caputo', 'Toby Perrett', 'Chiara Plizzari'] | 2023-06-14 | null | null | null | null | ['action-recognition-in-videos'] | ['computer-vision'] | [ 4.73784059e-01 -2.98434850e-02 -2.28943124e-01 -3.46820414e-01
-9.16483819e-01 -8.32651019e-01 9.74008799e-01 -6.15583241e-01
-1.39726192e-01 6.37466192e-01 8.03041220e-01 2.98048351e-02
1.22812530e-02 -9.72438082e-02 -8.39886725e-01 -5.94662011e-01
-4.34904248e-02 5.01041591e-01 2.32650980e-01 2.61876106e-01
6.54326081e-02 3.28140140e-01 -1.54160643e+00 8.44225407e-01
1.73674092e-01 9.31422830e-01 -9.20427218e-03 8.04512084e-01
6.55265033e-01 1.13029182e+00 -8.31344843e-01 -3.07193786e-01
5.05136490e-01 -6.09716237e-01 -9.38085318e-01 5.40528774e-01
9.29848909e-01 -8.24526966e-01 -9.70814943e-01 5.46635747e-01
3.70846540e-01 4.14906174e-01 8.69170725e-01 -1.64643693e+00
-6.19350910e-01 2.75987744e-01 -1.94208220e-01 5.39112866e-01
7.73729563e-01 6.79795384e-01 6.40546918e-01 -6.77683175e-01
9.95670438e-01 1.21675527e+00 6.14244580e-01 8.82145882e-01
-1.09474325e+00 -5.95359385e-01 3.52643639e-01 5.74639678e-01
-1.10426009e+00 -8.63265276e-01 4.57465887e-01 -5.67980587e-01
1.18664598e+00 1.36302501e-01 5.17337739e-01 2.33649302e+00
4.82149124e-02 1.46554005e+00 8.05976450e-01 6.91655651e-02
2.32976139e-01 -3.00121188e-01 -2.53199995e-01 9.67615396e-02
1.52681574e-01 1.17904775e-01 -8.01129878e-01 1.41543061e-01
7.07946360e-01 5.10149412e-02 -5.72867990e-01 -5.75351238e-01
-1.53259337e+00 4.26962048e-01 1.39342070e-01 3.19581144e-02
-4.42199826e-01 3.69589776e-01 5.05271196e-01 2.44718671e-01
1.71068713e-01 3.72198373e-01 -5.80284238e-01 -7.20006585e-01
-8.49647284e-01 4.09735560e-01 5.68483949e-01 1.16354692e+00
4.14924145e-01 1.53540775e-01 -9.25483555e-02 4.15950328e-01
-1.13704264e-01 5.51725626e-01 6.49946988e-01 -1.37458229e+00
7.61264741e-01 4.05299544e-01 2.09988847e-01 -5.67173302e-01
-1.31278470e-01 -6.52480721e-02 -3.02045405e-01 8.37085992e-02
5.59731126e-01 -2.87157208e-01 -1.09687841e+00 1.68530738e+00
2.32978657e-01 6.45823836e-01 3.75906080e-01 8.32868993e-01
9.50815082e-01 3.58889967e-01 2.90185183e-01 2.98322558e-01
1.04972303e+00 -1.00517368e+00 -4.30077612e-01 -6.60983860e-01
8.38000417e-01 -3.04440767e-01 8.69667292e-01 3.79475683e-01
-7.82802939e-01 -6.06249690e-01 -7.93784261e-01 1.08614028e-01
-4.02043790e-01 4.02541786e-01 3.35867077e-01 1.41982630e-01
-8.73058319e-01 5.60546577e-01 -8.06959093e-01 -9.15740252e-01
7.38936484e-01 -1.09216452e-01 -9.99787867e-01 -2.97068506e-01
-9.17979538e-01 9.53500032e-01 6.47387087e-01 -1.55237019e-01
-1.78551972e+00 -5.43863893e-01 -1.02159059e+00 -3.00269514e-01
5.98069847e-01 -4.40810204e-01 1.36586988e+00 -1.23986697e+00
-9.45544124e-01 1.12376773e+00 3.36566307e-02 -7.31165171e-01
6.65173054e-01 -6.08424783e-01 -8.15318346e-01 6.47005081e-01
4.12033588e-01 8.18833530e-01 9.68189061e-01 -1.01411664e+00
-6.44807935e-01 -2.55335689e-01 3.72394145e-01 2.87887216e-01
1.23625234e-01 -1.94040343e-01 -2.79249609e-01 -8.34401786e-01
-7.75078833e-02 -1.11202300e+00 1.72737181e-01 -1.45638324e-02
-3.37466747e-01 -3.59928259e-03 9.06295657e-01 -8.37962210e-01
8.56970191e-01 -2.42411041e+00 3.50168318e-01 -4.72369909e-01
-1.54120162e-01 2.21343085e-01 -4.48354423e-01 5.21999538e-01
-4.60289329e-01 -1.77521259e-02 -2.72168294e-02 -1.79860547e-01
1.39294893e-01 4.03535962e-01 -4.58352089e-01 6.41195595e-01
4.11769062e-01 8.52918744e-01 -9.46875334e-01 -2.94886768e-01
4.33315694e-01 2.98412025e-01 -4.18645203e-01 2.49449447e-01
-1.89314663e-01 6.07398152e-01 -4.33700919e-01 9.05835867e-01
4.13471639e-01 -7.57399797e-02 1.77163944e-01 -2.84122825e-01
4.01492357e-01 8.21392015e-02 -1.04301941e+00 2.10257339e+00
-2.90783234e-02 8.22973311e-01 -1.63631842e-01 -1.12586749e+00
6.07212305e-01 3.47883582e-01 5.46944082e-01 -5.28898895e-01
2.61390898e-02 -5.14647029e-02 -2.83377349e-01 -8.49037647e-01
2.09831819e-01 5.41892573e-02 -4.18306768e-01 2.56240904e-01
5.62467635e-01 2.38110974e-01 1.64957240e-01 2.82479286e-01
1.84357274e+00 9.11128461e-01 4.73270118e-01 1.45565346e-01
2.58569568e-01 1.77061185e-01 6.23007834e-01 9.48301136e-01
-9.11707819e-01 9.40344632e-01 4.50994819e-01 -5.45424759e-01
-7.43636131e-01 -1.26296473e+00 -9.26727336e-03 1.14731622e+00
4.93222177e-02 -4.56703424e-01 -6.37898684e-01 -1.27954388e+00
8.39623213e-02 9.36726749e-01 -9.62199986e-01 -3.19841862e-01
-4.37797844e-01 -9.49204341e-02 7.72655606e-01 8.26403260e-01
7.71995604e-01 -1.29412138e+00 -8.44393015e-01 -1.80300385e-01
-3.96664083e-01 -1.41420686e+00 -1.20926373e-01 2.32303768e-01
-6.41982973e-01 -1.37098813e+00 -7.29969203e-01 -3.95455778e-01
2.94711947e-01 3.30030978e-01 1.27333283e+00 -4.63374734e-01
-1.52075201e-01 1.20961869e+00 -9.34642494e-01 -1.09398030e-01
-2.97392100e-01 -1.26629323e-01 2.77553052e-01 1.23956881e-01
8.88792574e-01 -6.34379327e-01 -5.65504551e-01 6.92367196e-01
-7.74748266e-01 -1.58016697e-01 6.70446396e-01 5.25239646e-01
3.27912897e-01 -2.88448811e-01 3.77200276e-01 -4.49484527e-01
-2.61332598e-02 -1.02882636e+00 1.24995492e-01 2.94641078e-01
3.41725796e-01 -1.23295031e-01 3.37896675e-01 -7.03955472e-01
-1.08740795e+00 4.02330518e-01 2.57154912e-01 -9.13708270e-01
-1.00458527e+00 6.10414855e-02 -5.36982179e-01 4.21817064e-01
9.84605849e-01 4.32092577e-01 -2.62291402e-01 -5.20401120e-01
3.58345360e-01 5.72291374e-01 8.51161718e-01 -6.67210281e-01
7.05301106e-01 5.95732272e-01 -3.40644717e-01 -7.64491796e-01
-8.08812320e-01 -4.77619618e-01 -9.43883598e-01 -5.36080182e-01
1.03834021e+00 -1.33046257e+00 -3.12170178e-01 4.29949075e-01
-8.49612534e-01 -8.16451252e-01 -4.39046055e-01 4.90302235e-01
-1.17194974e+00 2.94028908e-01 7.71051571e-02 -4.52596307e-01
5.57218313e-01 -8.78390610e-01 1.26843584e+00 6.05691969e-02
-6.08102739e-01 -9.03307378e-01 1.38361961e-01 6.50497019e-01
-1.06948204e-01 6.63924932e-01 9.22662616e-02 -1.07825363e+00
-3.81953210e-01 -3.52553606e-01 2.62178034e-02 4.80806619e-01
8.59795585e-02 -2.81269878e-01 -1.08684099e+00 -4.02647555e-01
-4.43408526e-02 -8.23200405e-01 9.55635965e-01 1.34482399e-01
1.15052891e+00 -2.93670356e-01 -4.10193503e-01 5.62956870e-01
1.03840292e+00 1.92228302e-01 1.07609236e+00 6.79910362e-01
5.24940073e-01 3.09426814e-01 9.33920503e-01 6.01683199e-01
1.72507226e-01 5.64134002e-01 5.35416067e-01 3.44452649e-01
-3.82150888e-01 -4.59538788e-01 9.22828853e-01 -2.52066851e-01
-1.59836575e-01 -4.98854429e-01 -8.60191584e-01 9.50096846e-01
-1.96488941e+00 -1.59941006e+00 1.97616562e-01 1.83339489e+00
3.20547134e-01 8.24858202e-04 3.21637273e-01 -1.39360026e-01
5.31227410e-01 5.69135904e-01 -9.44858789e-01 5.40720671e-02
-2.75880814e-01 -2.34557763e-02 3.37231457e-01 4.59683910e-02
-1.59285557e+00 9.43238437e-01 6.43929291e+00 7.23871052e-01
-7.53260434e-01 8.97818655e-02 3.28025758e-01 -4.94788080e-01
2.25637421e-01 5.17328940e-02 -6.89699233e-01 5.57623088e-01
1.15851021e+00 5.80602176e-02 1.09604500e-01 1.21065378e+00
1.59623966e-01 -2.74308950e-01 -1.62283373e+00 1.17526031e+00
4.83381182e-01 -1.05537593e+00 2.76977092e-01 -3.95130664e-02
6.22473717e-01 1.41156897e-01 1.02199987e-01 7.29613662e-01
3.38190585e-01 -9.64637816e-01 7.40134120e-01 5.12573659e-01
7.77492166e-01 -1.57793492e-01 2.93913424e-01 2.05376327e-01
-1.00455856e+00 -4.16244864e-01 -1.78134814e-01 -8.95410180e-02
1.47784829e-01 -3.01049769e-01 -8.09359789e-01 5.37411809e-01
9.19622481e-01 1.53909373e+00 -8.57438385e-01 8.46880496e-01
-4.15435970e-01 5.87422252e-01 -1.23473965e-01 5.38306236e-01
4.30064261e-01 2.29333952e-01 8.76339018e-01 1.33422244e+00
3.80921453e-01 1.90278620e-01 6.71581775e-02 3.64447385e-01
1.07573487e-01 -6.32764757e-01 -1.05589557e+00 -3.33599597e-02
2.92333454e-01 8.33748043e-01 -2.90935606e-01 -4.33084339e-01
-4.50064421e-01 1.49920797e+00 1.26962945e-01 7.50637531e-01
-1.19382346e+00 6.58947304e-02 1.01786304e+00 2.47581407e-01
5.93465149e-01 -1.57746702e-01 3.02314609e-01 -1.62481952e+00
2.30674133e-01 -1.18470407e+00 8.06797385e-01 -1.24768996e+00
-1.14320672e+00 3.06057960e-01 4.38952416e-01 -1.93223524e+00
-5.98198116e-01 -8.19860518e-01 -4.51646388e-01 2.38949001e-01
-1.07466018e+00 -1.20032704e+00 -5.92841685e-01 7.91455030e-01
1.11804533e+00 -5.10790348e-01 7.58617163e-01 1.32403359e-01
-5.29057026e-01 4.70830411e-01 2.03427169e-02 4.25231397e-01
9.75885749e-01 -1.01320267e+00 4.52733189e-01 9.85839248e-01
2.21390933e-01 2.29598269e-01 7.04549849e-01 -7.24939764e-01
-1.10922790e+00 -1.27509475e+00 3.78726393e-01 -1.27848327e+00
7.31714189e-01 -2.66436189e-01 -8.01139772e-01 1.53923190e+00
1.03653893e-01 9.85135734e-02 7.39610195e-01 -2.27239743e-01
-5.24275124e-01 1.84864387e-01 -1.13756156e+00 5.66380560e-01
1.63897288e+00 -5.79540908e-01 -1.03436124e+00 4.31940019e-01
2.14721397e-01 -3.25948775e-01 -7.84499526e-01 2.96607107e-01
5.60908020e-01 -1.17016494e+00 1.05311441e+00 -1.15764284e+00
7.16714799e-01 -3.38765800e-01 -6.67981207e-01 -1.17866600e+00
-1.59898654e-01 -5.24701357e-01 -4.89587277e-01 1.04737294e+00
1.32443979e-01 -3.53482038e-01 7.21428752e-01 6.43210113e-01
-1.42170519e-01 -2.28720888e-01 -1.09424520e+00 -1.25850093e+00
-4.06668372e-02 -5.79959810e-01 3.86283040e-01 9.55681443e-01
-8.42058435e-02 1.28435373e-01 -6.20901465e-01 1.73309147e-01
5.89760780e-01 -3.33987951e-01 1.16188741e+00 -8.14541340e-01
-3.60915422e-01 -1.06188938e-01 -1.04598403e+00 -1.41730082e+00
4.12131876e-01 -6.18683338e-01 -1.05674781e-01 -1.49581170e+00
2.11164430e-01 1.60544008e-01 -9.21921730e-02 7.57460296e-01
1.34119555e-01 2.82903939e-01 2.59597749e-01 3.79623413e-01
-1.19523072e+00 6.30884767e-01 7.99879491e-01 -2.76849210e-01
2.26665840e-01 -1.37771010e-01 -5.52033246e-01 7.96056032e-01
6.74683690e-01 -3.63701195e-01 -6.33501291e-01 -6.19286060e-01
-3.19025367e-01 2.83287764e-02 8.47448707e-01 -1.61796045e+00
-1.76819205e-01 -2.99172759e-01 7.88587987e-01 -5.96187949e-01
7.00392663e-01 -8.89409244e-01 4.55342591e-01 4.94848415e-02
-3.56734782e-01 -2.43777752e-01 3.54272634e-01 8.01595628e-01
-1.12617612e-01 -9.24715698e-02 4.79591012e-01 -2.97275335e-01
-1.54148817e+00 7.35050216e-02 -4.46816057e-01 3.44300508e-01
1.44747269e+00 -6.85260713e-01 -5.46405494e-01 -6.29483581e-01
-1.26278973e+00 1.68904632e-01 7.07507312e-01 7.92815149e-01
5.76952994e-01 -1.48505127e+00 -7.62926936e-01 -2.20962316e-02
4.41844910e-01 -2.85901874e-01 5.50860286e-01 5.36908329e-01
-3.27136219e-01 3.94847214e-01 -5.16583383e-01 -7.30431020e-01
-1.07983088e+00 5.08899689e-01 4.31262106e-01 9.51905996e-02
-7.61108994e-01 6.23293400e-01 4.84421343e-01 -2.47144878e-01
1.75680518e-01 -2.51509756e-01 4.65301424e-03 8.07672963e-02
7.39846766e-01 4.38948691e-01 -2.95309514e-01 -1.00371718e+00
-7.37028062e-01 4.01080161e-01 3.99152786e-02 -7.74871036e-02
1.29856765e+00 -3.39384899e-02 8.02370489e-01 4.45519716e-01
1.24059510e+00 -6.83541834e-01 -2.03771949e+00 -4.02112380e-02
-1.78389549e-01 -8.15002918e-01 -4.58328784e-01 -1.09288275e+00
-7.94235349e-01 6.25897765e-01 5.22720635e-01 -1.62064433e-01
1.05303168e+00 3.59761834e-01 3.99471700e-01 5.51842153e-01
3.12703729e-01 -1.12977028e+00 5.53765595e-01 5.01978517e-01
1.20442176e+00 -1.33301175e+00 1.01642467e-01 2.05852911e-01
-1.29599023e+00 9.43345368e-01 9.16143835e-01 -1.50957748e-01
2.90762305e-01 -2.91775703e-01 2.14176951e-03 -2.15632856e-01
-7.34859526e-01 -3.85698646e-01 7.82472342e-02 1.21875477e+00
-6.25704750e-02 -1.79555323e-02 5.11347055e-01 5.53542912e-01
6.97738752e-02 4.38670740e-02 6.61411524e-01 1.14921403e+00
-1.28832847e-01 -4.92768645e-01 -2.84402937e-01 2.29299217e-01
-1.55938193e-01 3.84059876e-01 -7.66831100e-01 1.12512934e+00
1.36044219e-01 8.80604088e-01 2.67559201e-01 -6.50434792e-01
4.56667006e-01 3.83858770e-01 5.13210654e-01 -5.81812561e-01
-1.68497879e-02 -3.14059943e-01 4.48526531e-01 -1.08936441e+00
-6.88763380e-01 -1.29033399e+00 -8.76684248e-01 -2.89189741e-02
3.94956917e-01 -2.64486998e-01 1.89718142e-01 9.02904928e-01
6.53761446e-01 3.82589906e-01 3.41786474e-01 -1.26047111e+00
-4.65840787e-01 -1.05211449e+00 -4.83443499e-01 8.99371743e-01
4.17854130e-01 -1.09161353e+00 -7.31556535e-01 3.97771269e-01] | [8.31993579864502, 0.6658830046653748] |
6b251fa5-9e3e-4bc5-8a4a-14eca96a32f4 | emotional-speech-driven-animation-with | 2306.08990 | null | https://arxiv.org/abs/2306.08990v1 | https://arxiv.org/pdf/2306.08990v1.pdf | Emotional Speech-Driven Animation with Content-Emotion Disentanglement | To be widely adopted, 3D facial avatars need to be animated easily, realistically, and directly, from speech signals. While the best recent methods generate 3D animations that are synchronized with the input audio, they largely ignore the impact of emotions on facial expressions. Instead, their focus is on modeling the correlations between speech and facial motion, resulting in animations that are unemotional or do not match the input emotion. We observe that there are two contributing factors resulting in facial animation - the speech and the emotion. We exploit these insights in EMOTE (Expressive Model Optimized for Talking with Emotion), which generates 3D talking head avatars that maintain lip sync while enabling explicit control over the expression of emotion. Due to the absence of high-quality aligned emotional 3D face datasets with speech, EMOTE is trained from an emotional video dataset (i.e., MEAD). To achieve this, we match speech-content between generated sequences and target videos differently from emotion content. Specifically, we train EMOTE with additional supervision in the form of a lip-reading objective to preserve the speech-dependent content (spatially local and high temporal frequency), while utilizing emotion supervision on a sequence-level (spatially global and low frequency). Furthermore, we employ a content-emotion exchange mechanism in order to supervise different emotion on the same audio, while maintaining the lip motion synchronized with the speech. To employ deep perceptual losses without getting undesirable artifacts, we devise a motion prior in form of a temporal VAE. Extensive qualitative, quantitative, and perceptual evaluations demonstrate that EMOTE produces state-of-the-art speech-driven facial animations, with lip sync on par with the best methods while offering additional, high-quality emotional control. | ['Timo Bolkart', 'Michael J. Black', 'Yandong Wen', 'Shashank Tripathi', 'Kiran Chhatre', 'Radek Daněček'] | 2023-06-15 | null | null | null | null | ['disentanglement'] | ['methodology'] | [-4.51964000e-03 2.70437121e-01 1.82597642e-03 -1.47220850e-01
-3.72192681e-01 -4.30889845e-01 5.07696807e-01 -8.32028389e-01
4.29042280e-02 2.95833260e-01 4.93481696e-01 2.97943950e-01
4.94045347e-01 -3.99492979e-01 -6.31099820e-01 -8.33449066e-01
-1.61564332e-02 -5.40975742e-02 -8.66186693e-02 -2.76511073e-01
-3.80148739e-01 5.80316186e-01 -1.90577364e+00 2.20735744e-01
5.07749856e-01 1.07361114e+00 -1.50140852e-01 7.94514656e-01
1.25541598e-01 8.87907863e-01 -6.66433871e-01 -4.55272704e-01
1.44050255e-01 -7.53066659e-01 -4.01717812e-01 4.27279830e-01
4.33797747e-01 -2.88182139e-01 -4.30351079e-01 9.33472693e-01
6.70899451e-01 1.18648008e-01 5.34722745e-01 -1.56878018e+00
-3.72505367e-01 -8.90575647e-02 -5.58789730e-01 -4.26704019e-01
5.40989280e-01 5.80993056e-01 6.80092931e-01 -8.53281856e-01
1.07058787e+00 1.57853258e+00 5.36696255e-01 1.09729075e+00
-1.45145214e+00 -6.81584477e-01 7.20065758e-02 5.22821471e-02
-1.21675241e+00 -1.07426548e+00 1.20304477e+00 -3.59387815e-01
4.77392405e-01 4.38557863e-01 1.00510395e+00 1.64035404e+00
1.37144774e-01 6.01483524e-01 8.33575130e-01 -3.01361948e-01
1.51899576e-01 7.48451725e-02 -6.05804801e-01 5.44234753e-01
-8.08310628e-01 4.36311424e-01 -8.32990110e-01 1.15023321e-03
6.70163453e-01 -5.37328601e-01 -5.34475029e-01 -3.53306234e-01
-9.56487358e-01 5.44479370e-01 -5.37306778e-02 1.35836229e-01
-6.19415462e-01 3.35318714e-01 5.72793126e-01 2.34994590e-01
5.59858799e-01 -7.51589984e-02 -1.48626938e-01 -4.31167930e-01
-9.70477104e-01 1.46416977e-01 6.60857320e-01 7.57681191e-01
5.66653192e-01 6.41024649e-01 -1.63566500e-01 7.09519029e-01
4.04355526e-01 8.02220464e-01 2.87093937e-01 -1.36357832e+00
-1.23873584e-01 7.09447451e-03 -8.88066590e-02 -1.30007160e+00
-2.03839108e-01 1.47674251e-02 -7.98961878e-01 6.47416413e-01
1.93761781e-01 -2.87438065e-01 -7.34231114e-01 2.34409118e+00
5.96153915e-01 6.11526728e-01 7.15646222e-02 1.20059168e+00
7.83447266e-01 7.39850521e-01 1.76993385e-01 -6.04707003e-01
1.34155262e+00 -7.18934119e-01 -1.45614183e+00 9.45091248e-02
3.37262422e-01 -7.91401327e-01 1.24836874e+00 3.12800258e-01
-1.38589215e+00 -6.51781201e-01 -6.69655502e-01 5.51776821e-03
1.58700049e-01 -1.12270778e-02 1.78014889e-01 6.37459636e-01
-1.21650910e+00 3.18824679e-01 -7.72216439e-01 8.49601440e-03
1.11124635e-01 8.10256526e-02 -6.06414855e-01 4.55690265e-01
-1.26936269e+00 8.14986825e-01 -4.47348863e-01 3.71532841e-03
-9.32098925e-01 -8.72748137e-01 -1.08155429e+00 1.81876950e-03
1.77419230e-01 -5.74559808e-01 1.23710072e+00 -1.75820446e+00
-2.32810187e+00 8.12278926e-01 -3.11868936e-01 -7.55263418e-02
6.82970762e-01 -8.55540633e-02 -5.31944215e-01 5.85754097e-01
-2.26685256e-01 1.09536552e+00 1.34014213e+00 -1.39027429e+00
-5.88083789e-02 -5.69929443e-02 -3.90091449e-01 2.28926510e-01
-3.58463913e-01 2.93763429e-01 -6.84382975e-01 -8.22695613e-01
-3.57905537e-01 -1.10729206e+00 2.31599763e-01 5.39637208e-01
-2.42631599e-01 2.01942056e-01 1.40434587e+00 -7.16791451e-01
9.26845372e-01 -2.44896317e+00 4.26393926e-01 -9.24391225e-02
1.34671032e-01 2.10397214e-01 -3.85077596e-01 1.61219046e-01
-3.92323554e-01 5.45408055e-02 -1.90104637e-02 -8.12442183e-01
2.66547259e-02 1.82912230e-01 -3.71815741e-01 5.94993055e-01
3.96508545e-01 8.58682930e-01 -8.11394095e-01 -5.43293118e-01
2.76097119e-01 1.14463389e+00 -8.14776957e-01 3.43551368e-01
-3.33082825e-01 8.27769816e-01 -1.28742322e-01 4.30892646e-01
5.94680786e-01 2.28144914e-01 1.10044435e-01 -5.46698987e-01
-1.25994403e-02 5.97857684e-02 -1.03135478e+00 1.65647924e+00
-6.81349635e-01 8.92792821e-01 7.93862581e-01 -5.78229010e-01
7.07285285e-01 8.45032513e-01 8.64540339e-01 -6.78424418e-01
3.88676286e-01 -6.93864152e-02 -1.17399842e-01 -7.46030152e-01
2.18651757e-01 -3.64492118e-01 1.61177367e-01 2.86125809e-01
1.73849970e-01 -4.04769182e-01 -2.06993818e-01 -1.67455506e-02
8.33588719e-01 3.68614525e-01 -3.17174017e-01 3.72351799e-03
4.13011134e-01 -4.40329403e-01 5.57460606e-01 -5.27844168e-02
-3.82348806e-01 7.12959349e-01 6.81077480e-01 6.35210276e-02
-1.11535919e+00 -8.73336971e-01 1.40805230e-01 9.03954208e-01
1.40535653e-01 -3.58819753e-01 -1.06851590e+00 -3.65640014e-01
-2.84023196e-01 5.71421921e-01 -6.79675221e-01 -3.00944030e-01
-6.17487371e-01 -2.03592256e-02 7.92496562e-01 9.02564153e-02
3.44850540e-01 -1.18261158e+00 -7.34535456e-01 1.75499961e-01
-4.13472444e-01 -1.36540461e+00 -8.15429151e-01 -3.17745447e-01
-2.68020123e-01 -7.36114919e-01 -8.69331002e-01 -4.29154664e-01
2.66802013e-01 -1.27666980e-01 9.16123807e-01 -1.68784574e-01
-2.14100495e-01 7.00425446e-01 -1.66248888e-01 -1.31467268e-01
-7.81874657e-01 -5.39842308e-01 2.99011558e-01 5.66841125e-01
-3.16196173e-01 -8.56603384e-01 -5.28436720e-01 3.34552288e-01
-9.61261332e-01 2.30744869e-01 4.48858626e-02 8.56355786e-01
2.40638822e-01 -3.07498395e-01 6.26650572e-01 -4.49931882e-02
4.34901923e-01 -2.47833505e-01 -2.89007813e-01 -1.88693866e-01
4.70566005e-02 -2.23864496e-01 6.70466602e-01 -1.01851833e+00
-1.25021696e+00 9.01252031e-02 -5.24568796e-01 -1.24012840e+00
-2.19011098e-01 -6.33184761e-02 -4.06848729e-01 2.33677011e-02
5.01944661e-01 1.21905498e-01 6.37817025e-01 -3.46572064e-02
5.57721138e-01 5.27683258e-01 6.93756461e-01 -4.71760064e-01
6.77097321e-01 7.34556556e-01 5.53291216e-02 -1.14247906e+00
-2.14047328e-01 1.25077501e-01 -2.22761750e-01 -7.95512736e-01
1.02516484e+00 -9.09911096e-01 -1.17066121e+00 6.99070156e-01
-1.26846981e+00 -6.13779306e-01 -4.51754928e-01 4.65616763e-01
-9.17177439e-01 4.11585987e-01 -7.88259029e-01 -9.31069076e-01
-1.32808000e-01 -1.31904030e+00 1.25127304e+00 -1.49600003e-02
-5.73864937e-01 -7.70595014e-01 2.08506696e-02 1.00112960e-01
4.15669620e-01 5.73825300e-01 6.87183738e-01 1.10074803e-01
-1.89637393e-01 1.83900148e-01 1.98174074e-01 3.31587344e-01
2.77111173e-01 6.14416897e-01 -1.18870735e+00 1.46846306e-02
1.96486004e-02 -3.82442564e-01 3.16851169e-01 4.74774808e-01
6.85913026e-01 -6.12191498e-01 -2.01102695e-03 6.88140035e-01
8.20177615e-01 6.50613233e-02 7.03345716e-01 -3.79053295e-01
5.06022871e-01 1.14399159e+00 5.06912827e-01 4.22634393e-01
-2.56718900e-02 1.20572329e+00 5.34112632e-01 -3.09735268e-01
-6.12725675e-01 -3.19089115e-01 7.92456329e-01 7.41920650e-01
8.74274671e-02 -2.52513438e-01 -4.34085429e-01 3.82871151e-01
-1.65915442e+00 -1.18388879e+00 1.07324526e-01 1.81988525e+00
1.00645816e+00 -2.39560515e-01 1.62659481e-01 1.51751503e-01
6.55696332e-01 2.54133403e-01 -4.74665910e-01 -4.33065534e-01
-2.54339337e-01 2.09547982e-01 -2.25456372e-01 7.63429582e-01
-7.74852574e-01 1.09241736e+00 5.68193007e+00 8.57872665e-01
-1.88102067e+00 1.08736798e-01 5.95261633e-01 -4.93579179e-01
-4.61331218e-01 -3.18809330e-01 -1.75586611e-01 5.27253568e-01
9.51229513e-01 2.99090780e-02 3.70446533e-01 7.17886865e-01
9.82473135e-01 2.39148602e-01 -9.11490381e-01 1.00154996e+00
-6.15491532e-02 -1.17153239e+00 -1.17626384e-01 1.06010539e-02
5.05781114e-01 -5.07244647e-01 2.93763041e-01 7.54801109e-02
-5.38604148e-02 -1.02724791e+00 1.25262904e+00 5.81549823e-01
1.25669181e+00 -6.94451809e-01 9.93013680e-02 1.54514909e-01
-1.22962117e+00 2.61465490e-01 3.16944301e-01 2.29183584e-01
7.00943947e-01 2.90941775e-01 -2.93810010e-01 2.14840531e-01
6.34799719e-01 6.06195390e-01 2.15353027e-01 2.48045817e-01
-2.39853114e-01 4.72402781e-01 -3.08067024e-01 3.27243090e-01
-1.45748220e-02 -1.54407755e-01 9.65061247e-01 1.05982780e+00
2.86954612e-01 1.97614878e-01 -1.70800298e-01 1.06710887e+00
-1.19348519e-01 3.27188186e-02 -7.23569930e-01 1.40832346e-02
3.23501617e-01 1.16962039e+00 -1.42941818e-01 -9.85517278e-02
-2.75476247e-01 1.23435926e+00 -2.26384059e-01 5.42766333e-01
-1.15984654e+00 4.75740544e-02 1.43705416e+00 1.85366392e-01
6.21901080e-02 -1.57034278e-01 5.66044822e-02 -9.71023381e-01
-3.53001393e-02 -1.11825025e+00 -3.64567578e-01 -1.12677896e+00
-8.88220370e-01 7.60102987e-01 -2.91474164e-01 -1.21726537e+00
-5.83960712e-01 -2.12697387e-01 -6.87132537e-01 7.26484239e-01
-1.44493437e+00 -1.12882972e+00 -3.28475535e-01 8.14440429e-01
4.96647894e-01 2.44847298e-01 7.22076774e-01 4.20847833e-01
-4.19740975e-01 6.39467359e-01 -4.95695144e-01 -2.00852263e-03
9.73090768e-01 -5.05999446e-01 7.54397735e-02 5.58716774e-01
-9.92178991e-02 2.03248024e-01 1.04595816e+00 -4.51175064e-01
-1.49065351e+00 -8.62611115e-01 5.67048073e-01 -1.33251399e-01
5.23083687e-01 -5.05035818e-01 -8.93377841e-01 2.79040396e-01
3.85656774e-01 2.45571598e-01 4.18287307e-01 -5.63273370e-01
-1.98146477e-01 4.99233864e-02 -1.20819449e+00 9.45218682e-01
1.03603673e+00 -7.12781250e-01 -1.28293827e-01 -1.33306161e-01
8.28914464e-01 -3.92744899e-01 -7.79249489e-01 4.62481767e-01
7.09223568e-01 -1.17130280e+00 8.30254555e-01 -4.92464244e-01
5.03553092e-01 -2.81219661e-01 1.71120390e-02 -1.37770486e+00
1.89805239e-01 -1.19317496e+00 -2.28825226e-01 1.38115597e+00
-1.08558629e-02 -2.63103366e-01 7.64993310e-01 5.77595174e-01
-1.18420027e-01 -5.30353487e-01 -1.14873397e+00 -6.64547384e-01
-1.33801714e-01 -5.21232605e-01 3.86035442e-01 1.08328462e+00
6.91397861e-02 1.83595613e-01 -7.04741418e-01 1.81091964e-01
2.61316657e-01 -2.06866667e-01 9.62500334e-01 -8.73627543e-01
-3.91841717e-02 -5.59341967e-01 -3.44029307e-01 -9.26291347e-01
7.42933154e-01 -3.53211761e-01 2.80429512e-01 -7.29579270e-01
-3.65048945e-01 -1.51727274e-01 4.54216987e-01 3.90130788e-01
1.34604439e-01 2.08748475e-01 3.71013343e-01 9.81479660e-02
5.86455092e-02 1.15673244e+00 1.52731037e+00 6.56656399e-02
-2.70458937e-01 -3.15383047e-01 -4.12445255e-02 8.48965645e-01
4.03506547e-01 -2.61204481e-01 -4.51532990e-01 -1.22246370e-01
-1.33391634e-01 5.36815286e-01 6.07535481e-01 -8.81658256e-01
7.50919580e-02 -1.05755530e-01 5.62533662e-02 -8.31314847e-02
1.02188540e+00 -9.93366420e-01 4.62982148e-01 2.44669169e-01
-3.55254859e-01 -1.49833575e-01 4.09431338e-01 3.07042748e-01
-3.72898281e-01 2.84559548e-01 1.19574702e+00 2.36556560e-01
-2.67625928e-01 4.34491724e-01 -6.03793442e-01 -3.04388627e-02
9.25180376e-01 -2.00158939e-01 2.45386250e-02 -1.02203798e+00
-1.00283110e+00 -1.93956465e-01 7.89064705e-01 4.20628279e-01
4.68892157e-01 -1.46601319e+00 -6.30303264e-01 3.80096346e-01
-2.62024313e-01 -2.65913397e-01 5.14817834e-01 8.47233772e-01
-1.84487492e-01 -4.69335169e-02 -2.82294989e-01 -7.81391740e-01
-1.40646636e+00 5.36024928e-01 5.86110830e-01 1.15573533e-01
-6.51538312e-01 5.90751350e-01 3.64776194e-01 -2.33566344e-01
3.19055378e-01 4.21484001e-02 -3.65502685e-02 2.42579550e-01
3.45893860e-01 3.68716232e-02 -1.40565276e-01 -1.01492321e+00
-1.76870093e-01 7.60762930e-01 6.36562109e-01 -4.96458262e-01
1.05143201e+00 -2.76203603e-01 2.52793789e-01 1.90821543e-01
1.33926916e+00 5.02966046e-01 -1.83593464e+00 1.86768427e-01
-5.43181896e-01 -4.22446877e-01 3.36987823e-02 -4.33338255e-01
-1.49603713e+00 9.57212806e-01 5.68623602e-01 5.77344112e-02
1.42235696e+00 -1.81788042e-01 9.27128792e-01 -3.82739514e-01
5.00346813e-03 -9.62265670e-01 5.42632341e-01 2.63314009e-01
1.07200718e+00 -8.62511218e-01 -5.95300913e-01 -6.77245557e-01
-8.47573519e-01 9.81899083e-01 4.57885921e-01 1.91861302e-01
6.49403214e-01 7.26064622e-01 4.20736343e-01 5.40799578e-04
-9.44836617e-01 -9.67667252e-02 1.31607026e-01 8.51836979e-01
2.65618920e-01 -3.22816581e-01 2.09327787e-01 3.57811451e-01
-1.79172009e-01 -1.83291323e-02 3.65757704e-01 3.74416262e-01
4.62012924e-02 -8.45386922e-01 -3.82558703e-01 -4.98658210e-01
-4.52974111e-01 4.20282222e-02 -2.80476332e-01 7.97574043e-01
-4.61203419e-02 9.99513388e-01 1.50043800e-01 -3.62233907e-01
3.63653004e-01 2.55813420e-01 3.33468616e-01 -1.30011827e-01
-5.55818737e-01 4.84868407e-01 1.64385453e-01 -8.34269404e-01
-4.55004454e-01 -5.39368927e-01 -1.31320310e+00 -6.04043245e-01
-1.61066070e-01 -3.64229269e-02 6.82167590e-01 5.74392200e-01
7.40901113e-01 5.50999343e-01 9.24231112e-01 -1.32550764e+00
-1.33888498e-01 -7.58813322e-01 -4.83537704e-01 6.55537248e-01
5.96138954e-01 -7.21140385e-01 -6.76955760e-01 3.91138285e-01] | [13.18896198272705, -0.4441170394420624] |
dd5f9e3a-a8c3-4589-90d2-9f641c294c2d | linguistic-cues-of-deception-in-a | 2111.03913 | null | https://arxiv.org/abs/2111.03913v3 | https://arxiv.org/pdf/2111.03913v3.pdf | Linguistic Cues of Deception in a Multilingual April Fools' Day Context | In this work we consider the collection of deceptive April Fools' Day(AFD) news articles as a useful addition in existing datasets for deception detection tasks. Such collections have an established ground truth and are relatively easy to construct across languages. As a result, we introduce a corpus that includes diachronic AFD and normal articles from Greek newspapers and news websites. On top of that, we build a rich linguistic feature set, and analyze and compare its deception cues with the only AFD collection currently available, which is in English. Following a current research thread, we also discuss the individualism/collectivism dimension in deception with respect to these two datasets. Lastly, we build classifiers by testing various monolingual and crosslingual settings. The results showcase that AFD datasets can be helpful in deception detection studies, and are in alignment with the observations of other deception detection works. | ['Dimitris Plexousakis', 'Giorgos Flouris', 'Panagiotis Papadakos', 'Katerina Papantoniou'] | 2021-11-06 | null | null | null | null | ['deception-detection'] | ['miscellaneous'] | [-2.89438516e-01 -1.89420164e-01 -3.96188587e-01 -5.81654787e-01
-8.96285772e-01 -1.00381327e+00 1.49923933e+00 3.39952558e-01
-4.69110370e-01 6.71965241e-01 8.20953369e-01 -2.60817289e-01
-6.18366338e-02 -2.64936447e-01 -3.49345893e-01 -3.80022258e-01
4.14732903e-01 2.21435994e-01 -2.83923179e-01 -4.41080749e-01
1.00706983e+00 5.38975954e-01 -1.17085147e+00 4.77900445e-01
1.13760209e+00 8.00584912e-01 -6.01663172e-01 3.62618625e-01
2.91380286e-01 1.10063815e+00 -1.05083382e+00 -1.32364213e+00
4.86661270e-02 -4.26659197e-01 -8.67466211e-01 -2.71909833e-02
9.63480473e-01 -5.44115961e-01 -5.82142889e-01 9.88179922e-01
3.89583319e-01 -1.18640654e-01 8.74057710e-01 -1.23943973e+00
-1.17974758e+00 6.54243946e-01 -3.33254874e-01 6.00937068e-01
8.54097486e-01 -4.96038757e-02 9.49473321e-01 -7.15826690e-01
9.01850402e-01 1.39409399e+00 5.25519133e-01 7.18736529e-01
-8.39509904e-01 -5.98110437e-01 -2.83410043e-01 4.99171078e-01
-9.44377244e-01 -1.11861908e+00 9.34540451e-01 -4.89131302e-01
5.92644095e-01 3.17530096e-01 6.16880476e-01 2.37909842e+00
1.75035134e-01 1.12449419e+00 1.48608434e+00 -6.28513634e-01
-8.21823105e-02 3.09491217e-01 6.27697229e-01 5.64131260e-01
4.72670734e-01 3.05582941e-01 -9.97285068e-01 -4.44706649e-01
-2.13457178e-02 -4.24236387e-01 -4.84879524e-01 3.32750976e-02
-9.13702667e-01 1.13774729e+00 -4.32917960e-02 5.74701250e-01
8.94444883e-02 -2.33903155e-01 7.82280505e-01 7.04003751e-01
7.86234081e-01 5.32504380e-01 6.71806708e-02 -6.56516194e-01
-9.66715276e-01 3.59603405e-01 1.29974842e+00 5.19388795e-01
-6.45316243e-02 1.23423502e-01 8.44423771e-02 1.11876595e+00
1.76927164e-01 4.28773522e-01 7.76336849e-01 -5.25547147e-01
6.23473942e-01 2.79747307e-01 1.89819768e-01 -1.47495711e+00
-1.01321355e-01 -1.81869879e-01 -3.06254417e-01 -1.44335136e-01
7.18362153e-01 1.40353322e-01 -3.77758980e-01 1.22918653e+00
-2.11565234e-02 -4.88287449e-01 2.74444614e-02 1.01824987e+00
6.64249539e-01 4.05381441e-01 -2.92400718e-01 -1.37306333e-01
1.33258653e+00 -6.62944615e-01 -5.61712325e-01 -2.91544169e-01
7.44753182e-01 -7.59944797e-01 9.75966036e-01 8.46536875e-01
-7.97594190e-01 2.29232624e-01 -1.27504408e+00 -4.02113885e-01
-6.20039821e-01 6.56429306e-02 8.12066615e-01 1.14796102e+00
-5.23786843e-01 3.99493754e-01 -4.81492639e-01 -5.02140224e-01
4.33854848e-01 -5.95003486e-01 -4.29385632e-01 -3.54044467e-01
-1.17577648e+00 1.62006509e+00 3.14019710e-01 1.96387589e-01
-1.11833441e+00 -3.99558485e-01 -9.35968757e-01 -5.37918329e-01
2.85995960e-01 -1.49800047e-01 9.87061858e-01 -1.17862785e+00
-1.38681698e+00 1.55642927e+00 1.70806810e-01 -4.27935183e-01
6.79050803e-01 -3.83832276e-01 -1.21359968e+00 2.14294374e-01
2.10731570e-02 -2.85811067e-01 1.17871451e+00 -1.19730270e+00
-2.68982232e-01 -7.52366006e-01 7.51037076e-02 -1.31458938e-01
-2.52719611e-01 6.69241130e-01 6.18664861e-01 -7.81410813e-01
-2.33792633e-01 -4.90388632e-01 9.44785535e-01 -3.38396490e-01
-5.33131838e-01 -2.00581148e-01 7.51950860e-01 -1.18852437e+00
1.35103357e+00 -2.24720764e+00 -1.22335590e-02 7.50515088e-02
5.36352813e-01 2.49928117e-01 1.76358640e-01 8.51020336e-01
2.34908402e-01 3.37128669e-01 -1.10502243e-01 -3.20488632e-01
5.53940117e-01 2.56617039e-01 -3.35271001e-01 1.04961157e+00
-1.61573425e-01 9.76532221e-01 -8.26003075e-01 -2.82312006e-01
-1.14218771e-01 9.15120542e-02 -1.76110744e-01 -3.29369038e-01
3.12857896e-01 2.26013847e-02 -2.15933084e-01 9.52595115e-01
4.66191739e-01 4.28930014e-01 -7.43409060e-03 1.54015690e-01
-3.29236537e-01 9.33740139e-01 -1.48681715e-01 1.28480875e+00
-3.19721699e-01 1.14299464e+00 2.89877176e-01 -8.63457084e-01
8.61753345e-01 -2.52188798e-02 -4.50890988e-01 -6.62265658e-01
4.05768007e-01 6.95454657e-01 2.97236115e-01 -5.71891725e-01
7.64237702e-01 -3.18759769e-01 -4.68010962e-01 5.73569357e-01
6.52062297e-02 -3.33530784e-01 2.09797248e-01 3.96915287e-01
7.73705721e-01 -2.86278397e-01 5.19944072e-01 -6.13861918e-01
5.99377453e-01 -1.13682437e-03 3.39165300e-01 7.93129027e-01
-5.13463557e-01 4.16025698e-01 8.89497817e-01 -4.42771375e-01
-9.18720007e-01 -1.02550113e+00 -5.31399667e-01 7.11987197e-01
-2.66779602e-01 -3.67816210e-01 -5.58019519e-01 -1.03596532e+00
1.05047531e-01 1.37284088e+00 -7.23577142e-01 -3.58883768e-01
-3.94702226e-01 -6.60714149e-01 1.25167060e+00 -1.52336016e-01
4.90640074e-01 -6.19528055e-01 -3.82569104e-01 -2.97096401e-01
-2.05694214e-01 -9.13630188e-01 -4.23680604e-01 -2.19043419e-01
-1.76261052e-01 -1.23793137e+00 -2.14769006e-01 -5.60206711e-01
-1.42882473e-03 3.76031995e-02 1.16529918e+00 1.82823554e-01
1.59647129e-02 2.95162737e-01 -8.07925165e-01 -5.27850211e-01
-9.49482739e-01 -1.03186294e-01 1.88207030e-01 1.15391547e-02
6.40394807e-01 -4.19167906e-01 -3.77067104e-02 -5.18699959e-02
-8.88905764e-01 -5.75301826e-01 2.26236865e-01 9.59295452e-01
-5.76723099e-01 -5.92723906e-01 6.33245945e-01 -7.82379329e-01
1.18328238e+00 -8.27779770e-01 -3.65359634e-01 9.51977000e-02
-4.31903869e-01 -3.28031003e-01 7.17065513e-01 -9.29577127e-02
-1.04533267e+00 -1.07160795e+00 -1.49586186e-01 2.71241635e-01
-3.18192631e-01 5.32958508e-01 2.58867089e-02 -1.45531967e-01
9.45789754e-01 3.25857759e-01 1.87426969e-01 -4.60416615e-01
9.43178385e-02 1.20738935e+00 4.04963434e-01 -5.55653393e-01
5.42522252e-01 2.53031015e-01 -5.84134519e-01 -1.14480066e+00
-1.10669601e+00 -2.76239842e-01 -6.17595434e-01 -1.34060681e-01
2.53561050e-01 -5.39999843e-01 -6.68321550e-01 9.47348177e-01
-1.11275160e+00 6.47339672e-02 1.88743547e-01 5.54228783e-01
-3.26365858e-01 7.42334366e-01 -1.01924789e+00 -8.91065478e-01
-1.08789146e-01 -6.09037161e-01 8.01463664e-01 -3.91943216e-01
-4.05218720e-01 -1.29178357e+00 8.27131048e-02 7.40897775e-01
1.10591203e-01 5.49049556e-01 9.29900944e-01 -1.34004402e+00
2.81220138e-01 -3.91586006e-01 9.27654579e-02 6.18149281e-01
-6.10211715e-02 7.29765743e-02 -6.96268022e-01 -1.22142009e-01
5.25537312e-01 -9.69339192e-01 8.90943706e-01 -2.36476228e-01
3.67666841e-01 -7.86275625e-01 3.79381068e-02 1.47268102e-01
1.00816536e+00 -1.17018513e-01 6.30860031e-01 4.33719575e-01
3.44962448e-01 6.27295136e-01 2.73548931e-01 5.47243536e-01
5.81059039e-01 6.14331067e-01 1.35554969e-01 7.60092199e-01
-5.47267534e-02 -4.42807764e-01 8.48080873e-01 1.00664270e+00
2.01072305e-01 -5.92554927e-01 -1.10946608e+00 6.81811869e-01
-1.39025712e+00 -1.33143091e+00 -4.45255131e-01 1.94898307e+00
6.83538079e-01 -9.02310610e-02 4.23767209e-01 2.05332458e-01
4.01469350e-01 5.81489980e-01 1.00034907e-01 -1.03919303e+00
-5.10699749e-01 -1.03086464e-01 -1.15883037e-01 6.09860599e-01
-7.46242404e-01 9.01822090e-01 6.69857788e+00 8.60843658e-01
-8.84738505e-01 1.75819889e-01 3.48596096e-01 -4.22578931e-01
-3.22360128e-01 -3.11217904e-01 -5.31850159e-01 8.08528304e-01
1.07831955e+00 -1.77203864e-01 5.00027597e-01 5.47805727e-01
1.54076546e-01 -4.31579262e-01 -1.33635247e+00 1.03883684e+00
1.15418911e+00 -8.31645966e-01 -9.01992097e-02 5.61895408e-02
3.46775323e-01 -5.54659367e-02 1.25353351e-01 1.52643964e-01
4.14118916e-01 -1.09029150e+00 1.16404307e+00 3.93997073e-01
3.19602519e-01 -9.09534156e-01 9.74530399e-01 6.71002209e-01
3.15597296e-01 -1.50878057e-01 -3.39742899e-01 -2.77103692e-01
2.32972696e-01 5.91111064e-01 -5.90977967e-01 5.93283236e-01
4.35757309e-01 9.06087160e-01 -7.76868224e-01 5.84502280e-01
-6.27984524e-01 7.68452942e-01 2.31475118e-04 -4.38950241e-01
5.02615452e-01 -3.96687806e-01 9.32711244e-01 1.27002549e+00
2.51853913e-01 -9.54643786e-02 -2.95070052e-01 8.50902677e-01
-1.75397679e-01 1.12320893e-02 -7.74841785e-01 -4.49086636e-01
4.93865639e-01 1.13930047e+00 -2.61331592e-02 -1.54051036e-01
-3.84713560e-01 1.29586983e+00 7.09504962e-01 3.75016630e-02
-6.69451237e-01 -3.56185347e-01 5.50217569e-01 6.85855141e-03
-3.05738866e-01 -3.66343707e-01 -5.60254082e-02 -1.72733569e+00
2.03317791e-01 -1.26993012e+00 3.63173991e-01 -6.97051764e-01
-2.04065347e+00 2.69032627e-01 1.86129175e-02 -7.33544350e-01
-4.25701946e-01 -7.60227799e-01 -1.82774097e-01 5.40466905e-01
-1.29484403e+00 -1.47001791e+00 2.44574957e-02 5.61738610e-01
3.37950319e-01 -3.28186154e-01 6.15207374e-01 1.19485468e-01
-6.26039445e-01 6.98167741e-01 2.44415730e-01 3.22649539e-01
8.18383515e-01 -9.64674413e-01 3.26778591e-01 9.54589844e-01
4.04873133e-01 7.78109848e-01 5.83896637e-01 -6.63620353e-01
-1.20806468e+00 -2.13752598e-01 1.32214046e+00 -1.05130064e+00
1.35884130e+00 -7.73529112e-01 -7.76102781e-01 8.44085872e-01
4.71731186e-01 -4.20630962e-01 7.86783159e-01 3.70354563e-01
-9.84432101e-01 3.02749515e-01 -1.48142183e+00 6.08974099e-01
1.29100287e+00 -7.74312854e-01 -1.22432792e+00 3.84743989e-01
-1.50223479e-01 -2.25984305e-01 -7.81976700e-01 -4.59210128e-01
9.10024822e-01 -1.31077862e+00 5.22458613e-01 -8.38854313e-01
9.73680139e-01 3.73355925e-01 -2.09532693e-01 -1.81845462e+00
-2.48571634e-01 -6.34885013e-01 -5.80568612e-02 1.49901557e+00
3.44636112e-01 -9.58471060e-01 3.06600034e-01 4.91770685e-01
-4.96986181e-01 -2.19837576e-01 -1.27584553e+00 -1.03094840e+00
6.22297525e-01 -4.09419537e-01 1.99903473e-01 1.51074207e+00
6.33185148e-01 2.95878083e-01 -6.43275499e-01 -2.71055579e-01
4.87388909e-01 3.53783481e-02 4.88712579e-01 -1.10152268e+00
-2.27339771e-02 -4.52838361e-01 -6.22612357e-01 -8.67126763e-01
7.33731449e-01 -1.17471099e+00 -2.62971848e-01 -9.75670338e-01
3.25246364e-01 -6.17168993e-02 5.10503411e-01 1.66013896e-01
2.29869694e-01 3.18273216e-01 9.06896740e-02 1.55288845e-01
-3.29832435e-01 5.42612910e-01 9.33174491e-01 -2.78515997e-03
2.73967236e-01 -3.65993649e-01 -9.21479762e-01 8.43828797e-01
8.98917675e-01 -4.47458327e-01 2.00408250e-02 -5.61139941e-01
1.45706594e-01 -1.95566148e-01 7.56751001e-01 -3.39013606e-01
-6.97133914e-02 -9.63082165e-02 2.81326145e-01 -1.38717592e-01
4.15570349e-01 -5.02934575e-01 -3.83718491e-01 1.59590468e-01
-1.56638041e-01 4.12674807e-02 -1.16018221e-01 3.76518875e-01
-2.83700734e-01 -7.41071403e-01 6.81918383e-01 -2.49639273e-01
-4.22354341e-01 -3.04926991e-01 -7.25211799e-01 5.52615821e-01
8.39356601e-01 -3.38075817e-01 -9.74095166e-01 -6.39008820e-01
-1.90486118e-01 -1.85801581e-01 7.97130466e-01 6.85472310e-01
5.49474120e-01 -1.31932485e+00 -1.19240773e+00 4.41003367e-02
3.48693520e-01 -1.25112021e+00 -3.47806141e-02 1.03489077e+00
-5.47271132e-01 4.65025991e-01 -4.19591308e-01 1.77089155e-01
-1.06305325e+00 5.32891691e-01 1.60702348e-01 1.85197473e-01
-2.45075062e-01 4.90260035e-01 -4.68619406e-01 -4.25979704e-01
-3.90618950e-01 3.16913307e-01 -9.74994805e-03 3.56688470e-01
5.99372804e-01 8.66609454e-01 1.92219213e-01 -1.19772637e+00
-4.62074816e-01 -3.38976175e-01 -4.03118730e-01 -3.89219314e-01
1.29175723e+00 -2.60950118e-01 -4.43702459e-01 8.22891831e-01
1.34358609e+00 9.62772787e-01 -3.44734967e-01 3.06324214e-01
3.72852296e-01 -1.00231874e+00 -2.21024826e-01 -1.23134637e+00
-5.13512194e-01 4.79706556e-01 -4.01653677e-01 6.28360331e-01
4.59933758e-01 5.58715463e-02 5.46573877e-01 7.18271658e-02
4.00852889e-01 -1.22060311e+00 -1.60024557e-02 6.01521194e-01
1.54532659e+00 -1.15144312e+00 1.81720302e-01 -2.83745348e-01
-7.64801443e-01 1.01142275e+00 2.67541796e-01 -2.80114859e-01
8.18540603e-02 -1.41420677e-01 -2.16149874e-02 -4.99934316e-01
-4.64598626e-01 3.09625357e-01 4.06995684e-01 8.28206301e-01
4.55867827e-01 -5.56275062e-02 -1.15332520e+00 8.81498754e-01
-9.92516160e-01 -4.97648269e-01 9.65642750e-01 6.07355833e-01
-4.79241014e-02 -1.11270070e+00 -3.40095967e-01 6.96604490e-01
-9.53711689e-01 -3.69763821e-01 -1.40704048e+00 1.29024899e+00
-2.52538323e-01 1.41015136e+00 -2.54770219e-01 -2.36267656e-01
7.59547874e-02 2.40382597e-01 7.94485569e-01 -3.61996293e-01
-1.00090146e+00 -6.42559469e-01 1.03545380e+00 -2.64735758e-01
-1.13156892e-01 -1.03624427e+00 -1.26459524e-01 -9.46073472e-01
-3.57893437e-01 1.08623415e-01 4.76274818e-01 9.91277575e-01
1.19687982e-01 -3.72958332e-01 3.87187779e-01 -4.47187543e-01
-8.01178813e-01 -1.15838289e+00 -9.31840718e-01 6.10085070e-01
4.38239366e-01 -4.84724253e-01 -9.22160923e-01 -3.29434693e-01] | [8.179289817810059, 10.38051700592041] |
e582d2e8-fe50-4589-9e75-8234ae57b4b3 | a-two-stage-framework-with-self-supervised | 2304.09820 | null | https://arxiv.org/abs/2304.09820v1 | https://arxiv.org/pdf/2304.09820v1.pdf | A Two-Stage Framework with Self-Supervised Distillation For Cross-Domain Text Classification | Cross-domain text classification aims to adapt models to a target domain that lacks labeled data. It leverages or reuses rich labeled data from the different but related source domain(s) and unlabeled data from the target domain. To this end, previous work focuses on either extracting domain-invariant features or task-agnostic features, ignoring domain-aware features that may be present in the target domain and could be useful for the downstream task. In this paper, we propose a two-stage framework for cross-domain text classification. In the first stage, we finetune the model with mask language modeling (MLM) and labeled data from the source domain. In the second stage, we further fine-tune the model with self-supervised distillation (SSD) and unlabeled data from the target domain. We evaluate its performance on a public cross-domain text classification benchmark and the experiment results show that our method achieves new state-of-the-art results for both single-source domain adaptations (94.17% $\uparrow$1.03%) and multi-source domain adaptations (95.09% $\uparrow$1.34%). | ['Wanxiang Che', 'Xiao Xu', 'Libo Qin', 'Bohan Li', 'Yunlong Feng'] | 2023-04-18 | null | null | null | null | ['cross-domain-text-classification'] | ['natural-language-processing'] | [ 1.69708654e-01 -8.16413090e-02 -4.10895258e-01 -7.12801635e-01
-9.38807607e-01 -7.98641741e-01 7.23465681e-01 3.75623375e-01
-7.02773392e-01 9.22820628e-01 1.56036988e-01 -1.26651630e-01
2.18678668e-01 -5.55958748e-01 -3.56388956e-01 -3.36224020e-01
4.50079054e-01 9.05176997e-01 4.09335345e-01 -5.01057804e-01
9.54387486e-02 5.74401952e-02 -1.02754664e+00 6.86017990e-01
1.02467465e+00 8.63057017e-01 8.45441967e-02 4.06005919e-01
-5.93911588e-01 5.30156672e-01 -7.11891294e-01 -5.12391210e-01
3.34727675e-01 -4.37692314e-01 -1.05573332e+00 -6.62951097e-02
4.00334716e-01 -7.81505182e-03 -2.35722736e-01 8.85097086e-01
4.08124596e-01 1.07807621e-01 9.06285644e-01 -1.11304319e+00
-8.08829129e-01 7.98717678e-01 -6.63620055e-01 1.41924426e-01
7.58658424e-02 3.66675155e-03 7.82746017e-01 -1.01565850e+00
5.38090646e-01 1.04151237e+00 6.30730033e-01 7.12123454e-01
-1.41836965e+00 -1.07367432e+00 5.86109817e-01 -7.08807856e-02
-1.31655395e+00 -5.54704309e-01 8.37661803e-01 -5.37618220e-01
9.47274625e-01 -1.69188559e-01 -1.76164538e-01 1.45026827e+00
-1.17930159e-01 8.19999099e-01 1.13267791e+00 -7.23654389e-01
1.84327826e-01 5.52607000e-01 4.20463502e-01 3.08571875e-01
1.15180001e-01 -1.44288093e-01 -4.81192410e-01 -2.31674582e-01
7.20893070e-02 -2.08202183e-01 -5.31840660e-02 -3.43807340e-01
-1.15341747e+00 8.54423702e-01 1.21674664e-01 5.60048699e-01
-2.89706867e-02 -6.38964355e-01 6.86266899e-01 5.85447252e-01
8.94250333e-01 5.78227937e-01 -1.16023028e+00 -6.70782849e-02
-9.98253882e-01 1.77257985e-01 8.84456813e-01 1.18428671e+00
8.01894367e-01 -1.68362811e-01 -1.86568826e-01 1.39699578e+00
1.28479257e-01 3.36102664e-01 1.05935645e+00 -3.40446770e-01
1.01570189e+00 9.63867188e-01 2.25812718e-02 -2.94692427e-01
-4.44441408e-01 -3.66435975e-01 -8.04377735e-01 -5.04015945e-02
5.34244537e-01 -2.17484847e-01 -1.00704741e+00 1.76029086e+00
3.06491554e-01 -2.00819105e-01 4.35019314e-01 4.14785057e-01
9.57957029e-01 5.12011349e-01 4.11871821e-01 8.70838314e-02
1.29636133e+00 -1.33419251e+00 -3.11193973e-01 -7.34128296e-01
9.49426889e-01 -7.97437608e-01 1.40905964e+00 1.20174915e-01
-6.50397897e-01 -7.09094107e-01 -9.86384630e-01 -1.15420677e-01
-7.55652606e-01 3.08696806e-01 -1.57933533e-01 5.46249926e-01
-5.64019918e-01 3.12680751e-01 -3.65419060e-01 -5.64023376e-01
3.40226531e-01 1.73135325e-01 -4.65633571e-01 -3.92510295e-01
-1.40006948e+00 8.51389050e-01 7.94052243e-01 -8.75042319e-01
-5.50082028e-01 -9.76692736e-01 -7.28179693e-01 -1.17931655e-02
2.31054246e-01 -4.54118550e-01 1.42781973e+00 -1.20210099e+00
-1.41691768e+00 1.27462578e+00 -1.64378390e-01 -3.53072673e-01
6.55609667e-01 -1.56925544e-01 -7.97784150e-01 -1.78466037e-01
4.83340889e-01 5.67562461e-01 8.97895694e-01 -1.19103062e+00
-1.05713820e+00 -5.76393783e-01 -1.29598379e-01 2.09514529e-01
-7.22147048e-01 -1.30560026e-01 -3.54737610e-01 -1.04672384e+00
-1.89716920e-01 -9.35632467e-01 3.76929268e-02 -3.35569263e-01
-3.34473044e-01 -2.85684496e-01 9.28818882e-01 -6.57924831e-01
1.26976418e+00 -2.10217166e+00 -4.13105749e-02 -6.52575716e-02
9.31870788e-02 5.07492602e-01 -3.24085832e-01 3.95090282e-01
-1.97739884e-01 1.08456895e-01 -4.66787517e-01 -5.73647916e-01
5.01647629e-02 9.30572748e-02 -2.78866649e-01 1.34432748e-01
1.45379007e-01 6.61318719e-01 -6.71517432e-01 -4.66779679e-01
-9.22412872e-02 -7.36962631e-02 -4.84257370e-01 2.23894969e-01
-3.48962396e-01 5.24370193e-01 -4.75980163e-01 4.70009387e-01
7.91748166e-01 -2.11257935e-01 3.67741972e-01 -1.30031884e-01
7.65246227e-02 5.42904854e-01 -1.01027369e+00 1.95598400e+00
-6.20151758e-01 3.48128945e-01 -8.47315639e-02 -1.12535346e+00
1.06290388e+00 1.84023634e-01 5.20288587e-01 -7.84293115e-01
3.12540270e-02 2.45203406e-01 -2.51140147e-01 -1.31285116e-01
4.49775875e-01 -2.33355656e-01 -5.58960855e-01 7.11496294e-01
3.94893318e-01 3.91458608e-02 1.40330806e-01 6.08180277e-02
1.03144956e+00 1.22116022e-01 8.00548911e-01 -3.65985423e-01
7.71279633e-01 3.75039160e-01 4.21818554e-01 6.82897806e-01
-2.31971294e-01 3.94291610e-01 1.70451969e-01 -3.40492100e-01
-9.91090000e-01 -8.52501452e-01 -1.92934886e-01 1.90864336e+00
-3.23658437e-02 -2.45609581e-01 -5.66843092e-01 -1.51286423e+00
2.70200551e-01 8.38789701e-01 -7.19238579e-01 -4.26207066e-01
-4.89458084e-01 -7.34714329e-01 7.93169677e-01 5.55318177e-01
6.54655695e-01 -6.62605941e-01 2.07445100e-02 2.69925267e-01
-3.32734287e-01 -1.27524400e+00 -6.97091162e-01 5.63493133e-01
-7.19822049e-01 -9.16567862e-01 -8.03700209e-01 -1.01404583e+00
6.26391947e-01 2.52953887e-01 1.35803783e+00 -5.43319285e-01
3.05486172e-01 1.89484790e-01 -7.05005646e-01 -2.28595987e-01
-8.12626421e-01 7.53576279e-01 7.60152638e-02 -8.71353503e-03
8.79869103e-01 -3.49046826e-01 -9.20177251e-02 6.00937247e-01
-8.60730588e-01 -5.71370162e-02 4.92650062e-01 1.12529540e+00
4.72056448e-01 3.73722538e-02 8.49077284e-01 -1.42216241e+00
6.82548225e-01 -7.96044052e-01 -3.47959220e-01 3.20945859e-01
-7.72287130e-01 2.28307754e-01 1.13907027e+00 -7.88426995e-01
-1.40020192e+00 1.67922601e-01 -1.11708991e-01 -1.02665789e-01
-4.62383628e-01 4.06623453e-01 -4.68755931e-01 3.23654115e-01
1.26581597e+00 2.01682702e-01 -4.49849486e-01 -8.49963963e-01
4.59272265e-01 1.08183372e+00 4.04829383e-01 -8.19097579e-01
8.14517379e-01 3.30959588e-01 -6.39071703e-01 -4.14000392e-01
-1.21300757e+00 -6.91853225e-01 -1.13416827e+00 4.93802726e-01
5.37547231e-01 -1.06275320e+00 8.64414647e-02 3.85197759e-01
-8.89694393e-01 -5.17441213e-01 -4.01514113e-01 2.27569610e-01
-2.88455844e-01 3.00710261e-01 -2.54788250e-01 -2.91517824e-01
-4.57705200e-01 -8.19615960e-01 1.02288270e+00 8.92340317e-02
-4.87259865e-01 -1.18819785e+00 1.91301912e-01 3.28701973e-01
3.65390778e-01 -2.45816320e-01 1.09679258e+00 -1.63969433e+00
3.32576483e-01 -1.88474149e-01 -2.56319463e-01 5.34752071e-01
4.51695174e-01 -6.01335466e-01 -1.18086004e+00 -3.84450138e-01
-7.34349415e-02 -4.23060805e-01 8.46834838e-01 -9.73255187e-02
9.95649159e-01 -2.90320039e-01 -5.17867744e-01 5.49782038e-01
1.09053063e+00 2.15798810e-01 5.41068204e-02 6.47081375e-01
5.68845451e-01 6.24092996e-01 7.66245425e-01 6.02701187e-01
5.25102317e-01 7.58001566e-01 -1.92672119e-01 -1.33092552e-01
-1.86788425e-01 -5.15868068e-01 2.87082165e-01 4.84564483e-01
7.54849017e-01 -3.11397821e-01 -1.15046334e+00 6.72973394e-01
-1.76279402e+00 -5.12661040e-01 1.16345197e-01 1.99843824e+00
1.26415241e+00 3.25770408e-01 2.93029785e-01 -3.99829075e-02
8.58630478e-01 -1.32256016e-01 -9.99777377e-01 -1.97066531e-01
-2.30703771e-01 5.04555702e-01 4.52798218e-01 2.67835319e-01
-1.47114229e+00 1.19183576e+00 5.55756950e+00 8.73753965e-01
-1.01448643e+00 2.96954721e-01 4.57262784e-01 -3.85568626e-02
5.49305119e-02 -2.07158148e-01 -1.15275085e+00 6.85590386e-01
9.49675977e-01 -4.75242227e-01 7.11777881e-02 1.07092702e+00
-2.69587338e-01 2.89934158e-01 -1.16034448e+00 7.98845530e-01
1.01913646e-01 -9.85524356e-01 1.01829387e-01 -2.34492734e-01
7.63268888e-01 1.79161042e-01 -8.69650673e-03 8.30310702e-01
8.01434934e-01 -5.40098727e-01 5.63693821e-01 -1.07231520e-01
1.15359271e+00 -6.26956582e-01 7.58830369e-01 6.32208705e-01
-1.17477727e+00 -1.13082685e-01 -1.67590320e-01 2.52016127e-01
-2.25241154e-01 4.90566671e-01 -1.05684030e+00 5.81739247e-01
8.17599714e-01 9.84567106e-01 -7.45016217e-01 3.55630755e-01
-6.64639100e-02 5.56339025e-01 -1.42419398e-01 2.73361892e-01
2.81028040e-02 1.55522451e-01 4.02226269e-01 1.61527562e+00
2.37579308e-02 -1.63256586e-01 4.15535957e-01 5.60306430e-01
-4.48558122e-01 3.36911976e-01 -4.87264693e-01 1.12239771e-01
4.61401999e-01 1.02566683e+00 -2.46251330e-01 -6.32946551e-01
-6.13524377e-01 1.27370858e+00 4.86395806e-01 3.34525526e-01
-6.81508243e-01 -4.62907463e-01 8.18863690e-01 2.82139685e-02
2.12496534e-01 7.76012912e-02 -4.51544821e-01 -1.61875701e+00
4.57079969e-02 -1.01916373e+00 9.10933554e-01 -2.97688931e-01
-2.00423169e+00 7.40065992e-01 -6.99335150e-03 -1.35597742e+00
-3.08108747e-01 -5.42503893e-01 -1.94829255e-01 1.22114420e+00
-1.82092547e+00 -1.43173528e+00 -3.32319885e-01 8.44608307e-01
9.44465101e-01 -6.73794508e-01 1.05318093e+00 4.06076640e-01
-4.05399531e-01 1.11303258e+00 6.13619745e-01 4.28005487e-01
1.50618315e+00 -1.11898291e+00 7.62411535e-01 6.75857365e-01
-1.09661467e-01 3.63463730e-01 2.34819695e-01 -6.17398024e-01
-6.79313004e-01 -1.51048505e+00 9.56684887e-01 -6.16393864e-01
7.16751456e-01 -5.20140529e-01 -1.28227568e+00 8.27300966e-01
9.75931704e-04 1.42024821e-02 1.06030619e+00 1.17667563e-01
-8.11309516e-01 -1.00368693e-01 -1.45501375e+00 4.26103055e-01
1.04750466e+00 -4.45964754e-01 -9.24233735e-01 1.25095546e-01
6.63244188e-01 -3.89486492e-01 -8.65909040e-01 2.23739117e-01
4.87874508e-01 -5.25630236e-01 8.70097399e-01 -1.09420276e+00
3.11997235e-01 -2.93777101e-02 -3.51069450e-01 -1.60475993e+00
-4.12269622e-01 -1.82324871e-01 5.00981621e-02 1.66380250e+00
6.72957003e-01 -8.23589861e-01 6.94424570e-01 5.09847522e-01
-2.29007527e-01 -1.00000426e-01 -9.15054917e-01 -1.01523149e+00
7.90472925e-01 -2.80314058e-01 5.77536345e-01 1.48049045e+00
9.36000720e-02 7.37449586e-01 -7.99618736e-02 -4.43374328e-02
3.58667791e-01 2.92120963e-01 8.58829260e-01 -1.47759044e+00
-1.52959198e-01 -4.47704881e-01 1.59845412e-01 -1.28827214e+00
6.43985868e-01 -1.20470452e+00 -1.16222598e-01 -1.24310327e+00
6.05320483e-02 -7.17340529e-01 -4.36243564e-01 7.57777274e-01
-3.66874516e-01 -6.57336563e-02 4.54550125e-02 3.54534417e-01
-5.51113009e-01 5.78805387e-01 7.76283622e-01 -4.06984001e-01
-4.04169917e-01 4.29141112e-02 -1.10740900e+00 6.42922759e-01
1.11599958e+00 -8.41515005e-01 -3.47244769e-01 -5.14224291e-01
-4.70863074e-01 -4.09317970e-01 -1.02476075e-01 -8.56389761e-01
1.23555705e-01 -2.96814978e-01 4.47994322e-01 -3.32927614e-01
1.27133071e-01 -8.88978779e-01 -4.36880797e-01 2.19271243e-01
-6.82472169e-01 -9.33059305e-02 4.99146968e-01 5.70342243e-01
-2.50483841e-01 -3.74289066e-01 1.28005731e+00 -2.03468576e-02
-8.31382513e-01 6.40168488e-02 -2.01371849e-01 6.50791705e-01
1.01505804e+00 1.31529067e-02 -4.88638103e-01 2.24864166e-02
-6.65225208e-01 2.18766302e-01 6.19213820e-01 8.24093461e-01
1.17324032e-01 -1.25099945e+00 -9.08476830e-01 2.17992052e-01
8.17096651e-01 -1.17291607e-01 7.51672685e-02 1.59219459e-01
1.32141531e-01 4.48962957e-01 -1.90675125e-01 -5.16818285e-01
-1.17672801e+00 6.21326447e-01 3.21140230e-01 -6.49686575e-01
-3.37002844e-01 7.76528001e-01 3.61233681e-01 -1.01054037e+00
-2.93578468e-02 -1.60998598e-01 -1.50380746e-01 8.66082460e-02
3.83469433e-01 2.37570211e-01 3.81927997e-01 -5.66813707e-01
-4.05736506e-01 3.82104874e-01 -5.21966040e-01 -7.10178986e-02
1.12523580e+00 -3.63403261e-01 2.76861787e-01 4.24691290e-01
1.28820908e+00 5.31695262e-02 -1.11245704e+00 -1.01684225e+00
3.07939529e-01 -3.16626042e-01 -2.63107926e-01 -1.38288343e+00
-6.64914489e-01 7.45073438e-01 5.36578596e-01 -5.77396713e-02
1.23637068e+00 2.02596799e-01 7.47403324e-01 3.03711504e-01
1.37149498e-01 -1.36064720e+00 1.53159142e-01 8.58312845e-01
6.37255073e-01 -1.57198453e+00 -1.40724853e-01 -1.59424439e-01
-9.96748328e-01 9.08896148e-01 9.61116552e-01 2.52591103e-01
7.92986631e-01 1.72006279e-01 2.11544365e-01 2.85430700e-01
-7.66607225e-01 -2.57270724e-01 3.55881512e-01 8.28750551e-01
5.84546208e-01 4.45708297e-02 -7.50805018e-03 1.18551230e+00
-8.12644139e-02 -1.35195807e-01 4.74022143e-02 1.08888090e+00
-3.30375791e-01 -1.61821771e+00 -2.80023456e-01 4.98489648e-01
-3.39519233e-01 -2.52092183e-01 -7.44722486e-01 8.43129098e-01
3.11499923e-01 9.62142408e-01 -1.49788380e-01 -3.31325710e-01
7.40597844e-01 6.70278668e-01 2.55901441e-02 -1.00745189e+00
-8.37988257e-01 -2.21051514e-01 1.26151457e-01 5.37296943e-03
-1.53293535e-01 -5.18737733e-01 -1.10250044e+00 -1.18706211e-01
5.25613315e-02 2.07793951e-01 5.09717405e-01 8.60861659e-01
7.43761599e-01 3.35264325e-01 5.11845052e-01 -3.61212522e-01
-5.59014022e-01 -1.18473887e+00 -3.81225109e-01 6.91542625e-01
3.13776851e-01 -7.70760715e-01 -1.70570463e-01 4.63879645e-01] | [10.67474365234375, 8.005663871765137] |
d0291c94-2fe6-4ed6-a486-ec840ecca956 | geoconv-geodesic-guided-convolution-for | 2003.03055 | null | https://arxiv.org/abs/2003.03055v1 | https://arxiv.org/pdf/2003.03055v1.pdf | GeoConv: Geodesic Guided Convolution for Facial Action Unit Recognition | Automatic facial action unit (AU) recognition has attracted great attention but still remains a challenging task, as subtle changes of local facial muscles are difficult to thoroughly capture. Most existing AU recognition approaches leverage geometry information in a straightforward 2D or 3D manner, which either ignore 3D manifold information or suffer from high computational costs. In this paper, we propose a novel geodesic guided convolution (GeoConv) for AU recognition by embedding 3D manifold information into 2D convolutions. Specifically, the kernel of GeoConv is weighted by our introduced geodesic weights, which are negatively correlated to geodesic distances on a coarsely reconstructed 3D face model. Moreover, based on GeoConv, we further develop an end-to-end trainable framework named GeoCNN for AU recognition. Extensive experiments on BP4D and DISFA benchmarks show that our approach significantly outperforms the state-of-the-art AU recognition methods. | ['Tat-Jen Cham', 'Yuedong Chen', 'Guoxian Song', 'Jianfei Cai', 'Zhiwen Shao', 'Jianming Zheng'] | 2020-03-06 | null | null | null | null | ['facial-action-unit-detection'] | ['computer-vision'] | [-2.80900896e-02 -1.45710111e-01 -1.47170931e-01 -4.33281362e-01
-5.23959100e-01 -4.01667207e-01 5.94020903e-01 -5.69020987e-01
-2.78713077e-01 7.32838139e-02 4.45606977e-01 3.55205685e-02
1.17060877e-01 -6.32841766e-01 -5.82859576e-01 -8.03780973e-01
1.43030420e-01 1.95377544e-02 -3.49018544e-01 -1.18762396e-01
9.05451179e-02 5.70902944e-01 -1.39709854e+00 -4.96206246e-02
4.77546602e-01 1.35567760e+00 -5.44120848e-01 2.31406510e-01
-1.27145126e-02 2.43799627e-01 5.19628171e-03 -5.53355396e-01
4.71664995e-01 -5.15907347e-01 -6.71077371e-01 2.74553835e-01
6.82488799e-01 -4.19147402e-01 -5.81534863e-01 1.03440607e+00
6.64253235e-01 9.39549655e-02 7.34648585e-01 -1.01402462e+00
-7.87397146e-01 1.50825366e-01 -8.16245317e-01 -4.17914018e-02
2.77826905e-01 1.07321911e-01 9.74539340e-01 -1.23932624e+00
3.35037053e-01 1.40013957e+00 5.74735224e-01 9.59962904e-01
-1.21687198e+00 -8.21892023e-01 1.63723335e-01 1.21809663e-02
-1.62795162e+00 -6.47865295e-01 9.94496346e-01 -3.24425817e-01
9.71686602e-01 -8.69762227e-02 4.59540814e-01 1.19665444e+00
-1.59296349e-01 8.00051212e-01 7.35542536e-01 -5.74541315e-02
-4.97126319e-02 -5.36441088e-01 -3.04301262e-01 1.11716187e+00
-2.25146964e-01 -5.02495617e-02 -4.99460459e-01 -2.22719610e-02
1.10690570e+00 5.10580428e-02 -3.22437614e-01 -3.00340265e-01
-9.71421421e-01 8.24425995e-01 3.65765154e-01 2.17086375e-01
-2.75652438e-01 2.43377775e-01 3.57381761e-01 2.46184394e-01
8.44559133e-01 -1.36512354e-01 -2.17813864e-01 -4.37903196e-01
-4.74660605e-01 -8.36400241e-02 3.70394230e-01 6.60601735e-01
8.03023815e-01 1.01542860e-01 -8.04970041e-02 9.56545770e-01
4.13289219e-01 3.41647327e-01 4.80403721e-01 -7.18362451e-01
3.41313034e-01 8.83038998e-01 -3.40228587e-01 -1.27602947e+00
-4.30223525e-01 -1.20233148e-01 -8.19527388e-01 2.55297929e-01
4.95001853e-01 -9.54207629e-02 -5.47112167e-01 1.87148488e+00
4.75503325e-01 6.49674535e-01 -8.63753483e-02 1.12976503e+00
9.13557172e-01 3.74469310e-01 -2.32049897e-01 7.81885087e-02
1.05929744e+00 -1.06394053e+00 -2.86865562e-01 2.04958111e-01
9.23900306e-01 -7.00523555e-01 8.34134996e-01 1.54619306e-01
-9.08797562e-01 -3.56062025e-01 -6.66391492e-01 2.69024726e-02
-8.73591751e-02 3.25416595e-01 7.45000064e-01 6.22784793e-01
-9.72501218e-01 6.23076856e-01 -1.09154940e+00 -3.48811567e-01
8.03102732e-01 4.95126247e-01 -7.98716068e-01 -1.11919776e-01
-7.18108118e-01 4.95187908e-01 -9.21455920e-02 4.89892483e-01
-9.59589660e-01 -5.21819711e-01 -1.21729875e+00 -1.57920167e-01
3.52707803e-01 -4.34694976e-01 9.86305475e-01 -5.88629186e-01
-1.87397552e+00 1.13752842e+00 -6.46517426e-02 6.54561743e-02
5.04356742e-01 -1.01789720e-01 -2.61135876e-01 2.18892768e-01
-2.71419495e-01 5.34271896e-01 9.79572415e-01 -6.45369053e-01
-1.30984962e-01 -7.57602811e-01 8.15069154e-02 1.35486603e-01
-5.27393222e-01 1.41931668e-01 -6.17040396e-01 -8.06031764e-01
2.13925019e-01 -8.56512427e-01 1.79037735e-01 5.23871303e-01
-2.47004643e-01 -6.03813291e-01 7.28627443e-01 -3.89097631e-01
1.04057765e+00 -2.26438594e+00 4.88898337e-01 7.77834803e-02
2.89465785e-01 3.70133132e-01 -5.84687591e-01 3.56808901e-02
-1.40755042e-01 -5.74019440e-02 -2.22002372e-01 -6.00436449e-01
1.67102948e-01 1.27956182e-01 6.51204512e-02 7.93145955e-01
4.65131909e-01 9.71951485e-01 -8.26166272e-01 -1.96246773e-01
1.08996071e-01 7.63756812e-01 -6.82476103e-01 2.03871816e-01
-3.35527817e-03 5.09746373e-01 -6.38104320e-01 9.38453436e-01
7.89328456e-01 -8.08326751e-02 -2.63906986e-01 -3.41205299e-01
1.25385329e-01 1.50471270e-01 -8.43196273e-01 2.13370562e+00
-4.64793593e-01 2.23345935e-01 6.84545189e-02 -1.08392942e+00
1.14727986e+00 3.33407968e-01 6.71765864e-01 -5.74876785e-01
5.86362422e-01 2.95079231e-01 -2.46150956e-01 -2.83388823e-01
-2.35290274e-01 -8.87153894e-02 2.55191773e-01 7.04697728e-01
2.47916043e-01 2.94148028e-01 -3.11742604e-01 -3.14923108e-01
9.61267352e-01 4.78803664e-01 -6.68355869e-03 -1.71263948e-01
7.57390499e-01 -6.48583174e-01 6.72291338e-01 -2.04541832e-02
-3.64793271e-01 8.70696187e-01 6.66156292e-01 -4.93781924e-01
-4.63863909e-01 -8.61719191e-01 -1.76359385e-01 7.49282360e-01
8.99853371e-03 -4.61007118e-01 -1.09535205e+00 -1.15287709e+00
7.16745034e-02 -1.99489091e-02 -1.01558876e+00 -4.10712004e-01
-4.71228719e-01 -6.57998800e-01 7.69013584e-01 5.08872688e-01
6.30351484e-01 -8.24214160e-01 -1.58671051e-01 1.50381804e-01
1.46935746e-01 -1.26425493e+00 -1.17096221e+00 -5.64040303e-01
-7.77913928e-01 -1.15598810e+00 -8.86704743e-01 -7.37584949e-01
7.36315966e-01 5.64293146e-01 6.48746192e-01 8.66799504e-02
-4.83983308e-01 4.27667141e-01 -3.66938263e-01 3.30697224e-02
1.12152427e-01 -7.54956976e-02 3.59797746e-01 8.91902089e-01
9.17798758e-01 -8.13879013e-01 -7.72831261e-01 7.32450902e-01
-7.71533132e-01 -4.39512953e-02 4.73222047e-01 7.80675888e-01
6.07294261e-01 -3.98071200e-01 4.41638857e-01 -3.48037034e-01
5.21538913e-01 -2.89215118e-01 -5.45645177e-01 1.46959648e-01
-3.04024100e-01 4.20175828e-02 3.25722933e-01 -4.34890389e-01
-7.94976115e-01 1.74936295e-01 -4.54583764e-01 -1.05697858e+00
-1.89095482e-01 3.63205492e-01 -4.24381882e-01 -6.33681893e-01
3.66596580e-01 3.14714640e-01 5.23145616e-01 -6.94939911e-01
2.56804109e-01 6.15405202e-01 3.47169697e-01 -6.13068759e-01
6.95523500e-01 6.90985918e-01 1.46919275e-02 -8.99348617e-01
-9.80599165e-01 -3.48765373e-01 -9.57313716e-01 -2.28757769e-01
9.44679976e-01 -9.76901829e-01 -9.11054492e-01 8.88165236e-01
-1.09736109e+00 -3.61690700e-01 9.27372426e-02 5.68299592e-01
-5.74950814e-01 5.23408473e-01 -5.44455409e-01 -4.99427050e-01
-4.55924630e-01 -1.21342051e+00 1.49610984e+00 1.28858656e-01
-8.07918832e-02 -9.70920861e-01 8.22274089e-02 3.97952497e-01
2.28139937e-01 5.42367339e-01 5.66097736e-01 -1.71948522e-01
-2.46453166e-01 -2.97137737e-01 -4.70657676e-01 6.26289248e-01
5.22914469e-01 -1.69026077e-01 -9.22927856e-01 -1.46131173e-01
-1.89524397e-01 -5.84983170e-01 8.28613281e-01 -7.30805248e-02
1.35909903e+00 -3.68468404e-01 -6.95882514e-02 9.62898493e-01
8.96540344e-01 -2.90199459e-01 5.22190273e-01 -2.39263967e-01
1.02417326e+00 5.57263434e-01 3.65092367e-01 5.91638982e-01
2.08300143e-01 9.89179850e-01 5.44746459e-01 -6.92002475e-03
-1.51121825e-01 -2.47942641e-01 5.99048436e-01 8.62954736e-01
-5.09798527e-01 4.52965856e-01 -4.20804352e-01 3.90565276e-01
-1.79083550e+00 -7.48408139e-01 2.54319042e-01 2.08582187e+00
7.80534506e-01 -9.09159258e-02 3.64469886e-01 -1.22579448e-01
4.16310370e-01 3.56669843e-01 -6.35585487e-01 -2.35638276e-01
-1.24059275e-01 3.76151532e-01 8.83693919e-02 3.58099282e-01
-1.08303142e+00 1.18735242e+00 5.21131086e+00 8.55718970e-01
-1.28972828e+00 1.46471009e-01 4.71764445e-01 -1.90415785e-01
-1.17512532e-01 -3.93669963e-01 -8.62192035e-01 3.87189388e-01
4.21395779e-01 4.58745174e-02 4.09581214e-01 9.51507092e-01
1.56468093e-01 5.75479150e-01 -1.19210064e+00 1.44784772e+00
4.41727638e-01 -1.14811873e+00 2.50722587e-01 4.10139441e-01
6.92288578e-01 5.36875576e-02 2.32563168e-02 1.07470434e-02
8.20285752e-02 -1.34117496e+00 3.30391943e-01 4.39426094e-01
1.04439080e+00 -9.22916114e-01 5.20875812e-01 -6.54805228e-02
-1.28908181e+00 3.52132082e-01 -5.74547946e-01 2.80584246e-02
-8.37558955e-02 1.68017268e-01 -3.64576578e-01 2.16493338e-01
4.98596728e-01 1.16549981e+00 -3.43352735e-01 7.44083703e-01
-3.51538092e-01 6.67786300e-01 -4.08289254e-01 -1.59496702e-02
6.40790343e-01 -6.09037519e-01 4.72626954e-01 9.80716527e-01
3.81844968e-01 3.14035296e-01 -1.44865409e-01 9.75134552e-01
-4.36913639e-01 2.78074205e-01 -6.80119514e-01 -2.30958015e-01
-1.06363110e-01 1.41450620e+00 -2.33805746e-01 3.65909278e-01
-8.46437216e-01 1.35967445e+00 5.17240226e-01 1.70889273e-01
-6.43441737e-01 -1.98127791e-01 1.43224239e+00 -2.47469008e-01
2.98913121e-01 -3.95421952e-01 2.21538723e-01 -1.45491731e+00
2.01288000e-01 -7.40068555e-01 3.50090116e-01 -2.67717093e-01
-1.41577935e+00 6.60064638e-01 -4.89863396e-01 -1.31154609e+00
8.23633373e-03 -9.81563449e-01 -7.04385579e-01 7.74900615e-01
-1.44507098e+00 -1.28881967e+00 -4.62906778e-01 1.04515910e+00
3.43844086e-01 -3.00866544e-01 1.02892792e+00 4.24561501e-01
-8.63331854e-01 1.22711110e+00 -1.59613356e-01 6.30841017e-01
4.60003912e-01 -7.25856483e-01 4.75162327e-01 5.46941280e-01
3.80273402e-01 4.75278139e-01 -6.16832636e-02 -1.84310809e-01
-1.75557411e+00 -1.32848787e+00 7.30641782e-01 -5.37924588e-01
7.73723006e-01 -5.40029109e-01 -9.06397104e-01 5.67441285e-01
-3.32647502e-01 5.20246923e-01 8.97367895e-01 -1.93344876e-02
-8.74214113e-01 -2.84799337e-01 -8.97071183e-01 7.32341945e-01
1.66749132e+00 -6.75351024e-01 -1.37117386e-01 3.44889820e-01
3.15645099e-01 -2.29627386e-01 -1.02074599e+00 3.51430625e-01
6.83607221e-01 -7.07870841e-01 6.61485016e-01 -8.85153174e-01
3.21184427e-01 -2.44118556e-01 -1.80960178e-01 -1.24800801e+00
-1.86011463e-01 -9.29524243e-01 -2.45190486e-01 1.01424110e+00
3.35238457e-01 -7.00462401e-01 1.01960361e+00 2.57402390e-01
-7.34828562e-02 -9.09484029e-01 -1.18225765e+00 -8.24133337e-01
1.23855881e-01 -5.26214838e-01 5.30254960e-01 9.36142206e-01
6.93121552e-02 2.74355531e-01 -4.46461499e-01 -9.77310687e-02
7.06679702e-01 3.20443928e-01 8.23532343e-01 -1.12003660e+00
-1.50973469e-01 -8.58896315e-01 -1.05008078e+00 -1.31019926e+00
5.41378975e-01 -1.11501348e+00 -1.05857447e-01 -1.01341116e+00
7.63596818e-02 -1.98801979e-01 -2.24631891e-01 7.85395980e-01
3.17104459e-02 7.60381877e-01 -1.93862826e-01 9.68665630e-02
-4.34824079e-01 1.22711408e+00 1.34824383e+00 -1.60094306e-01
-1.90322823e-03 -1.48619130e-01 -6.84064984e-01 7.70237803e-01
5.61985016e-01 -1.10171176e-01 -2.06009910e-01 -7.15468884e-01
-1.81632340e-01 -2.78951555e-01 2.75037348e-01 -7.57611156e-01
-2.45317444e-02 -1.94514764e-03 9.69092846e-02 -1.22620538e-01
4.51733410e-01 -7.21941710e-01 -4.19422776e-01 -3.47706489e-02
5.63183520e-03 -1.93456382e-01 7.30383396e-02 5.58656156e-01
-3.31591994e-01 3.06912154e-01 8.64368975e-01 1.51228353e-01
-5.03736913e-01 1.15887988e+00 -6.82903603e-02 -1.73432633e-01
1.04559195e+00 -2.20014721e-01 1.96707934e-01 -3.52983087e-01
-6.13640606e-01 1.43052965e-01 3.40561897e-01 7.51799226e-01
8.77689242e-01 -1.79084170e+00 -8.89154732e-01 3.74255449e-01
3.90730649e-01 3.30179022e-03 2.33715221e-01 1.13679338e+00
-4.03525084e-01 2.88095981e-01 -2.40164801e-01 -5.91392279e-01
-1.35342181e+00 3.03166479e-01 6.24709249e-01 2.74236053e-01
-7.35882759e-01 1.12568486e+00 4.29142565e-01 -6.55304790e-01
4.17464465e-01 4.76383902e-02 -1.67510659e-01 1.17705045e-02
7.88112104e-01 6.32559136e-02 7.55140930e-02 -1.21717083e+00
-3.46909970e-01 1.14827347e+00 -1.36287227e-01 2.21618816e-01
1.40584540e+00 7.32574239e-02 -1.09444506e-01 -3.42694856e-02
1.76895344e+00 -2.45759085e-01 -1.57031298e+00 -5.93062699e-01
-2.28584081e-01 -6.54401481e-01 1.30103724e-02 -1.10370219e-01
-1.61639893e+00 1.12060702e+00 3.40661377e-01 -4.32605386e-01
1.14142096e+00 1.22065164e-01 9.93468583e-01 2.40073904e-01
3.68025661e-01 -8.13178182e-01 3.07569325e-01 4.01164919e-01
1.05675697e+00 -1.22022569e+00 -3.99570107e-01 -5.42352080e-01
-3.83890957e-01 1.33321083e+00 7.86867797e-01 -1.70883819e-01
1.01174390e+00 -2.24280879e-01 1.98750973e-01 -2.69432753e-01
-4.43078101e-01 -2.77312964e-01 5.24820268e-01 3.54282171e-01
3.95587891e-01 -2.13644534e-01 -3.04058433e-01 6.57165468e-01
4.23319787e-02 -4.32752445e-02 -6.15951121e-02 7.37251878e-01
6.12590602e-03 -1.03299332e+00 2.95870360e-02 1.25848100e-01
-4.41962302e-01 5.47647439e-02 -6.26320899e-01 5.68246782e-01
-6.28718808e-02 7.07436025e-01 1.91859543e-01 -7.13390946e-01
3.26565742e-01 -3.09329061e-03 7.47434556e-01 -5.64461470e-01
-4.88281921e-02 2.22841650e-01 -1.35178521e-01 -1.00201929e+00
-3.50496590e-01 -7.69100666e-01 -1.20147932e+00 -3.41389775e-01
-7.56625906e-02 -8.74092728e-02 5.34648836e-01 9.71242189e-01
6.90084994e-01 -1.26580849e-01 9.08182144e-01 -8.98961544e-01
-5.60715139e-01 -1.07403731e+00 -6.55849516e-01 4.71754551e-01
1.59449846e-01 -9.71363366e-01 -3.59587729e-01 -2.23040506e-01] | [13.70309066772461, 1.4013217687606812] |
a6dc4f8c-cc1a-4bc5-ba39-42221a5ea136 | generating-adjacency-matrix-for-video-query | 2008.08977 | null | https://arxiv.org/abs/2008.08977v2 | https://arxiv.org/pdf/2008.08977v2.pdf | Generating Adjacency Matrix for Video Relocalization | In this paper, we continue our work on video relocalization task. Based on using graph convolution to extract intra-video and inter-video frame features, we improve the method by using similarity-metric based graph convolution, whose weighted adjacency matrix is achieved by calculating similarity metric between features of any two different time steps in the graph. Experiments on ActivityNet v1.2 and Thumos14 dataset show the effectiveness of this improvement, and it outperforms the state-of-the-art methods. | ['Shuwei Huo', 'Yuan Zhou', 'Mingfei Wang', 'Ruolin Wang'] | 2020-08-19 | null | null | null | null | ['moment-retrieval'] | ['computer-vision'] | [ 1.00178503e-01 -2.39817888e-01 -4.09408152e-01 -2.21403360e-01
2.43896455e-01 -3.14918637e-01 5.00889361e-01 1.55501693e-01
-6.65535271e-01 6.08915210e-01 4.22180176e-01 -7.13484660e-02
-2.32325360e-01 -8.25213671e-01 -5.22221625e-01 -3.79788160e-01
-6.53919876e-01 -4.32610184e-01 7.45963097e-01 -1.33883236e-02
2.84901381e-01 4.34213698e-01 -1.07466698e+00 4.60286260e-01
2.39410266e-01 1.15980959e+00 1.35143269e-02 6.40542388e-01
-5.67777874e-03 1.33572865e+00 -3.63276362e-01 -8.76835883e-02
1.22441813e-01 -5.48690259e-01 -7.65458345e-01 8.40857551e-02
4.35540348e-01 -3.15912604e-01 -1.21569717e+00 1.01300061e+00
3.41910571e-01 4.46062297e-01 2.58664966e-01 -1.36026978e+00
-8.05199206e-01 7.64675558e-01 -6.34916306e-01 1.02530503e+00
7.13496685e-01 -1.42922536e-01 9.68493044e-01 -8.17292869e-01
7.30345964e-01 9.05278027e-01 6.99922740e-01 2.52477467e-01
-7.71066606e-01 -7.12304711e-01 2.39577740e-01 8.70947361e-01
-1.78851175e+00 -2.37266943e-01 8.03690612e-01 -2.95878917e-01
1.39137614e+00 -9.70851034e-02 9.47723806e-01 8.02532732e-01
3.53450447e-01 6.60005271e-01 6.57352805e-01 -1.26825720e-01
-4.89064567e-02 -6.99281216e-01 -2.79220678e-02 1.15492964e+00
9.76253375e-02 -1.71011180e-01 -6.84556484e-01 3.29737961e-02
9.37921464e-01 3.86435062e-01 -7.58480191e-01 -5.08565843e-01
-1.57465136e+00 6.95071042e-01 5.93465328e-01 6.31739497e-01
-3.58411610e-01 5.76064765e-01 7.05482543e-01 4.64304477e-01
3.24172974e-01 8.43900070e-03 -8.38864520e-02 -2.63376564e-01
-1.02109420e+00 -2.51803905e-01 6.64818048e-01 9.01243031e-01
7.54147947e-01 8.07027966e-02 -3.95448714e-01 5.24240613e-01
2.70893455e-01 -1.16451517e-01 6.04127586e-01 -8.83458197e-01
3.25480640e-01 7.01819479e-01 -6.08430088e-01 -1.17847276e+00
-3.40842932e-01 1.01936825e-01 -8.05130720e-01 -1.90614145e-02
6.46126792e-02 -4.57254723e-02 -9.37161326e-01 1.42347848e+00
3.41064781e-02 1.15698254e+00 -2.47079264e-02 9.04126048e-01
1.14254963e+00 6.40251219e-01 1.38031505e-03 -3.49562854e-01
8.74303877e-01 -1.09413898e+00 -8.19589436e-01 3.91948521e-01
6.95335984e-01 -4.71848339e-01 6.74532473e-01 1.96823448e-01
-8.32942009e-01 -6.97578669e-01 -1.24389541e+00 2.70781308e-01
-4.00533795e-01 -2.37254590e-01 7.77473450e-01 2.10979208e-01
-1.34773803e+00 9.36319292e-01 -9.02490139e-01 -6.90270603e-01
5.23424089e-01 4.87605721e-01 -8.16585362e-01 -2.86077689e-02
-1.19784498e+00 4.00315046e-01 7.09866405e-01 -7.99984261e-02
-8.85353982e-01 -5.16379297e-01 -1.11515522e+00 1.78630456e-01
5.33235550e-01 -4.64371741e-01 6.88759148e-01 -1.04874313e+00
-1.40320814e+00 5.25963068e-01 1.17155164e-01 -5.91981173e-01
1.36557505e-01 3.72161418e-02 -8.52698743e-01 5.62516451e-01
-2.16195270e-01 5.88083446e-01 7.14208663e-01 -4.77448076e-01
-7.13099778e-01 -1.11107782e-01 4.90432531e-01 3.26790720e-01
-5.25712848e-01 1.57087713e-01 -9.15430605e-01 -8.03476334e-01
-6.19185567e-02 -6.57843828e-01 1.10133821e-02 2.35379431e-02
-1.17719464e-01 -4.00752544e-01 1.20981455e+00 -5.00376821e-01
1.61405587e+00 -2.11154032e+00 1.05899997e-01 2.65177608e-01
6.84011817e-01 4.59101737e-01 -3.21687281e-01 5.02179801e-01
-4.06105548e-01 -3.94918956e-03 7.34452158e-02 1.48851737e-01
-3.42134804e-01 1.39565289e-01 1.88143402e-01 7.88538396e-01
8.88516679e-02 8.73554707e-01 -1.10403740e+00 -7.06493914e-01
5.13033390e-01 6.43299162e-01 -3.46787542e-01 1.68199077e-01
2.18259454e-01 2.12441087e-01 -3.83269221e-01 5.46509504e-01
4.51553345e-01 -2.89528072e-01 3.07906628e-01 -5.14827907e-01
1.88397393e-02 7.40470886e-02 -1.17724919e+00 2.34569812e+00
-9.62052494e-02 9.46109593e-01 -4.56203043e-01 -1.30855644e+00
5.25104582e-01 3.44485968e-01 9.77996826e-01 -6.25918865e-01
3.24116170e-01 -4.38279688e-01 1.77774519e-01 -4.53714550e-01
4.63061512e-01 5.36524355e-01 3.21138173e-01 3.56918275e-01
5.26003778e-01 4.93381202e-01 6.46610677e-01 5.90645254e-01
1.59884536e+00 2.16815442e-01 4.36856419e-01 -4.43761408e-01
9.11520958e-01 -5.57295918e-01 3.89350742e-01 5.52456737e-01
-4.80105489e-01 3.51753652e-01 5.38789272e-01 -5.30283093e-01
-6.32822216e-01 -1.11173141e+00 3.72236192e-01 9.58183646e-01
4.04274732e-01 -1.11917806e+00 -7.75035203e-01 -1.22966552e+00
-2.09968954e-01 1.22589424e-01 -8.66620839e-01 -2.02034220e-01
-7.38263845e-01 -3.40389371e-01 5.40120363e-01 8.21914375e-01
8.56306970e-01 -9.71221924e-01 -4.12352711e-01 3.38891894e-01
-7.10656494e-02 -1.35212398e+00 -9.38799441e-01 -2.56116360e-01
-6.90346301e-01 -1.38525689e+00 -4.50025439e-01 -8.81079674e-01
5.03772557e-01 6.50976837e-01 1.07350719e+00 4.02302712e-01
-3.15242946e-01 7.04500318e-01 -6.02083087e-01 3.53455693e-01
7.54155889e-02 -1.26467988e-01 -4.01659831e-02 2.52722144e-01
4.39353138e-01 -6.61107600e-01 -7.30665624e-01 2.45431036e-01
-9.46965635e-01 -1.18441820e-01 2.34406069e-01 5.54864049e-01
6.55399859e-01 2.48561166e-02 2.89745748e-01 -6.28550231e-01
6.74195826e-01 -5.03832638e-01 -5.40830910e-01 3.86542767e-01
-5.76310635e-01 -4.69926707e-02 4.96027112e-01 -5.22068083e-01
-5.59049189e-01 5.45140691e-02 2.41015196e-01 -8.49080443e-01
1.48578629e-01 6.02590978e-01 2.74004564e-02 -4.18650001e-01
3.16996902e-01 2.44054824e-01 -2.73516566e-01 9.36016254e-03
4.21130210e-01 3.61462057e-01 4.90145773e-01 -4.43040840e-02
4.02572036e-01 6.08724296e-01 1.01856351e-01 -8.27919543e-01
-4.45936769e-01 -7.38119423e-01 -6.64303541e-01 -5.79915941e-01
1.08448708e+00 -9.06005621e-01 -1.02958262e+00 2.91705519e-01
-1.06796992e+00 -3.69152218e-01 -1.36772469e-01 9.74276960e-01
-4.09019917e-01 9.02183175e-01 -5.71694553e-01 -1.76219791e-01
-2.08060741e-01 -1.09368849e+00 5.70673406e-01 1.74333662e-01
9.07289051e-03 -1.27252769e+00 3.29906046e-01 -1.47396848e-01
2.59887666e-01 2.50921309e-01 2.54485309e-01 -4.93822843e-01
-6.12685323e-01 -1.90529637e-02 -4.69868481e-01 4.53975648e-01
2.77929634e-01 1.19494751e-01 -4.39833343e-01 -5.30663669e-01
-3.96487713e-01 3.51395905e-02 1.09030485e+00 4.32126939e-01
1.56000018e+00 -4.97346818e-02 -4.63372469e-01 9.59377527e-01
1.52923143e+00 1.35117233e-01 9.70215261e-01 9.28345099e-02
1.10824025e+00 -7.62890130e-02 4.98595595e-01 3.71725321e-01
3.93313318e-01 6.70216739e-01 4.35360163e-01 -1.31904095e-01
-5.13503492e-01 -4.44948599e-02 3.74212593e-01 8.76257896e-01
-5.58209181e-01 -5.30006170e-01 -6.03610635e-01 4.59148675e-01
-2.37449050e+00 -1.29725158e+00 -1.03124782e-01 2.04115391e+00
2.65547663e-01 1.09482549e-01 1.11962050e-01 -8.88868943e-02
9.26584840e-01 5.86059690e-01 -2.29428500e-01 -2.67116100e-01
5.44993614e-04 3.86442661e-01 8.47854495e-01 3.27716529e-01
-1.28642523e+00 1.05305183e+00 6.81828594e+00 7.35271990e-01
-1.03587079e+00 1.31468832e-01 1.65603325e-01 -4.33423035e-02
3.24501395e-01 5.28862588e-02 -2.87105620e-01 4.55161840e-01
7.69536853e-01 -2.04420000e-01 7.87893891e-01 4.80885834e-01
8.47924128e-02 -1.69678479e-01 -1.27049148e+00 1.43814588e+00
5.89886665e-01 -1.65915477e+00 9.11862031e-03 -1.94383264e-01
6.82958066e-01 2.84308523e-01 -3.66469234e-01 1.93352118e-01
2.23035634e-01 -9.33577776e-01 2.91954488e-01 4.96699065e-01
6.97819889e-01 -7.26471305e-01 6.37252927e-01 -3.35054219e-01
-1.98791265e+00 1.73103735e-01 -3.48733485e-01 -1.07634731e-01
3.85815054e-02 3.57410461e-01 -6.26657367e-01 6.97748125e-01
9.17827666e-01 1.43631005e+00 -5.36225140e-01 1.02163398e+00
-4.04399812e-01 6.24121785e-01 1.36720762e-02 1.39095187e-01
3.23298037e-01 -2.11603880e-01 2.97560453e-01 1.40668297e+00
2.40007758e-01 7.56555498e-02 3.40432853e-01 5.02522826e-01
-4.61726040e-01 1.24893308e-01 -8.64549220e-01 -3.79457735e-02
2.52341449e-01 1.36838639e+00 -8.39851916e-01 -6.33522451e-01
-1.00376844e+00 1.37330019e+00 4.86892819e-01 3.93296063e-01
-1.22337031e+00 -6.81251943e-01 7.22975135e-01 -4.63629775e-02
5.96083105e-01 -5.10980666e-01 5.08137763e-01 -1.20306981e+00
-1.70119956e-01 -3.86159003e-01 7.47182846e-01 -7.22693264e-01
-8.61553907e-01 4.88992453e-01 -1.45483417e-02 -1.25369036e+00
2.36281771e-02 -5.48603654e-01 -7.85064578e-01 3.83847326e-01
-1.44572186e+00 -9.73909676e-01 -4.97956127e-01 1.36030138e+00
4.12107825e-01 -2.59883016e-01 4.10370916e-01 4.90343720e-01
-3.94122154e-01 5.89925766e-01 -2.64737338e-01 5.27383506e-01
6.26247346e-01 -9.13324058e-01 4.61606354e-01 1.01316476e+00
5.48832655e-01 3.72441262e-01 8.03413019e-02 -7.20780134e-01
-1.57650518e+00 -1.30240929e+00 5.03313720e-01 -9.25845355e-02
1.11908984e+00 -2.24018574e-01 -7.46904850e-01 8.88563335e-01
5.69112599e-01 6.38513803e-01 7.11434007e-01 -2.27887496e-01
-4.36636925e-01 3.06424238e-02 -1.00072551e+00 5.92182577e-01
1.73916328e+00 -5.74488580e-01 -4.87811565e-01 2.47107297e-01
8.15367460e-01 -2.54233539e-01 -1.16812932e+00 4.61929381e-01
5.24264276e-01 -9.62132037e-01 9.53722417e-01 -6.12501919e-01
8.15450996e-02 -5.83071172e-01 -2.61185765e-01 -1.20780504e+00
-6.73140168e-01 -6.90354407e-01 -4.10871029e-01 8.55188489e-01
7.25547671e-02 -4.64688599e-01 7.72628367e-01 -2.41405264e-01
-2.32373431e-01 -4.76072580e-01 -8.80611539e-01 -9.92376506e-01
-7.42051899e-01 -4.10763949e-01 4.03223217e-01 1.20613420e+00
4.33351785e-01 3.74973655e-01 -3.39219153e-01 4.05529737e-02
3.13415676e-01 -1.20013453e-01 5.91617882e-01 -8.25593531e-01
-3.55020106e-01 -4.34248239e-01 -1.17693686e+00 -7.64220178e-01
3.75155121e-01 -1.03041255e+00 -3.42920244e-01 -1.49374092e+00
3.36835176e-01 3.08852851e-01 -8.91059041e-01 5.69647789e-01
1.75509637e-03 5.72730839e-01 2.75424093e-01 -1.43449366e-01
-1.29402792e+00 3.93266261e-01 1.18779922e+00 -1.79835081e-01
-6.69643432e-02 -6.35469019e-01 -2.53066391e-01 6.51966453e-01
7.77576685e-01 -3.68251383e-01 -7.10761666e-01 -3.66217613e-01
-2.66754091e-01 -1.04286531e-02 2.77321875e-01 -1.36520565e+00
2.44798973e-01 -2.05523577e-02 3.67857873e-01 -4.83955830e-01
2.53914028e-01 -9.34462845e-01 3.78904119e-02 5.64426303e-01
-1.51630580e-01 4.50302571e-01 -2.61043687e-03 8.28938246e-01
-3.46781790e-01 1.35618195e-01 6.31343961e-01 1.96291059e-02
-1.28928399e+00 8.12851369e-01 -4.68946397e-01 -7.57928193e-02
1.33885109e+00 -3.90246570e-01 -2.66117781e-01 -3.71253222e-01
-7.95350611e-01 -4.31101769e-03 4.82512176e-01 5.14186800e-01
9.90801334e-01 -1.62339425e+00 -5.47884166e-01 1.85483217e-01
2.85204381e-01 -7.27834821e-01 1.76320717e-01 1.05121994e+00
-6.53155625e-01 1.39646038e-01 -4.16410863e-01 -5.33716202e-01
-1.51412559e+00 6.97152972e-01 2.37412393e-01 -3.15101266e-01
-6.60513520e-01 6.38949633e-01 2.12330118e-01 2.89064974e-01
2.05424115e-01 -3.91137868e-01 -4.10349220e-01 -2.77415693e-01
6.83770657e-01 4.78236377e-01 -4.05678973e-02 -7.09079206e-01
-6.13859832e-01 4.81762052e-01 -2.27384362e-02 1.30085677e-01
1.14244032e+00 -2.03826465e-02 -2.33282909e-01 1.50579318e-01
1.49622607e+00 -1.91707864e-01 -1.14295948e+00 -2.14119658e-01
-1.99334785e-01 -6.94941401e-01 2.52660275e-01 -2.44776249e-01
-1.63784790e+00 5.69355071e-01 8.53344679e-01 1.60740674e-01
1.14626217e+00 5.72669543e-02 7.25569725e-01 3.65272224e-01
2.10891694e-01 -1.13241434e+00 3.80417019e-01 5.63193440e-01
6.56261563e-01 -1.02383173e+00 2.07331493e-01 -3.94480228e-01
-2.75451928e-01 1.26028156e+00 6.58801496e-01 -4.69655901e-01
9.44357276e-01 8.09179991e-02 -3.35147142e-01 -4.85008687e-01
-5.22283018e-01 -2.86709875e-01 3.69480401e-01 6.64753616e-01
6.69543505e-01 -6.20283559e-02 -6.54094040e-01 -1.55837787e-03
4.60937172e-01 1.82626054e-01 5.36896110e-01 1.04980421e+00
-2.39993423e-01 -9.69947398e-01 7.57079273e-02 5.20050704e-01
-4.12995726e-01 -3.61579835e-01 -2.81009406e-01 8.53005111e-01
-2.17613578e-02 9.46033835e-01 3.35398108e-01 -8.11368346e-01
1.50583327e-01 -3.88388008e-01 8.10261428e-01 -4.91565675e-01
-3.58376354e-01 -2.74702292e-02 1.01146810e-01 -1.19520009e+00
-9.52711403e-01 -5.36975324e-01 -1.58027959e+00 -5.15707016e-01
-2.43690938e-01 -1.33944750e-01 4.34179872e-01 6.60284817e-01
4.48110789e-01 7.96105564e-01 5.67293048e-01 -8.41472566e-01
2.64306009e-01 -9.87466931e-01 -6.84464216e-01 6.95168376e-01
5.91760091e-02 -8.31932604e-01 -6.21173531e-02 1.48370564e-01] | [9.047072410583496, 0.6159029603004456] |
76562407-532f-43c4-91a0-a91fc325c341 | extended-feature-pyramid-network-for-small | 2003.07021 | null | https://arxiv.org/abs/2003.07021v2 | https://arxiv.org/pdf/2003.07021v2.pdf | Extended Feature Pyramid Network for Small Object Detection | Small object detection remains an unsolved challenge because it is hard to extract information of small objects with only a few pixels. While scale-level corresponding detection in feature pyramid network alleviates this problem, we find feature coupling of various scales still impairs the performance of small objects. In this paper, we propose extended feature pyramid network (EFPN) with an extra high-resolution pyramid level specialized for small object detection. Specifically, we design a novel module, named feature texture transfer (FTT), which is used to super-resolve features and extract credible regional details simultaneously. Moreover, we design a foreground-background-balanced loss function to alleviate area imbalance of foreground and background. In our experiments, the proposed EFPN is efficient on both computation and memory, and yields state-of-the-art results on small traffic-sign dataset Tsinghua-Tencent 100K and small category of general object detection dataset MS COCO. | ['Yong liu', 'Chunfang Deng', 'Liang Liu', 'Mengmeng Wang'] | 2020-03-16 | null | null | null | null | ['small-object-detection'] | ['computer-vision'] | [ 1.39871836e-01 -3.67512524e-01 -1.17673442e-01 -2.30887190e-01
-6.63326621e-01 -1.97537661e-01 9.53810960e-02 -3.65226537e-01
-3.25727433e-01 4.06750470e-01 -1.72873124e-01 6.50605783e-02
9.39631555e-03 -8.78779829e-01 -6.27216637e-01 -6.54548764e-01
1.21145815e-01 5.29129803e-03 1.31670368e+00 -9.10185128e-02
3.52837861e-01 7.09796309e-01 -1.70908034e+00 4.88907218e-01
9.66673493e-01 1.22767246e+00 3.82948369e-01 4.47408855e-01
-1.35168403e-01 7.06482887e-01 -6.12638414e-01 -2.05564767e-01
3.75488639e-01 -1.03050113e-01 -2.45977074e-01 2.49960870e-01
9.74766731e-01 -6.48515821e-01 -4.17492181e-01 1.12452531e+00
5.88599026e-01 -2.27685228e-01 5.67270160e-01 -1.30442894e+00
-5.92830300e-01 3.19409013e-01 -1.23906112e+00 7.03587770e-01
-4.60965097e-01 2.33167380e-01 6.93352818e-01 -1.04701114e+00
4.58104372e-01 1.33381486e+00 6.60670877e-01 3.15546989e-01
-9.45786893e-01 -1.10201371e+00 6.65482208e-02 4.82367367e-01
-1.46234965e+00 -3.29067767e-01 6.70921743e-01 -1.34170443e-01
5.44437051e-01 2.72416860e-01 5.48354208e-01 4.04303461e-01
2.31967583e-01 1.07348633e+00 1.15493190e+00 -2.03980505e-01
-1.69139877e-01 1.18990742e-01 3.37709665e-01 8.88214588e-01
7.53277957e-01 -8.87293443e-02 -5.57313800e-01 1.50166079e-01
1.08056021e+00 1.49057195e-01 -1.63810179e-01 -2.55640805e-01
-1.08158863e+00 5.70609212e-01 4.78323787e-01 2.01648414e-01
-2.91450411e-01 1.87716812e-01 3.01195830e-01 1.68134212e-01
2.07356006e-01 -3.75531226e-01 -5.07507503e-01 2.16175526e-01
-8.48913014e-01 2.52503037e-01 3.29358995e-01 9.47010756e-01
6.50557756e-01 1.77043781e-01 -4.60530102e-01 7.01208711e-01
6.73320442e-02 7.51155853e-01 2.32137531e-01 -7.62213945e-01
5.14040291e-01 7.40946949e-01 -1.07420340e-01 -1.18076944e+00
-4.99628365e-01 -5.09668708e-01 -8.31025481e-01 4.63969886e-01
5.99600434e-01 2.60955170e-02 -1.09690702e+00 1.29000962e+00
6.43296242e-01 2.33084828e-01 -1.70947477e-01 9.88683283e-01
1.10152709e+00 6.62935853e-01 -2.55765300e-02 -1.06435403e-01
1.84516692e+00 -1.09152627e+00 -5.37357867e-01 -1.10908985e-01
1.75301448e-01 -7.62131453e-01 1.02356267e+00 3.15263122e-01
-9.66660678e-01 -8.03230882e-01 -9.37545300e-01 -2.00672507e-01
-2.48333886e-01 6.43538952e-01 7.71446586e-01 7.40869820e-01
-5.66618323e-01 3.56596380e-01 -6.40598893e-01 -3.05882722e-01
7.94148386e-01 4.33197558e-01 -2.68571228e-01 -1.68684557e-01
-7.62550414e-01 5.62322617e-01 4.93242890e-01 3.64509046e-01
-6.01882696e-01 -7.40138471e-01 -5.94967902e-01 1.95713490e-01
6.56649053e-01 -4.11422253e-01 7.52347887e-01 -7.20622301e-01
-1.17460895e+00 6.26794755e-01 8.08023438e-02 -9.76546928e-02
4.94330108e-01 -2.98148133e-02 -5.74592233e-01 2.76772261e-01
1.81614250e-01 5.26674867e-01 1.11616874e+00 -9.16351616e-01
-1.36678767e+00 -4.96337861e-01 -1.44215599e-01 6.40599197e-03
-4.45627511e-01 4.75866705e-01 -6.61893964e-01 -6.03057802e-01
1.37074903e-01 -4.70805228e-01 -1.57872796e-01 4.25468862e-01
-2.49027029e-01 -3.77382219e-01 1.35744905e+00 -7.79213548e-01
1.23309088e+00 -2.17069864e+00 -4.72806245e-01 2.05371752e-01
2.96341628e-01 4.69153255e-01 -1.54927254e-01 -3.87938917e-01
3.64946425e-01 -3.26839089e-01 9.06077400e-02 1.33480653e-01
-2.09852472e-01 4.70418409e-02 -1.48725659e-01 4.73680168e-01
4.69279140e-01 8.89504671e-01 -4.89209563e-01 -1.04557824e+00
1.95816532e-01 4.74003464e-01 -5.13800800e-01 -3.09200078e-01
1.72407687e-01 -8.31009299e-02 -8.18300426e-01 1.18073475e+00
1.34699667e+00 -3.46577317e-01 -2.49055326e-01 -6.66545749e-01
-4.63203996e-01 -4.39780742e-01 -1.65893388e+00 1.17337787e+00
5.03647961e-02 5.40987670e-01 2.90840477e-01 -6.34267807e-01
8.48210990e-01 -2.77567476e-01 1.41176507e-01 -7.03118861e-01
2.47711584e-01 3.41453880e-01 6.67276531e-02 -4.68533784e-01
6.17204428e-01 3.68286632e-02 5.47829345e-02 -9.32221189e-02
-1.82704598e-01 2.28453904e-01 4.37243760e-01 -6.57792715e-03
1.01965141e+00 -7.27788955e-02 2.02343673e-01 -3.00838619e-01
7.34804928e-01 3.99260558e-02 1.00947714e+00 6.72854960e-01
-4.35220599e-01 6.13350451e-01 6.38986886e-01 -3.61599714e-01
-8.23865116e-01 -1.09746027e+00 -3.05008948e-01 1.13368714e+00
5.46374500e-01 -8.69334191e-02 -5.73000073e-01 -8.15750659e-01
2.45085254e-01 1.93146795e-01 -5.29404104e-01 1.40539736e-01
-1.04314637e+00 -9.09666657e-01 5.36475718e-01 7.68605709e-01
9.88719940e-01 -1.09419537e+00 -7.49787688e-01 1.40535653e-01
-5.88869713e-02 -1.23817992e+00 -9.70687389e-01 -4.00943868e-02
-7.99063802e-01 -1.14474523e+00 -7.72593677e-01 -1.09851408e+00
6.83323622e-01 5.90423048e-01 7.91479945e-01 7.87752680e-03
-7.39068389e-01 -2.97529936e-01 -2.06854120e-01 -3.42415303e-01
1.74549036e-02 -1.77168876e-01 -1.32886723e-01 2.84095913e-01
2.30970830e-01 -3.21550041e-01 -8.69662464e-01 5.41078508e-01
-8.52753878e-01 -7.60709047e-02 1.08237076e+00 7.76276112e-01
5.84469616e-01 2.85010964e-01 5.37021518e-01 -3.98515493e-01
2.19935074e-01 1.39431983e-01 -8.96983147e-01 3.81392390e-01
-1.79486275e-01 -2.07814634e-01 6.04293525e-01 -6.21833265e-01
-1.09516120e+00 1.65664777e-02 1.72604159e-01 -3.98152918e-01
5.14247380e-02 -3.22229058e-01 -2.92854518e-01 -5.77003956e-01
3.14980149e-01 4.79228675e-01 -1.06892340e-01 -6.61323249e-01
1.51759937e-01 8.44203889e-01 5.23854852e-01 -2.92936772e-01
9.04804826e-01 6.47619426e-01 3.02058041e-01 -8.90275657e-01
-7.37049222e-01 -4.35031295e-01 -6.40621483e-01 -2.35546902e-01
6.88133836e-01 -8.40180814e-01 -9.84818876e-01 6.31676376e-01
-9.18859720e-01 1.51071683e-01 -3.32102865e-01 2.64945239e-01
-6.49307966e-02 4.92936879e-01 -5.99419832e-01 -4.62421626e-01
-6.15795553e-01 -9.41569388e-01 1.16857886e+00 7.50121593e-01
6.65857375e-01 -1.37624905e-01 -5.56429863e-01 3.95441234e-01
5.68666756e-01 2.46692523e-01 7.22882152e-01 -1.53499156e-01
-1.10471320e+00 -1.66507199e-01 -1.07878721e+00 5.68309009e-01
-2.61005899e-03 -8.09634402e-02 -8.39182436e-01 -2.28352338e-01
-3.38083565e-01 -1.61997497e-01 1.25869358e+00 5.33645034e-01
1.25342667e+00 3.34137841e-03 -6.06938422e-01 4.18516099e-01
1.28646302e+00 -1.06140725e-01 5.15851438e-01 1.72064766e-01
7.68380761e-01 4.54489559e-01 8.83343220e-01 3.03819537e-01
1.91694736e-01 6.11086071e-01 6.56234659e-03 -2.63521433e-01
-6.64082468e-01 3.07264607e-02 3.89775306e-01 4.88246709e-01
-2.40957111e-01 1.29392177e-01 -4.61119562e-01 4.98496652e-01
-1.61387789e+00 -9.44777906e-01 -3.63979995e-01 1.83936453e+00
6.17879152e-01 3.56692433e-01 1.94512784e-01 -4.97389510e-02
9.54350889e-01 4.87546511e-02 -6.16756320e-01 4.33031797e-01
-4.11293954e-01 6.29281485e-03 7.62435913e-01 9.65194218e-03
-1.17751086e+00 9.56777036e-01 5.41293621e+00 1.64587617e+00
-1.09410954e+00 2.95226395e-01 5.11442780e-01 -5.28639071e-02
2.21962318e-01 -2.52713323e-01 -1.33789706e+00 5.40342569e-01
2.99102545e-01 1.57558233e-01 3.95044759e-02 9.73212838e-01
1.55312732e-01 -1.48658395e-01 -5.61526954e-01 8.77062738e-01
4.44913991e-02 -1.25825036e+00 7.57494718e-02 3.11862882e-02
4.84066963e-01 -4.26347405e-02 -4.10071723e-02 4.62312728e-01
-3.55336636e-01 -5.23945987e-01 6.37557983e-01 2.07392544e-01
8.35851669e-01 -7.93679714e-01 6.05087101e-01 2.15741664e-01
-1.78489602e+00 -3.67054075e-01 -6.51583016e-01 4.25570220e-01
2.26995736e-01 5.41450799e-01 -3.80811512e-01 3.06935459e-01
7.40999758e-01 3.39145064e-01 -9.02385116e-01 1.29417598e+00
1.43481508e-01 4.27784681e-01 -5.25236487e-01 -7.87991807e-02
1.88955605e-01 3.45058031e-02 5.16750336e-01 1.26377833e+00
3.01643401e-01 8.80009308e-02 4.37072635e-01 7.82586455e-01
-7.83338398e-02 1.56345814e-01 -1.20078810e-01 3.48369420e-01
2.81129062e-01 1.64827251e+00 -1.22884047e+00 -4.69219178e-01
-5.99001706e-01 9.41216528e-01 9.82890874e-02 5.68047948e-02
-1.03137803e+00 -9.13825929e-01 4.73898262e-01 4.18391526e-01
8.82659972e-01 7.76513591e-02 8.75027999e-02 -1.26452994e+00
3.07806939e-01 -6.44759059e-01 4.49653178e-01 -5.77842772e-01
-1.07327259e+00 3.40557069e-01 -1.44405976e-01 -1.32399631e+00
5.27089775e-01 -8.27611864e-01 -6.04753733e-01 4.24889296e-01
-1.91269028e+00 -1.57863522e+00 -5.70315540e-01 6.56008184e-01
6.39612734e-01 -1.87799130e-02 -1.13578169e-02 7.64152586e-01
-7.11060524e-01 7.44428873e-01 2.14642555e-01 1.51113778e-01
6.26679301e-01 -9.54889357e-01 2.25338057e-01 8.90729606e-01
-2.65119225e-01 2.37916514e-01 3.49695146e-01 -7.95096576e-01
-1.30656588e+00 -1.26962781e+00 4.27065581e-01 -1.57161221e-01
4.88414824e-01 -3.11660796e-01 -7.93952763e-01 2.91145414e-01
-2.34396234e-01 7.40571499e-01 1.83756894e-03 -3.77193868e-01
-2.36831456e-01 -5.77588618e-01 -1.30064356e+00 3.36277157e-01
1.12912297e+00 4.56839576e-02 -2.89737523e-01 8.28260258e-02
7.17393160e-01 -4.00758505e-01 -6.12704933e-01 7.61909664e-01
7.58746445e-01 -8.52710307e-01 1.17431641e+00 -3.36691529e-01
-9.08236429e-02 -9.12247241e-01 -1.99333027e-01 -5.05705416e-01
-5.45888901e-01 -3.72593224e-01 -2.26314619e-01 1.47471821e+00
8.61356333e-02 -4.27650183e-01 1.06472611e+00 2.34164391e-02
-9.54211131e-02 -6.06142819e-01 -1.00914311e+00 -8.90773654e-01
-3.88795078e-01 -9.68662575e-02 2.61305779e-01 4.80601639e-01
-6.31842971e-01 2.11171582e-01 -4.32537258e-01 3.28909457e-01
1.10075712e+00 5.44437945e-01 7.80502260e-01 -1.36190426e+00
-1.32286429e-01 -3.34146172e-01 -6.51626825e-01 -1.19325364e+00
-2.23701924e-01 -5.06424069e-01 1.02418669e-01 -1.29726434e+00
6.40417874e-01 -3.65577340e-01 -4.31213289e-01 4.85100806e-01
-2.59657890e-01 7.82900512e-01 2.21739635e-01 -2.33081654e-02
-9.03451085e-01 5.51051915e-01 1.68646920e+00 -8.01618695e-02
-1.31176159e-01 -4.01177094e-04 -5.66205144e-01 9.12912190e-01
5.30401468e-01 -4.96221244e-01 8.16374123e-02 1.30423820e-02
-4.60341483e-01 -2.56862015e-01 5.88510931e-01 -1.07368875e+00
3.56365621e-01 -2.01337278e-01 8.36588383e-01 -1.33765662e+00
2.68782407e-01 -6.02086484e-01 -3.29345822e-01 7.18713343e-01
2.80447543e-01 -3.14207762e-01 3.48002106e-01 6.64248526e-01
-7.35810325e-02 -4.00395170e-02 1.05964494e+00 4.31765802e-02
-1.15056765e+00 2.94487596e-01 -1.50119394e-01 2.91510392e-02
1.09749353e+00 -3.84977400e-01 -5.94360828e-01 3.56134981e-01
-3.33027095e-01 5.00300646e-01 1.36757970e-01 4.93233174e-01
7.54119813e-01 -1.41357911e+00 -8.31748903e-01 4.07331765e-01
5.72039299e-02 -7.29001537e-02 6.96629584e-01 9.98183012e-01
-6.22513592e-01 3.33909571e-01 -4.01997000e-01 -6.12212896e-01
-1.74135017e+00 4.04775918e-01 1.34182453e-01 -1.17291398e-01
-9.66621995e-01 7.92030513e-01 3.97653133e-01 -2.25436404e-01
1.94404110e-01 -6.53390527e-01 -2.27930218e-01 2.28067581e-02
8.30057561e-01 7.28511453e-01 -1.88194200e-01 -7.54758537e-01
-4.61974502e-01 1.12912679e+00 -3.70614499e-01 5.83829224e-01
1.16559374e+00 -1.35600075e-01 -5.94486259e-02 -1.49827376e-01
8.99680495e-01 2.01164931e-01 -1.27768457e+00 -4.02091384e-01
-2.95410246e-01 -7.54024625e-01 2.32651740e-01 -4.33182448e-01
-1.28465259e+00 6.16059899e-01 1.01748955e+00 -5.87150976e-02
1.27295339e+00 1.24378406e-01 1.07815051e+00 4.81637657e-01
2.04603642e-01 -1.13825047e+00 2.07311556e-01 3.37967366e-01
6.97914422e-01 -1.34992516e+00 1.69511124e-01 -6.87557220e-01
-3.75363559e-01 1.27633750e+00 1.07422817e+00 -2.05318466e-01
7.57070005e-01 4.41701502e-01 -2.05208912e-01 -2.58968264e-01
-3.47956508e-01 -5.12532651e-01 3.47064793e-01 3.91641438e-01
-6.31171763e-02 -5.26162423e-02 -3.29783231e-01 6.65010989e-01
2.84392059e-01 2.29071245e-01 4.15451467e-01 9.11232352e-01
-9.91312802e-01 -6.76967621e-01 -7.02921450e-01 7.15009987e-01
-5.84836662e-01 1.38807511e-02 1.25621989e-01 1.00039279e+00
6.37763023e-01 6.63512051e-01 2.16802165e-01 -2.04434142e-01
5.57666123e-01 -5.17929137e-01 5.62869668e-01 -2.32025474e-01
-3.40873480e-01 3.29125673e-01 2.22586636e-02 -6.94715559e-01
-2.23905087e-01 -5.96055090e-01 -1.32329977e+00 -2.70965219e-01
-9.13886309e-01 -7.61928931e-02 2.97029704e-01 5.17657578e-01
3.05294931e-01 6.67417049e-01 4.35311079e-01 -8.71063173e-01
-5.68177402e-01 -1.06668711e+00 -1.10059631e+00 9.68933925e-02
2.60614365e-01 -6.94483042e-01 -9.98952165e-02 -4.46644723e-02] | [8.807865142822266, -0.5077728033065796] |
4138a743-b97a-4457-9e88-40ef2229c7e6 | empathetic-response-generation-with-state | 2205.03676 | null | https://arxiv.org/abs/2205.03676v2 | https://arxiv.org/pdf/2205.03676v2.pdf | Empathetic Response Generation with State Management | A good empathetic dialogue system should first track and understand a user's emotion and then reply with an appropriate emotion. However, current approaches to this task either focus on improving the understanding of users' emotion or on proposing better responding strategies, and very few works consider both at the same time. Our work attempts to fill this vacancy. Inspired by task-oriented dialogue systems, we propose a novel empathetic response generation model with emotion-aware dialogue management. The emotion-aware dialogue management contains two parts: (1) Emotion state tracking maintains the current emotion state of the user and (2) Empathetic dialogue policy selection predicts a target emotion and a user's intent based on the results of the emotion state tracking. The predicted information is then used to guide the generation of responses. Experimental results show that dynamically managing different information can help the model generate more empathetic responses compared with several baselines under both automatic and human evaluations. | ['Ruifeng Xu', 'Lanjun Zhou', 'Jiachen Du', 'Jun Gao', 'YuHan Liu'] | 2022-05-07 | null | null | null | null | ['empathetic-response-generation', 'dialogue-management'] | ['natural-language-processing', 'natural-language-processing'] | [-1.73042700e-01 4.59336847e-01 -2.03472272e-01 -7.70970404e-01
-4.53168809e-01 -4.13466364e-01 5.27301550e-01 2.30104774e-01
-4.15332496e-01 7.21585572e-01 8.31437230e-01 1.86841294e-01
3.81730407e-01 -6.45863175e-01 5.48190475e-01 -4.11440521e-01
3.83122325e-01 7.71163881e-01 2.95942370e-02 -9.30806637e-01
6.35374606e-01 1.54530182e-01 -1.20174587e+00 5.67860544e-01
9.01158094e-01 8.01579237e-01 -9.36062727e-03 8.84309053e-01
-4.81077135e-01 1.28394461e+00 -8.22827220e-01 -1.74906895e-01
-2.37582609e-01 -1.04817939e+00 -1.55709553e+00 -1.96589917e-01
-7.26265252e-01 -2.63613254e-01 2.03949392e-01 7.27202415e-01
8.69233727e-01 6.56117916e-01 4.68366444e-01 -1.32345390e+00
-9.68757793e-02 6.91387177e-01 8.81124735e-02 -4.52793464e-02
9.24643397e-01 1.02279559e-01 6.60149455e-01 -4.04635489e-01
4.96687800e-01 1.39495623e+00 4.74580079e-01 1.26287055e+00
-7.18997061e-01 -2.84409910e-01 2.30017886e-01 2.88924009e-01
-7.28173137e-01 -5.89542508e-01 1.10064518e+00 -2.75396734e-01
9.02582407e-01 2.36642629e-01 6.36564374e-01 1.10736585e+00
-1.30631879e-01 7.21563756e-01 1.15372193e+00 -5.23378015e-01
4.63869572e-01 7.35835969e-01 3.66414189e-01 2.99322426e-01
-1.08143842e+00 -3.61416698e-01 -7.44241297e-01 -5.90838492e-01
2.29003385e-01 -7.48584509e-01 -2.57560730e-01 1.01591684e-01
-8.30473721e-01 1.09944153e+00 -1.68154500e-02 5.15536368e-01
-8.15499306e-01 -5.91127396e-01 9.04085040e-01 4.97032434e-01
5.53909421e-01 9.31461096e-01 -4.55088854e-01 -8.60235155e-01
-3.84633869e-01 2.76840210e-01 1.57364678e+00 4.04042065e-01
5.20544946e-01 -1.78592339e-01 -4.16246563e-01 1.37424040e+00
2.14207709e-01 -7.34637156e-02 6.45372272e-01 -1.12125719e+00
7.89119974e-02 8.34905744e-01 6.11306965e-01 -9.39976096e-01
-7.78187096e-01 3.87222141e-01 -4.36133772e-01 -3.11745089e-02
3.12405348e-01 -7.61958480e-01 1.76117439e-02 1.97031033e+00
5.91290832e-01 -4.66332972e-01 4.98515248e-01 1.02407300e+00
8.82936656e-01 9.98943448e-01 4.72005546e-01 -7.02387631e-01
1.51981473e+00 -9.76673186e-01 -1.01424909e+00 -2.25605994e-01
8.39149415e-01 -8.78284276e-01 1.01610041e+00 3.99247020e-01
-1.13583100e+00 -3.65750223e-01 -5.72973490e-01 1.69148922e-01
-1.32451862e-01 5.23622520e-02 3.84635448e-01 5.83257437e-01
-9.75869179e-01 2.00798213e-01 -1.40398130e-01 -7.24597573e-01
-5.62855661e-01 2.28392631e-01 5.39197354e-03 5.07922232e-01
-1.75943911e+00 1.43855929e+00 1.75714239e-01 -3.48798931e-02
7.79947788e-02 -1.78231373e-01 -6.85307682e-01 4.00571600e-02
-5.56823723e-02 -6.03198349e-01 1.73717356e+00 -1.23321772e+00
-2.39707255e+00 7.11881340e-01 -1.51567876e-01 -2.47835681e-01
3.14012825e-01 -1.18020035e-01 -3.35242659e-01 2.12085709e-01
-2.21088186e-01 8.53119433e-01 3.77016783e-01 -1.13071835e+00
-8.10733557e-01 -2.03951403e-01 1.18614823e-01 9.05762792e-01
-4.53097194e-01 5.72182953e-01 -1.25576481e-01 -1.17278710e-01
-2.68967479e-01 -7.45882690e-01 -3.76608461e-01 -4.10788774e-01
-1.04001708e-01 -7.64299989e-01 5.96488118e-01 -4.83695269e-01
1.35900342e+00 -1.70401883e+00 1.41198754e-01 1.14500448e-01
-1.26235202e-01 3.93579841e-01 -1.90072402e-01 8.22777152e-01
7.10325018e-02 -1.62299663e-01 2.11558312e-01 -3.76044989e-01
2.18120977e-01 -7.40275383e-02 -3.24732512e-01 -9.25715193e-02
-1.21848069e-01 4.95511562e-01 -9.10298824e-01 -6.96678281e-01
2.38168567e-01 2.44304031e-01 -5.83381772e-01 9.27747548e-01
-3.16816956e-01 7.12444067e-01 -6.79404795e-01 2.02972945e-02
8.94590318e-02 1.01784825e-01 4.01464969e-01 1.20566143e-02
-2.08919793e-01 6.48350954e-01 -8.68031025e-01 1.13756156e+00
-6.13495350e-01 1.68141782e-01 2.61012375e-01 -6.72514796e-01
1.29767680e+00 6.68515146e-01 7.17496097e-01 -7.15995014e-01
4.07059431e-01 -1.59742773e-01 9.37180370e-02 -7.22211957e-01
7.72780001e-01 -3.29581439e-01 -4.42935795e-01 1.11923349e+00
-3.54535013e-01 -4.44341525e-02 -1.24197483e-01 2.20254853e-01
9.02248800e-01 2.10638326e-02 5.29711068e-01 5.02564423e-02
8.80569696e-01 1.63589865e-01 5.47623575e-01 5.57409942e-01
-7.41279423e-01 -2.09727511e-01 6.89016521e-01 -5.82357764e-01
-5.71978807e-01 -1.54111907e-01 3.49911124e-01 1.74317729e+00
2.03536645e-01 -4.46196914e-01 -9.90780115e-01 -6.76245630e-01
-7.34896779e-01 1.03611410e+00 -5.53092659e-01 -4.25614536e-01
-4.90953386e-01 -5.86642146e-01 4.65970546e-01 1.81250066e-01
7.05076456e-01 -1.74318504e+00 -8.78092110e-01 5.25548637e-01
-9.13674235e-01 -8.18418980e-01 -2.69217432e-01 -4.20014374e-02
-3.42625499e-01 -6.45068288e-01 -2.22280815e-01 -7.16031969e-01
2.67088950e-01 -8.13284442e-02 9.26267445e-01 5.04405499e-02
3.71509939e-01 5.40136933e-01 -8.28133404e-01 -3.13914180e-01
-1.06131124e+00 1.20508246e-01 6.95601031e-02 9.78341978e-03
4.93480951e-01 -3.32047373e-01 -6.30696952e-01 5.43761671e-01
-3.63885432e-01 2.49084920e-01 9.85009521e-02 6.81956768e-01
-2.51521170e-01 -2.71882087e-01 1.13319099e+00 -8.61795843e-01
1.64062083e+00 -6.34661734e-01 2.82474875e-01 3.85675192e-01
-5.63573956e-01 -9.62443557e-03 7.97543943e-01 -5.37860990e-01
-1.57044721e+00 3.84616889e-02 -5.84277451e-01 2.85109758e-01
-4.34039801e-01 3.28969806e-01 1.67963747e-02 1.47309244e-01
7.90199578e-01 4.65756515e-03 2.60257095e-01 -2.22480983e-01
2.55150944e-01 1.16327345e+00 3.41443866e-01 -8.32237661e-01
5.20993536e-03 -1.42308995e-01 -7.52661645e-01 -6.73594236e-01
-7.28694022e-01 -6.81828260e-01 -4.06248212e-01 -7.56833911e-01
8.16932619e-01 -5.12126327e-01 -1.29945385e+00 3.98842126e-01
-1.37373590e+00 -6.03809476e-01 4.13583517e-02 2.86082804e-01
-7.80418992e-01 4.18316424e-01 -7.79053032e-01 -1.22281706e+00
-8.02534461e-01 -9.68208849e-01 6.93763196e-01 4.33931530e-01
-1.21097887e+00 -1.03884721e+00 5.21252811e-01 4.73838538e-01
6.08171523e-01 -1.54454008e-01 1.14734924e+00 -1.12672281e+00
5.59977949e-01 -2.66924292e-01 1.52692467e-01 7.10138977e-02
2.89790332e-02 -1.98951334e-01 -8.93994391e-01 3.18542272e-01
2.38607734e-01 -8.66525829e-01 -1.04915395e-01 -2.70680338e-02
4.70050633e-01 -7.45409548e-01 9.05680005e-03 -1.76270619e-01
5.46063304e-01 4.80411261e-01 6.85358286e-01 2.41504773e-01
2.17363853e-02 1.28893530e+00 1.14005673e+00 9.72896218e-01
8.71344328e-01 9.40659106e-01 1.27632946e-01 -2.32532863e-02
5.81018925e-01 -8.82012695e-02 4.54493493e-01 6.49773657e-01
1.43301889e-01 -2.29862913e-01 -6.99944556e-01 1.97793305e-01
-2.24802852e+00 -1.13619888e+00 -4.32176562e-03 1.75522256e+00
1.24077308e+00 -1.84718981e-01 5.76980472e-01 -3.38233151e-02
7.22986639e-01 2.11184531e-01 -3.81886363e-01 -1.29841423e+00
3.00782889e-01 -5.21565527e-02 -5.03432035e-01 7.82531321e-01
-7.21370101e-01 1.31326723e+00 5.75372410e+00 1.23417988e-01
-1.17744803e+00 -1.16139829e-01 6.90548718e-01 1.78856909e-01
3.31713520e-02 -3.45863216e-02 -4.88683641e-01 3.05888444e-01
9.31969225e-01 -3.86917561e-01 2.73236960e-01 8.89635026e-01
6.14034235e-01 -3.04694474e-01 -7.96217978e-01 7.44919538e-01
1.28131494e-01 -7.08933353e-01 -2.32070968e-01 -3.34372491e-01
6.89740777e-02 -6.89592659e-01 -4.72520173e-01 7.83152521e-01
4.25096124e-01 -6.35820925e-01 3.12450171e-01 6.80426121e-01
5.06742150e-02 -9.80204284e-01 8.03723693e-01 7.93985009e-01
-6.35564923e-01 -4.36352529e-02 -1.08512394e-01 -3.51687968e-01
4.02561724e-01 -3.36274266e-01 -1.18363130e+00 -2.90250070e-02
3.24405879e-01 2.43445545e-01 -1.30270585e-01 5.79597771e-01
-2.81674892e-01 4.64671701e-01 1.67376958e-02 -4.28040922e-01
1.13064334e-01 -2.75106102e-01 4.88224745e-01 1.19842267e+00
-1.79532513e-01 6.60066545e-01 5.40534854e-01 5.17692208e-01
2.13654578e-01 8.69314671e-01 -2.72901177e-01 1.82348520e-01
6.83982551e-01 1.53709304e+00 -4.74561751e-01 -3.74662220e-01
-5.96140453e-04 1.13188136e+00 3.92959118e-01 6.13037534e-02
-6.17269278e-01 -2.17653304e-01 5.99780917e-01 -3.45713377e-01
-4.62346345e-01 3.05512846e-01 -2.61816114e-01 -7.59421349e-01
-5.28472185e-01 -1.09204364e+00 6.06717229e-01 -9.99165595e-01
-1.08384621e+00 8.69073451e-01 -2.90435076e-01 -7.77668238e-01
-8.87777090e-01 -1.79379970e-01 -1.01645839e+00 8.88664663e-01
-9.66834009e-01 -8.50172758e-01 -3.15099597e-01 4.22663897e-01
7.51844704e-01 1.15679558e-02 1.55533171e+00 -2.59574175e-01
-6.14041805e-01 3.93489867e-01 -6.95069551e-01 2.27082595e-02
1.24204051e+00 -1.11796224e+00 -5.69949038e-02 1.74154297e-01
-6.47069931e-01 4.28746045e-01 1.07110703e+00 -4.55785811e-01
-8.76329064e-01 -4.88897502e-01 1.36078370e+00 -2.41616279e-01
4.00455058e-01 6.98405951e-02 -1.02388525e+00 6.06701709e-02
6.00099266e-01 -8.16188693e-01 1.05851209e+00 3.05026114e-01
-3.62535790e-02 3.56757432e-01 -1.35096073e+00 7.76801527e-01
4.53409821e-01 -2.50634015e-01 -6.69423997e-01 2.71681458e-01
3.35324883e-01 -3.66324931e-01 -9.78725910e-01 1.35102585e-01
5.59289694e-01 -1.05098176e+00 4.62761760e-01 -1.01025224e+00
4.15046901e-01 2.49021143e-01 2.36713201e-01 -1.55298984e+00
-9.95844305e-02 -1.10543096e+00 1.09511964e-01 1.26217842e+00
1.88786507e-01 -4.24409479e-01 5.67504942e-01 1.45361340e+00
-3.45239230e-02 -7.86688924e-01 -5.66126883e-01 6.86424598e-02
-2.07403935e-02 -1.23912163e-01 5.50891519e-01 1.10386765e+00
1.11335492e+00 9.94445384e-01 -7.58949995e-01 -1.92687079e-01
-1.38193280e-01 1.97486758e-01 1.02049959e+00 -1.21419644e+00
-1.62775200e-02 -6.38829231e-01 1.53889328e-01 -9.36436474e-01
6.06377423e-01 -2.58690357e-01 4.99485970e-01 -1.51981997e+00
-6.08561598e-02 -3.91543806e-01 2.16510221e-01 7.11416066e-01
-3.74128550e-01 -3.61069292e-01 -1.69188101e-02 1.56953514e-01
-8.36268663e-01 6.49976194e-01 7.60423601e-01 4.64540660e-01
-8.63595009e-01 3.72547060e-01 -9.43658412e-01 8.57921958e-01
1.18739247e+00 -1.04841210e-01 -4.91918743e-01 3.65786463e-01
2.20364690e-01 7.68048644e-01 -5.75516522e-02 -6.17591798e-01
5.15720844e-01 -6.04818285e-01 -5.73020764e-02 -1.98056176e-01
4.73715901e-01 -4.96208102e-01 -4.76997167e-01 2.87689388e-01
-9.00463939e-01 8.88382941e-02 7.90447928e-03 1.94539353e-01
-1.92163289e-01 -3.61178994e-01 1.06946814e+00 -1.50422782e-01
-7.78048635e-01 -1.79669797e-01 -1.12673485e+00 -2.09567414e-04
9.26726878e-01 -1.79185420e-01 -3.82172525e-01 -1.24207819e+00
-7.85284996e-01 5.14504671e-01 3.26704413e-01 6.56575859e-01
4.57400084e-01 -1.01071703e+00 -5.62088490e-01 -2.72927731e-01
2.36495614e-01 -6.48149550e-01 3.88776213e-01 6.88208699e-01
-2.10621916e-02 2.06567228e-01 -3.79115403e-01 -6.99151382e-02
-1.51868343e+00 1.40773430e-01 6.16923928e-01 -4.35497373e-01
-2.54426509e-01 6.50454104e-01 -4.22506593e-02 -6.24414682e-01
4.98633981e-01 4.98179644e-01 -1.03415918e+00 4.00216430e-01
6.39250875e-01 2.46024117e-01 -1.35648847e-01 -9.64798391e-01
-1.19724929e-01 -1.11708185e-02 -1.89278901e-01 -4.94925410e-01
1.14181113e+00 -3.17536056e-01 -2.13663086e-01 5.08418083e-01
7.24239767e-01 -2.79008169e-02 -7.60840952e-01 -2.27381080e-01
1.09288633e-01 -1.91560820e-01 -3.90258670e-01 -1.18755794e+00
-3.12735975e-01 5.48541665e-01 2.03932613e-01 5.29949188e-01
1.19674802e+00 -2.17527971e-01 9.64005351e-01 4.94021982e-01
3.47610474e-01 -1.66496885e+00 4.46625233e-01 9.22033727e-01
8.82944882e-01 -1.00324368e+00 -3.48614156e-01 -1.83418036e-01
-1.57546663e+00 1.20236969e+00 1.24635398e+00 3.67398202e-01
2.36984268e-01 6.53286204e-02 7.69315064e-01 -1.60580888e-01
-1.34470618e+00 -5.25168851e-02 -3.23200691e-03 4.93453413e-01
7.33032048e-01 -9.77938920e-02 -6.33483648e-01 8.85968089e-01
-3.80470067e-01 -1.82074547e-01 5.28775990e-01 8.52491915e-01
-7.67119825e-01 -1.40862167e+00 -2.44630098e-01 1.13616154e-01
-7.93693215e-02 2.65585989e-01 -1.28482699e+00 2.68199265e-01
-4.35157806e-01 1.47729850e+00 -2.25284651e-01 -4.88370061e-01
6.45499885e-01 6.74211562e-01 -3.89679559e-02 -7.22242892e-01
-1.40277052e+00 -1.06834047e-01 6.81558430e-01 -4.44682032e-01
-4.63544250e-01 -4.85194534e-01 -1.48547459e+00 -2.96644211e-01
-1.61052927e-01 7.39128053e-01 6.03062570e-01 9.45554674e-01
4.55342501e-01 -6.72198981e-02 1.19511449e+00 -6.90898120e-01
-6.74805641e-01 -1.25302041e+00 -1.63202524e-01 6.84356391e-01
-2.80482411e-01 -3.09329629e-01 -1.75813437e-01 -2.66813904e-01] | [13.147075653076172, 7.696939945220947] |
ad779b17-567b-4b5d-87f9-82fc75f70977 | towards-view-invariant-and-accurate-loop | 2305.14885 | null | https://arxiv.org/abs/2305.14885v1 | https://arxiv.org/pdf/2305.14885v1.pdf | Towards View-invariant and Accurate Loop Detection Based on Scene Graph | Loop detection plays a key role in visual Simultaneous Localization and Mapping (SLAM) by correcting the accumulated pose drift. In indoor scenarios, the richly distributed semantic landmarks are view-point invariant and hold strong descriptive power in loop detection. The current semantic-aided loop detection embeds the topology between semantic instances to search a loop. However, current semantic-aided loop detection methods face challenges in dealing with ambiguous semantic instances and drastic viewpoint differences, which are not fully addressed in the literature. This paper introduces a novel loop detection method based on an incrementally created scene graph, targeting the visual SLAM at indoor scenes. It jointly considers the macro-view topology, micro-view topology, and occupancy of semantic instances to find correct correspondences. Experiments using handheld RGB-D sequence show our method is able to accurately detect loops in drastically changed viewpoints. It maintains a high precision in observing objects with similar topology and appearance. Our method also demonstrates that it is robust in changed indoor scenes. | ['Shaojie Shen', 'Chuhao Liu'] | 2023-05-24 | null | null | null | null | ['simultaneous-localization-and-mapping'] | ['computer-vision'] | [ 2.10682765e-01 -2.18298078e-01 2.57254974e-03 -4.90144283e-01
-3.65264893e-01 -8.77164900e-01 5.43865979e-01 3.15135717e-01
-1.85828768e-02 4.72566783e-01 -2.04979017e-01 -2.22850457e-01
-1.12513013e-01 -6.61473811e-01 -8.41270506e-01 -2.48414531e-01
-1.24936379e-01 6.43912613e-01 7.07573950e-01 -4.16385263e-01
4.86888081e-01 5.32135308e-01 -1.62511122e+00 -3.57014596e-01
8.19339693e-01 8.60901356e-01 8.57208788e-01 5.64382672e-01
3.86973694e-02 6.16627753e-01 -2.04325259e-01 3.06722254e-01
3.85489583e-01 -3.26684207e-01 -4.65787470e-01 2.38726333e-01
8.34198952e-01 -1.11642212e-01 -1.75361305e-01 1.22757339e+00
3.77923518e-01 1.43751642e-02 -6.17855154e-02 -1.34295642e+00
-7.64249116e-02 -1.83725268e-01 -5.94008446e-01 -3.25894393e-02
1.19059467e+00 -3.08102161e-01 8.57312083e-01 -1.04268587e+00
9.66579556e-01 1.15018296e+00 8.04327786e-01 -3.36529426e-02
-1.05570853e+00 -4.05342579e-01 4.00132447e-01 2.12846234e-01
-1.62923670e+00 -4.09715831e-01 1.04412627e+00 -2.87568212e-01
7.46448219e-01 4.28628415e-01 8.23224008e-01 6.39559388e-01
3.88411999e-01 3.29204142e-01 1.17836523e+00 -4.46596950e-01
4.27447051e-01 2.06343994e-01 -2.14988828e-01 1.20510912e+00
5.10762811e-01 -4.77004685e-02 -9.60199118e-01 6.69924356e-03
1.04028654e+00 2.40870968e-01 -4.45406407e-01 -1.43643939e+00
-1.69325817e+00 5.91506004e-01 7.64459014e-01 1.47394821e-01
5.36954477e-02 1.31841913e-01 2.54687457e-03 4.11039263e-01
2.27929920e-01 4.50639009e-01 -3.21452737e-01 1.09588161e-01
-8.10781598e-01 6.02047564e-03 6.39569163e-01 1.69072104e+00
1.22672939e+00 -3.65490764e-01 4.81932402e-01 2.28156984e-01
5.29202282e-01 9.03287172e-01 -4.19317232e-03 -8.79675329e-01
4.13833082e-01 9.03742731e-01 2.68883646e-01 -1.54638827e+00
-6.88846469e-01 -3.49858373e-01 -6.10060930e-01 2.77346909e-01
8.04870799e-02 5.55377901e-01 -7.89976776e-01 1.32139623e+00
6.60258412e-01 -1.11267492e-01 -1.98638365e-01 8.74684751e-01
6.52064979e-01 9.99983307e-03 -6.02098644e-01 -4.10538986e-02
1.31658816e+00 -9.98996675e-01 -6.82743549e-01 -8.66223276e-01
3.28076303e-01 -9.92489457e-01 8.97740245e-01 9.37910005e-02
-4.57566470e-01 -6.09320819e-01 -1.37082314e+00 -1.77947134e-01
-3.31620425e-01 4.81416062e-02 6.18855894e-01 3.89902353e-01
-1.20589066e+00 1.08298533e-01 -7.84006774e-01 -9.86444771e-01
-1.84571743e-01 2.57987827e-01 -5.91563463e-01 -2.49775216e-01
-7.81804621e-01 7.88189650e-01 3.30391407e-01 1.52530625e-01
-8.01110625e-01 -2.96014994e-01 -1.22802854e+00 -6.07716262e-01
7.02305853e-01 -8.59955311e-01 8.45169067e-01 -7.09946394e-01
-1.23032987e+00 1.06807363e+00 -6.57377481e-01 -1.95289761e-01
8.00617576e-01 -8.85165557e-02 -2.64019519e-01 9.84381735e-02
5.94762504e-01 3.29930812e-01 6.57104492e-01 -1.76549399e+00
-7.65923321e-01 -7.75671601e-01 1.35176152e-01 6.31772101e-01
5.91290772e-01 -6.25287592e-01 -4.77185458e-01 -7.16295540e-02
1.41053879e+00 -9.99104202e-01 -3.41031075e-01 4.11311150e-01
-2.79531062e-01 3.56145233e-01 1.00184894e+00 -2.88903445e-01
6.97628140e-01 -1.95871592e+00 -1.37582198e-01 5.06509662e-01
1.05370559e-01 -4.75383997e-01 2.59043962e-01 6.48681581e-01
5.81292331e-01 -5.29527128e-01 9.50891972e-02 -5.93705058e-01
-1.78454936e-01 4.89287168e-01 -2.88764447e-01 1.01546800e+00
-7.25334704e-01 7.69295216e-01 -1.40724707e+00 -4.39322740e-01
7.89980650e-01 2.40200505e-01 -4.08410877e-01 1.81860954e-01
-1.81672603e-01 7.53755569e-01 -4.74857330e-01 9.13284779e-01
8.73047888e-01 -2.74761230e-01 3.52184057e-01 -2.34775916e-01
-3.79391611e-01 2.69419491e-01 -1.55962670e+00 2.54781508e+00
-3.39724004e-01 7.30236471e-01 3.27160388e-01 -5.36496162e-01
1.13693094e+00 -3.19704264e-01 2.83779323e-01 -8.69991601e-01
-1.57278135e-01 4.99273866e-01 -8.29436183e-01 -1.22416891e-01
8.96904230e-01 2.80778199e-01 -6.95308745e-02 -1.46661103e-01
-2.22260773e-01 -4.99988675e-01 -2.73367792e-01 2.16610417e-01
9.77339327e-01 3.60036165e-01 8.32823753e-01 -4.21225101e-01
5.21510839e-01 3.22371691e-01 3.43467355e-01 8.57695997e-01
-2.35890269e-01 5.76594114e-01 -2.90445834e-01 -5.49591899e-01
-7.97809482e-01 -1.39426303e+00 1.05285443e-01 7.42087483e-01
1.44901347e+00 -5.26924431e-01 -9.73004848e-02 -3.64880174e-01
1.66428491e-01 1.66822955e-01 -3.84779334e-01 1.42228842e-01
-3.81800443e-01 1.17857950e-02 -1.07316442e-01 1.76307306e-01
7.97270417e-01 -3.28074604e-01 -1.22888076e+00 1.63927987e-01
-4.07865196e-01 -1.21979725e+00 -2.93120712e-01 1.91257402e-01
-8.00831497e-01 -1.33082688e+00 -7.22860321e-02 -1.03945851e+00
9.82368410e-01 1.03372812e+00 1.07009995e+00 -8.86025056e-02
-2.44813859e-01 8.39464188e-01 -2.87246555e-01 -1.90398008e-01
-1.55905351e-01 -2.38943979e-01 2.68634051e-01 -2.96641111e-01
-6.44410262e-03 -5.19346535e-01 -7.06528008e-01 7.08955705e-01
-5.46284392e-02 2.69514441e-01 2.10277557e-01 5.13533592e-01
9.00692403e-01 4.35036141e-03 -2.48754159e-01 -4.11113888e-01
-8.11742991e-02 -2.28349585e-02 -9.77527022e-01 2.63487488e-01
-8.03879082e-01 -1.56159803e-01 2.27403000e-01 2.14542732e-01
-8.05885017e-01 4.38762307e-01 3.89030337e-01 -2.63813794e-01
-1.35599643e-01 6.31996915e-02 -8.00964013e-02 -7.92336285e-01
6.85799599e-01 3.20406497e-01 1.93839800e-02 -2.70543963e-01
3.93607408e-01 1.56415537e-01 6.71224654e-01 -2.32902184e-01
1.01340246e+00 1.23463321e+00 3.76088291e-01 -7.77553499e-01
-5.63822210e-01 -1.08190763e+00 -1.00296378e+00 -4.28800374e-01
5.56117117e-01 -1.20076215e+00 -6.24818921e-01 2.17142090e-01
-1.06361663e+00 1.29940975e-02 -1.05989024e-01 4.66865718e-01
-7.52645552e-01 5.37464797e-01 9.60636362e-02 -8.19200933e-01
3.82144339e-02 -1.03921342e+00 1.45262730e+00 6.38713455e-03
2.20351038e-03 -1.18839693e+00 3.06792468e-01 1.03704125e-01
7.33087361e-02 5.20748973e-01 1.75393850e-01 4.06873412e-02
-1.32159352e+00 1.18777798e-02 -6.23371340e-02 -4.50826406e-01
5.12741446e-01 -4.28582132e-01 -8.49103034e-01 -6.86742485e-01
-3.26938182e-02 2.35131547e-01 5.23767114e-01 1.73897922e-01
3.64192039e-01 5.51167689e-02 -7.54823923e-01 8.73220026e-01
1.84252310e+00 1.53804377e-01 2.84485787e-01 7.61308789e-01
8.87017846e-01 4.73067880e-01 1.10073912e+00 4.01383281e-01
7.29584455e-01 8.28746319e-01 9.22857165e-01 -8.03128481e-02
7.23900944e-02 -8.17756712e-01 1.56531736e-01 5.84939897e-01
1.20917544e-01 4.82982546e-02 -9.77683783e-01 4.04662073e-01
-2.07344389e+00 -6.46771848e-01 -3.27598840e-01 2.27900362e+00
1.81726217e-01 6.23106807e-02 -3.88783991e-01 -1.03797428e-01
4.95691389e-01 3.90437514e-01 -3.86790186e-01 1.51754707e-01
-1.07098721e-01 -4.47508216e-01 1.04114068e+00 8.09802890e-01
-1.07192957e+00 9.57162261e-01 5.94331503e+00 1.49650186e-01
-8.78595531e-01 9.66567621e-02 -4.35908914e-01 3.58756244e-01
-4.32678789e-01 4.06919420e-01 -6.67317390e-01 6.96652159e-02
7.37982467e-02 5.15776947e-02 4.64855224e-01 1.06100166e+00
1.10302262e-01 -7.10943282e-01 -1.09462380e+00 1.42372191e+00
3.03000391e-01 -1.32759976e+00 -1.71227828e-01 -2.30134912e-02
8.13027561e-01 2.77260989e-01 -1.88925266e-01 -2.56881863e-01
1.99196264e-01 -4.76204693e-01 1.11350441e+00 1.79073602e-01
7.38442183e-01 -4.22981143e-01 4.07065570e-01 5.57236195e-01
-1.72079682e+00 5.56652620e-02 -4.33939219e-01 -2.60721922e-01
2.52716362e-01 5.56880534e-01 -1.18934679e+00 9.54845071e-01
7.05026329e-01 9.97764409e-01 -7.90567636e-01 9.22116220e-01
-1.85388893e-01 -3.37245673e-01 -5.11076987e-01 1.26763225e-01
7.68334717e-02 -3.42200160e-01 7.69366801e-01 6.96221113e-01
4.41309214e-01 -5.15834868e-01 6.67284250e-01 7.62178838e-01
4.59867030e-01 -3.79144102e-02 -1.03007233e+00 7.55161643e-01
5.89351475e-01 9.37133789e-01 -1.09671295e+00 -1.20286115e-01
-2.91540891e-01 1.36041260e+00 6.37051687e-02 2.79732376e-01
-6.81346774e-01 -3.86817381e-02 6.44453585e-01 2.75557429e-01
-2.17868341e-03 -7.61404157e-01 -1.78108796e-01 -1.25885212e+00
3.15815419e-01 -3.30351204e-01 5.99839278e-02 -8.27813029e-01
-5.46576142e-01 3.07506084e-01 -1.74686760e-01 -1.53071737e+00
-8.92290007e-03 -3.57136130e-01 -6.50491491e-02 4.75883633e-01
-1.67288697e+00 -1.39879847e+00 -1.03642368e+00 9.35484529e-01
6.30351365e-01 3.36276740e-01 7.54943669e-01 1.10103451e-01
2.89112419e-01 -8.47929681e-04 2.35363275e-01 -2.42197052e-01
8.06434393e-01 -1.33117270e+00 4.54008728e-01 1.08093727e+00
3.57628107e-01 7.14765966e-01 1.07187700e+00 -9.12312746e-01
-1.85497010e+00 -1.05932140e+00 8.32392752e-01 -6.57536983e-01
3.14817011e-01 -7.67475307e-01 -3.31085831e-01 9.76884842e-01
-2.57830232e-01 1.02044225e-01 -1.70281455e-02 -1.73255503e-01
-3.72842163e-01 -2.53594726e-01 -1.19085586e+00 3.54799598e-01
1.71144366e+00 -8.47753942e-01 -4.11948472e-01 4.38565522e-01
6.43693328e-01 -1.13439691e+00 -3.38751733e-01 5.67037642e-01
5.90780139e-01 -1.36853600e+00 1.19470990e+00 4.74032253e-01
-5.75667024e-01 -9.56666589e-01 -4.75154698e-01 -1.09391665e+00
-2.24434093e-01 -5.79645097e-01 5.30112199e-02 8.47847700e-01
-2.58454621e-01 -9.29771364e-01 6.96108520e-01 -9.36835036e-02
-3.98344547e-02 -6.48603663e-02 -1.11872196e+00 -9.77111638e-01
-1.12775517e+00 -1.97222158e-01 4.05708790e-01 9.01325583e-01
-2.40971074e-01 8.37142840e-02 -3.87896389e-01 7.39417315e-01
1.12668407e+00 5.88401079e-01 1.19908917e+00 -1.35747826e+00
1.42843634e-01 1.19520478e-01 -1.04130757e+00 -1.40912604e+00
-2.22513974e-01 -7.99576879e-01 2.62936920e-01 -1.87697721e+00
-2.21327960e-01 -6.14148915e-01 1.10826917e-01 -7.96983093e-02
1.85323805e-01 1.91457331e-01 -9.78340209e-02 4.06122774e-01
-1.03595507e+00 3.97863805e-01 9.25868332e-01 2.97545549e-02
-1.97760344e-01 -1.48571953e-01 -1.14233345e-01 8.89312923e-01
5.30391097e-01 -1.58200875e-01 -6.78501725e-01 -4.48286802e-01
4.91936743e-01 -7.28032887e-02 7.62801826e-01 -1.36539865e+00
5.55631697e-01 -1.89102232e-01 3.12147230e-01 -1.06882989e+00
3.94998103e-01 -1.28719079e+00 5.42459130e-01 6.49378479e-01
1.58596173e-01 3.99870068e-01 -1.19420931e-01 1.12047243e+00
-1.40472218e-01 3.31578195e-01 4.70610559e-01 -5.42041481e-01
-1.60277104e+00 1.83931261e-01 8.98020789e-02 -2.57066369e-01
1.24247372e+00 -6.37822032e-01 2.47002393e-02 -5.01584411e-01
-4.25854772e-01 3.43676358e-01 1.31684840e+00 6.04025483e-01
8.68041217e-01 -1.35952234e+00 3.69773433e-03 7.32068896e-01
7.20605552e-01 3.10513973e-01 1.11374691e-01 8.66930246e-01
-8.90878022e-01 4.85666215e-01 -5.80080673e-02 -1.28672326e+00
-1.50522113e+00 5.71844637e-01 2.62941062e-01 1.82705209e-01
-6.59241438e-01 7.43406951e-01 2.98602223e-01 -6.65741563e-01
2.66737282e-01 -6.15390837e-01 3.85160953e-01 -2.66168833e-01
2.58067548e-01 5.03508151e-01 1.32662594e-01 -8.19597661e-01
-9.13760781e-01 1.14883637e+00 5.78212917e-01 1.10405236e-01
7.90627539e-01 -1.08539140e+00 -3.67493570e-01 7.55856276e-01
1.07807016e+00 4.31627691e-01 -1.13227654e+00 -3.09825361e-01
3.96369137e-02 -8.77229452e-01 -2.64732689e-01 -3.45165402e-01
-4.39062864e-01 4.84134167e-01 8.28015864e-01 -9.05653238e-02
8.03820729e-01 1.52050689e-01 2.06569210e-01 4.38145965e-01
1.71591401e+00 -9.27112520e-01 -2.15887446e-02 5.86213112e-01
6.72821403e-01 -1.55122042e+00 1.93952188e-01 -7.61144459e-01
-7.16151372e-02 1.04349136e+00 5.69857121e-01 7.14718774e-02
3.69861692e-01 8.61103535e-02 1.47047222e-01 -4.55151528e-01
-2.19705161e-02 -2.34065101e-01 7.89126530e-02 8.76764774e-01
-1.54894486e-01 1.52541384e-01 1.97707295e-01 -5.85261464e-01
-2.09697738e-01 -3.67533743e-01 2.87790269e-01 1.32690895e+00
-8.45752001e-01 -7.56992400e-01 -6.31008327e-01 -1.62929714e-01
4.27542210e-01 1.62304565e-01 -2.66796321e-01 7.32330799e-01
1.39927387e-01 1.02260756e+00 -5.94475726e-03 -4.27301109e-01
4.89762098e-01 -2.74986625e-01 7.49979317e-01 -5.24606586e-01
6.62212819e-02 -2.24172220e-01 -1.65816143e-01 -1.35920966e+00
-6.81782424e-01 -4.71886337e-01 -1.28866506e+00 -1.85372323e-01
-2.60477751e-01 3.73077095e-02 9.25380647e-01 5.60749888e-01
4.71410573e-01 1.59317762e-01 6.29202127e-01 -9.28870142e-01
-1.64741516e-01 -2.48486042e-01 -7.69231260e-01 2.90454358e-01
9.77118909e-01 -9.56007600e-01 -4.13438439e-01 -1.86829358e-01] | [7.310600280761719, -2.2370786666870117] |
2c1b947a-c026-424a-93bd-f2e0272649a8 | channel-interaction-networks-for-fine-grained-1 | 2003.05235 | null | https://arxiv.org/abs/2003.05235v1 | https://arxiv.org/pdf/2003.05235v1.pdf | Channel Interaction Networks for Fine-Grained Image Categorization | Fine-grained image categorization is challenging due to the subtle inter-class differences.We posit that exploiting the rich relationships between channels can help capture such differences since different channels correspond to different semantics. In this paper, we propose a channel interaction network (CIN), which models the channel-wise interplay both within an image and across images. For a single image, a self-channel interaction (SCI) module is proposed to explore channel-wise correlation within the image. This allows the model to learn the complementary features from the correlated channels, yielding stronger fine-grained features. Furthermore, given an image pair, we introduce a contrastive channel interaction (CCI) module to model the cross-sample channel interaction with a metric learning framework, allowing the CIN to distinguish the subtle visual differences between images. Our model can be trained efficiently in an end-to-end fashion without the need of multi-stage training and testing. Finally, comprehensive experiments are conducted on three publicly available benchmarks, where the proposed method consistently outperforms the state-of-theart approaches, such as DFL-CNN (Wang, Morariu, and Davis 2018) and NTS (Yang et al. 2018). | ['Yu Gao', 'Weilin Huang', 'Xun Wang', 'Matthew R. Scott', 'Xintong Han'] | 2020-03-11 | channel-interaction-networks-for-fine-grained | https://www.semanticscholar.org/paper/Channel-Interaction-Networks-for-Fine-Grained-Image-Gao-Han/8585737f242285ad322acf23bc4943ddd50bf045 | https://pdfs.semanticscholar.org/8585/737f242285ad322acf23bc4943ddd50bf045.pdf | aaai-2020-2020-2 | ['image-categorization'] | ['computer-vision'] | [ 2.56521851e-01 -4.03428108e-01 -2.62584180e-01 -5.43082178e-01
-7.36292839e-01 -7.55999744e-01 8.35820436e-01 2.11134166e-01
-3.52869779e-01 3.01901758e-01 2.18354970e-01 -1.46262705e-01
-9.22765490e-03 -5.28033197e-01 -7.54997551e-01 -6.58850551e-01
-5.13397828e-02 -3.14928651e-01 7.48709589e-02 -2.12378576e-02
2.79513836e-01 1.37788683e-01 -1.38225234e+00 5.85581005e-01
7.41954327e-01 1.63965809e+00 2.69932717e-01 4.95166183e-01
-1.47922352e-01 7.37393618e-01 -2.57249437e-02 -2.41795734e-01
1.74963206e-01 -4.34751719e-01 -5.48832655e-01 2.07015753e-01
5.49098611e-01 -1.42183661e-01 -3.61514956e-01 1.15764701e+00
3.45067710e-01 -3.15270931e-01 6.32017374e-01 -1.29519296e+00
-7.07598746e-01 6.40465915e-01 -8.42512190e-01 5.29934987e-02
-4.94804531e-02 3.49641591e-01 1.31612480e+00 -9.29063380e-01
3.01121861e-01 1.40913701e+00 3.28857183e-01 2.91749865e-01
-1.40573215e+00 -9.96587813e-01 5.72227299e-01 2.94337630e-01
-1.35495782e+00 -2.00901225e-01 7.55629897e-01 -5.56504071e-01
4.56317306e-01 -9.25686806e-02 6.52584672e-01 1.09915805e+00
1.17711015e-01 1.05496824e+00 1.56564355e+00 -2.10420877e-01
1.05951801e-01 -1.13793731e-01 1.12066194e-01 6.73261285e-01
-1.06421217e-01 1.19195268e-01 -7.14320600e-01 8.02865699e-02
6.30674779e-01 3.20473373e-01 -2.47032493e-01 -4.00375485e-01
-1.38100159e+00 7.80499101e-01 9.79239583e-01 1.18087739e-01
-1.93180114e-01 1.88825756e-01 5.33001840e-01 1.46605507e-01
2.54540116e-01 1.56747147e-01 -5.61778367e-01 -1.00529253e-01
-5.76158106e-01 8.15792084e-02 4.27213699e-01 7.58689702e-01
1.01485133e+00 -5.35027146e-01 -4.90011543e-01 8.10070932e-01
4.27931279e-01 9.63678360e-02 2.99678981e-01 -6.60412610e-01
4.96573567e-01 6.96964383e-01 -3.80775422e-01 -7.98383832e-01
-1.85168773e-01 -6.11489832e-01 -1.15020478e+00 -1.29611045e-02
4.33585405e-01 1.80751517e-01 -9.47997570e-01 1.84599090e+00
1.44058064e-01 3.41266125e-01 -2.30660528e-01 9.28767025e-01
7.24405766e-01 3.65531534e-01 2.02063754e-01 1.90785572e-01
1.32838571e+00 -1.17411685e+00 -2.76768863e-01 -1.48825333e-01
4.07431215e-01 -6.33063555e-01 1.28801787e+00 3.63871187e-01
-7.73099840e-01 -6.77045047e-01 -1.06883264e+00 -7.03050345e-02
-4.62284774e-01 1.23923898e-01 6.62771583e-01 1.61809549e-01
-6.60948932e-01 3.89522135e-01 -6.59342945e-01 -1.61617801e-01
8.26780558e-01 3.41621116e-02 -2.88661212e-01 -3.45119059e-01
-1.04596448e+00 1.81138158e-01 2.19978377e-01 2.39764974e-01
-9.86306906e-01 -7.57238209e-01 -7.98193395e-01 1.46677479e-01
2.65938640e-01 -4.88568455e-01 9.07461286e-01 -1.08707392e+00
-1.29514825e+00 6.51101589e-01 -1.26702294e-01 -2.41604269e-01
5.40490627e-01 5.39138354e-02 -8.31260085e-02 2.21992940e-01
2.13922948e-01 9.39996958e-01 8.88898790e-01 -1.28971374e+00
-8.15976202e-01 -4.42488372e-01 2.76755035e-01 7.54040927e-02
-3.35293651e-01 -1.96188271e-01 -8.58738065e-01 -7.16275513e-01
-1.02761641e-01 -1.00395858e+00 -9.56213474e-02 2.15522617e-01
-7.16483772e-01 -6.16533458e-02 6.40991569e-01 -1.91384688e-01
1.05661368e+00 -2.48026252e+00 1.71643004e-01 2.63527811e-01
5.72100818e-01 -8.43699947e-02 -3.74891430e-01 2.32587636e-01
4.82208617e-02 2.99367607e-01 -2.23132238e-01 -3.80419582e-01
1.13022506e-01 -9.35491323e-02 -8.56888369e-02 4.12929356e-01
4.08235282e-01 1.16446114e+00 -6.25827491e-01 -3.09299678e-01
2.26760790e-01 5.99143684e-01 -4.64118809e-01 1.92629859e-01
-2.21411306e-02 7.29912281e-01 -3.29149485e-01 7.65347481e-01
8.22829545e-01 -5.18064857e-01 -1.95672692e-04 -5.10580480e-01
-5.96285090e-02 5.89082465e-02 -9.07182157e-01 1.78275776e+00
-7.50391662e-01 6.82489216e-01 2.30181031e-02 -8.54227841e-01
6.62278891e-01 -1.73677668e-01 1.39473304e-01 -1.08171833e+00
2.42993772e-01 1.40299305e-01 2.19616562e-01 -1.19684063e-01
8.03196952e-02 2.30164200e-01 -3.55104744e-01 2.66972572e-01
1.04032189e-01 3.82478565e-01 -2.32326519e-02 4.15993422e-01
8.81713569e-01 -2.85590500e-01 1.52729347e-01 -2.91991204e-01
4.94096518e-01 -5.52651584e-01 3.94547641e-01 9.04118538e-01
-3.74626219e-01 5.69169760e-01 5.51534057e-01 -2.19413996e-01
-6.52216196e-01 -1.12508559e+00 -2.63921589e-01 1.17028224e+00
5.96089244e-01 -4.20656860e-01 -6.51720941e-01 -6.53823614e-01
3.03138286e-01 -2.11588487e-01 -1.08863080e+00 -9.78125930e-02
-2.93862075e-02 -5.57168365e-01 4.13867474e-01 6.85282409e-01
8.90272796e-01 -6.65878356e-01 -4.10215825e-01 1.32191613e-01
-1.75537556e-01 -1.25959730e+00 -6.55638754e-01 4.86080855e-01
-5.78595221e-01 -9.56906557e-01 -5.35618007e-01 -6.36721969e-01
4.13269788e-01 6.04597211e-01 8.00915241e-01 4.95719202e-02
-5.22509336e-01 1.42658934e-01 -4.91762906e-01 -2.10660860e-01
3.02805364e-01 9.08628628e-02 -5.59599459e-01 4.26905692e-01
2.52305299e-01 -2.55888462e-01 -1.00104976e+00 3.41875941e-01
-8.29427481e-01 2.23340347e-01 8.56892586e-01 1.31969690e+00
9.73901391e-01 -2.74500214e-02 3.98428261e-01 -9.11232710e-01
4.84869987e-01 -5.74157536e-01 -3.83133531e-01 1.80101007e-01
-5.40550530e-01 1.17837258e-01 7.03972101e-01 -4.20122623e-01
-8.74194920e-01 1.41288444e-01 1.76914975e-01 -4.60105121e-01
-2.14709193e-01 5.66101968e-01 -4.86068338e-01 -3.13397080e-01
2.18576446e-01 2.55028963e-01 -2.15338901e-01 -4.92932975e-01
5.38107991e-01 7.02426672e-01 5.46375930e-01 -5.14712572e-01
5.63113987e-01 7.48428047e-01 -2.35454172e-01 -4.46122169e-01
-1.04184353e+00 -6.58777297e-01 -8.84285271e-01 -9.02687758e-02
8.84048820e-01 -1.33410871e+00 -5.42863011e-01 7.50325084e-01
-8.65839183e-01 -4.80541855e-01 1.86332285e-01 2.31998414e-01
-2.83721179e-01 1.15627974e-01 -6.59609139e-01 -3.10342938e-01
2.27993652e-02 -1.36011302e+00 1.40199351e+00 3.76594156e-01
2.29200095e-01 -9.69249129e-01 -3.71163905e-01 2.90798128e-01
3.57945621e-01 2.53863811e-01 7.78733432e-01 -2.17740670e-01
-7.86950111e-01 -5.00226915e-02 -9.47791219e-01 3.69614333e-01
9.36784968e-02 -6.49609836e-04 -1.21108449e+00 -3.07080567e-01
-6.03873909e-01 -4.90596056e-01 1.45094800e+00 2.96119660e-01
1.82471848e+00 -7.16444105e-02 -2.69746333e-01 8.91275287e-01
1.46439385e+00 -1.93907902e-01 4.65314627e-01 4.27268654e-01
1.14232588e+00 4.25245345e-01 5.62474847e-01 3.80615205e-01
7.64048576e-01 5.66574097e-01 4.33701813e-01 -4.26486641e-01
-1.85996518e-01 -3.56584966e-01 5.97388037e-02 4.51060086e-01
2.49070227e-01 -1.22705050e-01 -5.84615588e-01 3.60521555e-01
-1.88246667e+00 -6.68652773e-01 -1.25055283e-01 1.84282887e+00
5.79684556e-01 1.78682253e-01 -7.29265669e-03 4.58692946e-02
6.24651670e-01 2.75445163e-01 -7.84240484e-01 -1.11888349e-01
-3.01884711e-01 9.23613906e-02 5.93836904e-01 3.24857742e-01
-1.34209645e+00 9.68207657e-01 4.85835218e+00 9.63564634e-01
-1.52139056e+00 1.36108026e-01 1.00070322e+00 1.44682169e-01
-2.21272960e-01 -8.76104366e-03 -5.44893384e-01 7.21171081e-01
4.23973531e-01 1.14233330e-01 6.13824904e-01 5.67745984e-01
-7.74333403e-02 -2.43498117e-01 -1.02580714e+00 1.28269839e+00
7.04658171e-03 -1.30699277e+00 -8.65875557e-02 2.22049087e-01
7.59167254e-01 2.43282005e-01 3.48761261e-01 1.07303873e-01
4.81593497e-02 -1.10866523e+00 9.95460570e-01 3.98217261e-01
8.49755406e-01 -8.10343146e-01 5.21551728e-01 1.22559398e-01
-1.70721281e+00 -2.97572553e-01 -2.49269962e-01 -4.01626937e-02
-3.00232112e-01 6.41367972e-01 -2.19262108e-01 3.93743783e-01
9.85009432e-01 1.09732902e+00 -8.82903457e-01 8.97626102e-01
-1.69935822e-01 5.20835459e-01 1.29821241e-01 7.57688507e-02
4.64383513e-01 7.45396391e-02 -4.60035279e-02 1.45517862e+00
-8.62564892e-03 -2.42885619e-01 3.15665424e-01 8.03305626e-01
-2.96965569e-01 3.77545469e-02 -2.56346136e-01 -5.39033674e-02
5.23408115e-01 1.60775185e+00 -8.45450163e-01 -2.27903172e-01
-6.22077346e-01 1.33208418e+00 4.00340557e-01 4.39445615e-01
-6.92065179e-01 -3.82258713e-01 8.30197215e-01 -2.18331695e-01
6.18119299e-01 -7.46042877e-02 -4.72758740e-01 -1.17712903e+00
4.17256951e-02 -7.10241139e-01 4.00016636e-01 -3.90356511e-01
-1.74605286e+00 5.60091138e-01 -3.53262812e-01 -1.31276214e+00
2.77894408e-01 -6.34755492e-01 -5.99418581e-01 9.74411368e-01
-2.03222227e+00 -1.50589478e+00 -7.65131116e-01 9.30531383e-01
4.26548600e-01 4.98237908e-02 4.11054462e-01 3.90974522e-01
-5.75855136e-01 9.56464231e-01 8.04238468e-02 3.16771448e-01
8.10635626e-01 -1.33905637e+00 2.27184623e-01 6.18672252e-01
1.43989129e-02 5.46682000e-01 8.13629031e-02 -1.44229129e-01
-1.44351840e+00 -1.52450800e+00 4.94354337e-01 -1.90050364e-01
7.88019657e-01 -8.65559816e-01 -8.63025367e-01 2.41407335e-01
1.09226309e-01 6.35809124e-01 8.12274158e-01 1.20879985e-01
-1.12584925e+00 -2.74020225e-01 -6.63934231e-01 2.39117086e-01
1.05448186e+00 -1.10161698e+00 2.07743898e-01 -4.97287549e-02
6.18237436e-01 -9.11994427e-02 -7.08375275e-01 2.12419838e-01
8.62867177e-01 -1.05871737e+00 7.80675709e-01 -2.53367156e-01
6.88277721e-01 -4.00685579e-01 -2.84571618e-01 -1.47349572e+00
-4.60527241e-01 -1.82971433e-01 1.83903173e-01 1.26982260e+00
4.26755667e-01 -5.04006565e-01 2.85606861e-01 1.54083595e-01
9.11067426e-02 -9.40858424e-01 -6.24394298e-01 -6.80479288e-01
1.92768142e-01 -5.71463645e-01 6.30880415e-01 7.70686269e-01
-3.06436438e-02 3.49654198e-01 -3.01111162e-01 1.14179179e-01
7.28068590e-01 4.46587116e-01 6.70018196e-01 -1.22012985e+00
-4.88373697e-01 -6.01018369e-01 -5.29436290e-01 -1.21946955e+00
2.08330646e-01 -9.96992528e-01 2.01282218e-01 -1.17437875e+00
7.83834338e-01 -5.36285281e-01 -6.95606947e-01 3.52420092e-01
-4.92061108e-01 5.87427795e-01 5.14993846e-01 3.74939501e-01
-9.57581580e-01 5.96404135e-01 1.33438444e+00 -4.36316460e-01
9.65466648e-02 -4.86457735e-01 -1.17532074e+00 5.31812012e-01
4.90983188e-01 -1.76280037e-01 -3.32988709e-01 -3.59594643e-01
-1.20246097e-01 -2.79672414e-01 7.16476321e-01 -7.76126087e-01
2.24044859e-01 8.20420757e-02 6.34203911e-01 -3.31117958e-01
-3.10642961e-02 -6.97190225e-01 -3.39242160e-01 2.18062252e-01
-6.55025363e-01 -2.53811002e-01 7.27327839e-02 8.35847318e-01
-5.84020138e-01 5.50958574e-01 9.79051948e-01 7.71958828e-02
-7.28239596e-01 7.64151096e-01 -1.62976347e-02 2.51828786e-02
1.00029135e+00 -8.52890611e-02 -3.58619750e-01 -2.75821328e-01
-3.79604578e-01 4.05768663e-01 4.38358784e-01 5.94820380e-01
4.09540921e-01 -1.46779680e+00 -4.11415488e-01 2.71052420e-01
6.92828774e-01 -7.77932629e-03 6.35688901e-01 9.61208403e-01
1.00267537e-01 2.22933248e-01 -2.41866872e-01 -1.02614188e+00
-1.09925675e+00 3.59581143e-01 2.08925486e-01 -1.74485385e-01
-2.34435782e-01 1.03724802e+00 1.00727940e+00 -2.72836089e-01
3.20967168e-01 -5.15231609e-01 -2.86961310e-02 2.38860354e-01
6.84486151e-01 -4.23439369e-02 -4.80093993e-02 -6.98171556e-01
-4.17043895e-01 6.71791375e-01 -3.59635979e-01 3.28850374e-02
1.03789544e+00 -4.47876066e-01 9.18518081e-02 4.77496445e-01
1.83139253e+00 -2.50675917e-01 -1.77825296e+00 -4.70899522e-01
-3.01460445e-01 -7.54554391e-01 2.89124221e-01 -1.03086841e+00
-1.25624728e+00 1.35069692e+00 8.90222073e-01 8.74631107e-02
1.44171143e+00 2.74089158e-01 7.65697300e-01 -4.72468436e-02
2.34860882e-01 -9.18640912e-01 3.29115301e-01 4.66528952e-01
6.24483466e-01 -1.54255676e+00 -4.68237370e-01 -4.50618714e-01
-7.67691851e-01 1.05186939e+00 5.22141159e-01 -1.04015525e-02
9.45890367e-01 2.37276092e-01 5.04730865e-02 -3.65592837e-02
-7.49743223e-01 -4.09516513e-01 3.09451193e-01 4.26005751e-01
4.48189110e-01 4.41944361e-01 -7.90865975e-04 7.00475276e-01
2.26741508e-01 -5.19204885e-02 1.47158921e-01 6.89978898e-01
-1.64885551e-01 -1.03715265e+00 8.00482929e-02 5.45707524e-01
-3.08226377e-01 -3.50682229e-01 -6.24134004e-01 5.27155101e-01
4.16491479e-01 9.16784346e-01 2.54031539e-01 -8.07801664e-01
1.33178502e-01 -4.81683850e-01 4.07248765e-01 -4.16889220e-01
-4.59512681e-01 3.54808092e-01 -4.43374068e-01 -9.11039174e-01
-4.02761221e-01 -6.67841554e-01 -1.16758585e+00 -8.04717615e-02
-1.97306111e-01 -2.13648275e-01 7.07961142e-01 8.46041620e-01
6.35999203e-01 6.24548972e-01 1.02955234e+00 -8.59540880e-01
-2.30992138e-01 -8.12475801e-01 -7.29388177e-01 5.00382721e-01
5.68719029e-01 -7.77828097e-01 -3.09683681e-01 4.62600440e-02] | [9.59562873840332, 2.074218273162842] |
a953741d-ea63-4d90-a2df-c6e94c402b77 | entropy-driven-sampling-and-training-scheme | 2206.11474 | null | https://arxiv.org/abs/2206.11474v5 | https://arxiv.org/pdf/2206.11474v5.pdf | Entropy-driven Sampling and Training Scheme for Conditional Diffusion Generation | Denoising Diffusion Probabilistic Model (DDPM) is able to make flexible conditional image generation from prior noise to real data, by introducing an independent noise-aware classifier to provide conditional gradient guidance at each time step of denoising process. However, due to the ability of classifier to easily discriminate an incompletely generated image only with high-level structure, the gradient, which is a kind of class information guidance, tends to vanish early, leading to the collapse from conditional generation process into the unconditional process. To address this problem, we propose two simple but effective approaches from two perspectives. For sampling procedure, we introduce the entropy of predicted distribution as the measure of guidance vanishing level and propose an entropy-aware scaling method to adaptively recover the conditional semantic guidance. For training stage, we propose the entropy-aware optimization objectives to alleviate the overconfident prediction for noisy data.On ImageNet1000 256x256, with our proposed sampling scheme and trained classifier, the pretrained conditional and unconditional DDPM model can achieve 10.89% (4.59 to 4.09) and 43.5% (12 to 6.78) FID improvement respectively. The code is available at https://github.com/ZGCTroy/ED-DPM. | ['Xi Li', 'Shoudong Ding', 'Yang Chen', 'Taiping Yao', 'Hui Wang', 'Guangcong Zheng', 'Shengming Li'] | 2022-06-23 | null | null | null | null | ['conditional-image-generation'] | ['computer-vision'] | [ 3.35203350e-01 2.87341662e-02 1.80652320e-01 -5.24538994e-01
-9.02715862e-01 -1.39078230e-01 4.49503988e-01 -1.10836916e-01
-3.32233995e-01 5.59362471e-01 3.03375065e-01 -1.68973301e-02
-1.02243409e-01 -8.77876878e-01 -5.13646126e-01 -1.11958301e+00
-7.94723704e-02 -6.40737042e-02 2.94670314e-01 -4.54151854e-02
3.54365438e-01 1.01691544e-01 -1.42514384e+00 3.81220192e-01
1.24595225e+00 1.20328248e+00 6.07812047e-01 6.93188667e-01
-1.00442469e-01 6.08278394e-01 -5.14345765e-01 -3.54582071e-01
4.16258752e-01 -3.75933737e-01 -3.66353691e-01 1.29472390e-01
1.86857894e-01 -3.74534488e-01 -2.77245313e-01 1.59815419e+00
7.56718278e-01 1.88726738e-01 7.63156414e-01 -9.40385282e-01
-8.20088565e-01 5.66655695e-01 -7.59812295e-01 2.85481036e-01
3.10266055e-02 1.73512474e-01 5.73193550e-01 -8.83151531e-01
5.85985839e-01 1.34202158e+00 4.82487649e-01 7.38016784e-01
-1.04770708e+00 -7.19241440e-01 1.50773972e-01 2.18427598e-01
-1.31033921e+00 -3.58966082e-01 9.81818080e-01 -3.58467191e-01
3.47153008e-01 1.66155800e-01 3.84360909e-01 1.12766171e+00
3.77477854e-01 7.47822583e-01 1.60480917e+00 -1.15314193e-01
2.79482812e-01 1.16996929e-01 1.10807031e-01 4.76502061e-01
7.55016059e-02 3.32538247e-01 -5.00615656e-01 -5.93175665e-02
6.96710289e-01 -1.85727522e-01 -4.83255416e-01 2.00753763e-01
-7.24260271e-01 6.23659134e-01 4.26775485e-01 1.63481578e-01
-3.80388349e-01 -3.90430354e-02 1.93536237e-01 6.71472028e-02
6.26454294e-01 -1.60527259e-01 -4.42221612e-01 -5.46350181e-02
-1.13786101e+00 1.04135282e-01 5.89020371e-01 7.37504005e-01
6.88516736e-01 2.29799703e-01 -3.90966982e-01 8.94997060e-01
5.39016306e-01 5.75827718e-01 6.69706702e-01 -1.29142070e+00
2.31393531e-01 2.70356923e-01 -1.37108177e-01 -1.09255016e+00
2.29014568e-02 -8.51724923e-01 -1.39697349e+00 2.43075147e-01
1.55059695e-01 -1.64422318e-01 -1.21229064e+00 1.65858614e+00
2.80593216e-01 2.95610249e-01 -2.18689423e-02 7.96583772e-01
7.86587834e-01 8.33947957e-01 9.92062092e-02 -4.09930378e-01
1.14695179e+00 -7.62541294e-01 -8.21156919e-01 -1.25324860e-01
1.94447845e-01 -8.78365219e-01 9.72252607e-01 7.11774945e-01
-8.86243582e-01 -6.67721808e-01 -1.16344750e+00 1.89135611e-01
5.76557815e-02 1.44276246e-01 2.52270937e-01 6.58472359e-01
-1.15977859e+00 9.08955634e-01 -7.99536407e-01 -6.72867615e-03
6.78312659e-01 2.86225323e-02 -1.80172205e-01 -3.31691265e-01
-1.10769820e+00 5.48263311e-01 4.36645091e-01 2.18829557e-01
-1.03964090e+00 -5.32008827e-01 -4.48480397e-01 -1.66346103e-01
3.80427465e-02 -3.95016760e-01 7.70555258e-01 -9.73089397e-01
-1.48817194e+00 5.75388789e-01 -1.34585217e-01 -4.99909312e-01
7.89482296e-01 -1.29023299e-01 -2.48121321e-01 3.26692492e-01
1.66084796e-01 7.08433032e-01 1.24987543e+00 -1.61224079e+00
-4.05964106e-01 -4.29572642e-01 -2.82514364e-01 2.72209108e-01
-2.74342448e-01 -1.78229079e-01 -3.29112291e-01 -9.69687641e-01
4.92083967e-01 -6.36154532e-01 -5.04528999e-01 7.87134841e-02
-4.51961040e-01 9.17178541e-02 8.31206322e-01 -1.07527721e+00
1.11979210e+00 -2.26835465e+00 -9.04692560e-02 1.44834027e-01
2.62672454e-01 1.15483172e-01 6.96777478e-02 -1.49638265e-01
2.19839197e-02 3.52373391e-01 -6.79058731e-01 -4.65576470e-01
-2.66571581e-01 3.16954672e-01 -1.12557918e-01 4.09781367e-01
1.96570054e-01 3.66661280e-01 -8.74381006e-01 -4.83630866e-01
1.17762655e-01 7.17074037e-01 -4.62720364e-01 4.64738719e-02
1.02999754e-01 6.39782608e-01 -4.59924579e-01 6.80951774e-01
1.45175636e+00 3.38819437e-02 -8.96053687e-02 -3.93465757e-01
1.27108976e-01 -3.08656275e-01 -1.39529610e+00 1.57502472e+00
-2.24568680e-01 3.86987954e-01 3.95605981e-01 -9.67295229e-01
1.05674338e+00 2.00146288e-01 3.57416481e-01 -4.65330362e-01
1.78784594e-01 2.57223547e-01 -1.41842321e-01 -5.11933744e-01
2.99055338e-01 -2.79259443e-01 3.47465336e-01 -1.11838998e-02
2.03079134e-01 -7.09561929e-02 -2.21386980e-02 1.81721970e-01
1.09283447e+00 6.84645995e-02 -6.22835942e-02 -4.95058984e-01
5.98649502e-01 -4.06406373e-01 8.90976489e-01 8.08333874e-01
-6.48575366e-01 9.19136107e-01 2.43635029e-01 -7.88299274e-03
-9.00224626e-01 -1.08756042e+00 -3.95770699e-01 7.74820685e-01
4.05280173e-01 -1.01117454e-01 -9.46274519e-01 -7.07904458e-01
-3.12629521e-01 6.68243289e-01 -4.23401207e-01 -2.82774985e-01
-4.07494456e-01 -1.27216291e+00 3.78903359e-01 2.16440722e-01
1.10415363e+00 -6.80816531e-01 -1.30016506e-01 1.87187403e-01
-4.33002502e-01 -9.28375423e-01 -4.02643472e-01 2.46948093e-01
-1.09802961e+00 -7.34113455e-01 -7.79614806e-01 -6.56460404e-01
7.40589678e-01 1.00719087e-01 9.07156944e-01 1.38736993e-01
-7.20919594e-02 1.64435506e-01 -4.14749682e-01 -3.34413648e-02
-5.41000605e-01 -3.73000562e-01 1.00430315e-02 1.94008842e-01
1.26273677e-01 -8.86851668e-01 -1.00977981e+00 3.13211411e-01
-1.00768006e+00 5.71225025e-02 6.62275910e-01 1.00198996e+00
7.94606864e-01 3.34677458e-01 4.05468553e-01 -4.27529275e-01
7.42682457e-01 -5.58790922e-01 -3.96804154e-01 -1.39157906e-01
-8.22006941e-01 7.62412623e-02 5.06133556e-01 -4.51829672e-01
-1.55218339e+00 6.00569621e-02 -5.28649569e-01 -2.71269381e-01
-2.30360478e-01 3.05968225e-01 -2.19295084e-01 -2.27096472e-02
8.50260198e-01 5.63702226e-01 -1.05895258e-01 -6.72292292e-01
3.33671629e-01 6.99370384e-01 5.37652552e-01 -6.25204980e-01
6.99030936e-01 6.27106547e-01 -8.01740810e-02 -8.54835629e-01
-7.28720486e-01 -3.40814978e-01 -3.67561191e-01 -2.83684492e-01
8.93661559e-01 -9.62357044e-01 -1.91168383e-01 6.68638229e-01
-1.09318006e+00 -2.13688388e-01 -2.82332808e-01 2.90744215e-01
-3.69396538e-01 7.26838350e-01 -8.33014846e-01 -9.92050052e-01
-5.78381300e-01 -1.26463735e+00 7.77158499e-01 2.65361845e-01
3.46555978e-01 -8.20267320e-01 -4.13676322e-01 3.23724180e-01
4.54223990e-01 3.02647144e-01 5.35121560e-01 -4.26160842e-01
-5.43303728e-01 -2.04735577e-01 -2.57224679e-01 1.14249277e+00
-5.76733574e-02 -3.49854259e-03 -1.09296453e+00 -1.54231340e-01
7.84889996e-01 -2.12350748e-02 1.24551392e+00 7.36319840e-01
1.31593120e+00 -2.77258694e-01 -7.43196905e-02 7.19387591e-01
1.62840426e+00 1.60076499e-01 9.21474934e-01 1.36109874e-01
4.26595002e-01 4.05440450e-01 5.83306849e-01 6.04013979e-01
1.99488297e-01 2.20861211e-01 4.86251444e-01 1.91622719e-01
-6.51361287e-01 -2.29141355e-01 5.36287665e-01 9.66766238e-01
-8.84833783e-02 -2.53036499e-01 -5.94366133e-01 3.50255609e-01
-1.58713162e+00 -9.03139114e-01 -2.94960201e-01 2.06915092e+00
1.05193543e+00 3.26698720e-01 -5.80815554e-01 2.39087209e-01
9.30132866e-01 1.51538447e-01 -4.80222106e-01 9.63792019e-03
-3.80210966e-01 3.35110426e-02 5.25277078e-01 6.63665771e-01
-1.16054165e+00 6.04787111e-01 5.39545631e+00 1.51359069e+00
-9.14288998e-01 3.57110173e-01 1.16734815e+00 2.86904305e-01
-2.15505049e-01 2.58497261e-02 -7.17730403e-01 8.68591249e-01
6.17573202e-01 -2.29284782e-02 3.28172326e-01 7.25416064e-01
4.58179802e-01 -3.79457980e-01 -5.87787807e-01 1.05787301e+00
-1.51016921e-01 -1.05565953e+00 1.64295256e-01 -2.33516730e-02
9.33197021e-01 -1.16900735e-01 2.93053210e-01 8.10227320e-02
1.32747859e-01 -8.35882902e-01 7.52006829e-01 7.06765056e-01
5.48313677e-01 -4.53419298e-01 8.49589407e-01 5.77321351e-01
-9.80706632e-01 -1.58561856e-01 -5.11982977e-01 1.24510840e-01
8.33748952e-02 1.24475777e+00 -2.90126979e-01 4.85471308e-01
9.32151020e-01 6.29959106e-01 -4.20977592e-01 1.04174995e+00
-4.18818563e-01 8.38393688e-01 -3.68610799e-01 3.57199788e-01
-7.82787651e-02 -3.71532202e-01 8.58205199e-01 1.16329384e+00
6.76312625e-01 1.49066135e-01 9.96380299e-02 7.62287080e-01
-5.84051237e-02 -6.03325814e-02 -1.92202896e-01 5.28418481e-01
3.60924989e-01 1.26776385e+00 -8.51807952e-01 -4.20303524e-01
5.17431609e-02 1.23147678e+00 -8.78040642e-02 5.06438613e-01
-7.45767951e-01 -1.67728916e-01 3.45174521e-01 8.94259512e-02
1.65177047e-01 -1.50044516e-01 -7.55422413e-01 -1.20544267e+00
2.06439972e-01 -7.48762846e-01 2.12507024e-01 -7.20419705e-01
-1.48633730e+00 7.29681194e-01 -2.13077247e-01 -1.11809933e+00
2.81138510e-01 -3.95435661e-01 -6.63570523e-01 8.82263303e-01
-1.59042645e+00 -9.18902934e-01 -5.22170067e-01 4.48211432e-01
7.89230824e-01 4.47429083e-02 4.60279047e-01 5.05137324e-01
-4.24423397e-01 5.33938885e-01 4.05554116e-01 -1.17631853e-01
8.27953458e-01 -1.25551999e+00 -6.55600652e-02 1.14978516e+00
-2.79359400e-01 4.22899425e-01 9.96897638e-01 -7.56630301e-01
-9.20240521e-01 -1.11366701e+00 3.75092268e-01 -1.25388131e-01
4.66219634e-01 -7.95687884e-02 -1.01368344e+00 4.47021760e-02
2.03164667e-01 -2.75088269e-02 3.00703824e-01 -4.68298435e-01
-1.21123418e-01 -3.10349494e-01 -1.61122429e+00 5.18323541e-01
9.92660224e-01 -1.43415555e-01 -2.43678436e-01 2.54021287e-01
6.63517416e-01 -3.44421446e-01 -9.42284882e-01 5.64639747e-01
1.66746780e-01 -1.17463231e+00 9.22638297e-01 3.12043846e-01
7.04479039e-01 -3.89199287e-01 -4.51414734e-01 -1.11713445e+00
-3.26083571e-01 -6.25946522e-01 -2.26022556e-01 1.35339594e+00
3.26859117e-01 -6.20305538e-01 7.38158941e-01 4.02196467e-01
-2.47257054e-01 -8.75256419e-01 -1.06678545e+00 -6.20966613e-01
1.06269352e-01 -6.62919044e-01 2.60021001e-01 6.48858011e-01
-4.72985029e-01 -4.29297499e-02 -4.88295496e-01 3.22417676e-01
1.07222593e+00 -3.84188384e-01 2.02206582e-01 -9.58289683e-01
-3.19785714e-01 -4.17080671e-01 -3.10543954e-01 -1.42411089e+00
-3.71608049e-01 -8.32701564e-01 3.24161202e-01 -1.43166006e+00
2.77073443e-01 -3.79501224e-01 -4.58737761e-01 1.21828392e-01
-3.07173431e-01 2.89965779e-01 1.08379357e-01 3.16868335e-01
-2.15255201e-01 7.89282143e-01 1.52918708e+00 -1.79352671e-01
5.11780903e-02 7.82618672e-02 -7.75212228e-01 7.75720656e-01
1.12120354e+00 -5.71268260e-01 -4.69138145e-01 -3.76423925e-01
-1.44731715e-01 -7.72726983e-02 5.02806723e-01 -1.15447354e+00
2.06276283e-01 5.77441566e-02 5.46701908e-01 -4.82880414e-01
2.87942708e-01 -5.71437001e-01 -1.30111068e-01 4.97202218e-01
-1.82913229e-01 -4.05906200e-01 -2.50335503e-02 9.25127149e-01
-3.60672116e-01 -4.89779949e-01 1.16442251e+00 -2.52385825e-01
-6.93294227e-01 4.31440532e-01 -3.68619412e-01 7.02604428e-02
7.11882353e-01 -2.89897114e-01 -2.32863307e-01 -5.10258973e-01
-9.98194098e-01 1.44538775e-01 2.50542134e-01 5.03407083e-02
8.92863095e-01 -1.23864007e+00 -7.99551070e-01 8.40828270e-02
-4.34394389e-01 6.31895214e-02 6.68282032e-01 7.11065710e-01
-4.27443027e-01 -3.87162924e-01 1.12935781e-01 -6.98369563e-01
-8.66073489e-01 2.38578647e-01 3.38614583e-01 -3.41958195e-01
-5.52868247e-01 1.12129354e+00 1.55537769e-01 -3.45887065e-01
2.19295874e-01 -2.69989669e-01 -1.53332531e-01 -1.59454256e-01
4.17983234e-01 5.26839852e-01 -3.80957462e-02 -5.74684978e-01
-3.60381156e-02 4.68081206e-01 6.44215867e-02 -1.85359299e-01
1.14887214e+00 -4.60555166e-01 -2.60586798e-01 4.73127887e-03
1.21219540e+00 -1.48506865e-01 -1.74272263e+00 -6.44767880e-02
-1.36272222e-01 -4.79711741e-01 3.87807667e-01 -9.49212134e-01
-1.25014985e+00 8.18950415e-01 1.19695520e+00 1.37904316e-01
1.44442272e+00 -2.79450417e-01 8.13701868e-01 -1.88417137e-01
1.12687238e-01 -1.26971364e+00 1.28082186e-01 4.02041227e-01
9.58097398e-01 -1.33221984e+00 -1.78216007e-02 -5.91512263e-01
-6.15541697e-01 1.06547892e+00 5.81778824e-01 -2.36172810e-01
1.07060480e+00 3.71435225e-01 -2.76347380e-02 1.47533072e-02
-5.29073119e-01 7.65709206e-02 9.85513628e-02 7.62724638e-01
2.13286430e-01 6.79649264e-02 -4.66240764e-01 8.42545509e-01
-1.01161323e-01 -2.51930449e-02 4.72984761e-01 6.63770735e-01
-7.26028085e-01 -8.59738946e-01 -4.74259734e-01 5.35398364e-01
-6.76897168e-01 -2.96445638e-01 2.91767061e-01 1.64805129e-01
5.11097908e-01 1.23882949e+00 -2.23033041e-01 -4.76362497e-01
1.00661188e-01 -1.62293807e-01 1.68872744e-01 -2.12151542e-01
-9.35002118e-02 3.62296820e-01 -1.77589446e-01 -4.34958607e-01
-3.87421221e-01 -5.27067423e-01 -1.11636126e+00 -1.41392499e-01
-3.60952735e-01 1.74582247e-02 6.13453388e-01 6.49039924e-01
3.56534511e-01 5.11313736e-01 6.14389777e-01 -8.69394958e-01
-7.68909991e-01 -1.07144570e+00 -6.53116405e-01 4.15150970e-01
1.53159320e-01 -4.29649413e-01 -7.31438577e-01 1.68595210e-01] | [11.464888572692871, -2.387092113494873] |
de221061-cd72-487d-8e22-f3611a18201e | cross-lingual-genqa-a-language-agnostic | 2110.07150 | null | https://arxiv.org/abs/2110.07150v3 | https://arxiv.org/pdf/2110.07150v3.pdf | Cross-Lingual Open-Domain Question Answering with Answer Sentence Generation | Open-Domain Generative Question Answering has achieved impressive performance in English by combining document-level retrieval with answer generation. These approaches, which we refer to as GenQA, can generate complete sentences, effectively answering both factoid and non-factoid questions. In this paper, we extend GenQA to the multilingual and cross-lingual settings. For this purpose, we first introduce GenTyDiQA, an extension of the TyDiQA dataset with well-formed and complete answers for Arabic, Bengali, English, Japanese, and Russian. Based on GenTyDiQA, we design a cross-lingual generative model that produces full-sentence answers by exploiting passages written in multiple languages, including languages different from the question. Our cross-lingual generative system outperforms answer sentence selection baselines for all 5 languages and monolingual generative pipelines for three out of five languages studied. | ['Alessandro Moschitti', 'Eric Lind', 'Rik Koncel-Kedziorski', 'Luca Soldaini', 'Benjamin Muller'] | 2021-10-14 | null | null | null | null | ['generative-question-answering'] | ['natural-language-processing'] | [-3.77762645e-01 6.94252104e-02 3.17858338e-01 -4.69164789e-01
-2.06191921e+00 -1.15694606e+00 6.62345946e-01 -3.34869474e-01
-1.24654628e-01 1.12042749e+00 6.06857836e-01 -4.29957807e-01
1.19363956e-01 -1.14561462e+00 -6.00746810e-01 -1.90720111e-01
6.08659744e-01 1.27610135e+00 1.17219305e-02 -7.71750271e-01
-1.36539921e-01 -3.63783479e-01 -9.42190528e-01 8.13503563e-01
1.46218622e+00 2.67853677e-01 4.53295261e-02 9.82648909e-01
-5.42298317e-01 9.71864820e-01 -9.67064023e-01 -1.21764112e+00
-1.80485800e-01 -1.01699579e+00 -1.23623407e+00 -2.18756184e-01
6.02349877e-01 -2.36929223e-01 -2.73686126e-02 5.37390232e-01
8.34034681e-01 -1.00728720e-01 6.32019460e-01 -7.50724733e-01
-1.47972679e+00 1.09530830e+00 1.42800540e-01 1.54882222e-01
8.78877521e-01 1.52322188e-01 1.43614972e+00 -1.07376814e+00
9.27238703e-01 1.72176218e+00 4.74096507e-01 8.47059429e-01
-1.23204160e+00 -3.44007939e-01 -1.38038084e-01 9.89640504e-02
-1.42077315e+00 -3.35145801e-01 4.58292276e-01 -4.31929715e-02
1.24718475e+00 3.40241700e-01 1.25418097e-01 1.13994205e+00
2.29540691e-01 9.32304859e-01 1.08150685e+00 -6.08513653e-01
-1.18026817e-02 5.70623204e-02 -4.70288545e-02 5.53086519e-01
-1.20980993e-01 -4.63162690e-01 -5.10880888e-01 -3.55846025e-02
1.74675271e-01 -7.37403035e-01 -2.03002542e-01 4.19143170e-01
-1.10976756e+00 1.35955405e+00 -2.12765671e-03 1.44798577e-01
-1.71406940e-01 -1.06140263e-01 2.95839071e-01 6.13548934e-01
6.73331141e-01 6.85670674e-01 -5.34098208e-01 3.75370570e-02
-5.79692602e-01 6.50501847e-01 1.35873961e+00 1.22603726e+00
7.49018312e-01 1.19176298e-01 -7.88761437e-01 9.88679409e-01
2.05028132e-01 1.11124575e+00 4.80783552e-01 -8.22948515e-01
7.41514087e-01 5.45532227e-01 -9.28296968e-02 -3.96497071e-01
-4.81185652e-02 -4.19645727e-01 -2.82800943e-01 -7.76377916e-01
3.48163426e-01 -5.22037983e-01 -8.48332644e-01 1.72330153e+00
4.12935227e-01 -7.74699986e-01 8.02521288e-01 6.04409695e-01
1.63967669e+00 1.04164493e+00 5.48740998e-02 2.29305968e-01
1.72826314e+00 -1.12784660e+00 -8.08185935e-01 -4.43411559e-01
5.08988202e-01 -1.11181355e+00 1.43675840e+00 -5.36189154e-02
-1.19640803e+00 -6.77054346e-01 -5.63464344e-01 -6.64413929e-01
-4.51330692e-01 3.34979475e-01 3.43289375e-01 7.34221756e-01
-1.25597763e+00 -4.55302835e-01 -3.29506248e-01 -3.08766872e-01
-2.40059778e-01 -1.48459211e-01 8.04432109e-03 -5.82296848e-01
-1.65891922e+00 9.29702699e-01 4.22736347e-01 -2.91207403e-01
-1.08057702e+00 -4.94463235e-01 -1.06698537e+00 -1.12527072e-01
2.23599732e-01 -1.11674249e+00 1.53377950e+00 -2.00992286e-01
-1.34310448e+00 1.02837074e+00 -4.16934401e-01 -5.17781794e-01
2.86645144e-02 -5.30435503e-01 -5.95070302e-01 1.59126386e-01
6.07267082e-01 8.02801609e-01 5.28743267e-01 -9.83541071e-01
-2.87746459e-01 -3.39120507e-01 3.97539377e-01 4.84415412e-01
1.02566235e-01 2.00097710e-01 -4.78390753e-01 -5.74134707e-01
-1.30159691e-01 -8.41626585e-01 2.43432890e-03 -1.18994379e+00
-5.82988679e-01 -7.16689289e-01 5.56028783e-01 -1.04165816e+00
1.20769525e+00 -1.55255377e+00 4.65540923e-02 -2.95470893e-01
-2.05736786e-01 -3.74576822e-03 -3.89880568e-01 9.21221316e-01
4.18710560e-01 3.27196009e-02 -1.30828246e-01 -2.89081544e-01
4.73396510e-01 5.37365556e-01 -6.91126883e-01 -4.13211793e-01
7.72485554e-01 1.64990509e+00 -9.70222712e-01 -5.49761653e-01
-3.20171475e-01 3.68707508e-01 -6.59520209e-01 3.03770900e-01
-8.57012689e-01 5.99044144e-01 -5.59521198e-01 6.54000580e-01
1.76521808e-01 -6.96605630e-03 1.10178173e-01 3.12799692e-01
2.35031560e-01 1.05727887e+00 -5.79005659e-01 1.81487083e+00
-7.95115709e-01 1.84048057e-01 -2.33811468e-01 -1.35433450e-01
1.14837432e+00 5.50829828e-01 -2.25278646e-01 -8.51303518e-01
-1.72627494e-01 7.74939060e-01 -1.30736396e-01 -7.44479120e-01
8.00303280e-01 -2.80605644e-01 -6.72745049e-01 7.89517283e-01
6.65348649e-01 -1.00448489e+00 8.77266645e-01 5.93381643e-01
8.55633616e-01 4.46882367e-01 9.41420197e-02 -3.08074832e-01
7.07919121e-01 2.62067229e-01 6.47960678e-02 8.16650212e-01
3.73830795e-01 7.21138835e-01 3.83421987e-01 2.79278178e-02
-7.19760239e-01 -1.27774537e+00 8.99843425e-02 1.33121848e+00
-4.36338633e-01 -6.83034956e-01 -1.02814007e+00 -8.41012597e-01
-1.74001575e-01 1.27237153e+00 -3.52288902e-01 -4.20046598e-02
-9.10508037e-01 -6.16455495e-01 7.92296231e-01 3.15153629e-01
3.98863971e-01 -1.23509920e+00 -2.83510759e-02 4.17540401e-01
-1.00630140e+00 -1.18579769e+00 -6.52434647e-01 -1.16154291e-01
-3.88111144e-01 -9.08226490e-01 -7.99721301e-01 -9.25314903e-01
1.76189557e-01 -3.21843565e-01 2.25899577e+00 -3.55230719e-01
1.33204430e-01 6.68659210e-01 -6.25557780e-01 -3.80719364e-01
-8.89306068e-01 6.76270783e-01 -5.09914339e-01 -2.62949049e-01
6.63060725e-01 5.29787596e-03 -2.90677458e-01 -8.89234617e-02
-9.78388429e-01 -3.29304099e-01 3.53064269e-01 5.99161804e-01
7.14527726e-01 -5.50278246e-01 1.18494236e+00 -1.31916320e+00
1.24623477e+00 -6.88564479e-01 -2.72632182e-01 7.19959736e-01
-1.16591930e-01 4.12417293e-01 6.50781810e-01 5.51473200e-02
-1.47726870e+00 -4.56636697e-01 -8.12873185e-01 4.45786893e-01
5.99496774e-02 7.48163462e-01 -3.69461745e-01 5.90453088e-01
9.64336574e-01 3.46877068e-01 -4.83825505e-01 -2.94492126e-01
1.00979269e+00 6.50685251e-01 5.70290267e-01 -8.99217665e-01
7.65699148e-01 4.77436790e-03 -6.65601790e-01 -5.12057483e-01
-1.32565868e+00 -2.04495877e-01 -3.63629699e-01 -9.76525620e-02
1.18134892e+00 -1.05610847e+00 -1.22133993e-01 2.98501253e-01
-1.51197362e+00 -2.76658922e-01 -5.44833064e-01 7.35404342e-02
-3.86726588e-01 -9.21021216e-03 -9.39660072e-01 -6.45153046e-01
-8.58139634e-01 -9.71753061e-01 1.53486931e+00 1.65406853e-01
-4.72571224e-01 -1.36367357e+00 5.66600442e-01 7.58902192e-01
5.09180605e-01 3.15520652e-02 1.14851809e+00 -8.56907785e-01
-7.13281989e-01 -8.44019726e-02 1.30357891e-01 4.09844548e-01
-4.10971697e-03 -3.37510705e-01 -7.17546165e-01 -8.19710121e-02
1.64873481e-01 -1.01002979e+00 8.31563354e-01 -4.21401579e-03
1.64028019e-01 -4.92836386e-01 3.57320428e-01 1.14280805e-01
1.12863052e+00 4.36681742e-03 7.07505465e-01 7.12082759e-02
4.83902663e-01 8.88801873e-01 2.85825968e-01 -9.93366241e-02
1.33956885e+00 2.61780113e-01 -2.98047010e-02 4.16649543e-02
-6.00419402e-01 -4.49412376e-01 6.02583647e-01 1.45823538e+00
5.60383320e-01 -6.12048268e-01 -6.66505456e-01 9.04650807e-01
-1.35283077e+00 -6.65241838e-01 -5.66376388e-01 1.84294784e+00
1.37476587e+00 -4.29369718e-01 -1.04503438e-01 -6.28809810e-01
2.72378713e-01 6.45068660e-02 -2.15764612e-01 -5.86530864e-01
-7.07679987e-01 8.21371913e-01 -3.33756268e-01 8.64879847e-01
-8.50260198e-01 1.33098495e+00 6.07231331e+00 9.09238338e-01
-5.42939663e-01 4.38405842e-01 4.53833461e-01 2.33233228e-01
-1.13192666e+00 2.30578437e-01 -1.29065371e+00 1.69024467e-01
1.18520176e+00 -3.09315205e-01 2.34822541e-01 5.67849874e-01
-4.40898746e-01 -3.11460719e-03 -9.74188805e-01 5.42549729e-01
8.29140961e-01 -1.03953147e+00 4.38897580e-01 -5.82783639e-01
1.15514588e+00 4.70276736e-02 2.44215503e-02 9.91967559e-01
1.05262899e+00 -1.05008626e+00 6.38621807e-01 3.16644877e-01
7.87079215e-01 -8.52767169e-01 7.61269689e-01 2.18601227e-01
-1.01059043e+00 3.21161121e-01 -2.28374481e-01 3.67872000e-01
6.63452148e-01 4.18176055e-01 -8.71955037e-01 9.26424086e-01
5.81628442e-01 2.37198725e-01 -9.62269127e-01 4.02517140e-01
-1.07962561e+00 8.71635377e-01 -1.69851974e-01 -1.72398120e-01
4.07675982e-01 -2.08981797e-01 3.69065136e-01 1.17311060e+00
4.21897858e-01 4.91385236e-02 2.03043804e-01 9.55717146e-01
-2.08951667e-01 4.07075763e-01 -4.15514886e-01 -1.33755505e-01
2.64432222e-01 1.18012333e+00 -2.18580112e-01 -4.88271117e-01
-3.37878823e-01 1.15471625e+00 3.57781172e-01 4.23533559e-01
-4.76784557e-01 -5.02750576e-01 3.46771181e-02 -2.70760596e-01
4.72160801e-02 -3.08834553e-01 9.95951146e-03 -1.48140991e+00
-2.77326684e-02 -1.50068486e+00 8.63620102e-01 -9.67858672e-01
-1.62778616e+00 1.05528390e+00 -1.06603205e-01 -6.45698488e-01
-1.11652017e+00 -3.87398571e-01 -3.88752133e-01 1.27154064e+00
-1.57462251e+00 -1.53194368e+00 8.99668410e-02 6.86607599e-01
8.88654649e-01 -8.11693966e-02 1.19660687e+00 4.52426732e-01
-1.30095109e-01 4.45396125e-01 -1.07594050e-01 2.85017520e-01
9.11210299e-01 -1.44674456e+00 8.45355809e-01 9.85300303e-01
6.45924389e-01 7.93092072e-01 4.68665481e-01 -7.50627995e-01
-1.28022826e+00 -1.28434086e+00 1.75518668e+00 -1.10518348e+00
7.12106049e-01 -5.81504345e-01 -8.86790156e-01 8.55582356e-01
9.56872880e-01 -7.78110266e-01 8.78481328e-01 1.98002815e-01
-3.43760520e-01 8.38971436e-02 -7.86681473e-01 6.38572633e-01
5.56336045e-01 -7.89753497e-01 -8.40845823e-01 7.60123312e-01
1.11668372e+00 -5.82730770e-01 -8.05442750e-01 2.58116752e-01
7.10478947e-02 -6.48103416e-01 8.21918309e-01 -6.43139541e-01
6.28066242e-01 -2.12087512e-01 -3.16149354e-01 -1.59258997e+00
-9.96029004e-03 -5.60432971e-01 -5.69033474e-02 1.61363041e+00
9.21433926e-01 -6.83631837e-01 1.86227977e-01 1.34434342e-01
-4.43691790e-01 -4.87685859e-01 -7.91844726e-01 -6.36890769e-01
6.96600199e-01 -3.07220250e-01 8.45440149e-01 4.94375944e-01
-4.37832952e-01 1.30260634e+00 -9.74646956e-02 -2.17760324e-01
2.86735713e-01 5.44194579e-01 8.30431104e-01 -6.44544959e-01
-5.67960799e-01 4.78262380e-02 3.39008898e-01 -1.14259946e+00
3.56768340e-01 -1.20066345e+00 2.57074833e-01 -1.98553741e+00
6.48086891e-02 -1.60621703e-01 6.31910503e-01 2.43092135e-01
-7.38020122e-01 6.09088421e-01 -1.41283244e-01 -6.00171015e-02
-7.52342343e-01 8.56255829e-01 1.39568794e+00 -2.85438485e-02
1.38453558e-01 -2.47871056e-01 -9.80119944e-01 2.59187102e-01
6.69450641e-01 -3.75046372e-01 -4.96609479e-01 -1.22647679e+00
5.50615668e-01 8.78567845e-02 5.53862453e-02 -5.79502225e-01
-5.65197691e-03 2.78400242e-01 7.74847865e-02 -7.72886813e-01
2.93799937e-01 1.52302399e-01 -8.66587535e-02 1.39139459e-01
-3.52201700e-01 3.53413701e-01 1.59798816e-01 6.73611313e-02
-5.96655488e-01 -4.12283868e-01 3.07910413e-01 -4.72655386e-01
-4.30619150e-01 -6.22696914e-02 -4.34018612e-01 1.15266287e+00
4.91272479e-01 2.48074651e-01 -8.90938699e-01 -9.19819057e-01
-4.20085132e-01 5.60288012e-01 -8.89076442e-02 6.36093378e-01
5.63214839e-01 -1.34229898e+00 -1.66080296e+00 5.10357097e-02
3.62905651e-01 -1.27812237e-01 3.86930883e-01 4.56868351e-01
-6.09988868e-01 9.99482214e-01 4.83522087e-01 -3.42535138e-01
-7.26151645e-01 1.60839215e-01 3.10077220e-01 -7.29722500e-01
1.23385593e-01 1.25751972e+00 1.70991540e-01 -1.42081594e+00
-4.79015708e-01 2.12577600e-02 -3.40871125e-01 8.42305571e-02
3.63382041e-01 6.27270192e-02 3.72270375e-01 -7.49005020e-01
-1.24147438e-01 1.96956336e-01 -6.99122250e-02 -4.93570924e-01
7.61206031e-01 -2.91865826e-01 -6.11795604e-01 2.08542913e-01
9.47518945e-01 5.23417413e-01 -5.40166438e-01 -3.50753397e-01
7.71577060e-02 1.48473874e-01 -6.48965657e-01 -1.08645141e+00
-5.83821237e-01 7.64764547e-01 -8.01411867e-02 2.31615141e-01
9.73193645e-01 6.45198345e-01 1.30853999e+00 3.98256391e-01
5.07946432e-01 -9.33411837e-01 2.29580328e-01 1.29529667e+00
1.13740098e+00 -1.07306361e+00 -6.51654422e-01 -2.67674804e-01
-9.19959366e-01 7.54357338e-01 7.31041253e-01 6.00134581e-02
2.63803359e-02 -1.89673118e-02 6.03195190e-01 -3.57795686e-01
-9.19242144e-01 -5.88991404e-01 5.92060685e-01 5.07403910e-01
8.81130636e-01 1.69294715e-01 -5.18723309e-01 7.59369254e-01
-9.86553013e-01 -5.49015522e-01 2.95595199e-01 5.97312570e-01
-2.29235947e-01 -1.46917570e+00 -3.89490813e-01 2.43787393e-01
-7.13075340e-01 -8.19723129e-01 -7.83957899e-01 6.26660585e-01
-1.65843517e-02 1.51598918e+00 -1.69175088e-01 4.21958640e-02
3.50129128e-01 4.97601032e-01 5.61250627e-01 -1.12724090e+00
-8.40714753e-01 -2.81298757e-02 6.25294626e-01 -2.63818577e-02
-1.38997495e-01 -5.97084522e-01 -9.80406880e-01 6.29801005e-02
-2.36121789e-01 7.06847966e-01 3.75777423e-01 1.04570508e+00
4.01193231e-01 5.19415736e-01 2.23026037e-01 2.03540310e-01
-5.42374909e-01 -1.30002511e+00 -1.20777473e-01 1.26581877e-01
-5.55272326e-02 1.62828729e-01 -1.41971335e-01 1.86614141e-01] | [11.385845184326172, 8.355509757995605] |
f7410ef4-6261-464e-bda5-8e5c615fcf90 | gan-for-vision-kg-for-relation-a-two-stage | 2105.11789 | null | https://arxiv.org/abs/2105.11789v1 | https://arxiv.org/pdf/2105.11789v1.pdf | GAN for Vision, KG for Relation: a Two-stage Deep Network for Zero-shot Action Recognition | Zero-shot action recognition can recognize samples of unseen classes that are unavailable in training by exploring common latent semantic representation in samples. However, most methods neglected the connotative relation and extensional relation between the action classes, which leads to the poor generalization ability of the zero-shot learning. Furthermore, the learned classifier incline to predict the samples of seen class, which leads to poor classification performance. To solve the above problems, we propose a two-stage deep neural network for zero-shot action recognition, which consists of a feature generation sub-network serving as the sampling stage and a graph attention sub-network serving as the classification stage. In the sampling stage, we utilize a generative adversarial networks (GAN) trained by action features and word vectors of seen classes to synthesize the action features of unseen classes, which can balance the training sample data of seen classes and unseen classes. In the classification stage, we construct a knowledge graph (KG) based on the relationship between word vectors of action classes and related objects, and propose a graph convolution network (GCN) based on attention mechanism, which dynamically updates the relationship between action classes and objects, and enhances the generalization ability of zero-shot learning. In both stages, we all use word vectors as bridges for feature generation and classifier generalization from seen classes to unseen classes. We compare our method with state-of-the-art methods on UCF101 and HMDB51 datasets. Experimental results show that our proposed method improves the classification performance of the trained classifier and achieves higher accuracy. | ['Xiaonan Luo', 'BaoCai Yin', 'Jinghua Li', 'Shaofan Wang', 'Dehui Kong', 'Bin Sun'] | 2021-05-25 | null | null | null | null | ['zero-shot-action-recognition'] | ['computer-vision'] | [ 3.44879001e-01 1.07239902e-01 -2.77243286e-01 -2.60164827e-01
-1.24877371e-01 4.12523374e-02 5.80557346e-01 -3.89708430e-01
-1.12279974e-01 5.69229364e-01 2.27597520e-01 1.40603989e-01
1.83208678e-02 -1.48219252e+00 -4.72285360e-01 -8.76216233e-01
5.26602387e-01 2.71123290e-01 6.47516012e-01 -1.59399644e-01
-4.65536974e-02 2.28028432e-01 -1.81447995e+00 6.35215878e-01
7.79516816e-01 1.25613928e+00 9.38790441e-02 4.13398147e-01
-5.64517915e-01 9.98178363e-01 -7.87823796e-01 -3.21093917e-01
-1.12577609e-03 -9.20502663e-01 -7.30350733e-01 4.26568151e-01
-4.95223217e-02 -3.93158883e-01 -7.34134674e-01 1.08852160e+00
4.94224250e-01 5.81603587e-01 7.78639853e-01 -1.72733319e+00
-1.41771758e+00 4.61775273e-01 -1.09418094e-01 2.22258568e-01
2.41064087e-01 4.10159260e-01 6.00095153e-01 -7.65085220e-01
4.90524381e-01 1.43669188e+00 2.96885312e-01 1.07672405e+00
-6.83183193e-01 -7.98747778e-01 1.78814128e-01 7.44150996e-01
-1.36269927e+00 -3.47152382e-01 9.34538603e-01 -3.08620334e-01
9.75929976e-01 2.71382518e-02 7.35868096e-01 1.48118436e+00
1.40944809e-01 9.17271614e-01 5.72744846e-01 -3.17703694e-01
4.43917394e-01 2.37943083e-02 1.42306671e-01 5.15363395e-01
1.73298027e-02 4.89366725e-02 -2.68029481e-01 2.02169325e-02
7.92987347e-01 6.40014350e-01 -3.80789727e-01 -4.66358453e-01
-7.86646187e-01 9.69740272e-01 6.47529364e-01 3.81659657e-01
-1.55112132e-01 -1.84461530e-02 3.91371310e-01 2.72729874e-01
4.33318406e-01 -1.13978632e-01 -1.49763048e-01 3.04474793e-02
-1.55898049e-01 -1.85252443e-01 5.39178669e-01 1.09681666e+00
7.31197000e-01 3.07216197e-01 -5.08859396e-01 9.22858000e-01
3.20295572e-01 3.15508276e-01 1.23470271e+00 -1.76327899e-01
4.48014081e-01 1.11858547e+00 -3.96892279e-01 -1.00626719e+00
9.57497656e-02 -3.20810735e-01 -8.85093629e-01 7.95722008e-04
-1.31902754e-01 5.25655560e-02 -1.38777053e+00 1.63773394e+00
3.67852479e-01 5.71475446e-01 5.21633029e-01 8.31609070e-01
1.18827558e+00 7.44283259e-01 2.65838772e-01 -2.77925283e-01
1.13864231e+00 -1.06524181e+00 -7.61715353e-01 -2.63853043e-01
6.96470201e-01 -2.43886963e-01 1.16944492e+00 -1.35468468e-02
-3.34972203e-01 -9.39115405e-01 -1.15859902e+00 3.59870568e-02
-8.97103071e-01 -2.68873990e-01 5.15674829e-01 4.31390882e-01
-4.01334643e-01 7.26403952e-01 -5.82463861e-01 -4.61871356e-01
9.47211683e-01 1.07532538e-01 -2.08682045e-01 -4.32496637e-01
-1.45930362e+00 4.49139237e-01 7.91109383e-01 -2.05610394e-02
-1.20100093e+00 -3.28096271e-01 -1.15822124e+00 3.30370724e-01
5.75940669e-01 -4.44209725e-01 7.68904269e-01 -1.22264647e+00
-1.34090996e+00 5.03061891e-01 2.99866199e-01 -9.15793255e-02
1.82189360e-01 2.07110539e-01 -9.31820452e-01 -8.01833495e-02
4.86256927e-02 4.63996559e-01 9.87547278e-01 -8.75084460e-01
-6.09538078e-01 -3.55734617e-01 2.54195612e-02 2.47254878e-01
-5.84266961e-01 -5.30390978e-01 -2.04689875e-01 -5.41360259e-01
1.26760796e-01 -4.33618426e-01 7.81492889e-02 -1.69823915e-01
-2.25684226e-01 -4.61337835e-01 1.30900431e+00 -1.92514256e-01
1.15514159e+00 -2.47829008e+00 7.20707998e-02 -1.51095375e-01
-1.88174322e-02 4.41962332e-01 -3.79443705e-01 2.43932202e-01
-3.07104558e-01 -1.01113431e-01 -1.52847230e-01 2.51354069e-01
-9.78473425e-02 6.28264964e-01 -4.71622050e-01 8.86332989e-02
2.30106249e-01 1.19489229e+00 -1.19051015e+00 -3.07308257e-01
3.68669480e-01 3.56272310e-01 -2.43388563e-01 4.14594382e-01
-2.26241365e-01 1.28915282e-02 -7.08220541e-01 6.33846402e-01
4.11833316e-01 -1.92278683e-01 -6.87104762e-02 -1.46642715e-01
6.73477769e-01 -1.65696159e-01 -1.13175964e+00 1.59272051e+00
-3.06074232e-01 8.69946107e-02 -7.42764235e-01 -1.15641546e+00
1.11560905e+00 2.87091285e-01 6.44706562e-03 -6.26579821e-01
4.10275459e-01 -1.83809012e-01 6.31554350e-02 -8.16605270e-01
-1.57932434e-02 -4.23727304e-01 -4.63404097e-02 2.69119173e-01
5.00290334e-01 1.64875254e-01 -1.06457934e-01 1.36871114e-01
1.10777926e+00 2.39111736e-01 3.17282885e-01 8.41958448e-02
5.67534506e-01 -1.60253257e-01 6.50600970e-01 3.31635773e-01
-1.98230088e-01 3.76390994e-01 5.35441875e-01 -5.45542061e-01
-6.66716576e-01 -1.06235552e+00 6.89776465e-02 1.04614067e+00
3.93131167e-01 -1.90887094e-01 -6.46797836e-01 -1.18672967e+00
-1.05466954e-01 1.01033700e+00 -1.04700315e+00 -1.11929083e+00
3.98479626e-02 -6.25727177e-01 1.18422344e-01 8.31647515e-01
7.49913812e-01 -1.63192344e+00 -2.55269676e-01 2.24741191e-01
2.09124133e-01 -7.79548883e-01 -3.32924873e-01 8.88296962e-02
-7.28450179e-01 -1.40386832e+00 -5.15153229e-01 -8.42661679e-01
8.54258180e-01 2.29401976e-01 6.40023470e-01 1.48701444e-01
-4.27194685e-01 3.06431890e-01 -8.60029399e-01 -2.53870100e-01
-3.15875679e-01 -1.85985819e-01 4.16873321e-02 5.18750489e-01
8.89339864e-01 -6.37965918e-01 -4.25199181e-01 4.15158808e-01
-1.26790249e+00 -1.11606114e-01 4.62901175e-01 1.02653611e+00
4.29024577e-01 1.25702590e-01 7.81032085e-01 -9.61744070e-01
5.18392622e-01 -8.73353899e-01 -1.55131027e-01 3.77000064e-01
-5.48935175e-01 -4.37325612e-02 9.53312874e-01 -8.96434128e-01
-1.12375069e+00 -1.52617386e-02 1.06187902e-01 -8.93504918e-01
-2.26414412e-01 2.33912602e-01 -8.61037314e-01 1.94432005e-01
7.01595426e-01 6.18115604e-01 6.58633485e-02 -3.23740870e-01
5.83256543e-01 1.01352382e+00 4.02751684e-01 -2.61515141e-01
6.14831805e-01 2.34225661e-01 -2.21687078e-01 -6.02974057e-01
-9.63185787e-01 -2.06521392e-01 -6.26072943e-01 -1.02928102e-01
1.22598147e+00 -6.63407028e-01 -2.05171183e-01 6.12942278e-01
-8.43225539e-01 -3.26495767e-01 -7.21498847e-01 3.24091136e-01
-5.94081879e-01 2.28476822e-01 -5.02185106e-01 -5.77429771e-01
-1.70987859e-01 -9.38315988e-01 9.57419455e-01 5.69145441e-01
1.94233045e-01 -9.93753135e-01 -1.83415547e-01 1.84158996e-01
1.79226652e-01 3.55085939e-01 1.08255851e+00 -1.04311371e+00
-4.58675981e-01 -4.34941769e-01 -1.03417717e-01 5.73710322e-01
4.45213526e-01 -3.57966870e-01 -1.02760112e+00 -1.86944589e-01
4.52372544e-02 -6.47179842e-01 8.25405419e-01 5.46365773e-05
1.45030797e+00 -3.44351858e-01 -4.92196530e-01 5.50745070e-01
1.40973425e+00 6.46181703e-01 1.17450094e+00 -1.37069508e-01
8.74425054e-01 2.83980787e-01 6.35960937e-01 3.61642480e-01
-8.98846537e-02 3.22461218e-01 3.96618992e-01 2.37961575e-01
-3.39902043e-01 -6.00521207e-01 2.09465638e-01 7.02850521e-01
2.55371258e-02 -4.61289704e-01 -5.66517949e-01 4.40721750e-01
-1.92970622e+00 -1.10644627e+00 2.24680513e-01 2.00436687e+00
5.72993398e-01 1.94774732e-01 -2.13306740e-01 1.21710397e-01
8.47550094e-01 3.23386699e-01 -7.16777503e-01 -1.75179422e-01
2.10816249e-01 2.96847045e-01 1.18167643e-02 1.07339129e-01
-9.07057047e-01 1.11646783e+00 4.74407196e+00 1.42792761e+00
-9.92524743e-01 2.96866119e-01 4.85373080e-01 -2.41994224e-02
-1.42263100e-01 -1.10315673e-01 -7.24640310e-01 6.67500675e-01
7.21699834e-01 -2.63683110e-01 4.11351383e-01 1.28512156e+00
-3.66353035e-01 4.64845449e-01 -1.12769222e+00 1.00248504e+00
5.04058182e-01 -1.13593519e+00 6.34131849e-01 -7.46071562e-02
5.69116175e-01 -3.68170440e-01 -2.22097039e-01 1.03439939e+00
2.57149339e-01 -1.02029586e+00 1.97528228e-01 7.25148320e-01
9.40732241e-01 -8.89158428e-01 6.99195802e-01 4.93585676e-01
-1.30591834e+00 -3.24557066e-01 -1.02522600e+00 -1.03314601e-01
-2.32077777e-01 9.23120752e-02 -4.33565080e-01 6.05151951e-01
4.54324961e-01 9.90387559e-01 -5.68785429e-01 6.20302558e-01
-4.83945310e-01 3.08684051e-01 2.67173618e-01 -1.71726808e-01
2.90955186e-01 -1.77034363e-01 1.89776942e-01 5.33550501e-01
2.58910924e-01 6.10387206e-01 3.35806906e-01 8.72860670e-01
-2.12295696e-01 9.51861814e-02 -8.76445770e-01 -2.90211111e-01
2.97176540e-01 1.17153215e+00 -7.23168433e-01 -7.98802853e-01
-5.29363930e-01 1.25134778e+00 4.76040602e-01 4.71470922e-01
-9.27061796e-01 -8.97454500e-01 3.70692790e-01 -4.46666442e-02
4.46288139e-01 4.22732443e-01 3.52192044e-01 -1.33750892e+00
-1.01824448e-01 -6.24219358e-01 6.97903037e-01 -9.57955897e-01
-1.50343406e+00 6.14080429e-01 -1.27756476e-01 -1.50476933e+00
-3.06520704e-02 -5.46671987e-01 -9.62097764e-01 7.34202564e-01
-1.01893723e+00 -1.22823548e+00 -6.16481721e-01 8.76046538e-01
8.26256990e-01 -5.63093543e-01 9.84189391e-01 1.23068027e-01
-6.46525800e-01 5.57165086e-01 -7.51321986e-02 4.42309231e-01
2.75675684e-01 -7.26818264e-01 1.92801759e-01 6.69816077e-01
1.10074192e-01 3.25210512e-01 3.55186537e-02 -9.09313023e-01
-1.04321790e+00 -1.52172577e+00 4.73781139e-01 -2.63815701e-01
5.84154129e-01 -3.66331667e-01 -1.06691945e+00 8.35113525e-01
-1.87023267e-01 4.26215798e-01 8.80118430e-01 -1.83884069e-01
-3.15677464e-01 -1.57139406e-01 -1.09846580e+00 5.36100030e-01
1.38990712e+00 -5.36071301e-01 -1.07336617e+00 4.69344169e-01
1.03706205e+00 -1.18693866e-01 -6.23723030e-01 3.78654033e-01
2.49308690e-01 -6.38655066e-01 7.63397872e-01 -1.27107453e+00
6.09917402e-01 -1.74456537e-01 -2.13623926e-01 -1.42410123e+00
-6.18540645e-01 1.24437064e-01 -2.97315359e-01 1.36261415e+00
1.37507886e-01 -7.48344481e-01 7.77011573e-01 4.95108336e-01
-3.42957318e-01 -9.79053259e-01 -9.70936000e-01 -1.04737318e+00
-1.93986684e-01 -1.90476611e-01 7.91867733e-01 1.20930374e+00
-6.23670518e-02 7.12696135e-01 -3.09696764e-01 -1.53736874e-01
4.10304040e-01 1.72083855e-01 5.46889782e-01 -1.07313812e+00
-4.39405978e-01 -2.10561037e-01 -8.94653499e-01 -7.13778257e-01
2.70166665e-01 -1.12848580e+00 -1.40237123e-01 -1.52578449e+00
3.11852127e-01 -1.42550930e-01 -6.66281223e-01 7.99215496e-01
-3.30326110e-01 1.44347638e-01 -7.02177808e-02 2.80029234e-02
-6.69520199e-01 1.13087690e+00 1.60448778e+00 -3.83484811e-01
-1.87587902e-01 -1.22184999e-01 -5.48470080e-01 5.92548907e-01
5.21263301e-01 -4.80465055e-01 -1.03205490e+00 -2.02565774e-01
-3.29324633e-01 -3.67045775e-02 4.73884523e-01 -1.19321287e+00
7.29492214e-03 -3.23334813e-01 8.56204271e-01 -2.55084336e-01
4.08490747e-01 -8.15114558e-01 -3.44363251e-03 6.59492970e-01
-2.40042359e-01 -6.26863182e-01 -5.32457978e-02 1.01783228e+00
-1.00405701e-01 -4.06569451e-01 6.80288017e-01 -2.36187831e-01
-1.10738480e+00 8.40441883e-01 6.35273308e-02 2.25793466e-01
1.43069828e+00 -4.20060039e-01 -5.53609550e-01 -7.82703459e-02
-9.26183820e-01 1.87624648e-01 2.94735998e-01 8.36355805e-01
9.51228917e-01 -1.86560345e+00 -3.00196409e-01 6.24675512e-01
4.25519675e-01 7.10924435e-03 4.90074098e-01 3.55699241e-01
-2.20983014e-01 -1.42773598e-01 -4.40680057e-01 -2.69457906e-01
-8.52396607e-01 1.20574856e+00 3.81593406e-01 1.11851737e-01
-6.97661102e-01 8.97486210e-01 3.52129519e-01 -2.51593918e-01
8.74751657e-02 3.20755597e-03 -4.61363286e-01 1.53066635e-01
7.34053850e-01 2.32786283e-01 -3.82912219e-01 -5.84178746e-01
-2.84002513e-01 3.34604502e-01 -4.58123572e-02 4.18635666e-01
1.21249485e+00 2.95566231e-01 1.44545704e-01 5.91458738e-01
1.34489048e+00 -5.58656812e-01 -1.08217835e+00 -2.08379954e-01
-4.89975363e-01 -5.90975702e-01 -1.61872745e-01 -5.65873384e-01
-1.24859560e+00 9.87564743e-01 7.05419660e-01 2.36723170e-01
1.00128841e+00 1.41823702e-02 1.01980650e+00 2.00782314e-01
3.21651936e-01 -9.83922303e-01 3.64784747e-01 3.17602277e-01
7.42536485e-01 -1.07623386e+00 -2.56554276e-01 -4.49507982e-01
-6.80236220e-01 1.10004139e+00 1.10057116e+00 -3.32369506e-01
6.71647966e-01 -3.50813657e-01 -1.15635514e-01 -2.35278234e-01
-7.16039658e-01 -3.18065971e-01 3.17941040e-01 8.36019516e-01
-1.44242719e-01 -5.15934601e-02 -2.15812653e-01 1.10044789e+00
1.62658766e-01 8.35561901e-02 1.51402906e-01 9.23953116e-01
-6.92164123e-01 -9.88533556e-01 9.51800495e-02 6.62975252e-01
1.14381522e-01 -1.32123619e-01 -3.26666445e-01 7.48802006e-01
6.15372896e-01 8.13175082e-01 2.58174300e-01 -8.55892420e-01
3.56178820e-01 5.13846338e-01 3.23333949e-01 -1.01884377e+00
-5.09444550e-02 -3.87976289e-01 -1.58486649e-01 -5.87780535e-01
-4.70368005e-02 -1.20919965e-01 -1.33915651e+00 1.23746030e-01
-3.99903029e-01 2.26762176e-01 5.08609824e-02 1.12751818e+00
2.98775971e-01 9.39822435e-01 6.47813916e-01 -3.55278820e-01
-6.30311370e-01 -1.09222698e+00 -8.06784213e-01 8.83092821e-01
-2.20651522e-01 -9.26694930e-01 -5.28447032e-01 8.28836784e-02] | [9.942127227783203, 2.5872435569763184] |
69ab957b-5bbc-479a-828c-da085a8f1d86 | screening-for-rem-sleep-behaviour-disorder | 1910.11702 | null | https://arxiv.org/abs/1910.11702v1 | https://arxiv.org/pdf/1910.11702v1.pdf | Screening for REM Sleep Behaviour Disorder with Minimal Sensors | Rapid-Eye-Movement (REM) sleep behaviour disorder (RBD) is an early predictor of Parkinson's disease, dementia with Lewy bodies, and multiple system atrophy. This study investigates a minimal set of sensors to achieve effective screening for RBD in the population, integrating automated sleep staging (three state) followed by RBD detection without the need for cumbersome electroencephalogram (EEG) sensors. Polysomnography signals from 50 participants with RBD and 50 age-matched healthy controls were used to evaluate this study. Three stage sleep classification was achieved using a Random Forest (RF) classifier and features derived from a combination of cost-effective and easy to use sensors, namely electrocardiogram (ECG), electrooculogram (EOG), and electromyogram (EMG) channels. Subsequently, RBD detection was achieved using established and new metrics derived from ECG and EMG metrics. The EOG and EMG combination provided the best minimalist fully automated performance, achieving $0.57\pm0.19$ kappa (3 stage) for sleep staging and an RBD detection accuracy of $0.90\pm0.11$, (sensitivity, and specificity $0.88\pm0.13$, and $0.92\pm0.098$). A single ECG sensor allowed three state sleep staging with $0.28\pm0.06$ kappa and RBD detection accuracy of $0.62\pm0.10$. This study demonstrated the feasibility of using signals from a single EOG and EMG sensor to detect RBD using fully-automated techniques. This study proposes a cost-effective, practical, and simple RBD identification support tool using only two sensors (EMG and EOG), ideal for screening purposes. | ['Maarten De Vos', 'Michele T. M. Hu', 'Mkael Symmonds', 'Christine Lo', 'Fernando Andreotti', 'Navin Cooray'] | 2019-10-24 | null | null | null | null | ['sleep-staging'] | ['medical'] | [ 1.30037084e-01 -1.20926075e-01 -1.43282086e-01 -1.70446590e-01
-6.03456557e-01 -2.06527099e-01 -2.52547655e-02 -6.34715036e-02
-9.13820326e-01 1.16389930e+00 -1.03808500e-01 -3.84187728e-01
-3.93112868e-01 -2.97106415e-01 1.54920980e-01 -4.55898732e-01
-4.15058732e-01 4.29563075e-01 2.21157163e-01 6.69951588e-02
-3.87160107e-02 2.85568833e-01 -1.31635296e+00 -2.70116389e-01
1.08352876e+00 1.23044729e+00 7.19126046e-01 5.10059118e-01
5.78469336e-01 1.35017216e-01 -7.34634697e-01 2.98262164e-02
1.51793763e-01 -5.76565564e-01 -3.85692179e-01 -1.49717331e-01
-1.16935775e-01 -4.98928159e-01 2.21649006e-01 7.21550882e-01
8.90515089e-01 -5.21567613e-02 3.94691437e-01 -9.50072229e-01
-1.45040333e-01 -8.96989107e-02 -3.37773979e-01 7.51856148e-01
4.95086581e-01 5.06950080e-01 8.08092773e-01 -2.43846729e-01
2.86266118e-01 3.31038594e-01 9.96289015e-01 9.34026062e-01
-1.45772815e+00 -9.72209871e-01 -4.61413622e-01 5.99827290e-01
-1.24137998e+00 -4.67562854e-01 4.08853561e-01 -4.11016315e-01
1.42544222e+00 4.58912998e-01 1.63454747e+00 7.79085636e-01
7.77937531e-01 -4.22650613e-02 1.39346004e+00 1.99079197e-02
6.10420525e-01 1.32510424e-01 4.61190969e-01 4.46682572e-01
7.07981527e-01 -1.41179949e-01 -1.70146629e-01 -2.32679203e-01
4.64836657e-01 2.77732402e-01 -3.22460115e-01 3.31745595e-01
-7.32825458e-01 5.56973338e-01 -9.98694077e-02 2.97064096e-01
-5.75307727e-01 -1.39407068e-01 1.85230538e-01 3.99291903e-01
1.34361103e-01 5.39227188e-01 -3.91169041e-01 -6.49018705e-01
-1.20010769e+00 -1.73271567e-01 4.39283252e-01 4.78749663e-01
2.04229936e-01 -1.52027145e-01 -1.97183862e-02 9.36225355e-01
6.91330492e-01 8.48243117e-01 9.08457100e-01 -1.13322449e+00
5.52449636e-02 9.85674679e-01 2.71824241e-01 -3.44204128e-01
-1.12197530e+00 -2.88272947e-01 -7.23951995e-01 3.56212407e-01
7.77919888e-02 -7.02561587e-02 -6.19388402e-01 1.36045182e+00
-2.63035834e-01 -5.02624452e-01 -2.18919367e-01 8.52757573e-01
4.44129586e-01 -9.39724967e-02 -5.91300800e-02 -6.62003815e-01
1.77593708e+00 -2.03359440e-01 -4.35325891e-01 -6.09288096e-01
4.14555699e-01 -1.81765616e-01 1.28616023e+00 5.97842038e-01
-1.11639404e+00 -2.73099184e-01 -1.34631073e+00 2.58721292e-01
1.73861980e-01 2.47784436e-01 1.76887244e-01 1.04224694e+00
-1.17191410e+00 5.29081702e-01 -1.37924552e+00 -6.57452047e-01
6.32252872e-01 1.04341054e+00 -2.27443963e-01 2.84411788e-01
-8.58503282e-01 1.33656442e+00 -3.29577208e-01 3.38954851e-02
-6.48173511e-01 -2.56430477e-01 -4.19919938e-01 -4.83014762e-01
-4.26740825e-01 -9.40655053e-01 7.76128769e-01 -4.59328771e-01
-1.32005131e+00 1.11222064e+00 -3.33338767e-01 -6.31249607e-01
2.86449462e-01 -2.17690930e-01 -5.97756684e-01 5.84686399e-01
5.53221822e-01 2.57856786e-01 6.44643962e-01 -3.02972972e-01
-5.21305740e-01 -1.10433149e+00 -3.04586828e-01 1.08572669e-01
-8.55067223e-02 3.12526703e-01 3.76116395e-01 -2.82685101e-01
1.27273470e-01 -1.09302807e+00 -2.08199963e-01 -1.47706240e-01
-6.13231696e-02 -1.75653622e-01 2.39883512e-01 -1.09957993e+00
1.23313463e+00 -1.91367412e+00 -2.71127999e-01 1.65797412e-01
6.35966718e-01 1.16600238e-01 5.58075786e-01 -1.14013746e-01
1.18163511e-01 -2.42254771e-02 -3.69881332e-01 -3.46635580e-01
-2.10725456e-01 1.56855911e-01 6.20703101e-01 9.15547192e-01
-1.74644366e-01 8.30667615e-01 -6.81308031e-01 -1.92130223e-01
2.54367679e-01 4.09849733e-01 -1.46896675e-01 4.06953730e-02
5.30702055e-01 4.27356958e-01 5.01640029e-02 8.47743988e-01
1.15897603e-01 -1.68225229e-01 7.69165903e-02 6.52350532e-03
-1.93818390e-01 6.60588145e-01 -8.12143445e-01 1.39791942e+00
-1.09071359e-01 5.83592832e-01 1.87451378e-01 -6.53485060e-01
1.02909756e+00 2.85030097e-01 5.43975055e-01 -8.72069120e-01
3.25201750e-01 4.45647389e-01 3.87276590e-01 -5.95050573e-01
-2.45070145e-01 -5.54920495e-01 2.08887964e-01 5.57965100e-01
-1.34204254e-01 9.89229605e-02 1.90273479e-01 -2.28307247e-01
1.65016830e+00 -2.78275818e-01 7.15314984e-01 -4.02976811e-01
2.63901263e-01 -1.77685216e-01 6.24050081e-01 5.61575532e-01
-4.80368316e-01 3.57780188e-01 1.67811349e-01 5.43140806e-02
-4.82151091e-01 -1.02929604e+00 -3.68876040e-01 2.53212214e-01
6.24852479e-02 -4.85714406e-01 -5.46846509e-01 -3.22333962e-01
-3.72072197e-02 6.08995616e-01 -3.51325840e-01 -4.25136149e-01
-2.51316279e-01 -1.13900626e+00 4.84303832e-01 7.17048228e-01
4.14302081e-01 -8.83463025e-01 -1.34773803e+00 3.03496689e-01
2.69003175e-02 -5.04829705e-01 -8.03428739e-02 5.90683877e-01
-1.33630407e+00 -1.10261142e+00 -9.08877134e-01 -3.80990416e-01
5.14770269e-01 -3.12044378e-02 6.26205385e-01 -1.98510438e-01
-4.21348006e-01 4.54945922e-01 -1.70364678e-01 -1.11351110e-01
2.55325697e-02 -2.29958445e-01 6.10889912e-01 -4.33677733e-01
9.03289020e-01 -1.16964161e+00 -9.67321277e-01 4.62725520e-01
-1.68593153e-02 -2.19602510e-01 8.91918302e-01 5.09848714e-01
8.20786893e-01 -3.97163004e-01 6.45214081e-01 -1.67214461e-02
8.42834413e-01 -3.33692521e-01 -2.47804806e-01 -2.35411629e-01
-1.42963934e+00 -4.08052534e-01 4.54486132e-01 -2.27355361e-01
-3.13555479e-01 -2.66892016e-01 -1.78343043e-01 -3.19025032e-02
-2.38273114e-01 -1.59916133e-02 -2.15312973e-01 1.97867438e-01
6.70668900e-01 3.88323575e-01 4.44288850e-01 -5.67800403e-01
-4.23168331e-01 1.10401297e+00 4.58179325e-01 1.96429729e-01
1.13240786e-01 5.33798099e-01 -1.50469497e-01 -1.06967068e+00
-1.03633121e-01 -7.38596082e-01 -5.23468256e-01 -5.52243777e-02
9.84013498e-01 -9.34702218e-01 -8.00579488e-01 3.17282617e-01
-4.33771282e-01 -4.52135235e-01 -2.82896459e-01 1.07954395e+00
-6.42686129e-01 3.40378970e-01 -2.25508347e-01 -9.12025332e-01
-1.19490111e+00 -1.02481329e+00 5.76766610e-01 7.36007690e-02
-1.20146930e+00 -4.09106642e-01 2.28659421e-01 5.47433794e-01
2.97963113e-01 9.79464650e-02 4.64589536e-01 -5.06289780e-01
1.27591208e-01 -2.72223294e-01 -5.13439551e-02 6.10605478e-01
6.06734157e-01 -7.36740589e-01 -3.51916671e-01 -2.87983567e-01
7.77938068e-01 1.15941904e-01 3.54474783e-01 5.55896163e-01
1.85780704e-01 -8.50973129e-02 -3.92287225e-01 4.32293236e-01
1.17865753e+00 9.88774121e-01 7.29528964e-01 7.72221982e-01
1.82165623e-01 -6.79552108e-02 1.27744347e-01 2.64815599e-01
4.30977702e-01 5.82579732e-01 2.39539176e-01 4.04426038e-01
-8.47261995e-02 5.27638972e-01 8.70890677e-01 3.54493320e-01
-7.17381299e-01 4.08815324e-01 -7.89772749e-01 5.06264269e-01
-1.15911472e+00 -7.31936157e-01 -2.96750933e-01 2.17650723e+00
6.79707766e-01 5.36097229e-01 6.53719366e-01 4.49539065e-01
4.65213269e-01 -4.34153080e-01 -8.64197314e-01 -3.55713814e-01
3.97524834e-02 7.96370983e-01 5.06405234e-01 -1.52606890e-01
-3.21268857e-01 -9.64353420e-03 5.51600790e+00 -2.32038330e-02
-8.88694346e-01 4.86652315e-01 -5.89420646e-02 -7.83544660e-01
2.86371410e-01 -1.77403331e-01 -8.50893795e-01 1.10582936e+00
1.42716765e+00 1.98515225e-02 5.40535271e-01 6.44284427e-01
8.22964728e-01 -7.50408113e-01 -8.20513248e-01 1.20681107e+00
3.17542888e-02 -7.84319520e-01 -7.10330784e-01 3.15712959e-01
-7.29377717e-02 5.59548080e-01 -4.69728351e-01 -2.03824311e-01
-5.97353876e-01 -8.34861517e-01 5.13926268e-01 7.94056535e-01
1.33494353e+00 -3.55470300e-01 9.15958226e-01 -4.07903008e-02
-1.00892842e+00 -3.55128437e-01 1.57602206e-01 -5.36237597e-01
2.45000914e-01 4.41234857e-01 -7.18451083e-01 -1.26481012e-01
1.15562785e+00 9.14808869e-01 -4.56962526e-01 9.99448657e-01
-4.68773752e-01 7.63662338e-01 -6.36132300e-01 -3.88046294e-01
-2.28533402e-01 -2.42217958e-01 8.67649913e-01 6.40233219e-01
4.10096824e-01 1.49657890e-01 -7.07040727e-01 1.00292432e+00
3.36775959e-01 -2.26900473e-01 -9.08184499e-02 1.17935553e-01
2.68365890e-01 1.00005817e+00 -9.37922597e-01 6.81122690e-02
-5.21971822e-01 8.78404498e-01 -4.05703604e-01 -1.63100302e-01
-4.30687308e-01 -6.14915729e-01 5.98283827e-01 6.54817879e-01
-1.02645472e-01 -1.47605196e-01 -6.79667532e-01 -9.43896830e-01
3.43818694e-01 -7.47883797e-01 3.82258266e-01 -7.26611674e-01
-1.04565227e+00 3.34152907e-01 -2.38865167e-01 -1.36501372e+00
-5.10906316e-02 -4.02577907e-01 -6.76737785e-01 1.04724586e+00
-8.04804683e-01 -4.57411170e-01 -3.23314846e-01 7.19385803e-01
1.76629826e-01 -1.50486082e-01 9.01978493e-01 2.48448491e-01
-7.24095404e-01 3.57151628e-01 3.79935727e-02 -2.96033591e-01
4.23923552e-01 -1.24796820e+00 -1.02226786e-01 5.16436279e-01
-6.46534026e-01 9.87080157e-01 3.63117516e-01 -9.31978166e-01
-1.14318252e+00 -7.52175570e-01 9.17771757e-01 -4.16503012e-01
4.45972890e-01 7.39463489e-04 -2.15444043e-01 3.66790444e-01
-3.00912738e-01 -7.59805024e-01 1.07515478e+00 -1.81697726e-01
6.30524933e-01 -4.24469918e-01 -1.61025143e+00 3.45490932e-01
1.02161252e+00 -5.27223885e-01 -1.25138831e+00 -1.81542143e-01
-7.87537098e-02 1.25743672e-01 -1.21118438e+00 3.03968132e-01
1.10333502e+00 -1.00465035e+00 6.56755805e-01 6.59680814e-02
-2.78960079e-01 -2.32110739e-01 9.17652398e-02 -7.70819068e-01
-2.72059351e-01 -8.43193769e-01 -2.11432517e-01 7.34586239e-01
3.11359823e-01 -9.84604776e-01 6.12335861e-01 8.32338810e-01
-4.04810429e-01 -8.55769038e-01 -1.18999648e+00 -8.43400717e-01
-5.72915435e-01 -5.00800490e-01 2.73139160e-02 1.37575373e-01
5.20125091e-01 5.33513367e-01 -1.59307905e-02 -1.91465110e-01
3.86714876e-01 -1.63809687e-01 5.87340258e-02 -1.49327767e+00
-1.14344001e-01 -3.24495196e-01 -9.24971640e-01 -5.27644515e-01
-4.71297532e-01 -9.70800519e-01 -3.95101100e-01 -2.08687258e+00
2.79766560e-01 -1.65499955e-01 -5.19926846e-01 5.71639538e-01
1.04874372e-01 5.24636924e-01 -2.56408423e-01 4.57558423e-01
-3.26118022e-01 1.79920480e-01 5.12063265e-01 1.36929974e-02
-7.35914886e-01 4.09540325e-01 -9.54564750e-01 7.40110040e-01
1.38979757e+00 -6.86559796e-01 -5.51679373e-01 1.15562670e-01
3.47288609e-01 -1.47637010e-01 5.30323923e-01 -1.28196466e+00
-3.97114120e-02 3.70112628e-01 7.46477127e-01 -5.01177013e-01
6.97700918e-01 -6.15150630e-01 4.19325858e-01 9.95894432e-01
5.15751779e-01 2.77423486e-02 1.53340489e-01 3.04128051e-01
5.26300251e-01 -1.96205124e-01 8.30242276e-01 -3.77428651e-01
-4.11648989e-01 -1.41972721e-01 -8.59230161e-01 -1.00944318e-01
9.17496383e-01 -1.19392359e+00 -1.27772227e-01 2.02747453e-02
-1.32646418e+00 5.42676784e-02 6.14784718e-01 -1.06367655e-01
6.89016700e-01 -7.82995522e-01 -1.63713336e-01 3.91123086e-01
-1.67616680e-01 -4.39758807e-01 3.85727808e-02 1.80257201e+00
-2.61441410e-01 4.13303703e-01 -7.66957104e-01 -5.28892934e-01
-1.34868240e+00 -9.52694938e-02 5.14789701e-01 4.02153954e-02
-8.89948368e-01 8.12867045e-01 -8.16965520e-01 4.43018138e-01
1.01668976e-01 -6.27110898e-01 -2.81615078e-01 3.25185210e-01
6.45560145e-01 1.05431271e+00 2.30813533e-01 -3.37329358e-01
-8.58730018e-01 5.29052019e-01 2.70334691e-01 -1.93771824e-01
1.42417669e+00 -5.41641593e-01 -3.30266923e-01 7.65739441e-01
8.22668672e-01 -9.43502933e-02 -7.62158871e-01 7.61144400e-01
-3.38925198e-02 2.98914015e-01 7.47795478e-02 -1.07218146e+00
-4.63664502e-01 5.39090216e-01 1.36246479e+00 4.11269605e-01
1.37236404e+00 8.80929753e-02 1.06017101e+00 1.94668129e-01
6.00591838e-01 -1.19081295e+00 -5.99013627e-01 -4.35617387e-01
5.96918941e-01 -6.33188426e-01 2.34024256e-01 4.56683248e-01
-3.79537553e-01 8.90551269e-01 2.63679981e-01 -1.79471970e-01
6.39581680e-01 -4.51677814e-02 -2.78504699e-01 -3.88816595e-01
-3.69998246e-01 -2.78156072e-01 1.17914788e-01 6.71317697e-01
6.66353106e-02 2.66098708e-01 -1.08296669e+00 1.34853613e+00
-3.95031989e-01 5.50176501e-01 2.78046548e-01 1.00205874e+00
-5.92622936e-01 -8.57473969e-01 -1.77095115e-01 1.55988848e+00
-6.17179990e-01 -4.74811308e-02 -3.27677876e-01 6.54289663e-01
4.11618084e-01 1.24052024e+00 1.38788775e-01 -4.66109753e-01
4.98459399e-01 2.36900896e-01 4.72388446e-01 -7.76179492e-01
-4.98731583e-01 -6.87921746e-03 2.80471325e-01 -6.95124090e-01
-7.18968928e-01 -1.02626240e+00 -1.50957155e+00 2.34743997e-01
-1.49231866e-01 1.56493902e-01 4.40142930e-01 9.70922947e-01
5.94833851e-01 2.59708583e-01 1.30819246e-01 -4.79340196e-01
-1.78791791e-01 -1.38684618e+00 -1.11476922e+00 -2.67894328e-01
4.62398738e-01 -8.64027202e-01 -5.36682248e-01 -1.04916103e-01] | [13.536670684814453, 3.388672113418579] |
c323d6de-bd1c-4c1d-8f55-225006f59f6d | on-the-role-of-supervision-in-unsupervised | 2010.02423 | null | https://arxiv.org/abs/2010.02423v2 | https://arxiv.org/pdf/2010.02423v2.pdf | On the Role of Supervision in Unsupervised Constituency Parsing | We analyze several recent unsupervised constituency parsing models, which are tuned with respect to the parsing $F_1$ score on the Wall Street Journal (WSJ) development set (1,700 sentences). We introduce strong baselines for them, by training an existing supervised parsing model (Kitaev and Klein, 2018) on the same labeled examples they access. When training on the 1,700 examples, or even when using only 50 examples for training and 5 for development, such a few-shot parsing approach can outperform all the unsupervised parsing methods by a significant margin. Few-shot parsing can be further improved by a simple data augmentation method and self-training. This suggests that, in order to arrive at fair conclusions, we should carefully consider the amount of labeled data used for model development. We propose two protocols for future work on unsupervised parsing: (i) use fully unsupervised criteria for hyperparameter tuning and model selection; (ii) use as few labeled examples as possible for model development, and compare to few-shot parsing trained on the same labeled examples. | ['Kevin Gimpel', 'Karen Livescu', 'Haoyue Shi'] | 2020-10-06 | null | https://aclanthology.org/2020.emnlp-main.614 | https://aclanthology.org/2020.emnlp-main.614.pdf | emnlp-2020-11 | ['constituency-parsing'] | ['natural-language-processing'] | [ 3.41933072e-01 7.72906721e-01 -2.63516694e-01 -7.51754344e-01
-1.30276239e+00 -7.89703786e-01 6.24510288e-01 3.96756202e-01
-6.90228045e-01 5.72599292e-01 3.77165616e-01 -5.63767016e-01
2.34355852e-01 -7.40460277e-01 -6.75767004e-01 -4.14293855e-01
5.97792007e-02 6.05699778e-01 3.62528056e-01 -3.53030376e-02
2.24694144e-02 -8.88816193e-02 -1.54177070e+00 1.72773772e-03
7.95662344e-01 3.79741043e-01 3.30198705e-01 1.12025714e+00
-4.10098761e-01 8.55945885e-01 -6.04252696e-01 -9.02655840e-01
1.01224050e-01 -4.76632565e-01 -9.90177512e-01 2.22561620e-02
5.13971388e-01 -2.44825751e-01 2.36330733e-01 9.61776495e-01
5.48115015e-01 3.69469970e-01 4.06268895e-01 -4.88380373e-01
-5.74812472e-01 1.33017147e+00 -7.02718720e-02 5.73614776e-01
7.34581426e-02 1.35793477e-01 1.39958739e+00 -6.74986959e-01
9.97581542e-01 1.03728187e+00 4.34358478e-01 1.05080891e+00
-1.29949462e+00 -5.36587715e-01 3.59461993e-01 -3.72159272e-01
-4.47582573e-01 -8.00367713e-01 6.20638132e-01 -3.30346137e-01
1.14653170e+00 5.76061159e-02 3.07702959e-01 1.30775058e+00
-3.81390035e-01 7.33976662e-01 1.00511229e+00 -1.04976749e+00
6.03853643e-01 -3.33963893e-03 1.04111242e+00 5.02031565e-01
2.84794569e-01 1.15426689e-01 -4.64933038e-01 -1.25351027e-01
1.77427873e-01 -8.39173377e-01 4.05992568e-01 1.84032977e-01
-8.72353435e-01 1.15979612e+00 -1.20105870e-01 4.42169607e-01
6.76891953e-02 2.85625979e-02 5.31824946e-01 3.12261313e-01
7.30235934e-01 8.08951497e-01 -8.46099019e-01 -4.28103983e-01
-8.55445981e-01 4.46706675e-02 8.55926633e-01 1.19816697e+00
7.91191161e-01 -1.75015837e-01 -2.53256917e-01 1.12792480e+00
9.62098092e-02 2.50515379e-02 5.19744039e-01 -1.16985381e+00
8.49241793e-01 1.92424864e-01 2.77584360e-04 -2.38701422e-02
-4.98708010e-01 -2.30165496e-02 -1.80319414e-01 -2.55866647e-01
6.94647551e-01 -6.52340293e-01 -1.10634995e+00 2.09035063e+00
3.48971039e-01 -2.36319174e-04 3.42882752e-01 2.38528728e-01
1.01300561e+00 5.92262685e-01 7.87415385e-01 -3.93125087e-01
1.52075708e+00 -1.06446350e+00 -5.88211775e-01 -7.66312361e-01
1.30013454e+00 -7.58510113e-01 1.37831032e+00 1.90607965e-01
-1.09209371e+00 -4.79402155e-01 -9.66804922e-01 -2.16589183e-01
-3.59011322e-01 5.19399121e-02 1.02228367e+00 1.05509388e+00
-9.85643446e-01 7.51420438e-01 -1.04965842e+00 -6.23714864e-01
1.74034119e-01 3.09698611e-01 -2.93681592e-01 1.99277085e-04
-1.10134995e+00 6.81386471e-01 5.33973277e-01 -3.46692771e-01
-5.53210676e-01 -4.14328516e-01 -1.12452519e+00 -1.88046414e-02
1.66069776e-01 -3.62165868e-01 1.56310737e+00 -6.07459903e-01
-1.66473246e+00 1.26493096e+00 -8.41698125e-02 -4.26830053e-01
-4.97865118e-03 -3.65899116e-01 -1.09463647e-01 6.16972987e-03
4.35507745e-01 7.90568054e-01 2.84655869e-01 -9.96984482e-01
-6.99511707e-01 -2.53605306e-01 3.60941678e-01 -4.19776663e-02
-2.80829489e-01 6.50682807e-01 -5.87831676e-01 -6.44808292e-01
1.16285145e-01 -9.27991807e-01 -6.03728235e-01 -8.92604530e-01
-6.44058108e-01 -3.06336790e-01 5.96897341e-02 -6.44839764e-01
1.12864995e+00 -2.12515807e+00 -3.48396488e-02 -3.07019562e-01
-2.08604842e-01 1.60865381e-01 -4.52124774e-01 1.98592946e-01
-2.89610736e-02 5.25014400e-01 -3.85807753e-01 -9.67768133e-01
-3.83189991e-02 2.21345127e-01 -5.85630462e-02 1.36814220e-02
4.84485626e-01 7.06435144e-01 -1.05511582e+00 -6.92808092e-01
-4.22949344e-02 -1.59179479e-01 -7.08484352e-01 3.91465932e-01
-4.52003449e-01 5.51556289e-01 -4.94828254e-01 6.27844214e-01
4.05891955e-01 -1.19482502e-01 2.91255862e-01 2.67332137e-01
-1.54973611e-01 6.46532118e-01 -8.32177103e-01 1.76746762e+00
-4.95795399e-01 4.76426721e-01 -9.50385854e-02 -8.67970228e-01
7.88721323e-01 3.56839299e-01 1.90178037e-01 -4.44526523e-01
3.08798671e-01 1.29989833e-01 1.82053313e-01 -3.38015378e-01
3.02555799e-01 -3.24034840e-01 -6.02975845e-01 6.14791214e-01
6.69238627e-01 -1.44422844e-01 9.04703379e-01 3.21656615e-01
1.37042511e+00 3.50813828e-02 1.37662202e-01 -3.31700712e-01
3.44878510e-02 1.86341628e-01 9.15316463e-01 1.03471386e+00
-2.67032325e-01 6.34094536e-01 7.07067728e-01 -3.39503050e-01
-1.10555100e+00 -8.68416369e-01 -4.14444566e-01 1.71321404e+00
-3.45677048e-01 -6.00846589e-01 -1.25055981e+00 -8.97693336e-01
-6.44142687e-01 1.23833621e+00 -7.91026771e-01 1.58882678e-01
-8.04132104e-01 -1.18413150e+00 5.39523840e-01 6.90660536e-01
1.22279100e-01 -1.39261150e+00 -4.42278117e-01 4.33526814e-01
1.21012360e-01 -1.26537609e+00 3.23622338e-02 9.06773150e-01
-1.03669858e+00 -7.79155195e-01 -5.35454035e-01 -1.12726164e+00
5.76642752e-01 -2.51422167e-01 1.40354288e+00 1.43748179e-01
1.35174682e-02 3.18749219e-01 -8.72147858e-01 -4.02746916e-01
-8.24667454e-01 4.49214160e-01 -2.17243776e-01 -6.57117486e-01
5.74669003e-01 -4.87231076e-01 -1.64345771e-01 -2.02258974e-01
-5.56352794e-01 8.21569562e-02 3.69884282e-01 9.19508219e-01
4.40523088e-01 -5.62926412e-01 5.56829929e-01 -1.83960557e+00
5.69854081e-01 -4.44374025e-01 -6.78456247e-01 4.74668771e-01
-4.26291823e-01 1.46455228e-01 7.22045362e-01 -3.94399703e-01
-1.25871873e+00 2.08229169e-01 -4.52349603e-01 1.48520455e-01
-4.04473126e-01 3.08224678e-01 -8.78524259e-02 3.16515952e-01
7.42241919e-01 -3.75803262e-01 -5.39790928e-01 -7.93109477e-01
6.87212229e-01 4.02041882e-01 3.65709573e-01 -9.32079792e-01
5.89642227e-01 -1.02134280e-01 -5.63624680e-01 -6.70176804e-01
-1.21002781e+00 -2.71165788e-01 -8.15935850e-01 2.11503521e-01
1.18696845e+00 -8.48237932e-01 1.90594286e-01 1.58987403e-01
-1.25851357e+00 -7.12543905e-01 -3.87351632e-01 5.31110764e-01
-5.33411026e-01 3.47698420e-01 -1.06139076e+00 -7.54459262e-01
-3.99345279e-01 -1.01083589e+00 1.05466628e+00 2.97711402e-01
-4.17675078e-01 -9.78173971e-01 5.76430798e-01 4.59085494e-01
-5.77954315e-02 4.41231430e-02 1.19471276e+00 -1.20538318e+00
-1.08258054e-01 -1.90728322e-01 1.85532272e-01 2.45599851e-01
-6.97165579e-02 3.95790696e-01 -1.25623381e+00 -6.59601241e-02
-3.17743793e-02 -6.46520555e-01 8.52729142e-01 5.37475765e-01
8.17699969e-01 4.92966250e-02 -1.10404491e-01 4.04751956e-01
1.05755162e+00 1.55989438e-01 2.94974625e-01 5.18347859e-01
5.23953915e-01 8.57609212e-01 7.56993771e-01 1.32677302e-01
4.05246198e-01 2.17306063e-01 4.85960627e-03 4.53794003e-02
-5.14217094e-02 -1.81647837e-01 2.06269637e-01 1.19939339e+00
-2.59304163e-03 -1.66761309e-01 -1.02001822e+00 6.49823725e-01
-1.56679308e+00 -6.01125956e-01 1.95613042e-01 2.15121269e+00
8.72370839e-01 6.79689527e-01 1.05536394e-01 -2.31090277e-01
9.26672757e-01 3.04060966e-01 -3.02348018e-01 -7.73913324e-01
-7.98936039e-02 6.56788170e-01 2.86850125e-01 5.90310812e-01
-1.31811893e+00 1.55683136e+00 7.05260134e+00 4.17348295e-01
-6.31494164e-01 4.22588527e-01 1.18399107e+00 -1.48100033e-01
-3.29524547e-01 5.12409747e-01 -1.20182145e+00 5.68769753e-01
1.51426852e+00 6.51828274e-02 1.64238676e-01 1.05525315e+00
-1.12669773e-01 -6.26261905e-02 -1.20972502e+00 3.63783598e-01
-3.17906260e-01 -1.20479679e+00 -3.72140080e-01 -2.26466879e-01
6.61927879e-01 2.88881570e-01 -3.24370205e-01 9.59285200e-01
9.35499132e-01 -7.20914960e-01 4.95313913e-01 -1.57296538e-01
6.01245582e-01 -4.91336673e-01 5.33330083e-01 5.50260842e-01
-1.02616775e+00 5.57554737e-02 -6.67716980e-01 1.27474656e-02
4.01028037e-01 4.61173654e-01 -4.65565860e-01 1.14124835e-01
7.28443265e-01 3.74339312e-01 -8.42627347e-01 5.55188119e-01
-4.67491180e-01 1.23000038e+00 -4.17202055e-01 -1.86402619e-01
3.96045417e-01 -1.26428545e-01 2.75429130e-01 1.41801405e+00
1.11685684e-02 3.41115683e-01 2.97015021e-03 5.04161596e-01
-1.83255985e-01 4.27387297e-01 -5.15686691e-01 -2.18477443e-01
7.45720387e-01 1.19250870e+00 -1.10465002e+00 -6.10653639e-01
-7.64662206e-01 4.79421467e-01 9.25544024e-01 1.60106406e-01
-4.63317841e-01 -2.10556582e-01 1.90771356e-01 -3.47028971e-01
2.51505733e-01 7.58524798e-03 -4.88783240e-01 -1.17910290e+00
-1.33122370e-01 -6.69373393e-01 6.69007957e-01 -2.56480455e-01
-1.25754631e+00 7.86596060e-01 8.23904648e-02 -9.75065708e-01
-4.72066045e-01 -7.24065304e-01 -1.09345353e+00 5.23413241e-01
-1.28868580e+00 -1.00192213e+00 2.72814721e-01 -1.33712515e-01
8.54562521e-01 -1.38204426e-01 1.23195362e+00 1.61706340e-02
-9.79176402e-01 7.52032518e-01 -2.96521303e-03 4.37955856e-01
6.51557803e-01 -1.49640977e+00 1.10906827e+00 1.13355994e+00
3.86669338e-01 7.61950016e-01 7.89087653e-01 -5.72390199e-01
-1.02580953e+00 -9.06124413e-01 1.13214982e+00 -6.84290707e-01
6.65596128e-01 -6.34862959e-01 -8.15312147e-01 1.04801142e+00
1.83596522e-01 -2.16060914e-02 1.09465873e+00 7.65780568e-01
-1.43279776e-01 2.97701955e-01 -1.08714592e+00 6.00847483e-01
9.99664307e-01 -1.53850779e-01 -8.35818946e-01 5.76184928e-01
1.13392353e+00 -1.69557884e-01 -1.02653456e+00 3.27342212e-01
2.01935381e-01 -9.62100685e-01 5.18395424e-01 -8.54985118e-01
5.97700596e-01 4.53557968e-01 -1.54791549e-01 -1.13724029e+00
-3.90407056e-01 -5.94449341e-01 1.06391370e-01 1.73019409e+00
9.55857038e-01 -4.17483836e-01 9.54788208e-01 1.16481376e+00
-2.86544174e-01 -6.71034694e-01 -9.44420874e-01 -6.61061943e-01
7.18984663e-01 -5.80847323e-01 3.79141420e-01 7.75134742e-01
1.72569171e-01 6.77606642e-01 -1.42791361e-01 5.71236610e-02
6.01904571e-01 -1.99903101e-01 7.82296240e-01 -1.17320931e+00
-6.43050909e-01 -2.40898296e-01 1.21104456e-02 -7.57793963e-01
3.42394263e-01 -6.82887018e-01 3.51471096e-01 -1.25276697e+00
3.31183761e-01 -6.39193416e-01 -2.46270463e-01 4.32031900e-01
-6.29338741e-01 -6.68441504e-02 2.18140855e-01 1.82699598e-02
-6.01742446e-01 1.68397337e-01 7.94984281e-01 1.32614136e-01
-2.34162912e-01 -1.86736807e-02 -9.00711596e-01 8.16836476e-01
8.23554158e-01 -6.99159920e-01 -1.69947162e-01 -6.76650703e-01
1.37294419e-02 1.23859301e-01 -5.08106947e-01 -7.93319225e-01
1.37983551e-02 2.64216680e-02 3.41486000e-02 -3.00615311e-01
1.96735248e-01 -7.11906552e-02 -3.88085485e-01 1.51981667e-01
-5.91441274e-01 -1.28534853e-01 2.00903714e-01 3.95654082e-01
-8.43951106e-02 -1.04101634e+00 8.48610759e-01 -5.24769723e-01
-7.78756440e-01 1.05199769e-01 -4.84686017e-01 4.50171202e-01
8.25925052e-01 -4.70755138e-02 -4.56112832e-01 4.28601913e-02
-1.06851530e+00 1.98977888e-01 5.24297953e-01 3.15924108e-01
-1.67759150e-01 -8.42703640e-01 -6.91027462e-01 4.05365899e-02
3.75439465e-01 4.79845740e-02 2.19806563e-02 2.72251189e-01
-4.44390446e-01 1.39351666e-01 2.07470506e-01 -2.53047049e-01
-1.12578142e+00 4.18198436e-01 -3.12500417e-01 -6.54686511e-01
-6.35530949e-01 1.16604149e+00 8.98966342e-02 -8.42568219e-01
1.97632611e-01 -2.07808718e-01 -4.20357317e-01 7.67743066e-02
5.11403024e-01 9.77712646e-02 3.60321291e-02 -3.39863390e-01
-2.57641017e-01 4.93081719e-01 -3.83488029e-01 -3.52094233e-01
1.56045091e+00 1.86647713e-01 1.90685913e-01 5.87085843e-01
1.00987530e+00 1.76613510e-01 -1.23986542e+00 -2.22986892e-01
5.80610693e-01 -7.16721592e-03 -2.17525423e-01 -6.49795473e-01
-7.18196929e-01 7.81186402e-01 3.32650930e-01 3.76992613e-01
7.31906176e-01 4.30130839e-01 5.38537443e-01 4.90222782e-01
2.77852207e-01 -1.41792750e+00 -9.09905434e-02 7.84292340e-01
1.28074899e-01 -1.54073000e+00 -1.04624413e-01 -5.80029547e-01
-6.81918681e-01 1.10298169e+00 6.70732558e-01 -3.14561129e-01
5.66382945e-01 4.15127128e-01 3.33214372e-01 -1.60814360e-01
-9.13985372e-01 -2.98395485e-01 -1.61426410e-01 4.50606346e-01
8.79864216e-01 1.23640522e-01 -5.31592429e-01 1.03017604e+00
-5.12391746e-01 -6.24422729e-01 3.53791356e-01 1.19862986e+00
-6.25803828e-01 -1.51729643e+00 -3.21902558e-02 6.32717907e-01
-7.38736451e-01 -4.95969057e-01 -4.12948012e-01 7.79688120e-01
-1.34663984e-01 1.26311290e+00 8.90552923e-02 -1.50558218e-01
1.85729593e-01 4.48225528e-01 3.74857128e-01 -1.51873696e+00
-6.56828463e-01 9.52464566e-02 6.83066487e-01 -1.47263303e-01
-3.33563983e-01 -7.31107652e-01 -1.04578686e+00 2.38601744e-01
-5.41882396e-01 2.56178617e-01 6.21992052e-01 1.15055597e+00
-9.20756068e-03 2.70186454e-01 4.93047535e-01 -8.59959066e-01
-5.53337753e-01 -1.41778195e+00 -4.47945148e-01 3.89686525e-01
-2.44207397e-01 -5.23610413e-01 -6.61615133e-01 2.05539435e-01] | [10.377419471740723, 9.637932777404785] |
5f776a23-0bdc-4ebd-b17e-2936ea4062c0 | semeval-2012-task-5-chinese-semantic | null | null | https://aclanthology.org/S12-1050 | https://aclanthology.org/S12-1050.pdf | SemEval-2012 Task 5: Chinese Semantic Dependency Parsing | null | ['Meishan Zhang', 'Yanqiu Shao', 'Ting Liu', 'Wanxiang Che'] | 2012-07-01 | null | null | null | semeval-2012-7 | ['semantic-dependency-parsing'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.2076497077941895, 3.744049549102783] |
849edd92-c846-4f84-a94e-9873885a778c | learning-to-generate-explanation-from-e | null | null | https://aclanthology.org/2022.coling-1.260 | https://aclanthology.org/2022.coling-1.260.pdf | Learning to Generate Explanation from e-Hospital Services for Medical Suggestion | Explaining the reasoning of neural models has attracted attention in recent years. Providing highly-accessible and comprehensible explanations in natural language is useful for humans to understand model’s prediction results. In this work, we present a pilot study to investigate explanation generation with a narrative and causal structure for the scenario of health consulting. Our model generates a medical suggestion regarding the patient’s concern and provides an explanation as the outline of the reasoning. To align the generated explanation with the suggestion, we propose a novel discourse-aware mechanism with multi-task learning. Experimental results show that our model achieves promising performances in both quantitative and human evaluation. | ['Hsin-Hsi Chen', 'Hen-Hsen Huang', 'An-Zi Yen', 'Wei-Lin Chen'] | null | null | null | null | coling-2022-10 | ['explanation-generation'] | ['natural-language-processing'] | [ 4.06118810e-01 1.26509428e+00 -5.82539499e-01 -7.80595958e-01
-6.13009393e-01 -4.46191579e-02 8.17015707e-01 4.01992947e-01
1.73540562e-01 9.21986878e-01 1.15171278e+00 -6.39439642e-01
-2.83752382e-01 -5.28455138e-01 -5.84188998e-01 -1.99663937e-02
1.53482288e-01 5.22498667e-01 -1.35982022e-01 -1.75844133e-01
4.00968254e-01 -2.57075459e-01 -9.42092896e-01 1.32376909e+00
1.19182217e+00 5.47821522e-01 2.21748173e-01 5.73154807e-01
-2.18172163e-01 1.48127639e+00 -6.92528248e-01 -7.52288043e-01
-4.70311016e-01 -8.29239964e-01 -1.39986885e+00 3.79304141e-02
-8.91342461e-02 -4.33829367e-01 6.95309713e-02 4.90198165e-01
1.29219115e-01 -5.01909107e-02 7.74436355e-01 -1.21965396e+00
-1.03898525e+00 1.32751954e+00 -1.13776132e-01 -9.59714234e-04
6.84704185e-01 -5.36643378e-02 1.16725552e+00 -6.11878395e-01
6.26300156e-01 1.46022546e+00 3.24565232e-01 1.14260399e+00
-1.05281115e+00 -4.64394480e-01 5.25114179e-01 8.78009975e-01
-5.53736031e-01 -1.55410126e-01 8.23891401e-01 -2.52542198e-01
1.21134973e+00 4.21958804e-01 8.93667042e-01 1.40319383e+00
6.24041855e-01 8.47874701e-01 7.52254307e-01 -3.65179002e-01
3.90623808e-01 1.60845503e-01 2.66432375e-01 7.65801191e-01
5.33545502e-02 6.15634620e-02 -7.24479556e-01 -2.86703348e-01
5.91341197e-01 -8.59252512e-02 -1.94402322e-01 2.49730930e-01
-1.21776640e+00 1.15336156e+00 6.35340750e-01 1.85546383e-01
-8.10467243e-01 3.22399616e-01 3.26776057e-01 -1.57355919e-01
5.16535461e-01 5.93799174e-01 -3.35256726e-01 3.62579413e-02
-6.86241508e-01 5.22018194e-01 9.58345592e-01 6.41630292e-01
-1.85624838e-01 6.28077686e-02 -6.06244504e-01 6.25332952e-01
5.33162057e-01 1.22337034e-02 5.91662705e-01 -1.13126552e+00
4.30224001e-01 7.87539303e-01 -9.61834379e-03 -1.01510024e+00
-6.16994321e-01 -3.95453364e-01 -9.07370090e-01 -2.33149901e-01
1.84608012e-01 -2.55397201e-01 -4.35745120e-01 1.53685081e+00
1.21061675e-01 1.14110671e-01 5.02970040e-01 1.01421356e+00
1.35614347e+00 6.11162364e-01 4.51000214e-01 -3.62675190e-01
1.56900656e+00 -1.34998822e+00 -1.10662258e+00 -5.74386716e-01
5.25857508e-01 -4.19851035e-01 8.73577297e-01 2.93188661e-01
-1.19096470e+00 -3.87204766e-01 -6.98337674e-01 -1.10403337e-01
3.55522722e-01 1.85802490e-01 8.52479160e-01 2.37023588e-02
-7.99831212e-01 2.75044084e-01 -5.43112159e-01 -3.23531240e-01
5.23958623e-01 8.94206092e-02 -1.17430948e-02 9.46489051e-02
-1.32173717e+00 1.26431501e+00 4.64192778e-01 -1.28286257e-01
-6.71431422e-01 -6.90393567e-01 -9.64479983e-01 4.41766292e-01
4.33772504e-01 -1.35068357e+00 1.70495021e+00 -6.91493511e-01
-1.42252684e+00 6.55514181e-01 -6.95782006e-01 -8.57188821e-01
4.04486954e-01 -2.28366539e-01 -3.85669500e-01 1.45382822e-01
6.74866289e-02 8.72160017e-01 4.25515860e-01 -1.32759631e+00
-4.82447088e-01 -1.71541736e-01 2.15602383e-01 7.78587013e-02
-3.65443788e-02 -1.68172225e-01 6.05907366e-02 -7.57483959e-01
5.07348999e-02 -5.85725963e-01 -4.79152083e-01 -3.00508529e-01
-8.44648302e-01 -5.96956849e-01 3.66542935e-01 -6.90970302e-01
1.37252593e+00 -1.56971252e+00 1.61669329e-01 -2.54431844e-01
5.88272929e-01 -5.42262457e-02 7.35159218e-02 4.87052023e-01
-3.37362349e-01 4.34908777e-01 -2.43505567e-01 -1.31424069e-01
-1.08947963e-01 3.05403143e-01 -6.78807199e-01 -2.42647663e-01
5.33336759e-01 1.16786730e+00 -9.90618646e-01 -7.62654424e-01
-2.08720807e-02 3.29194725e-01 -9.40157413e-01 5.46645939e-01
-6.26565814e-01 5.49730420e-01 -8.36423218e-01 3.75499815e-01
7.70590976e-02 -7.69997120e-01 3.24814498e-01 -1.71660677e-01
2.42807135e-01 6.64318740e-01 -3.98833752e-01 1.39245367e+00
-3.66244107e-01 6.03632450e-01 -5.98690569e-01 -9.18837428e-01
7.56231964e-01 6.31546080e-01 -1.15134813e-01 -4.33738768e-01
1.96113773e-02 -1.61745269e-02 2.85812974e-01 -1.00831068e+00
2.96232373e-01 -5.90044439e-01 4.85195452e-03 7.19020665e-01
-3.21175605e-01 -1.58558220e-01 -1.40391245e-01 4.98110086e-01
1.05165660e+00 1.61488637e-01 1.03920162e+00 3.64015512e-02
5.56478977e-01 3.48728180e-01 3.67215633e-01 6.80747867e-01
6.88869804e-02 3.71911168e-01 8.00782859e-01 -7.84777820e-01
-7.15408802e-01 -7.80091822e-01 2.41894156e-01 5.26073277e-01
-1.59453407e-01 -5.23334444e-01 -7.55374134e-01 -7.39555001e-01
-2.15536803e-01 1.78120625e+00 -9.02767241e-01 -2.59696901e-01
-5.54083347e-01 -3.93629640e-01 3.00105751e-01 7.36383796e-01
4.34536278e-01 -1.52686441e+00 -9.68253493e-01 2.14452773e-01
-7.28531659e-01 -9.71807957e-01 -2.21030205e-01 4.62403744e-02
-1.09338248e+00 -1.01799691e+00 -2.71953523e-01 -6.41121149e-01
7.00109601e-01 -3.76554318e-02 1.00936532e+00 4.54729527e-01
9.85493734e-02 9.67831239e-02 -3.04915041e-01 -5.74974895e-01
-7.85484731e-01 -1.88852012e-01 -4.31449771e-01 -4.64039892e-01
1.99045613e-01 -1.73310593e-01 -4.30783391e-01 3.98776606e-02
-7.07484245e-01 8.96649897e-01 5.28401673e-01 1.07250988e+00
3.20109248e-01 -4.36274350e-01 9.03408825e-01 -1.11817825e+00
1.25086975e+00 -7.32858241e-01 1.61528409e-01 4.75900441e-01
-8.25061858e-01 3.75107974e-01 6.38682246e-01 -2.84210801e-01
-1.52046573e+00 -2.61476994e-01 -1.74251616e-01 2.13324472e-01
-3.46785039e-01 8.29516232e-01 2.16309726e-01 7.58817911e-01
7.07375228e-01 -4.69143949e-02 -2.29237378e-01 -6.97644129e-02
5.74577332e-01 3.89728785e-01 6.03716373e-01 -4.71273631e-01
3.84918630e-01 1.82645887e-01 -2.03014106e-01 -2.42970660e-01
-1.23580945e+00 1.09252095e-01 -1.93917841e-01 -3.78081620e-01
9.80665684e-01 -5.53877115e-01 -1.01273775e+00 -5.08779228e-01
-2.03323793e+00 -3.22685003e-01 -1.02965258e-01 6.00839496e-01
-8.13236952e-01 -4.58042361e-02 -3.93196613e-01 -8.86021674e-01
-4.85900581e-01 -1.00293624e+00 7.74618328e-01 1.94296643e-01
-1.32004702e+00 -1.28061533e+00 -8.14088061e-02 6.88834548e-01
2.42638022e-01 2.65326381e-01 1.49243569e+00 -1.15156853e+00
-4.85328943e-01 1.27531171e-01 -2.54318237e-01 -5.40454090e-01
6.79900944e-02 -4.07215834e-01 -8.39589059e-01 7.14637995e-01
2.62473255e-01 -3.23413432e-01 7.52761602e-01 5.82808614e-01
1.60728598e+00 -9.50542808e-01 -5.89814842e-01 1.92382317e-02
7.49944508e-01 3.09985429e-01 5.01703084e-01 2.52107531e-01
3.34840506e-01 1.05712700e+00 6.32137716e-01 4.59020883e-01
7.57728994e-01 5.12529433e-01 4.59085763e-01 -7.55622610e-02
-3.94324139e-02 -4.05993044e-01 1.52826577e-01 3.81505340e-01
-2.21235920e-02 -5.44773936e-01 -9.96131420e-01 4.90695775e-01
-2.32751632e+00 -1.30709124e+00 -3.95891994e-01 1.31866872e+00
8.86752188e-01 7.39750937e-02 -3.96994710e-01 -1.59154683e-01
5.00123262e-01 -1.01048864e-01 -6.60196960e-01 -8.71711075e-01
6.63901865e-02 -2.13659987e-01 -3.96819830e-01 7.83207059e-01
-5.07468581e-01 7.66315699e-01 6.77835035e+00 4.32399243e-01
-5.64694881e-01 8.97152647e-02 1.03181851e+00 -2.19180360e-02
-9.46576893e-01 -1.10953994e-01 -4.19420689e-01 -2.85580810e-02
1.05045986e+00 -5.97914040e-01 1.36513542e-02 7.41660476e-01
7.38996863e-01 -8.03333148e-02 -1.59911633e+00 4.39482838e-01
3.04309130e-01 -1.93965173e+00 5.27759969e-01 -1.33086056e-01
5.66840768e-01 -7.89266407e-01 -1.25483528e-01 1.62633210e-01
2.53813297e-01 -1.27027988e+00 9.16857541e-01 6.15742564e-01
3.34603727e-01 -3.96425754e-01 9.15455282e-01 6.06081188e-01
-6.82717502e-01 -3.12048703e-01 4.43207063e-02 -4.48739141e-01
5.88042915e-01 3.94753337e-01 -1.58635163e+00 3.70783061e-01
1.74304783e-01 7.15883911e-01 -2.42184520e-01 6.03349745e-01
-1.00551891e+00 6.75696373e-01 5.93002796e-01 -3.92867923e-01
1.19103827e-01 2.91111648e-01 4.64265198e-01 1.13950455e+00
3.41345489e-01 7.73718476e-01 -6.96830750e-02 1.34058130e+00
-7.07719103e-02 1.18526489e-01 -6.01886511e-01 -1.28866866e-01
1.88532546e-01 9.97906983e-01 -5.40026665e-01 -6.53981924e-01
-9.90508422e-02 7.83009529e-01 4.03182447e-01 1.47762522e-01
-1.08194232e+00 4.40759778e-01 9.36313868e-02 4.46681641e-02
-2.36325398e-01 4.13295031e-01 -9.38128352e-01 -7.96433330e-01
-1.13978505e-01 -8.25997949e-01 4.81549889e-01 -1.24532533e+00
-1.10769165e+00 9.31839466e-01 2.90378451e-01 -8.18404317e-01
-7.83309519e-01 -2.49743596e-01 -1.01016665e+00 7.95419216e-01
-1.25716150e+00 -1.30858183e+00 -2.90116608e-01 2.30978206e-02
1.13423431e+00 -2.15065897e-01 1.14711165e+00 -3.89994234e-01
-3.56938094e-01 2.10928321e-01 -9.88617003e-01 -3.00936878e-01
5.16720474e-01 -1.28421080e+00 3.91752750e-01 4.35128152e-01
-1.59973964e-01 9.56756771e-01 1.12872779e+00 -1.01419544e+00
-8.29998672e-01 -9.24002886e-01 1.57114303e+00 -3.76726240e-01
4.69335854e-01 3.17246318e-01 -1.08985245e+00 7.09562898e-01
7.10930526e-01 -7.15534627e-01 1.16907430e+00 2.22421691e-01
-1.55690297e-01 4.81847346e-01 -1.12908494e+00 9.42833424e-01
9.23731387e-01 -9.07900184e-02 -1.35555363e+00 5.99587560e-01
1.09951389e+00 -5.02365828e-01 -5.25666475e-01 2.22539142e-01
6.67791188e-01 -7.39085376e-01 8.15598428e-01 -1.10882902e+00
1.47040701e+00 2.37466805e-02 1.51107520e-01 -1.53012526e+00
-3.47224921e-01 -3.75782400e-01 -3.18192512e-01 8.33504856e-01
9.15585458e-01 -2.27965489e-01 5.30534506e-01 9.83164966e-01
-4.24260765e-01 -1.12611103e+00 -6.34409904e-01 3.95249808e-03
-1.97291300e-01 -3.94755960e-01 6.86774611e-01 8.05546463e-01
9.10747826e-01 6.48233771e-01 -3.81524265e-01 5.29396348e-03
2.99531251e-01 3.26540560e-01 4.03939724e-01 -1.13806748e+00
-3.46716851e-01 -4.75569338e-01 3.69803995e-01 -7.91813076e-01
5.87562203e-01 -1.11794055e+00 5.28127551e-02 -2.31147170e+00
6.99732244e-01 3.26912582e-01 2.20320627e-01 7.37506926e-01
-4.57206547e-01 -4.71240312e-01 1.60179660e-01 2.56770104e-02
-5.84815264e-01 5.36352813e-01 1.37846363e+00 -1.79417267e-01
-1.10548638e-01 2.61336332e-03 -1.22948098e+00 9.49969888e-01
9.14857745e-01 -5.49983144e-01 -4.83586371e-01 -5.28180122e-01
3.69312227e-01 5.35626173e-01 6.06093407e-01 -3.13172847e-01
3.67790341e-01 -4.71549451e-01 3.19706619e-01 -5.50184369e-01
1.65269494e-01 -5.06660163e-01 1.41828448e-01 9.38796759e-01
-1.35368276e+00 7.58399442e-02 3.34949851e-01 5.60395300e-01
-2.10535020e-01 -2.23786727e-01 2.79401571e-01 -1.31415799e-01
-3.79292190e-01 -3.01642656e-01 -4.68396842e-01 -2.84768701e-01
9.71469641e-01 6.09710850e-02 -6.90827727e-01 -8.66312802e-01
-8.96266222e-01 5.16024649e-01 -2.17042536e-01 5.74643791e-01
1.40780377e+00 -1.22961152e+00 -9.21492457e-01 -2.75540948e-01
6.73253909e-02 -1.82334363e-01 3.15291584e-01 5.60462356e-01
-1.10535607e-01 8.79144430e-01 -1.31420299e-01 -2.32101962e-01
-1.31601989e+00 4.74122941e-01 2.10229024e-01 -4.68198091e-01
-7.06150949e-01 6.42667949e-01 3.74070913e-01 -2.52726614e-01
2.26187333e-01 -5.58040202e-01 -7.33412027e-01 -1.49889961e-01
7.74931252e-01 1.62353471e-01 -4.57477391e-01 -1.51366785e-01
-2.59610087e-01 -8.65884498e-02 -8.97930413e-02 -2.88367748e-01
1.45163703e+00 6.75350055e-02 -5.74368462e-02 6.39544189e-01
3.60585660e-01 -3.38174433e-01 -7.14333653e-01 4.80870577e-03
2.57920951e-01 -1.08556189e-01 -1.30651623e-01 -1.40032125e+00
-5.71891963e-01 9.13116574e-01 -1.30866438e-01 2.30185196e-01
7.91568458e-01 5.62050939e-01 6.65312052e-01 4.87229437e-01
-5.16568460e-02 -4.88256633e-01 3.89056891e-01 2.40295127e-01
1.62251532e+00 -1.24055994e+00 4.72919531e-02 -6.05703413e-01
-1.10266316e+00 1.33063853e+00 9.24213707e-01 3.60678047e-01
6.77161217e-02 -3.39782462e-02 1.07090399e-02 -6.24883235e-01
-1.47848201e+00 3.89002383e-01 5.76534271e-01 4.05431181e-01
1.02245533e+00 2.25930557e-01 -5.37076175e-01 1.26764429e+00
-4.08121735e-01 8.04834813e-02 6.67544663e-01 2.52399474e-01
-4.06576067e-01 -9.30191755e-01 -3.89101475e-01 4.81590003e-01
-2.74062216e-01 -2.28708506e-01 -8.01239133e-01 4.29197937e-01
-1.25327662e-01 1.30643284e+00 -2.16192156e-01 -7.24501312e-02
1.79541156e-01 2.92451173e-01 2.02136412e-01 -9.21647787e-01
-7.75767982e-01 -5.98046891e-02 5.92103601e-01 -5.89450777e-01
-5.44158399e-01 -5.13276577e-01 -2.00270867e+00 -4.24674675e-02
-4.97459993e-02 3.31377596e-01 4.52278644e-01 1.46806252e+00
1.74611226e-01 1.18890858e+00 5.97516187e-02 -2.50275970e-01
-4.05467004e-01 -1.04004943e+00 2.06526816e-01 2.12282881e-01
2.84183979e-01 -3.25657070e-01 -2.51847208e-02 3.70655954e-01] | [9.55998706817627, 6.944816589355469] |
4216cb30-f523-48de-b158-3d43764eb27d | deep-learning-applications-for-covid-19 | null | null | https://journalofbigdata.springeropen.com/articles/10.1186/s40537-020-00392-9 | https://journalofbigdata.springeropen.com/track/pdf/10.1186/s40537-020-00392-9.pdf | Deep Learning applications for COVID-19 | This survey explores how Deep Learning has battled the COVID-19 pandemic and provides directions for future research on COVID-19. We cover Deep Learning applications in Natural Language Processing, Computer Vision, Life Sciences, and Epidemiology. We describe how each of these applications vary with the availability of big data and how learning tasks are constructed. We begin by evaluating the current state of Deep Learning and conclude with key limitations of Deep Learning for COVID-19 applications. These limitations include Interpretability, Generalization Metrics, Learning from Limited Labeled Data, and Data Privacy. Natural Language Processing applications include mining COVID-19 research for Information Retrieval and Question Answering, as well as Misinformation Detection, and Public Sentiment Analysis. Computer Vision applications cover Medical Image Analysis, Ambient Intelligence, and Vision-based Robotics. Within Life Sciences, our survey looks at how Deep Learning can be applied to Precision Diagnostics, Protein Structure Prediction, and Drug Repurposing. Deep Learning has additionally been utilized in Spread Forecasting for Epidemiology. Our literature review has found many examples of Deep Learning systems to fight COVID-19. We hope that this survey will help accelerate the use of Deep Learning for COVID-19 research. | ['Borko Furht', 'Taghi M. Khoshgoftaar', 'Connor Shorten'] | 2021-01-11 | null | null | null | journal-of-big-data-2021-1 | ['epidemiology'] | ['medical'] | [ 5.82592823e-02 -3.37384790e-01 -3.16774964e-01 -3.07631761e-01
-2.28766978e-01 -4.96531755e-01 3.50476265e-01 5.89932740e-01
-7.04011142e-01 6.11256540e-01 1.95066079e-01 -5.96492887e-01
-4.49082479e-02 -7.13400006e-01 -4.60915864e-01 -6.94938123e-01
-3.40516895e-01 8.11153233e-01 -5.92111051e-01 -3.60809773e-01
-1.40655577e-01 4.92568791e-01 -7.09157288e-01 5.35484493e-01
4.90596116e-01 7.41599560e-01 -7.59152398e-02 8.73701572e-01
-4.63010333e-02 8.60455036e-01 -9.76944625e-01 -1.78066984e-01
-8.44776407e-02 7.07086995e-02 -7.92068183e-01 -5.96397042e-01
-8.80655572e-02 -3.82884294e-01 1.42755270e-01 7.13680804e-01
7.09796190e-01 -2.12196648e-01 9.08911884e-01 -1.16051924e+00
-9.64948535e-01 -6.25793934e-02 -5.22573054e-01 6.27478957e-01
2.30955750e-01 3.51433575e-01 6.83995724e-01 -7.66250134e-01
7.83916414e-01 1.29137444e+00 1.13846135e+00 7.63429940e-01
-7.45045066e-01 -8.45975161e-01 -8.51308033e-02 9.58613381e-02
-9.62960005e-01 1.23330154e-01 7.48743489e-02 -1.13492179e+00
1.39091277e+00 -8.77058879e-02 6.48381591e-01 1.36568797e+00
9.24010456e-01 7.82059371e-01 5.87403417e-01 1.90565661e-01
3.44627559e-01 -1.15794942e-01 5.38479090e-01 6.68202400e-01
4.02976185e-01 3.50613058e-01 -1.19656317e-01 -6.06984913e-01
9.69993025e-02 6.17078662e-01 2.00865582e-01 9.24476981e-02
-1.10660839e+00 1.40400898e+00 4.01720703e-01 1.99458867e-01
-5.33356309e-01 2.52134576e-02 9.40598130e-01 4.68452275e-01
7.22843468e-01 7.96356976e-01 -9.18652833e-01 2.78856486e-01
-5.06760001e-01 4.82101113e-01 6.08657181e-01 4.78471249e-01
5.06355822e-01 2.72773355e-01 -5.72448857e-02 7.77967453e-01
1.81369722e-01 9.57847297e-01 2.98470825e-01 -4.69914287e-01
-1.27819611e-03 2.24852219e-01 -5.33615649e-02 -1.18577254e+00
-1.19706261e+00 -1.67711511e-01 -1.27918458e+00 -1.96201146e-01
-7.58305714e-02 -7.70635486e-01 -1.04200864e+00 1.55329859e+00
1.19075775e-01 4.00689065e-01 2.66125530e-01 6.66369677e-01
1.24568188e+00 7.09602654e-01 2.50846028e-01 -2.04363123e-01
1.83621442e+00 -5.64049065e-01 -7.97542691e-01 -3.18205655e-01
7.48495579e-01 -4.96154904e-01 3.18535537e-01 3.78846943e-01
-4.34964031e-01 -1.78130761e-01 -6.81185961e-01 -1.85779348e-01
-1.15178597e+00 -4.89483804e-01 7.89264202e-01 6.65678382e-01
-9.02354300e-01 -7.59192109e-02 -7.89705932e-01 -6.08883619e-01
6.56436682e-01 5.06241560e-01 -1.47987250e-02 2.15529241e-02
-1.61591291e+00 9.51656461e-01 6.21713437e-02 -9.98142883e-02
-1.25370073e+00 -7.91940868e-01 -8.94477904e-01 -3.13434511e-01
-1.76316932e-01 -1.08566487e+00 9.49195802e-01 -4.35285389e-01
-6.85637116e-01 1.23119402e+00 -6.97703511e-02 -9.35376644e-01
-1.81468323e-01 -1.30572334e-01 -6.32617235e-01 -9.62486938e-02
1.80170581e-01 5.02654433e-01 6.50950313e-01 -5.07362127e-01
-5.12696087e-01 -6.33982658e-01 -1.58417314e-01 -3.39929432e-01
-2.41250843e-01 4.60425287e-01 1.69772625e-01 -6.23021483e-01
-9.47778225e-01 -9.91998255e-01 -4.60546851e-01 -3.12009215e-01
-1.94423705e-01 -4.66001779e-01 9.17202175e-01 -5.65566838e-01
6.91067278e-01 -1.89634156e+00 -1.76037505e-01 3.30152623e-02
7.63655543e-01 7.18847871e-01 -5.09532034e-01 4.51337099e-01
4.26076213e-03 2.33510628e-01 1.46579044e-03 1.97077170e-01
-5.55184543e-01 1.66420072e-01 -3.16310316e-01 6.49801791e-01
4.33820188e-01 1.26924670e+00 -1.11615217e+00 6.49156347e-02
2.07528189e-01 8.05494785e-01 -6.73694372e-01 1.18887797e-01
-6.03817225e-01 4.50471193e-01 -6.61155760e-01 7.17342913e-01
7.41682172e-01 -7.24501431e-01 7.35378265e-02 1.80145711e-01
2.91114092e-01 -1.98495165e-02 -1.08035848e-01 1.26156414e+00
-3.36242378e-01 8.89026701e-01 2.42676929e-01 -1.28991187e+00
6.94565594e-01 3.38737994e-01 1.00289524e+00 -3.35743308e-01
3.71565878e-01 -1.95310190e-01 -6.43757209e-02 -9.13598537e-01
2.77959079e-01 -1.40126243e-01 -6.07068762e-02 6.55024707e-01
-2.76137531e-01 1.27612622e-02 -2.57312685e-01 1.96343046e-02
1.07083678e+00 -6.56054735e-01 3.31032336e-01 -2.21770316e-01
3.19525778e-01 3.00966978e-01 4.99839485e-01 7.12098241e-01
-6.10710979e-01 4.81206272e-03 2.33798012e-01 -1.26180696e+00
-7.33915329e-01 -8.67667258e-01 -2.28015423e-01 1.66895819e+00
-2.87435412e-01 -4.13568735e-01 -7.01341033e-01 -5.44751227e-01
1.33796856e-01 3.82550329e-01 -8.36047113e-01 -1.62227377e-01
-4.25039530e-01 -1.47834325e+00 8.52079451e-01 4.75068688e-01
-1.14456646e-01 -1.20021069e+00 -5.02781689e-01 5.32804392e-02
-1.87087566e-01 -9.81898546e-01 -1.55644521e-01 3.13158840e-01
-3.91572505e-01 -1.25552261e+00 -5.72096229e-01 -1.01430953e+00
2.07167178e-01 2.36163482e-01 1.09513032e+00 1.81480199e-01
-7.17270911e-01 3.97596866e-01 -1.57502480e-02 -1.36720383e+00
-5.72072804e-01 2.11060554e-01 6.68211401e-01 -5.42722583e-01
1.31046116e+00 1.29303202e-01 -6.77974403e-01 -4.54358868e-02
-9.46041882e-01 -5.27506888e-01 1.65048555e-01 1.06983423e+00
3.85849535e-01 -2.75241107e-01 1.01468384e+00 -1.12853324e+00
1.08429956e+00 -1.09889257e+00 -3.44468683e-01 1.18492514e-01
-7.19658732e-01 -3.11239153e-01 4.32416767e-01 -6.06443062e-02
-5.64612865e-01 -3.84750307e-01 -5.52501261e-01 -2.33381525e-01
-1.85414806e-01 4.94640708e-01 7.13874996e-01 2.36962169e-01
1.01078856e+00 -1.14718508e-02 2.84476280e-01 -2.40117952e-01
2.64673531e-01 1.10654712e+00 9.69644934e-02 -6.86764419e-02
2.03137636e-01 6.15575910e-01 -1.19790524e-01 -1.32828116e+00
-9.11124885e-01 -6.09278202e-01 -1.33098751e-01 2.63794392e-01
1.61610031e+00 -1.04050112e+00 -1.10831916e+00 4.91650224e-01
-1.39061964e+00 -3.60928057e-03 1.30759358e-01 2.83653021e-01
-2.89740771e-01 9.39066410e-02 -1.08624423e+00 -4.87551391e-01
-1.09727931e+00 -1.34469223e+00 1.29611790e+00 -3.38079743e-02
-3.37001771e-01 -1.36514235e+00 6.43405437e-01 6.00721478e-01
6.49064243e-01 4.00008231e-01 1.13303685e+00 -1.20486450e+00
6.89435378e-02 -2.43837178e-01 -2.15824291e-01 2.08645850e-01
4.13049847e-01 -3.05692106e-01 -1.25449634e+00 -5.19042850e-01
-1.38777405e-01 -7.00663209e-01 1.12545359e+00 7.99310148e-01
1.23769784e+00 -3.93423796e-01 -6.37873828e-01 7.83247828e-01
8.78349781e-01 4.31974202e-01 2.59428229e-02 1.50195524e-01
7.58885622e-01 7.61385918e-01 4.41426128e-01 4.42981660e-01
4.35318470e-01 1.20223276e-01 4.43522453e-01 -4.43882346e-01
5.14989436e-01 3.10403198e-01 2.22331509e-01 5.67277491e-01
2.93819252e-02 -4.94930595e-01 -1.33175898e+00 4.47896361e-01
-1.56090772e+00 -1.06578112e+00 -4.24320884e-02 1.49529409e+00
5.62764287e-01 -4.34542924e-01 1.48444459e-01 -6.07761145e-01
4.74727541e-01 1.29859582e-01 -9.86862242e-01 -9.60752368e-01
-1.11196838e-01 8.24077874e-02 2.98883706e-01 4.44714814e-01
-1.50979829e+00 9.24287438e-01 7.57212591e+00 4.99849796e-01
-1.32276857e+00 1.50989592e-01 9.19567049e-01 2.18270603e-03
-2.31769487e-01 -8.59091341e-01 -6.61055088e-01 3.31514657e-01
1.06334245e+00 6.62776977e-02 3.60795021e-01 8.35152864e-01
3.41295272e-01 6.30857885e-01 -1.00532579e+00 1.00445652e+00
3.83155979e-03 -1.85347545e+00 -1.04255691e-01 3.01071614e-01
7.78650045e-01 9.86403883e-01 3.46032083e-01 2.78137147e-01
6.01829469e-01 -1.36459315e+00 -4.42502439e-01 2.45481163e-01
8.95391941e-01 -7.98625708e-01 1.13293695e+00 1.32514626e-01
-7.21586704e-01 -1.22874252e-01 -5.59318244e-01 -3.11200116e-02
-3.35118771e-02 5.93486726e-01 -1.54056942e+00 7.84517005e-02
7.80645013e-01 9.88662124e-01 -5.41854789e-03 4.42257881e-01
2.68325776e-01 5.19647896e-01 -8.47977474e-02 -3.52769524e-01
3.34323466e-01 3.95025611e-02 5.48273504e-01 1.80492032e+00
-3.25718224e-01 -4.67675105e-02 4.85136688e-01 5.42787254e-01
-5.87887093e-02 1.88633073e-02 -9.95962858e-01 -4.78474081e-01
6.76902905e-02 7.38846898e-01 -4.16114241e-01 -3.27822983e-01
-4.30292845e-01 5.93231857e-01 -6.06503710e-02 4.87603009e-01
-6.26172900e-01 -2.19831556e-01 1.45711207e+00 -6.61723688e-02
8.42680484e-02 5.73512772e-03 -3.93442631e-01 -9.98199165e-01
-8.98893178e-01 -1.14376366e+00 9.80601847e-01 -4.48731750e-01
-1.68468285e+00 6.46248162e-01 -1.41759306e-01 -6.84643984e-01
-5.63081205e-01 -1.05434191e+00 -3.06407899e-01 4.65548128e-01
-1.44697988e+00 -1.01401126e+00 1.18072078e-01 4.80578512e-01
5.87440789e-01 -7.84673035e-01 1.25530684e+00 6.55995384e-02
-5.99132776e-01 4.21228737e-01 5.60552776e-01 3.62484783e-01
7.23604977e-01 -8.63012135e-01 7.38888860e-01 3.16529542e-01
-3.23895931e-01 8.37610960e-01 5.63486814e-01 -8.29276800e-01
-1.42962384e+00 -1.50442433e+00 8.52362037e-01 -7.10304439e-01
6.41826332e-01 -4.92643803e-01 -6.37812495e-01 7.97949195e-01
4.29718405e-01 -2.48532414e-01 1.35463214e+00 1.83837876e-01
-5.49876928e-01 -6.32542670e-02 -1.57306170e+00 4.24081564e-01
5.07209897e-01 -5.05348921e-01 -3.44542593e-01 9.59524989e-01
1.09292257e+00 -9.73860770e-02 -8.00283253e-01 3.10660809e-01
5.77054918e-01 -3.54264021e-01 1.16482234e+00 -1.60174906e+00
3.43161583e-01 1.56769410e-01 4.44773808e-02 -1.25275445e+00
-5.09417593e-01 -5.37854552e-01 -8.43393207e-02 2.12259755e-01
2.65065521e-01 -7.26014972e-01 6.81982458e-01 2.41108865e-01
1.61195993e-01 -1.00155520e+00 -4.87719953e-01 -4.85343426e-01
5.09571433e-01 -5.30793607e-01 7.40491688e-01 1.24061108e+00
-7.03890994e-02 6.73439026e-01 -6.21972322e-01 5.35667352e-02
4.93193686e-01 2.34406330e-02 6.80210769e-01 -1.17482746e+00
-1.23298608e-01 -2.72842228e-01 -4.48384196e-01 -7.47857690e-01
1.97519153e-01 -8.54159296e-01 -2.44222194e-01 -1.54731488e+00
1.43893570e-01 -2.75729716e-01 -4.82461482e-01 6.27515614e-01
2.88946964e-02 2.77520269e-01 -3.40396278e-02 -7.07679465e-02
-3.09496909e-01 -3.53138819e-02 9.18545842e-01 -7.77815700e-01
-1.01355910e-01 1.01216950e-01 -8.70199919e-01 8.50619793e-01
1.14473414e+00 -3.89312685e-01 -3.71015340e-01 -7.22783267e-01
5.45233488e-01 -2.38047168e-01 -3.67497690e-02 -2.47521862e-01
-1.40708730e-01 -5.60439169e-01 5.29058814e-01 -6.30505741e-01
2.76172280e-01 -6.76997364e-01 -3.61175150e-01 1.07448661e+00
-3.12581420e-01 3.09631020e-01 3.72176319e-01 7.93655336e-01
2.74812788e-01 2.54328877e-01 8.84459078e-01 -2.12739855e-01
-6.32968009e-01 5.79194725e-01 -1.21944988e+00 4.35055971e-01
1.10579336e+00 2.66652524e-01 -5.44908166e-01 -4.19347674e-01
-5.08866072e-01 6.51310742e-01 2.43626889e-02 7.64061332e-01
9.05654669e-01 -8.16041350e-01 -1.01649582e+00 4.84820902e-01
3.68188262e-01 -8.10723007e-02 1.96423307e-01 6.63658857e-01
-7.26444185e-01 8.19165289e-01 -6.33054525e-02 -8.59338522e-01
-1.31317270e+00 1.16043353e+00 4.30831879e-01 -3.24382335e-01
-3.22500616e-01 9.33954239e-01 7.38092721e-01 -5.83553851e-01
4.11278754e-01 -3.92333150e-01 -3.89917940e-01 -3.62416133e-02
9.66360152e-01 3.12416106e-01 5.76224662e-02 -3.76609921e-01
-1.00400591e+00 4.42777693e-01 -4.71268505e-01 6.04810715e-01
1.58908212e+00 1.34153828e-01 -3.17372084e-01 1.78574160e-01
1.33665705e+00 -3.77220660e-01 -2.14099467e-01 -1.17137186e-01
1.09829276e-03 2.33628765e-01 -2.36642659e-01 -9.83647168e-01
-9.01431322e-01 1.15332055e+00 7.56657064e-01 9.34485942e-02
1.06855381e+00 2.33367886e-02 1.21331918e+00 8.20660353e-01
1.24188483e-01 -7.71273434e-01 8.20395947e-02 9.28942025e-01
6.28821373e-01 -1.46687615e+00 7.83469826e-02 3.45569074e-01
-6.41835451e-01 8.19323599e-01 3.69337440e-01 9.43106487e-02
1.29442894e+00 5.51995695e-01 3.10830891e-01 -7.35581219e-01
-8.60247791e-01 1.59766242e-01 -2.94820704e-02 1.17484748e+00
5.43240607e-01 4.06610966e-01 -6.17908351e-02 4.81365710e-01
-1.31679907e-01 5.72659485e-02 3.37758124e-01 6.49933398e-01
-4.67444360e-01 -8.37679625e-01 -3.57132614e-01 7.61891961e-01
-8.92284989e-01 -2.99830198e-01 -7.13030636e-01 6.51782155e-02
4.06870544e-01 1.15597057e+00 1.96220174e-01 -4.90729392e-01
-2.35095650e-01 -1.66881472e-01 -1.12952001e-01 -7.41439819e-01
-7.39393950e-01 -1.93736196e-01 1.14215619e-03 -3.85754108e-01
-3.61597776e-01 -2.18865246e-01 -1.30065930e+00 -5.46738863e-01
2.74358373e-02 1.83476042e-02 6.61864221e-01 9.35925603e-01
7.47864544e-01 2.85340458e-01 3.86184603e-01 -8.28187987e-02
-2.83554047e-01 -6.72312498e-01 -2.31960744e-01 5.72138950e-02
1.21753991e+00 -2.97909945e-01 1.27625972e-01 -1.99551992e-02] | [15.56959056854248, -1.6666680574417114] |
ad54e3d8-745a-4fcd-a723-b2cafbd9f092 | benign-malignant-lung-nodule-classification | 1605.08350 | null | http://arxiv.org/abs/1605.08350v1 | http://arxiv.org/pdf/1605.08350v1.pdf | Benign-Malignant Lung Nodule Classification with Geometric and Appearance Histogram Features | Lung cancer accounts for the highest number of cancer deaths globally. Early
diagnosis of lung nodules is very important to reduce the mortality rate of
patients by improving the diagnosis and treatment of lung cancer. This work
proposes an automated system to classify lung nodules as malignant and benign
in CT images. It presents extensive experimental results using a combination of
geometric and histogram lung nodule image features and different linear and
non-linear discriminant classifiers. The proposed approach is experimentally
validated on the LIDC-IDRI public lung cancer screening thoracic computed
tomography (CT) dataset containing nodule level diagnostic data. The obtained
results are very encouraging correctly classifying 82% of malignant and 93% of
benign nodules on unseen test data at best. | ['Tizita Nesibu Shewaye', 'Alhayat Ali Mekonnen'] | 2016-05-26 | null | null | null | null | ['lung-nodule-classification'] | ['medical'] | [-7.38980398e-02 2.51651227e-01 -4.86201227e-01 -9.12391394e-02
-9.27985787e-01 -5.05847275e-01 5.24975479e-01 -2.30830889e-02
-2.99834102e-01 5.42627871e-01 1.57040164e-01 -5.81710696e-01
-1.97970852e-01 -6.60715699e-01 6.76641017e-02 -7.50565469e-01
-7.18994439e-02 1.30862188e+00 7.97511041e-01 4.99467045e-01
-1.91149488e-01 8.29481125e-01 -1.01367152e+00 5.85134625e-01
3.33455980e-01 8.79258633e-01 3.48236918e-01 1.05992615e+00
2.26256624e-01 1.10958445e+00 8.47515538e-02 2.33967695e-02
4.44632262e-01 -8.32430720e-01 -1.09004796e+00 2.15361938e-01
4.18361962e-01 -3.76036465e-01 -2.14479148e-01 6.48196220e-01
3.67108256e-01 -4.43188339e-01 1.29984856e+00 -7.18420446e-01
-1.43216047e-02 2.62930959e-01 -2.34336913e-01 6.03632629e-01
-3.37253749e-01 5.79111511e-03 8.62564087e-01 -1.00242543e+00
3.09064627e-01 6.98810041e-01 7.07798719e-01 6.57140255e-01
-5.81021726e-01 -3.33728939e-01 -9.67526019e-01 2.34237179e-01
-1.32498372e+00 2.84841269e-01 -1.70417175e-01 -6.27938032e-01
4.31841731e-01 7.70237386e-01 8.29190671e-01 4.15750504e-01
4.49578524e-01 3.52277070e-01 1.02039742e+00 -2.11327881e-01
-1.34866312e-01 3.21702152e-01 -7.91981369e-02 1.24370265e+00
7.74645686e-01 2.37083867e-01 5.96552312e-01 -5.63723207e-01
9.19328392e-01 3.31052393e-01 -1.19603641e-01 -4.43416506e-01
-1.33941948e+00 8.36157382e-01 8.46535563e-01 9.56341863e-01
-3.66160333e-01 9.70927849e-02 2.48974532e-01 -1.65507421e-01
8.39467943e-02 1.83651552e-01 -1.46073818e-01 4.18654770e-01
-8.21760178e-01 -2.79314816e-01 8.42807114e-01 4.18953896e-01
1.80402428e-01 -3.04362178e-01 -6.09502554e-01 6.56234503e-01
2.60694176e-01 7.67557323e-01 9.99498785e-01 -4.89043862e-01
-5.01835831e-02 5.27674079e-01 -4.87719662e-02 -3.93264413e-01
-5.41211367e-01 -5.05378902e-01 -1.07591438e+00 -2.30974123e-01
2.36543342e-01 2.21974075e-01 -1.11295986e+00 6.44920468e-01
1.06079970e-04 5.30057997e-02 -2.31340498e-01 6.56607568e-01
9.64529991e-01 1.95125014e-01 3.01837415e-01 -9.57151353e-02
1.52958584e+00 -9.23598051e-01 -2.53415942e-01 1.18057625e-02
7.90450037e-01 -5.90323031e-01 6.76879346e-01 -2.14341685e-01
-8.03355157e-01 -4.16893095e-01 -3.11011940e-01 2.57490247e-01
1.50465101e-01 1.01372981e+00 2.15639696e-01 7.30930924e-01
-9.38536286e-01 2.08355218e-01 -1.08219838e+00 -9.02329803e-01
4.15576190e-01 6.40504241e-01 -2.81663358e-01 -1.73673287e-01
-6.14054680e-01 8.76289129e-01 6.59866214e-01 -4.28411514e-01
-9.53276575e-01 -5.79801202e-01 -1.46155179e-01 1.60299867e-01
4.39555347e-01 -8.66431832e-01 1.62371147e+00 -4.82288837e-01
-7.41471767e-01 9.57361758e-01 4.78392914e-02 -5.29999793e-01
6.95570350e-01 3.43659282e-01 -1.44872069e-01 3.64036083e-01
1.25647917e-01 6.48720205e-01 6.09591305e-01 -8.23838472e-01
-1.00631869e+00 -1.17181212e-01 -8.96272182e-01 2.24138349e-01
-2.72115797e-01 -3.17794591e-01 -3.54400754e-01 -2.19882876e-01
2.68064141e-01 -1.39123201e+00 -4.49652672e-01 6.93486854e-02
-3.94342184e-01 -4.15112883e-01 1.30053985e+00 -5.43597460e-01
9.48557913e-01 -1.67180228e+00 -3.52065593e-01 3.97473007e-01
4.33774859e-01 1.82768464e-01 3.86529505e-01 -1.34760931e-01
-1.48550421e-01 5.09434342e-01 -1.44409500e-02 6.31683469e-01
-5.09664953e-01 2.20480040e-01 3.97755891e-01 3.43531817e-01
5.78433415e-03 1.20921731e+00 -5.90054274e-01 -1.01350081e+00
4.74068284e-01 2.01471537e-01 -1.47689328e-01 3.36290002e-01
8.19680318e-02 5.90582609e-01 -8.04501891e-01 6.45062923e-01
7.63863251e-02 -7.17186928e-01 1.35174349e-01 -2.36789510e-01
1.27277404e-01 -9.19042751e-02 -4.26339746e-01 4.19339597e-01
-2.01429933e-01 3.08807731e-01 -3.70303303e-01 -4.31028426e-01
3.80295128e-01 9.31447864e-01 9.48845506e-01 -7.40598887e-02
1.38743475e-01 4.90742862e-01 5.33139169e-01 -7.07906544e-01
-3.57083768e-01 -4.00005758e-01 1.29169554e-01 3.16475242e-01
-2.92591393e-01 -4.67453390e-01 1.70927614e-01 -3.02566253e-02
1.61916852e+00 -8.66532803e-01 1.04017377e+00 -4.96951878e-01
9.84876454e-01 4.48883235e-01 -3.17860544e-02 6.67539418e-01
-3.43112200e-01 7.47856855e-01 2.77898073e-01 -2.25373670e-01
-1.14470673e+00 -1.17713714e+00 -3.86750519e-01 6.33001447e-01
-4.81629640e-01 3.21607709e-01 -5.34281313e-01 -1.10433555e+00
-1.47811428e-01 3.80737692e-01 -7.35588431e-01 1.91726521e-01
-6.63733006e-01 -8.39806139e-01 3.83311987e-01 5.11149645e-01
7.99747229e-01 -9.49879587e-01 -4.14055139e-01 -1.05339192e-01
-2.76103467e-01 -8.46379936e-01 -2.76747465e-01 2.59993911e-01
-1.24871230e+00 -1.60865307e+00 -1.07812130e+00 -1.10830879e+00
8.90755713e-01 5.98540604e-01 1.10707343e+00 4.62969780e-01
-1.02976429e+00 4.47701365e-01 9.20587685e-03 -2.43005708e-01
-9.36692953e-01 3.57298672e-01 -4.20270830e-01 -5.64908862e-01
9.27803963e-02 3.13370258e-01 -6.00323439e-01 6.70176983e-01
-8.03718209e-01 -2.56983399e-01 1.42911518e+00 7.56308854e-01
8.54330957e-01 5.08080125e-01 7.15038553e-02 -1.11856997e+00
1.76185966e-01 -8.55345190e-01 -2.97076464e-01 2.38848791e-01
-4.43639249e-01 -1.35539457e-01 5.66468120e-01 -1.48733184e-01
-7.98928022e-01 2.80981064e-01 6.49399087e-02 -2.44431093e-01
-2.92503238e-01 3.37281115e-02 4.18661624e-01 -2.10283980e-01
9.00733948e-01 6.73461035e-02 -1.76549837e-01 -1.59322187e-01
-4.18348074e-01 5.66006660e-01 4.53311890e-01 2.80871868e-01
1.04930520e+00 5.69819212e-01 6.03887200e-01 -1.01293075e+00
-8.64647627e-01 -1.28822422e+00 -9.12695825e-01 -2.46999457e-01
1.25169170e+00 -7.24844158e-01 -6.38193786e-02 -4.94296178e-02
-3.12820107e-01 -6.41586259e-02 -3.40608776e-01 8.69484544e-01
-2.99127787e-01 2.61572093e-01 -5.74322283e-01 -2.57768035e-01
-3.52499932e-01 -7.32900560e-01 9.45992529e-01 -9.94389504e-03
-1.64679348e-01 -1.18630958e+00 2.48802081e-01 2.62715787e-01
6.50369525e-01 -4.98903468e-02 1.24213231e+00 -1.04849195e+00
-6.48536742e-01 -8.20112586e-01 -4.82955605e-01 2.21647054e-01
5.22396863e-01 -5.64663019e-03 -5.72921455e-01 -3.73672038e-01
1.67528987e-01 -2.75559574e-01 1.05374146e+00 5.17687917e-01
1.15545654e+00 -1.80932462e-01 -9.20468867e-01 9.87414867e-02
1.45451534e+00 5.59544750e-02 3.08966488e-01 -9.77647901e-02
7.10835874e-01 1.11743815e-01 5.08973598e-01 1.10346869e-01
-3.23112756e-01 6.97845072e-02 6.03030205e-01 -1.49712265e-01
-5.22761941e-01 5.31407073e-02 -1.66953057e-01 4.50404495e-01
-3.37774783e-01 -5.19726515e-01 -1.39453650e+00 4.11018163e-01
-1.18439865e+00 -1.21072006e+00 -7.19343364e-01 2.05123472e+00
1.81063950e-01 -1.53489590e-01 9.22729746e-02 2.04720482e-01
7.92242348e-01 -3.28455329e-01 -1.87612116e-01 2.74074852e-01
4.34282601e-01 2.12686241e-01 8.28647375e-01 5.10468939e-03
-1.44102979e+00 2.12391257e-01 6.73670149e+00 7.80318618e-01
-1.28537667e+00 2.92303264e-01 9.26776886e-01 3.32569212e-01
2.23890811e-01 -4.45076197e-01 -5.43015540e-01 -1.21516071e-01
8.42465937e-01 -2.52718449e-01 -3.46967131e-01 1.09874308e+00
1.35549247e-01 -4.27414328e-01 -1.08052266e+00 5.96362889e-01
-9.79341194e-02 -1.14864349e+00 7.36845285e-02 2.75115639e-01
8.20893705e-01 4.55529481e-01 2.17231482e-01 2.03671068e-01
3.16996798e-02 -1.22826636e+00 -4.32097502e-02 3.79397750e-01
8.59973550e-01 -3.11063766e-01 1.29501629e+00 6.68406963e-01
-1.30986035e+00 -1.02879137e-01 -2.95357853e-01 6.45112753e-01
-5.34185946e-01 3.96429390e-01 -2.06955719e+00 3.19366127e-01
8.05848122e-01 5.43988109e-01 -1.22618949e+00 1.46519399e+00
1.84538424e-01 1.14070451e+00 -3.94714296e-01 -2.27544039e-01
2.56444216e-01 2.86536247e-01 2.20048815e-01 1.06859052e+00
7.70173550e-01 3.30958664e-01 2.76656002e-01 4.25348699e-01
2.48284126e-03 4.27291006e-01 -7.06362963e-01 5.20591140e-02
8.27104747e-02 1.59294951e+00 -1.29795170e+00 -4.74725038e-01
-3.67064506e-01 5.88783443e-01 -2.30901212e-01 -3.27807695e-01
-8.74593914e-01 6.08484328e-01 -6.01173043e-01 6.81208968e-01
2.24655196e-01 4.25120682e-01 1.84024964e-02 -7.13690102e-01
-6.41849279e-01 -3.16841722e-01 6.21197462e-01 -5.33258498e-01
-1.15337658e+00 7.69963264e-01 1.49820343e-01 -1.50387776e+00
-4.31190938e-01 -8.29234004e-01 -9.30492938e-01 3.42140794e-01
-8.89562905e-01 -1.01829183e+00 -8.12477827e-01 2.95937866e-01
3.47680748e-01 -2.95878172e-01 8.06800365e-01 -3.93954702e-02
-1.43085405e-01 -1.23398706e-01 2.14911461e-01 2.42217064e-01
6.36589050e-01 -1.50413024e+00 -3.39884847e-01 1.50227398e-01
-1.84908122e-01 -4.13720876e-01 1.50519028e-01 -7.41539717e-01
-6.60869598e-01 -1.78236222e+00 7.13746965e-01 -4.96118903e-01
4.18189466e-01 5.59286535e-01 -7.48799205e-01 5.27730465e-01
-1.39966118e-03 5.19960344e-01 6.72162533e-01 -1.03228998e+00
2.66027153e-01 2.93624759e-01 -1.36775970e+00 2.44275376e-01
3.19179714e-01 -8.16834271e-02 -2.60913342e-01 9.26322997e-01
-2.52986610e-01 -5.58864065e-02 -9.42426741e-01 9.91832018e-01
2.50298589e-01 -1.19239318e+00 1.10667670e+00 -1.26709476e-01
6.41373396e-02 -8.14621300e-02 1.76397488e-01 -6.48950815e-01
-7.67851710e-01 4.61569637e-01 4.72962528e-01 3.90371054e-01
5.67101777e-01 -1.77676111e-01 1.38999903e+00 2.34059677e-01
2.30986923e-01 -7.54382372e-01 -8.28267753e-01 -7.52452075e-01
2.00542152e-01 2.07082987e-01 -2.41381288e-01 5.08806765e-01
-7.83674777e-01 1.08583093e-01 3.12279075e-01 -1.40747413e-01
5.83089054e-01 -2.00637490e-01 2.79564232e-01 -1.43753910e+00
-2.11345270e-01 -5.24755657e-01 -4.79372412e-01 -1.55215561e-01
-3.61817390e-01 -1.18194687e+00 6.84389770e-02 -1.76674175e+00
1.01879632e+00 -3.14567238e-01 -9.39696729e-02 4.61127400e-01
-1.79068387e-01 6.42321527e-01 -1.13431312e-01 7.75612772e-01
-3.82240891e-01 -1.71944618e-01 1.43127799e+00 -6.32874444e-02
3.24599713e-01 9.14536893e-01 3.87242883e-02 9.71881151e-01
1.02621114e+00 -8.04495513e-01 -3.99869859e-01 2.18926281e-01
-5.99611700e-01 3.37330937e-01 6.63081944e-01 -1.52683663e+00
-9.37127769e-02 -3.00386816e-01 6.83021724e-01 -9.75210488e-01
7.90380761e-02 -1.23807156e+00 4.21516478e-01 1.73130679e+00
-3.01453799e-01 -2.77679473e-01 -1.07051216e-01 6.91195071e-01
-1.62648588e-01 -6.04112387e-01 1.27546668e+00 -6.45896733e-01
-4.31650996e-01 5.58419406e-01 -7.86380887e-01 -1.74111381e-01
1.59071279e+00 5.73649816e-02 -2.42927894e-02 -3.92419338e-01
-6.80713594e-01 -1.53240681e-01 2.15616107e-01 -1.05016693e-01
3.18011492e-01 -1.35612917e+00 -9.53502417e-01 -4.22894470e-02
1.50071651e-01 -1.35894284e-01 -3.08736295e-01 1.30272651e+00
-1.04838634e+00 1.14733493e+00 2.20330637e-02 -9.19404328e-01
-1.81607270e+00 2.55760700e-01 8.45231414e-01 -8.97178531e-01
-2.35140085e-01 8.07459056e-01 2.85588175e-01 -1.73732042e-01
-5.72953485e-02 -6.97698236e-01 -2.32780457e-01 -3.93517464e-01
-1.23904333e-01 6.09757900e-01 1.86989293e-01 -5.58030963e-01
-2.15315461e-01 4.90692526e-01 -1.68624118e-01 8.45527127e-02
6.51931226e-01 2.83690870e-01 -3.25784236e-01 3.05518001e-01
1.20023394e+00 -2.23164316e-02 -2.11163431e-01 1.53210433e-02
3.82738829e-01 -5.17045744e-02 8.02609250e-02 -8.89202774e-01
-9.56536949e-01 7.96358645e-01 1.08010709e+00 4.24994677e-01
8.88484240e-01 6.09212577e-01 5.07278621e-01 9.29217458e-01
-1.73484851e-02 -2.64254630e-01 4.54183280e-01 2.49518856e-01
6.42023504e-01 -1.62242424e+00 2.85661936e-01 -8.62999976e-01
-7.11484373e-01 1.43158174e+00 6.61868513e-01 -3.42310667e-01
9.16185081e-01 4.58141230e-02 1.29403725e-01 -1.86270669e-01
-9.34911191e-01 -4.11300123e-01 6.88441992e-01 2.10015610e-01
8.69788826e-01 2.52605945e-01 -3.80096316e-01 1.51987061e-01
2.29636118e-01 4.71460782e-02 2.18849108e-01 9.46631432e-01
-1.29693413e+00 -9.30983722e-01 -6.06541932e-01 1.23296726e+00
-8.05018485e-01 1.51652366e-01 -8.64468336e-01 1.35560966e+00
1.14071548e-01 4.04231459e-01 3.34526189e-02 7.40720630e-02
-1.57560095e-01 1.37419418e-01 3.67524117e-01 -1.03391623e+00
-6.87819481e-01 4.13138896e-01 -9.31435898e-02 1.09723032e-01
-3.09862316e-01 -5.51718593e-01 -1.53413594e+00 1.64832324e-01
-4.24361348e-01 4.00151700e-01 3.47779691e-01 6.13249481e-01
-3.08020890e-01 6.41816616e-01 7.44460881e-01 -4.56877291e-01
-7.97268212e-01 -1.03018951e+00 -4.99375433e-01 2.29244724e-01
-5.78790717e-02 -2.15190336e-01 -5.59207141e-01 1.65035337e-01] | [15.3897123336792, -2.1471614837646484] |
01b54a5c-b9c2-4a20-bad4-07869f251890 | spa-web-based-platform-for-easy-access-to | null | null | https://aclanthology.org/L16-1615 | https://aclanthology.org/L16-1615.pdf | SPA: Web-based Platform for easy Access to Speech Processing Modules | This paper presents SPA, a web-based Speech Analytics platform that integrates several speech processing modules and that makes it possible to use them through the web. It was developed with the aim of facilitating the usage of the modules, without the need to know about software dependencies and specific configurations. Apart from being accessed by a web-browser, the platform also provides a REST API for easy integration with other applications. The platform is flexible, scalable, provides authentication for access restrictions, and was developed taking into consideration the time and effort of providing new services. The platform is still being improved, but it already integrates a considerable number of audio and text processing modules, including: Automatic transcription, speech disfluency classification, emotion detection, dialog act recognition, age and gender classification, non-nativeness detection, hyper-articulation detection, dialog act recognition, and two external modules for feature extraction and DTMF detection. This paper describes the SPA architecture, presents the already integrated modules, and provides a detailed description for the ones most recently integrated. | ['Ricardo Ribeiro', "Eug{\\'e}nio Ribeiro", 'o', 'David Martins de Matos', 'Pedro Curto', 'Helena Moniz', 'Alberto Abad', 'Jaime Ferreira', 'Isabel Trancoso', 'Fern Batista'] | 2016-05-01 | spa-web-based-platform-for-easy-access-to-1 | https://aclanthology.org/L16-1615 | https://aclanthology.org/L16-1615.pdf | lrec-2016-5 | ['age-and-gender-classification'] | ['computer-vision'] | [-2.42707193e-01 3.05572927e-01 2.05937728e-01 -4.56424087e-01
-4.99098539e-01 -5.68732858e-01 6.05989456e-01 2.72794336e-01
-5.98939002e-01 3.46613735e-01 3.53108019e-01 -4.42219228e-01
-8.84213969e-02 -4.33323711e-01 4.88878995e-01 -5.58009744e-01
1.70599461e-01 7.93900132e-01 5.57513058e-01 -4.82537955e-01
-4.38287333e-02 5.45196414e-01 -1.95792735e+00 3.06450427e-01
7.37614512e-01 9.17790949e-01 3.56369972e-01 7.81260908e-01
-3.63215357e-01 5.59018016e-01 -9.09741580e-01 -3.72331440e-01
-2.63559133e-01 -1.86909333e-01 -9.92599368e-01 5.73739670e-02
-1.70688093e-01 -4.02758688e-01 2.02824473e-02 6.36212409e-01
7.27338195e-01 1.57023326e-01 2.43936718e-01 -1.05849838e+00
1.03400886e-01 5.68876982e-01 4.64025259e-01 2.22053826e-01
1.07253742e+00 1.70073491e-02 5.82699776e-01 -6.16931915e-01
4.66277570e-01 1.27364051e+00 4.23300534e-01 7.17099667e-01
-9.08277631e-01 -4.47094083e-01 -4.63015363e-02 2.26702750e-01
-1.18575585e+00 -7.64157057e-01 3.96071047e-01 -4.51208025e-01
1.47195256e+00 6.63639605e-01 6.10054910e-01 1.36921811e+00
-4.42777693e-01 4.13631916e-01 1.08178270e+00 -9.02770579e-01
4.70420986e-01 7.04013288e-01 6.26229525e-01 4.20432240e-01
-3.12794477e-01 -3.15465838e-01 -4.45607752e-01 -4.98286039e-01
1.83859855e-01 -4.59655911e-01 -1.62654549e-01 -9.12320986e-02
-7.25284398e-01 5.05285144e-01 -6.04034662e-01 1.07520306e+00
-1.79260612e-01 -8.59548151e-01 8.04321706e-01 3.84182245e-01
4.29314435e-01 -1.93231523e-01 -8.20476890e-01 -9.06111479e-01
-7.47946262e-01 6.65213466e-02 1.40187562e+00 8.64226878e-01
3.34293544e-01 7.36084953e-02 1.96891919e-01 1.16001761e+00
5.23080766e-01 3.59151453e-01 9.31097448e-01 -6.99028552e-01
2.74820060e-01 7.05295324e-01 -1.20747522e-01 -3.83771032e-01
-6.50817573e-01 2.77389795e-01 -1.93928316e-01 3.91472757e-01
5.38339257e-01 -2.26571426e-01 -3.93504202e-01 1.46204531e+00
3.41520667e-01 -5.57167590e-01 -4.41460758e-02 4.70442235e-01
1.01374829e+00 3.10028702e-01 1.56537935e-01 -6.96195364e-01
1.68099737e+00 -3.75953466e-01 -1.19677138e+00 5.44829927e-02
3.78164828e-01 -1.02585006e+00 1.23844624e+00 5.99646747e-01
-1.19751382e+00 -3.64001960e-01 -8.76911223e-01 5.95521517e-02
-9.09779608e-01 1.14810817e-01 4.77829695e-01 1.52195764e+00
-1.30058050e+00 3.28622401e-01 -8.38602066e-01 -9.71588373e-01
-5.56862831e-01 7.44334698e-01 -4.15786028e-01 7.47132659e-01
-1.19696164e+00 9.18260753e-01 3.61394554e-01 -4.68240172e-01
-1.97229758e-02 -6.72933310e-02 -8.52513313e-01 2.22754240e-01
8.70101899e-02 -3.47975016e-01 1.33226418e+00 -9.55033243e-01
-2.27031183e+00 8.99001479e-01 -1.12356074e-01 -1.38047531e-01
4.15168971e-01 7.35486746e-02 -9.38336730e-01 1.70934305e-01
-2.07795471e-01 6.10405542e-02 8.50410402e-01 -4.09473717e-01
-5.27732551e-01 -6.70899332e-01 -2.86172330e-01 2.20673829e-01
-5.99246264e-01 8.05195510e-01 -5.17222822e-01 -3.99325043e-01
-2.71742880e-01 -6.26153588e-01 3.58886540e-01 -6.81690395e-01
6.97155371e-02 -3.88104707e-01 9.28883672e-01 -1.12316871e+00
1.53607297e+00 -2.44211817e+00 -1.50384083e-01 3.11289519e-01
-1.08385645e-01 7.01073229e-01 4.52344030e-01 8.40747893e-01
-7.34694973e-02 -6.42877072e-02 2.02209294e-01 -4.71634179e-01
2.80216664e-01 2.60832518e-01 1.75766885e-01 1.78438976e-01
-2.82301337e-01 4.25985344e-02 -4.74016607e-01 -4.93165165e-01
8.64728451e-01 6.43732250e-01 -1.27382264e-01 1.30139038e-01
4.77309749e-02 3.37720543e-01 -1.11615486e-01 4.78204459e-01
6.85847878e-01 4.99085337e-01 5.66821992e-01 4.63508099e-01
-7.13191807e-01 7.35414088e-01 -1.47286427e+00 1.43951643e+00
-7.73751140e-01 4.56230938e-01 7.70516455e-01 -5.63510358e-01
1.01846921e+00 8.92114222e-01 3.08201402e-01 -2.74070263e-01
3.82491946e-01 4.87306565e-01 -2.07053959e-01 -1.02876103e+00
3.10505599e-01 3.16617578e-01 1.82406485e-01 2.46507183e-01
4.53899503e-01 1.96546242e-01 3.68100911e-01 -1.52242839e-01
9.54669535e-01 -1.17693953e-01 6.12961471e-01 -3.66041809e-02
1.23855042e+00 -4.10946906e-01 3.23280483e-01 2.85327202e-03
-3.31563979e-01 8.31295103e-02 4.66166407e-01 -9.21778828e-02
-6.51767790e-01 -8.90570402e-01 -2.76734442e-01 1.37243271e+00
-6.44393802e-01 -8.31880808e-01 -1.17810071e+00 -5.97033262e-01
-4.60912645e-01 6.72825873e-01 -4.81089987e-02 3.80618036e-01
-3.56775373e-01 -2.57978410e-01 4.99314249e-01 9.98295993e-02
5.78471899e-01 -1.31298447e+00 -3.71494770e-01 1.81105182e-01
-2.33897492e-01 -1.03555036e+00 -8.89699236e-02 9.32373852e-02
-7.38399327e-01 -7.77734995e-01 -5.21421552e-01 -9.83694494e-01
1.08534247e-01 -2.81590372e-01 8.77230644e-01 -1.11970282e-03
-2.15481177e-01 8.00918162e-01 -5.45462728e-01 -3.44452679e-01
-9.76552308e-01 4.42590743e-01 1.37301609e-01 -9.17291567e-02
6.51210129e-01 -7.32862175e-01 -2.53185760e-02 2.38579810e-01
-5.89157522e-01 -5.81947863e-01 1.22718573e-01 4.24867362e-01
-2.12571561e-01 -2.06318423e-01 7.12117612e-01 -7.02644706e-01
8.64141285e-01 -3.49451825e-02 -4.70633864e-01 6.04459234e-02
-6.53636813e-01 -3.42810094e-01 5.19551337e-01 -1.32070491e-02
-1.32843983e+00 3.21203947e-01 -1.14723337e+00 3.08538765e-01
-1.12675011e+00 4.47474904e-02 -8.67350042e-01 6.43798187e-02
3.79093140e-01 1.40170872e-01 3.58705312e-01 -1.04032707e+00
1.35807514e-01 1.58207536e+00 1.83792964e-01 -1.05853681e-03
1.75501078e-01 -2.55469959e-02 -8.44349682e-01 -1.46636105e+00
1.53694719e-01 -9.04754341e-01 -8.62046182e-01 -3.45273942e-01
9.42944825e-01 -6.32045925e-01 -1.09056282e+00 7.29414940e-01
-1.08669007e+00 -5.88960908e-02 -1.15587667e-01 5.24177611e-01
-4.39168453e-01 6.00093722e-01 -6.34813488e-01 -1.27366376e+00
-5.13319612e-01 -8.87508810e-01 5.78806877e-01 2.08130404e-01
-6.76850498e-01 -9.38832760e-01 1.28733352e-01 7.11535394e-01
3.67859483e-01 -2.31916830e-01 8.39049757e-01 -1.20717692e+00
2.18131542e-01 -3.44260067e-01 3.85509372e-01 6.54461861e-01
1.53536811e-01 2.69808561e-01 -1.30622780e+00 -8.97203460e-02
3.16892236e-01 4.50167246e-02 1.01297200e-01 6.47887811e-02
5.52507699e-01 -3.70060682e-01 -1.07131109e-01 1.66091323e-02
7.40027726e-01 6.61651194e-01 6.08052731e-01 2.83091277e-01
-5.59170768e-02 9.23547029e-01 4.11846429e-01 5.80884278e-01
2.26372838e-01 1.06265390e+00 2.71581680e-01 2.41503954e-01
6.74677938e-02 3.64389062e-01 7.12740064e-01 9.69141781e-01
-2.11250782e-01 1.10905528e-01 -8.86126161e-01 2.04281673e-01
-1.70191407e+00 -9.90560114e-01 -4.58342105e-01 2.47435999e+00
5.22504628e-01 6.34306073e-02 9.70903635e-01 7.57575929e-01
7.58329093e-01 -4.48808745e-02 1.76667631e-01 -9.84525681e-01
2.21723586e-01 5.65350056e-01 -8.61898437e-02 8.18368018e-01
-1.04691064e+00 7.64185607e-01 6.42530251e+00 6.94747090e-01
-1.05708611e+00 5.72591364e-01 1.47145048e-01 1.74405682e-03
3.32381725e-01 -4.29967165e-01 -7.09782660e-01 5.07869184e-01
1.25344193e+00 4.16204520e-02 5.64690948e-01 9.90084171e-01
3.90992910e-01 -2.42220312e-01 -5.69405138e-01 7.34093070e-01
9.76260379e-02 -8.96095216e-01 -4.16219980e-01 4.05818410e-02
-3.88632923e-01 7.20134601e-02 -2.32758716e-01 3.86553198e-01
-3.73798370e-01 -3.27327073e-01 5.68510115e-01 4.76471595e-02
5.90621352e-01 -8.87656331e-01 7.54306257e-01 3.34790170e-01
-1.01626098e+00 -2.30257675e-01 1.20613046e-01 -2.75507987e-01
7.78603405e-02 3.71149749e-01 -8.38339627e-01 5.38768291e-01
7.95226216e-01 -9.90021825e-02 -5.08339643e-01 7.72240639e-01
-1.37083298e-02 4.17211652e-01 -2.66634405e-01 -2.85362124e-01
-3.01700860e-01 -2.18270779e-01 7.37503350e-01 1.43750799e+00
3.01554888e-01 -2.69650340e-01 -5.73343225e-02 1.79343417e-01
7.43324339e-01 8.58456314e-01 -3.45537275e-01 2.23505929e-01
3.74700814e-01 1.54969716e+00 -7.85524189e-01 -1.32187203e-01
-6.62214875e-01 9.26017404e-01 -1.88377082e-01 -3.00299283e-02
-4.52739090e-01 -7.78322518e-01 8.02175820e-01 3.01848501e-01
1.44884363e-01 -2.66014993e-01 7.72535354e-02 -8.28811347e-01
6.35664240e-02 -1.12109482e+00 5.49306095e-01 -4.32088017e-01
-7.77632654e-01 8.72899532e-01 1.75603386e-03 -7.82750607e-01
-9.02708352e-01 -9.17407751e-01 -7.10947454e-01 9.49162602e-01
-5.28921664e-01 -9.07220960e-01 -5.66353798e-02 9.28518772e-01
6.69034064e-01 -4.21649933e-01 1.47766256e+00 7.23040819e-01
-7.23944724e-01 3.96299124e-01 -9.09339041e-02 -1.21721420e-02
7.26127088e-01 -1.41026497e+00 1.06524557e-01 5.32501578e-01
-1.10088497e-01 5.79374135e-01 6.76613808e-01 -3.36471498e-01
-7.97740281e-01 -2.76566416e-01 1.65374780e+00 -1.32594451e-01
6.52181208e-01 -8.27690780e-01 -7.33448386e-01 2.86132306e-01
6.02552593e-01 -8.24629545e-01 9.77985024e-01 4.82993931e-01
-1.19035907e-01 -2.92070091e-01 -1.30366230e+00 3.51968795e-01
5.76927125e-01 -5.57228625e-01 -6.41757786e-01 3.80520731e-01
3.30335885e-01 -3.47804725e-01 -8.97058129e-01 -1.96142703e-01
4.09536183e-01 -1.45345235e+00 7.32916832e-01 -3.12436931e-02
-6.13539636e-01 5.48563376e-02 3.03041279e-01 -9.00795817e-01
-4.76060696e-02 -1.12925923e+00 4.95807044e-02 1.90419257e+00
4.91172075e-01 -9.81437087e-01 4.76327419e-01 8.70577335e-01
-2.41894886e-01 1.43832341e-01 -1.40356457e+00 -7.01594472e-01
-4.77433473e-01 -5.82478046e-01 6.25965357e-01 5.79534531e-01
9.78556514e-01 5.52167356e-01 -5.42777441e-02 1.13840528e-01
-2.65387058e-01 -4.13090706e-01 4.49342608e-01 -1.38178539e+00
-5.40462792e-01 -6.24658287e-01 -5.75978279e-01 -2.93556243e-01
8.65072161e-02 -6.94717109e-01 -4.85684395e-01 -1.24555624e+00
-6.29454672e-01 1.63981497e-01 1.38489813e-01 4.28819418e-01
3.06706250e-01 -1.45373553e-01 1.05564497e-01 -9.74762663e-02
-1.42119199e-01 4.49039787e-02 2.70309806e-01 2.65283018e-01
-5.32401919e-01 6.79224491e-01 -1.89075872e-01 7.86562741e-01
1.04266822e+00 -1.93671271e-01 -3.58287692e-01 3.30971062e-01
-3.37171108e-02 -1.58763438e-01 8.73178467e-02 -1.28643620e+00
1.05063520e-01 4.68974292e-01 1.71328902e-01 -4.48232919e-01
5.21499515e-01 -1.00531971e+00 5.03610298e-02 5.26105702e-01
-8.47007707e-02 1.34894967e-01 1.58335313e-01 -3.00436243e-02
-2.31338173e-01 -7.60869563e-01 6.56152844e-01 -7.62957335e-02
-4.60820615e-01 -3.17505658e-01 -1.41106486e+00 -3.26230437e-01
1.14602554e+00 -1.85397401e-01 -1.63787529e-01 -4.82573956e-01
-1.37139940e+00 -1.87466711e-01 2.94609725e-01 4.62845504e-01
2.82127112e-02 -7.02752292e-01 -6.04095533e-02 6.06036305e-01
1.24267228e-01 -8.39489937e-01 3.26264471e-01 8.78980637e-01
-5.49073517e-01 4.69343215e-01 -3.49794894e-01 -2.51989067e-01
-2.21113896e+00 6.38128579e-01 2.16083422e-01 -1.42684683e-01
-4.93805587e-01 3.34478885e-01 -6.34750009e-01 -4.62432355e-01
7.29099035e-01 -2.93229729e-01 -7.71891415e-01 4.21828657e-01
8.19244981e-01 6.67111456e-01 6.63478076e-01 -6.17288888e-01
-6.28350914e-01 8.25484544e-02 1.81021497e-01 -6.37482703e-01
1.05415487e+00 -5.04794717e-01 -3.05595905e-01 7.55207419e-01
6.60373271e-01 3.77373189e-01 -4.43650246e-01 2.90304601e-01
2.89660305e-01 -8.07146430e-02 -1.35416508e-01 -9.64328110e-01
-5.87918162e-01 8.66271973e-01 8.19845855e-01 1.04707170e+00
1.16519964e+00 -2.90112607e-02 3.81733656e-01 1.98252544e-01
1.73226625e-01 -1.59198081e+00 -4.73631769e-01 8.29951465e-01
7.75479794e-01 -7.05891609e-01 -4.79675353e-01 -8.11817288e-01
-5.11401296e-01 1.47124088e+00 3.43628436e-01 6.22303486e-01
8.74080479e-01 5.46677470e-01 4.15242106e-01 4.28537913e-02
-6.15305543e-01 -6.59689724e-01 -5.21251634e-02 1.18959391e+00
7.91948318e-01 5.79520948e-02 -8.77168953e-01 7.54216075e-01
-3.68993640e-01 -1.74317196e-01 2.35579073e-01 7.18936980e-01
-4.55336481e-01 -1.87899733e+00 -6.40852392e-01 1.80284962e-01
-8.19293618e-01 -1.10362042e-02 -8.73237908e-01 9.52984214e-01
3.07574868e-01 1.22034025e+00 -1.17616795e-01 -4.22972918e-01
4.82895672e-01 9.27115977e-01 1.23504795e-01 -6.90935254e-01
-1.45844042e+00 3.23416710e-01 6.53504968e-01 -3.48692417e-01
-1.50296003e-01 -1.00002277e+00 -1.04348040e+00 -2.25258991e-01
-1.93299115e-01 3.73440683e-01 1.05654764e+00 9.06542659e-01
2.71463424e-01 2.63411224e-01 4.95387167e-01 -6.64102852e-01
-2.24808365e-01 -1.27294803e+00 -8.12864482e-01 1.20449131e-02
-5.08015007e-02 -3.03375661e-01 -3.82586956e-01 -8.86092111e-02] | [13.968788146972656, 6.271763324737549] |
0c6dbbf6-942e-4527-ae65-8b38fa245d7f | cross-lingual-information-retrieval-with-bert | 2004.13005 | null | https://arxiv.org/abs/2004.13005v1 | https://arxiv.org/pdf/2004.13005v1.pdf | Cross-lingual Information Retrieval with BERT | Multiple neural language models have been developed recently, e.g., BERT and XLNet, and achieved impressive results in various NLP tasks including sentence classification, question answering and document ranking. In this paper, we explore the use of the popular bidirectional language model, BERT, to model and learn the relevance between English queries and foreign-language documents in the task of cross-lingual information retrieval. A deep relevance matching model based on BERT is introduced and trained by finetuning a pretrained multilingual BERT model with weak supervision, using home-made CLIR training data derived from parallel corpora. Experimental results of the retrieval of Lithuanian documents against short English queries show that our model is effective and outperforms the competitive baseline approaches. | ['Amro El-Jaroudi', 'Damianos Karakos', 'Lingjun Zhao', 'Zhuolin Jiang', 'William Hartmann'] | 2020-04-24 | cross-lingual-information-retrieval-with-bert-1 | https://aclanthology.org/2020.clssts-1.5 | https://aclanthology.org/2020.clssts-1.5.pdf | lrec-2020-5 | ['cross-lingual-information-retrieval'] | ['natural-language-processing'] | [-3.11536014e-01 -3.63875836e-01 -4.52934921e-01 -6.85662389e-01
-1.63652515e+00 -3.68967146e-01 1.01346648e+00 3.87973189e-01
-1.12267804e+00 8.89929235e-01 5.57418108e-01 -4.81399238e-01
-3.89319181e-01 -5.61433911e-01 -7.06912100e-01 -6.49582669e-02
1.11425810e-01 1.02005041e+00 2.43112653e-01 -6.76888466e-01
2.54758596e-01 1.91707298e-01 -9.79660749e-01 7.72686005e-01
6.59058034e-01 9.62550044e-01 3.81187588e-01 6.03958547e-01
-3.53244752e-01 1.10003805e+00 -5.82068563e-01 -7.98624218e-01
-2.91100666e-02 1.60303101e-01 -1.34525979e+00 -9.03741181e-01
4.73823905e-01 -3.48789662e-01 -3.14718872e-01 7.00566411e-01
8.74243796e-01 2.96812981e-01 4.36533362e-01 -5.79326749e-01
-1.18059051e+00 8.88221502e-01 -2.04131126e-01 4.68377143e-01
4.36204135e-01 -5.79385757e-01 1.62579930e+00 -1.18411374e+00
7.55330384e-01 1.49072802e+00 4.59909230e-01 4.66737777e-01
-8.61948490e-01 -3.96794975e-01 9.56372023e-02 5.22980809e-01
-1.53319466e+00 -2.59139866e-01 4.90990669e-01 -1.51906814e-02
1.58959842e+00 3.94276470e-01 2.10665837e-02 1.07264888e+00
3.77438039e-01 1.06908166e+00 7.97157109e-01 -8.90371859e-01
-3.40906888e-01 3.96599889e-01 4.22707051e-01 3.74455601e-01
-2.42670506e-01 -1.06151953e-01 -5.56650043e-01 -4.25192982e-01
1.67894676e-01 -2.34764665e-01 -1.30011350e-01 1.81892902e-01
-1.00972211e+00 9.85362232e-01 6.12369478e-01 7.54875362e-01
-2.18795165e-01 8.90446380e-02 6.66415334e-01 8.30059767e-01
8.91934574e-01 6.14003897e-01 -8.13903511e-01 5.47223926e-01
-6.58802330e-01 2.24353477e-01 7.69786060e-01 9.55187321e-01
5.64542413e-01 -5.55539072e-01 -5.79319358e-01 1.31847918e+00
6.80371284e-01 6.26111209e-01 9.11342621e-01 -3.05018961e-01
7.56304264e-01 4.53618109e-01 -1.50635019e-01 -7.61637866e-01
-2.56604433e-01 -3.39576632e-01 -6.49132967e-01 -7.22793281e-01
-1.22338317e-01 2.51181006e-01 -5.10575593e-01 1.43958771e+00
7.63860065e-03 -5.27848542e-01 5.78990817e-01 8.11689138e-01
1.39339519e+00 7.97431648e-01 2.16928422e-01 8.79706666e-02
1.18747997e+00 -1.39857948e+00 -5.62638938e-01 -2.41750419e-01
7.95455992e-01 -1.16588306e+00 1.29651678e+00 -4.31172103e-02
-1.19686913e+00 -6.05472863e-01 -5.79844356e-01 -7.27389634e-01
-7.95487285e-01 2.17509270e-01 5.41400790e-01 5.54845929e-02
-1.52189791e+00 1.75523996e-01 -3.52129310e-01 -3.74738157e-01
-3.48392546e-01 5.48496544e-01 -3.18475991e-01 -4.59143013e-01
-2.04402351e+00 1.00416040e+00 4.98054564e-01 2.59952962e-01
-7.97040820e-01 -2.45897591e-01 -7.64071822e-01 1.49874464e-01
7.29734486e-04 -6.98208988e-01 1.41629648e+00 -7.93827772e-01
-1.21023583e+00 1.26385248e+00 -1.39882639e-01 -5.92965961e-01
2.34038591e-01 -4.64019775e-01 -6.45052493e-01 8.78127664e-02
3.53401095e-01 7.25094914e-01 1.99430674e-01 -6.83364511e-01
-4.97800469e-01 -2.83011049e-01 3.32245797e-01 5.12412846e-01
-5.41578591e-01 7.88397968e-01 -1.04465663e+00 -4.37321097e-01
-4.43415910e-01 -5.91096520e-01 -8.02190676e-02 -5.93241513e-01
-1.87782779e-01 -9.53633547e-01 3.87538105e-01 -9.68045652e-01
1.23836744e+00 -1.59664083e+00 -5.97288534e-02 1.10464618e-01
-3.53883028e-01 2.86083966e-01 -5.76661587e-01 8.14644754e-01
2.44402379e-01 1.48793645e-02 2.06929073e-01 -3.94492894e-01
7.65717924e-02 1.17622606e-01 -6.15566492e-01 -8.45850855e-02
-4.10895944e-02 1.43933213e+00 -9.69561398e-01 -6.34938478e-01
-2.97958791e-01 6.04834378e-01 -1.77153528e-01 4.52998847e-01
-1.31205305e-01 6.60905167e-02 -5.34105718e-01 5.45594871e-01
1.46368131e-01 -3.28003585e-01 1.68467909e-01 7.24105686e-02
1.22921146e-01 1.07706535e+00 -3.14645588e-01 1.86244667e+00
-8.01003695e-01 6.68456733e-01 -2.32919422e-03 -8.67603898e-01
1.01882350e+00 6.65636897e-01 8.57111253e-03 -1.43304038e+00
-4.40913923e-02 4.56995040e-01 -4.49507743e-01 -5.80679059e-01
9.43414032e-01 2.44429037e-01 -2.41898045e-01 5.08894324e-01
1.32294372e-01 -2.00958535e-01 2.48130947e-01 4.45460260e-01
8.20301771e-01 -3.23376879e-02 1.43073201e-01 -4.82630074e-01
1.11971903e+00 -7.85250962e-02 -9.22117531e-02 9.92868602e-01
3.77918221e-02 2.92245924e-01 3.03412438e-03 -3.88985276e-01
-7.04960227e-01 -8.06702733e-01 -1.50328234e-01 1.81549323e+00
-1.59526654e-02 -3.77481103e-01 -3.61957282e-01 -8.74706924e-01
-1.29255559e-02 5.59747577e-01 -3.44370961e-01 -1.95099324e-01
-8.65038395e-01 -6.95302188e-01 4.93542254e-01 3.94209325e-01
3.62538159e-01 -1.53872025e+00 2.68057525e-01 2.67264426e-01
-4.29424018e-01 -1.15571845e+00 -5.50203383e-01 1.46022201e-01
-6.43797517e-01 -6.42284095e-01 -6.94442928e-01 -1.36964178e+00
2.50605494e-01 2.28428274e-01 1.81750572e+00 3.36169273e-01
-6.70354441e-02 5.19231498e-01 -3.60549301e-01 -2.29280308e-01
-2.69301087e-01 6.02291524e-01 -3.24639939e-02 -3.06599259e-01
7.58654296e-01 1.39892116e-01 -4.27765697e-01 1.90095603e-01
-1.02437270e+00 -3.27912688e-01 6.01527631e-01 1.21187592e+00
6.50283456e-01 -5.09068966e-01 1.09510326e+00 -8.38368297e-01
1.29969537e+00 -4.02458578e-01 -6.91204548e-01 1.00707316e+00
-8.26045573e-01 1.80930108e-01 5.45914948e-01 1.30989440e-02
-1.19283974e+00 -7.32742250e-01 -5.96908033e-01 2.76172683e-02
2.31842190e-01 1.06474602e+00 1.45919085e-01 -7.84340277e-02
5.33053637e-01 5.67131042e-02 -7.69399464e-01 -7.32601166e-01
4.01976973e-01 1.14342213e+00 4.84249443e-01 -6.74571991e-01
2.23973602e-01 5.76098002e-02 -6.57911777e-01 -4.03237730e-01
-1.08136165e+00 -9.64390874e-01 -5.97192466e-01 1.08527832e-01
6.98908329e-01 -1.17589259e+00 -2.77760088e-01 1.65371791e-01
-1.47430241e+00 -1.89470381e-01 -1.14448227e-01 5.86384535e-01
-6.72970340e-02 1.85526654e-01 -1.24236560e+00 -4.84554529e-01
-1.12795186e+00 -1.07329631e+00 1.43122888e+00 -9.73990280e-03
2.17261836e-01 -1.53214324e+00 4.32694554e-01 5.93781114e-01
7.54718184e-01 -8.13783169e-01 1.01907885e+00 -1.02486050e+00
-4.35896099e-01 -5.65340996e-01 -3.73941541e-01 5.28327048e-01
-1.23598419e-01 -3.85537207e-01 -1.01273048e+00 -5.59136808e-01
-1.92932531e-01 -9.97170866e-01 1.11294985e+00 1.53686717e-01
9.64408517e-01 -2.44333714e-01 -2.31817201e-01 1.32059053e-01
1.33167434e+00 -6.01679161e-02 3.72759998e-01 7.06351995e-01
4.34339285e-01 8.61155570e-01 6.11722887e-01 -3.36430699e-01
7.70042539e-01 7.50745475e-01 -1.85263474e-02 -1.90077618e-01
-2.23352332e-02 -2.02548802e-01 2.45450854e-01 1.42686474e+00
2.41000637e-01 -3.60541403e-01 -9.53523040e-01 5.60564578e-01
-1.88192403e+00 -7.56134868e-01 2.74163157e-01 2.02424932e+00
1.18499196e+00 -4.62986939e-02 -5.00336826e-01 -7.20824897e-01
4.87443388e-01 6.51508495e-02 -2.04710588e-01 -5.11248887e-01
-4.82472926e-01 2.06983596e-01 2.47657314e-01 9.62616563e-01
-1.20250559e+00 1.34578538e+00 6.82437468e+00 1.02672923e+00
-1.10432291e+00 3.67597342e-01 7.24193931e-01 2.64957454e-02
-5.17734170e-01 -2.97373712e-01 -1.14924836e+00 -2.58703716e-02
9.73627090e-01 -3.76812190e-01 1.72313079e-01 6.52322888e-01
-2.07924217e-01 1.48274839e-01 -9.57530499e-01 6.54014170e-01
5.15541971e-01 -1.10811853e+00 3.56483310e-01 -4.20914352e-01
7.65104175e-01 8.14975500e-01 2.92962380e-02 9.26578581e-01
6.26772165e-01 -8.16094697e-01 2.15943247e-01 5.42249441e-01
5.66246331e-01 -5.84591508e-01 1.12172282e+00 4.32175606e-01
-8.90145540e-01 1.20513201e-01 -7.81310141e-01 2.77013808e-01
6.44246042e-02 2.22645849e-01 -9.36660469e-01 7.61309028e-01
9.26298738e-01 7.03899860e-01 -8.75866055e-01 7.59882987e-01
-4.24094737e-01 4.11246598e-01 -7.30284452e-02 -2.34240979e-01
5.31026304e-01 5.19207157e-02 2.60755181e-01 1.51762736e+00
-4.33872417e-02 -5.25068223e-01 1.60319328e-01 4.18162644e-01
-5.07260919e-01 7.21282959e-01 -6.71904087e-01 9.42401141e-02
9.19179246e-02 1.34187996e+00 6.46097064e-02 -6.07814193e-01
-5.64580739e-01 8.57074797e-01 8.92011404e-01 5.50779998e-01
-2.41345659e-01 -4.37081575e-01 -7.05782995e-02 -4.69591171e-01
-1.50441319e-01 2.21507903e-03 3.36354524e-01 -1.33663046e+00
1.94686010e-01 -8.59376431e-01 8.49496484e-01 -8.25668931e-01
-1.77663064e+00 1.22838247e+00 1.49208922e-02 -7.31937408e-01
-7.64433563e-01 -6.09396636e-01 -3.20290834e-01 1.33280849e+00
-2.27715111e+00 -1.30848134e+00 3.14943969e-01 8.23336780e-01
6.12580359e-01 -5.66823900e-01 1.00217843e+00 6.54724598e-01
-3.76229197e-01 7.66977727e-01 6.18733943e-01 1.78392038e-01
1.22042716e+00 -1.06448495e+00 5.38691223e-01 4.03011084e-01
4.05779839e-01 9.46389019e-01 -6.25055190e-03 -4.76504683e-01
-1.05952859e+00 -1.12450552e+00 2.03004074e+00 -4.43225473e-01
9.00981009e-01 -3.98772478e-01 -9.98749077e-01 7.46053159e-01
8.58134508e-01 -1.91956073e-01 7.71804810e-01 3.49822432e-01
-4.59849566e-01 -2.44461209e-01 -7.49296069e-01 4.17566478e-01
5.99973738e-01 -1.15265799e+00 -8.30417097e-01 9.29789066e-01
1.16792071e+00 -3.24975140e-02 -7.56046236e-01 6.29434466e-01
4.92242247e-01 -5.01591921e-01 1.48531699e+00 -1.05493724e+00
1.98267609e-01 2.55496085e-01 -3.48112971e-01 -1.17388415e+00
-2.68006384e-01 -9.70351845e-02 1.64612249e-01 1.14562917e+00
7.27605522e-01 -7.16490924e-01 1.93126634e-01 5.44405341e-01
-7.03779012e-02 -7.89766252e-01 -1.04698408e+00 -6.56225502e-01
3.84878159e-01 -3.39710116e-01 5.43339610e-01 8.66649270e-01
3.97881195e-02 1.14833975e+00 -2.41708562e-01 -3.08714539e-01
6.75887614e-02 1.92613631e-01 3.38899970e-01 -1.07136858e+00
-5.08120775e-01 -5.39967656e-01 1.19878344e-01 -1.22823787e+00
8.67881954e-01 -1.40522015e+00 -1.54048456e-02 -1.70966661e+00
3.53085458e-01 -4.76474881e-01 -6.83591843e-01 3.54117692e-01
-5.08336782e-01 3.46182644e-01 -2.01425508e-01 5.14817178e-01
-1.17646039e+00 6.94246829e-01 9.84348595e-01 -6.17384911e-01
5.34058958e-02 2.03761663e-02 -4.49484795e-01 2.84740061e-01
6.03404164e-01 -3.67676914e-01 -1.98743999e-01 -1.25671351e+00
5.95550299e-01 3.10749598e-02 1.15099326e-02 -3.09906632e-01
4.28592801e-01 3.94544423e-01 1.49052396e-01 -7.38951087e-01
2.61745632e-01 -6.91347480e-01 -6.41147912e-01 2.25050345e-01
-1.15746891e+00 6.08441889e-01 1.68819934e-01 3.78380120e-01
-7.51995921e-01 -5.53432584e-01 2.88455635e-02 -2.34090403e-01
-7.87498295e-01 2.08350182e-01 -1.45080239e-01 1.52624786e-01
1.41204238e-01 8.19732428e-01 -5.67813516e-01 -3.87832224e-01
-4.07439440e-01 6.74346030e-01 -2.83545464e-01 9.40343976e-01
5.34304857e-01 -1.40107858e+00 -1.06125760e+00 1.16970830e-01
4.82327968e-01 -3.38358164e-01 -8.38453248e-02 6.49665654e-01
-3.34383368e-01 1.51511669e+00 3.92409980e-01 -4.42579478e-01
-1.24022532e+00 5.22456408e-01 2.69057661e-01 -1.17358267e+00
-1.46016702e-01 1.10249019e+00 3.05327564e-01 -1.19943440e+00
4.21915770e-01 1.24322593e-01 -8.19508672e-01 -1.14915647e-01
6.89823329e-01 -8.85694772e-02 6.41449869e-01 -6.33024096e-01
-3.56677741e-01 4.37151104e-01 -5.34388483e-01 -3.75982642e-01
1.01077235e+00 -3.30678403e-01 -6.83370709e-01 3.76259357e-01
1.70573282e+00 -2.05076337e-01 -5.80834821e-02 -9.45219696e-01
6.68184221e-01 -1.25506118e-01 2.01457113e-01 -1.08566785e+00
-7.36585736e-01 1.02725410e+00 4.81428921e-01 -1.35260187e-02
1.03246927e+00 1.82620004e-01 7.86876559e-01 1.25998282e+00
4.77803588e-01 -1.16444647e+00 -1.07293256e-01 1.05824184e+00
1.19696081e+00 -1.58525479e+00 -4.06957954e-01 2.76265860e-01
-4.37516689e-01 8.96220148e-01 5.87865710e-01 1.18715711e-01
4.88035202e-01 -1.64539367e-01 4.83857840e-01 -2.82683969e-01
-1.20477772e+00 -2.39111111e-01 9.57202911e-01 6.95981160e-02
1.02309704e+00 -5.95187806e-02 -5.81131876e-01 2.74877012e-01
-6.15866035e-02 -4.92853373e-02 -2.36912489e-01 7.62950063e-01
-1.79542631e-01 -1.51160657e+00 -1.44114658e-01 2.30112389e-01
-9.41807687e-01 -9.07735348e-01 -6.29148424e-01 5.56848288e-01
-6.10999107e-01 9.41352606e-01 -5.39916120e-02 -8.88070390e-02
3.96359116e-01 1.73510015e-01 1.31402731e-01 -6.70646250e-01
-1.19193745e+00 6.96210414e-02 2.56819874e-01 -2.84729570e-01
-3.97034883e-01 -2.39979535e-01 -6.42584205e-01 2.90335387e-01
-5.70458889e-01 7.72919416e-01 7.16520965e-01 9.13255095e-01
3.08376074e-01 2.42926732e-01 5.07779241e-01 -3.40691626e-01
-8.46740901e-01 -1.23044419e+00 -2.16890931e-01 1.31064773e-01
3.26800466e-01 -1.27405142e-02 -1.82685465e-01 -4.21059906e-01] | [11.379855155944824, 9.77081298828125] |
dc6763a4-ef35-4a7e-9106-f0590a79aa62 | procedure-aware-pretraining-for-instructional | 2303.18230 | null | https://arxiv.org/abs/2303.18230v1 | https://arxiv.org/pdf/2303.18230v1.pdf | Procedure-Aware Pretraining for Instructional Video Understanding | Our goal is to learn a video representation that is useful for downstream procedure understanding tasks in instructional videos. Due to the small amount of available annotations, a key challenge in procedure understanding is to be able to extract from unlabeled videos the procedural knowledge such as the identity of the task (e.g., 'make latte'), its steps (e.g., 'pour milk'), or the potential next steps given partial progress in its execution. Our main insight is that instructional videos depict sequences of steps that repeat between instances of the same or different tasks, and that this structure can be well represented by a Procedural Knowledge Graph (PKG), where nodes are discrete steps and edges connect steps that occur sequentially in the instructional activities. This graph can then be used to generate pseudo labels to train a video representation that encodes the procedural knowledge in a more accessible form to generalize to multiple procedure understanding tasks. We build a PKG by combining information from a text-based procedural knowledge database and an unlabeled instructional video corpus and then use it to generate training pseudo labels with four novel pre-training objectives. We call this PKG-based pre-training procedure and the resulting model Paprika, Procedure-Aware PRe-training for Instructional Knowledge Acquisition. We evaluate Paprika on COIN and CrossTask for procedure understanding tasks such as task recognition, step recognition, and step forecasting. Paprika yields a video representation that improves over the state of the art: up to 11.23% gains in accuracy in 12 evaluation settings. Implementation is available at https://github.com/salesforce/paprika. | ['Juan Carlos Niebles', 'Silvio Savarese', 'Mubbasir Kapadia', 'Roberto Martín-Martín', 'Honglu Zhou'] | 2023-03-31 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Zhou_Procedure-Aware_Pretraining_for_Instructional_Video_Understanding_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Zhou_Procedure-Aware_Pretraining_for_Instructional_Video_Understanding_CVPR_2023_paper.pdf | cvpr-2023-1 | ['video-understanding'] | ['computer-vision'] | [ 6.81562424e-01 1.01078458e-01 -5.72479248e-01 -4.43644226e-01
-7.40024328e-01 -1.06784320e+00 4.66741890e-01 2.96679527e-01
-1.56577557e-01 6.11789048e-01 5.20152450e-01 -4.54766959e-01
-1.73138484e-01 -5.44244468e-01 -1.33751678e+00 -2.96548218e-01
-5.87828495e-02 1.90953597e-01 1.45349026e-01 2.71511376e-01
3.18095773e-01 2.30477303e-01 -1.77718925e+00 9.06821191e-01
8.37361157e-01 7.55174816e-01 4.19832438e-01 9.35253143e-01
-1.40195370e-01 1.40387237e+00 -3.92680138e-01 -1.01320975e-01
-8.04701913e-03 -5.62746704e-01 -1.35122454e+00 4.06557560e-01
5.69356680e-01 -4.55171138e-01 -3.84088397e-01 7.07915485e-01
-3.13749939e-01 6.35661423e-01 7.63259113e-01 -1.16949725e+00
-4.05689329e-01 9.19817805e-01 -1.76421151e-01 2.72471815e-01
8.87449145e-01 1.09962925e-01 6.74826682e-01 -6.04251385e-01
1.02681947e+00 8.58132005e-01 4.03157204e-01 8.32245708e-01
-1.05691183e+00 -3.64356130e-01 4.12688136e-01 4.44824070e-01
-1.07943881e+00 -4.21654135e-01 4.26932156e-01 -7.45308518e-01
9.45174098e-01 2.03646928e-01 7.90158391e-01 1.19307876e+00
-1.60526559e-01 1.30042708e+00 8.49060535e-01 -4.49910343e-01
1.68117091e-01 -4.78926823e-02 4.34717268e-01 1.14408004e+00
6.43522888e-02 -1.80513591e-01 -8.65197122e-01 4.01932538e-01
7.97487378e-01 2.70443074e-02 -5.34777164e-01 -4.67387497e-01
-1.25737512e+00 4.84442502e-01 1.20907582e-01 5.23510538e-02
-2.99751639e-01 2.03197241e-01 4.07533854e-01 4.21913445e-01
2.16025058e-02 5.17679214e-01 -4.05582756e-01 -6.98452711e-01
-8.12383652e-01 1.28489867e-01 1.13695168e+00 1.11320877e+00
9.53446209e-01 -1.74655952e-02 -3.55052143e-01 4.26031560e-01
-1.38128817e-01 -4.16859128e-02 5.37358165e-01 -1.27744603e+00
7.06998050e-01 5.94470918e-01 6.01545721e-02 -4.18884099e-01
-1.88784987e-01 1.17256097e-01 3.10632065e-02 -6.91400394e-02
6.63620889e-01 -1.91990927e-01 -1.26068676e+00 1.64545929e+00
1.32324785e-01 8.73312950e-01 6.44447803e-02 5.63410819e-01
1.09627092e+00 9.38751101e-01 3.21970910e-01 -1.10811599e-01
1.36292815e+00 -1.25769734e+00 -7.13067532e-01 -4.54914659e-01
1.20516551e+00 -4.56393421e-01 1.02914202e+00 3.45123798e-01
-1.07791221e+00 -6.73958242e-01 -6.82168663e-01 -3.49261165e-02
-3.29910427e-01 2.07797557e-01 7.72832811e-01 3.78507346e-01
-8.20606351e-01 7.04368234e-01 -9.98565376e-01 -2.84328282e-01
6.10379755e-01 2.17709363e-01 -7.18745410e-01 -5.04853010e-01
-5.71922779e-01 6.19206250e-01 7.24242747e-01 -7.89760798e-02
-1.41772604e+00 -1.08237815e+00 -1.41268718e+00 2.93002307e-01
9.94610012e-01 -5.27234375e-01 1.58081126e+00 -1.30777383e+00
-1.44320297e+00 9.21211839e-01 -2.40471438e-01 -2.20369905e-01
2.02664003e-01 -2.66990900e-01 -4.75403145e-02 3.80778432e-01
3.79447304e-02 6.48257554e-01 8.36003721e-01 -1.07574165e+00
-9.02829826e-01 -4.57773246e-02 5.25524139e-01 4.57070202e-01
3.44823562e-02 -1.63787335e-01 -8.82857382e-01 -4.37702000e-01
-1.61086306e-01 -1.16494238e+00 2.11479485e-01 -3.31763059e-01
-2.65422463e-01 -2.41034374e-01 5.16430616e-01 -8.78094196e-01
1.48329127e+00 -2.06686544e+00 4.51692671e-01 4.69594263e-02
1.34367794e-01 3.51357669e-01 -3.15773666e-01 4.06217098e-01
-4.23939139e-01 3.89794111e-02 -1.08917989e-01 -1.63036026e-02
-2.17500165e-01 3.47839862e-01 -2.72046290e-02 2.02047274e-01
1.47498086e-01 8.92075539e-01 -1.28255105e+00 -3.31021994e-01
3.55143785e-01 1.55659556e-01 -5.98151565e-01 5.22509813e-01
-3.83759230e-01 4.83840168e-01 -6.16374135e-01 6.49687827e-01
9.38778073e-02 -3.76271546e-01 4.24684078e-01 -2.11270273e-01
-4.06656861e-02 4.84721392e-01 -1.19524837e+00 2.06242156e+00
-4.85462397e-01 7.87162304e-01 -2.25393355e-01 -1.17640686e+00
3.05935293e-01 3.67040336e-01 4.37036783e-01 -3.04540575e-01
7.44093210e-03 -2.14011893e-01 -1.57021910e-01 -8.51803243e-01
3.66983175e-01 2.76640236e-01 -1.08255371e-01 4.84796286e-01
5.83458543e-01 -3.90859284e-02 7.19757497e-01 3.86957079e-01
1.37034571e+00 7.57643700e-01 4.82516289e-01 2.11846270e-02
6.36973143e-01 3.01693976e-01 3.12873721e-01 7.45789111e-01
-5.31983264e-02 3.11679363e-01 6.81588888e-01 -4.09146786e-01
-5.47859430e-01 -9.02759016e-01 4.68298584e-01 1.36144817e+00
-1.83046803e-01 -7.38661587e-01 -8.36034477e-01 -9.39213872e-01
-4.44384627e-02 7.62256384e-01 -7.93437123e-01 -2.92294443e-01
-8.45115483e-01 2.85530716e-01 2.04182118e-01 7.97574043e-01
1.57780722e-01 -1.25221074e+00 -5.74575722e-01 1.54975787e-01
-3.79582793e-01 -1.36554623e+00 -6.31781399e-01 2.46588156e-01
-8.08604836e-01 -1.53481257e+00 -1.94899037e-01 -1.05708277e+00
9.21962798e-01 4.35539097e-01 1.25982237e+00 1.14636369e-01
-1.13265775e-01 1.01026487e+00 -4.22924817e-01 -2.60124475e-01
-4.21231270e-01 -1.28596738e-01 -2.45414600e-01 3.05890758e-02
2.10967049e-01 -1.75064847e-01 -3.82178277e-01 2.71078330e-02
-8.90867770e-01 5.10695815e-01 3.87639821e-01 6.44438207e-01
6.53689921e-01 -1.14029162e-01 1.11623101e-01 -1.41539752e+00
4.60540533e-01 -4.95896906e-01 -5.67172110e-01 4.89307791e-01
-1.22146621e-01 2.37670213e-01 6.36302710e-01 -5.58631182e-01
-1.32488644e+00 2.96621859e-01 1.94211006e-01 -7.50780463e-01
-5.25272906e-01 9.46428359e-01 1.12844877e-01 1.00105897e-01
4.72591996e-01 2.52022833e-01 -3.07059556e-01 -2.19590530e-01
4.30103898e-01 1.80458188e-01 6.56446278e-01 -1.00045288e+00
5.22114933e-01 1.19941436e-01 -1.34472460e-01 -7.08446324e-01
-1.15288389e+00 -7.21069038e-01 -6.77231967e-01 -5.89619040e-01
1.05077600e+00 -1.01548958e+00 -9.95250523e-01 2.54669964e-01
-7.63995647e-01 -1.00625837e+00 -3.42739016e-01 5.46777368e-01
-7.54707098e-01 1.35969013e-01 -6.90136194e-01 -4.24543023e-01
2.15756327e-01 -1.25017643e+00 6.98619902e-01 4.32632864e-01
-3.24230045e-01 -1.19661129e+00 -2.11818680e-01 7.65047491e-01
-1.90848514e-01 9.03775990e-02 9.60576832e-01 -7.13630080e-01
-8.09830308e-01 -1.08773753e-01 7.60177076e-02 2.66115278e-01
2.67312557e-01 6.51827306e-02 -7.14343905e-01 -3.60301435e-02
-2.93869883e-01 -5.82858980e-01 5.70730865e-01 3.98935497e-01
1.54020000e+00 -3.04953337e-01 -4.19087678e-01 7.16134012e-01
1.16357231e+00 3.29298496e-01 5.85945070e-01 2.51287192e-01
1.04856420e+00 6.84028804e-01 7.14251637e-01 2.39143819e-02
4.46980834e-01 5.02728164e-01 5.13816886e-02 5.10604262e-01
-3.22244316e-01 -6.15697682e-01 6.20528758e-01 9.73227859e-01
-3.44007790e-01 -2.42750928e-01 -9.34614539e-01 6.18396699e-01
-1.90921271e+00 -1.27485025e+00 8.70882273e-02 2.25555277e+00
8.43481421e-01 7.97810778e-02 -7.13039562e-02 -1.64635390e-01
4.60338861e-01 1.96896419e-01 -4.66493160e-01 -5.10288954e-01
6.33748949e-01 4.58165199e-01 3.06177706e-01 4.99523371e-01
-1.31358540e+00 1.09531271e+00 5.56588125e+00 6.48193896e-01
-8.80961716e-01 -2.11793423e-01 3.54318768e-01 1.02988698e-01
-3.17349397e-02 6.62588179e-02 -8.44708502e-01 3.17279011e-01
1.31369054e+00 -4.26934361e-01 6.95873082e-01 7.41879582e-01
2.11868659e-01 -2.65558034e-01 -1.63842142e+00 8.94939125e-01
3.04445654e-01 -1.56287909e+00 2.10539371e-01 -4.90320295e-01
8.21898520e-01 -4.06625211e-01 -1.89996526e-01 8.50333512e-01
4.71230030e-01 -9.70279276e-01 5.93104959e-01 3.77082616e-01
7.63066530e-01 -3.99212807e-01 2.52338737e-01 4.55752611e-01
-1.40837812e+00 -9.45741013e-02 1.36679307e-01 -1.05163716e-01
-4.50290367e-02 -2.39535883e-01 -1.13403118e+00 7.05179811e-01
4.03703153e-01 1.07391834e+00 -4.56144392e-01 9.24693704e-01
-7.10616112e-01 1.00591397e+00 -2.14747954e-02 1.87565699e-01
4.22186226e-01 -2.38603324e-01 2.60556549e-01 1.25764191e+00
2.85511613e-01 3.85536253e-01 4.18349385e-01 5.90170026e-01
-3.76748681e-01 -1.28705382e-01 -8.02561164e-01 -3.91826332e-01
3.84677827e-01 1.15884459e+00 -8.21633577e-01 -6.46745443e-01
-7.52417624e-01 9.25943613e-01 5.41523814e-01 6.59421086e-01
-8.69122207e-01 2.08209828e-02 6.55203342e-01 -6.26738358e-04
5.23255587e-01 -1.47933334e-01 3.24742049e-01 -1.40011942e+00
-1.21352367e-01 -1.02143133e+00 5.83411157e-01 -1.14046526e+00
-5.49060285e-01 2.09967986e-01 3.13731402e-01 -1.26052833e+00
-4.49966788e-01 -7.08469570e-01 -6.36171639e-01 5.95779538e-01
-1.36066246e+00 -9.30209160e-01 -6.34034753e-01 7.19461322e-01
1.18874323e+00 1.33578137e-01 7.30710387e-01 1.03483617e-01
-6.86283052e-01 3.33980143e-01 -3.76256764e-01 2.75928497e-01
5.66932023e-01 -1.27609038e+00 2.14407481e-02 9.65203345e-01
3.85254651e-01 5.15967488e-01 4.32025880e-01 -7.08942950e-01
-1.81441808e+00 -1.24937451e+00 5.86767375e-01 -6.87923074e-01
7.26719856e-01 -2.12013409e-01 -1.10850441e+00 1.35399222e+00
-1.10162236e-03 -4.35126238e-02 8.02050829e-01 3.99035849e-02
-4.05959398e-01 1.50888294e-01 -5.48474312e-01 6.27329826e-01
1.22184098e+00 -7.00658977e-01 -9.02103007e-01 5.00191927e-01
4.79737431e-01 -1.03925371e+00 -8.39956224e-01 -2.96644811e-02
4.42460179e-01 -5.57020128e-01 9.47325766e-01 -1.08726096e+00
8.68029296e-01 -5.17964363e-02 2.47651622e-01 -1.37670588e+00
-1.19170569e-01 -5.68905175e-01 -4.05148745e-01 1.03160989e+00
3.81065637e-01 3.34915146e-02 8.96736443e-01 8.84852648e-01
-6.15492344e-01 -5.30605257e-01 -3.49455327e-01 -6.37542367e-01
-3.80071700e-01 -3.91958982e-01 2.40312934e-01 1.14295650e+00
2.04414830e-01 3.59802127e-01 -2.73110390e-01 9.84256417e-02
2.61260390e-01 1.37264252e-01 7.66515672e-01 -9.27129626e-01
-1.84770852e-01 -1.79745287e-01 -1.87507808e-01 -1.22792804e+00
6.98916137e-01 -1.24454248e+00 1.53737202e-01 -1.73252845e+00
2.59667516e-01 -1.08746164e-01 -2.93663412e-01 8.87795627e-01
-4.13977444e-01 -3.00686091e-01 2.74382472e-01 -1.09577328e-02
-9.46969688e-01 -2.55407989e-02 1.37798429e+00 -1.06694475e-01
-5.03289700e-01 -1.62824299e-02 -3.76998156e-01 8.03074837e-01
5.80584586e-01 -3.83667529e-01 -7.10880041e-01 -6.13629043e-01
3.60595733e-02 5.62438488e-01 9.87721607e-02 -1.13896132e+00
3.53966147e-01 -3.92859876e-01 1.55872375e-01 -2.07597986e-01
2.84654737e-01 -8.28785419e-01 2.32662007e-01 2.20448688e-01
-5.71786284e-01 -3.87291573e-02 4.98778939e-01 6.71537757e-01
-4.15743738e-01 -5.75035453e-01 1.61260813e-01 -3.12908769e-01
-1.47446108e+00 2.28413403e-01 -6.57489538e-01 1.19295239e-01
1.33663571e+00 -5.28570473e-01 -4.85079348e-01 -4.79357481e-01
-9.70486462e-01 3.36770236e-01 2.85236448e-01 5.17535508e-01
6.45152926e-01 -1.05102015e+00 -4.08981562e-01 1.79486096e-01
2.82460272e-01 5.14901280e-02 5.42933226e-01 7.25622237e-01
-7.22441077e-01 4.34632629e-01 -3.95706117e-01 -4.98370498e-01
-1.55582392e+00 5.94130218e-01 1.26059651e-02 -3.06545079e-01
-6.58712149e-01 8.95471573e-01 4.24895227e-01 -3.45461160e-01
3.45909774e-01 -7.28937745e-01 -5.07014990e-01 4.54425737e-02
6.02339804e-01 1.77682832e-01 -4.66013774e-02 -3.46773326e-01
-1.89753562e-01 4.22279894e-01 -1.49297833e-01 2.92085499e-01
1.27405167e+00 1.67364180e-01 2.24300712e-01 5.25901139e-01
1.06452906e+00 -1.36119947e-01 -1.59733295e+00 -1.26988187e-01
9.18567926e-02 -4.21256721e-01 -1.16318442e-01 -8.10624897e-01
-9.83819783e-01 6.56607926e-01 -2.31723841e-02 -1.55382052e-01
1.03975153e+00 -6.35927962e-03 2.74918199e-01 6.17717266e-01
3.22530538e-01 -8.34948957e-01 2.24535614e-01 6.95553303e-01
5.98791957e-01 -1.15035152e+00 6.64586648e-02 -8.18911374e-01
-9.23556864e-01 1.30029368e+00 7.45379865e-01 1.64658695e-01
2.71402508e-01 8.88351053e-02 -1.61853760e-01 -3.22553962e-01
-9.45735514e-01 -1.65330589e-01 5.62518775e-01 5.99406898e-01
4.81731594e-01 1.64858401e-01 4.39153090e-02 5.53886056e-01
1.63912028e-01 5.42659402e-01 7.81378090e-01 1.34875655e+00
-2.91276634e-01 -8.78195584e-01 -6.62579164e-02 8.06242764e-01
-3.03976327e-01 -1.29829673e-02 -2.37379652e-02 8.42599750e-01
-3.92425358e-02 7.08344102e-01 9.25147310e-02 -2.68518448e-01
4.92179394e-01 4.49603587e-01 8.48932266e-01 -1.16867900e+00
-7.05561221e-01 -1.87658101e-01 5.26036501e-01 -9.88875926e-01
-7.65580118e-01 -7.31828749e-01 -1.26689422e+00 -2.30845258e-01
-1.81933977e-02 2.03867972e-01 3.06391656e-01 1.01223993e+00
1.32425874e-01 9.03716564e-01 7.47260526e-02 -7.89686322e-01
-8.63751099e-02 -6.55350447e-01 -1.52971700e-01 8.36998701e-01
1.54872701e-01 -6.98868930e-01 -1.02487221e-01 8.49422097e-01] | [8.844502449035645, 0.6694371104240417] |
145d19f5-1567-41fb-b094-0a0edf565c9a | cluster-consistency-simple-yet-effect-robust | 2211.03333 | null | https://arxiv.org/abs/2211.03333v1 | https://arxiv.org/pdf/2211.03333v1.pdf | Cluster consistency: Simple yet effect robust learning algorithm on large-scale photoplethysmography for atrial fibrillation detection in the presence of real-world label noise | Obtaining large-scale well-annotated is always a daunting challenge, especially in the medical research domain because of the shortage of domain expert. Instead of human annotation, in this work, we use the alarm information generated from bed-side monitor to get the pseudo label for the co-current photoplethysmography (PPG) signal. Based on this strategy, we end up with over 8 million 30-second PPG segment. To solve the label noise caused by false alarms, we propose the cluster consistency, which use an unsupervised auto-encoder (hence not subject to label noise) approach to cluster training samples into a finite number of clusters. Then the learned cluster membership is used in the subsequent supervised learning phase to force the distance in the latent space of samples in the same cluster to be small while that of samples in different clusters to be big. In the experiment, we compare with the state-of-the-art algorithms and test on external datasets. The results show the superiority of our method in both classification performance and efficiency. | ['Xiao Hu', 'Cynthia Rudin', 'Gari Clifford', 'Amit Shah', 'Zhicheng Guo', 'Cheng Ding'] | 2022-11-07 | null | null | null | null | ['photoplethysmography-ppg', 'atrial-fibrillation-detection'] | ['medical', 'medical'] | [ 1.23215236e-01 1.21864259e-01 7.02524511e-03 -4.91436988e-01
-7.19037831e-01 -2.46203095e-01 -2.82171220e-02 -1.82696417e-01
-1.69820964e-01 7.50273228e-01 -2.98145935e-02 1.16349712e-01
7.82847255e-02 -3.94637704e-01 -2.39236087e-01 -1.01761007e+00
1.42375708e-01 4.48034853e-01 1.54705495e-01 4.71821755e-01
9.80510116e-02 8.71967301e-02 -1.33957767e+00 2.43138030e-01
9.49466288e-01 1.13749182e+00 6.06796108e-02 2.36001432e-01
-3.59650329e-02 8.03100288e-01 -7.18124390e-01 -9.49481651e-02
2.39317596e-01 -7.93279111e-01 -3.65218252e-01 3.77784699e-01
-6.78773969e-02 1.60351098e-02 -8.80302489e-02 1.26139379e+00
7.60455787e-01 -2.85097454e-02 4.32990283e-01 -1.09882212e+00
-2.11998656e-01 2.75584728e-01 -5.96357942e-01 8.97089541e-02
-7.69455284e-02 -1.05516717e-01 5.25400996e-01 -6.64740086e-01
3.94399852e-01 7.10545123e-01 6.36374891e-01 7.77985752e-01
-1.07659626e+00 -8.51598084e-01 -1.69419333e-01 2.44853735e-01
-1.61184287e+00 -4.75713789e-01 8.26238096e-01 -4.29111689e-01
2.81861037e-01 2.79185981e-01 5.69436312e-01 9.62691009e-01
5.88386059e-02 4.50065106e-01 1.27118206e+00 -3.85440558e-01
5.04563093e-01 5.52625597e-01 4.41961884e-02 5.66867352e-01
1.43638030e-01 -1.66146934e-01 -3.25069547e-01 -2.15074614e-01
7.12217093e-01 4.65219244e-02 -3.68901938e-01 -3.22928220e-01
-1.01765704e+00 8.05017650e-01 -1.14212595e-01 3.25541854e-01
-4.30389911e-01 -4.68618870e-02 4.27736908e-01 5.40599450e-02
3.27611566e-01 1.74730420e-01 -4.18903947e-01 -2.25319982e-01
-9.61765766e-01 -3.55441570e-01 8.41650307e-01 8.39489102e-01
5.45635879e-01 -1.85741335e-01 -9.36477408e-02 7.13885367e-01
3.41019243e-01 9.85852852e-02 9.28175867e-01 -8.61213565e-01
2.46354014e-01 6.87312424e-01 6.27803579e-02 -9.44446146e-01
-4.97447729e-01 -4.45863724e-01 -1.00672674e+00 -1.53665811e-01
3.83092880e-01 -2.92784333e-01 -7.99314439e-01 1.51539516e+00
4.44334567e-01 6.81916893e-01 -8.33402667e-03 1.07065463e+00
5.50648749e-01 6.15489006e-01 3.75663117e-02 -8.03217530e-01
1.38308787e+00 -8.40942681e-01 -1.00954771e+00 3.97347286e-02
6.93715632e-01 -6.07894480e-01 8.09740782e-01 5.32801747e-01
-5.87308586e-01 -7.76367664e-01 -9.19087946e-01 3.93462479e-01
-2.99528837e-02 5.07182062e-01 5.28753042e-01 6.45123661e-01
-5.40397227e-01 5.95527768e-01 -9.73614454e-01 -2.09743530e-01
3.52336287e-01 2.34532908e-01 -3.47508013e-01 1.47314956e-02
-1.26253819e+00 7.10892498e-01 5.14102519e-01 2.24240810e-01
-5.65267861e-01 -4.78037208e-01 -6.96575046e-01 -6.45045042e-02
3.08043092e-01 -1.57209501e-01 9.62333381e-01 -9.94603455e-01
-1.63490260e+00 8.71087551e-01 -5.60324788e-02 -2.86676168e-01
4.44724590e-01 -2.61847284e-02 -6.34991884e-01 2.44086310e-01
-3.97059182e-03 3.12275589e-01 9.27796721e-01 -9.53670979e-01
-5.38102925e-01 -3.15225840e-01 -7.34575987e-01 2.43223801e-01
-4.00135189e-01 -1.06541798e-01 -5.94073415e-01 -1.84574187e-01
2.96679139e-01 -1.03407955e+00 -2.17062011e-01 -3.11271369e-01
-4.21479672e-01 -3.68509948e-01 7.13831365e-01 -6.74617112e-01
1.50933611e+00 -2.33573961e+00 -3.18196774e-01 4.19718981e-01
2.99905866e-01 2.28988573e-01 4.98796195e-01 1.93268973e-02
-3.90382349e-01 -1.76549509e-01 -1.74451888e-01 -3.66571575e-01
-1.95126578e-01 2.13615090e-01 -7.75939077e-02 7.15850949e-01
-1.04291506e-01 4.89356011e-01 -9.31128860e-01 -7.78053045e-01
3.42553377e-01 1.83951646e-01 -1.25823602e-01 6.17994428e-01
3.26500475e-01 6.75599754e-01 -4.35944557e-01 4.56732452e-01
7.40549982e-01 -5.00950456e-01 1.62718937e-01 -3.79774332e-01
5.63504808e-02 1.35885701e-01 -1.40438688e+00 1.69017684e+00
-1.57626793e-01 2.83352047e-01 -1.85172200e-01 -1.18218267e+00
1.11256504e+00 6.54234350e-01 6.88296735e-01 -4.61992264e-01
1.23179711e-01 2.23449275e-01 2.30894998e-01 -8.34826112e-01
-1.51217163e-01 -4.90828216e-01 4.50154059e-02 4.56329405e-01
4.29359451e-02 2.35925913e-01 -9.94731486e-02 -5.18686213e-02
8.44201565e-01 -1.33227497e-01 2.43776187e-01 -2.18461752e-01
6.25081658e-01 -3.42169672e-01 9.19562161e-01 5.23337245e-01
-4.58671391e-01 8.24014366e-01 5.80169618e-01 -4.08916175e-01
-1.08628118e+00 -7.07215786e-01 -3.49687904e-01 2.27018282e-01
7.62244463e-02 -4.25850064e-01 -8.84990811e-01 -7.34031200e-01
-1.80930585e-01 4.35915053e-01 -5.12466431e-01 -2.60983169e-01
-4.68651712e-01 -9.53438103e-01 4.62365329e-01 4.24854517e-01
4.99912560e-01 -1.08804989e+00 -6.38162553e-01 3.91586155e-01
-3.10088396e-01 -1.05455256e+00 -3.72529387e-01 1.60802290e-01
-7.73481548e-01 -1.04932189e+00 -7.32629001e-01 -7.62995303e-01
6.96349680e-01 -2.71797657e-01 6.87144935e-01 -2.10758999e-01
-6.14170015e-01 -1.27616990e-02 -3.14998597e-01 -2.44466394e-01
-3.76266688e-01 -1.82890013e-01 2.93770522e-01 5.36575198e-01
7.67084062e-01 -3.44992787e-01 -7.70329416e-01 4.94270861e-01
-6.33774042e-01 7.90254921e-02 4.74860817e-01 8.62439632e-01
7.68314064e-01 2.16518074e-01 7.33156621e-01 -9.78641629e-01
4.76618499e-01 -4.67951685e-01 -5.73959053e-01 2.44169965e-01
-7.58342803e-01 -1.63278148e-01 6.93698049e-01 -6.42443776e-01
-7.17309833e-01 3.01751763e-01 1.35643184e-01 -8.71975720e-01
-1.67333141e-01 3.79702806e-01 -8.94601494e-02 2.23077416e-01
5.74379086e-01 4.42742437e-01 2.99049169e-01 -4.76522088e-01
1.46765783e-01 1.13471794e+00 5.40726483e-01 -1.73171744e-01
4.29436058e-01 3.55645806e-01 -6.82633594e-02 -5.40517628e-01
-7.38268435e-01 -6.37254477e-01 -4.55439121e-01 -2.34642789e-01
9.80780721e-01 -1.03807855e+00 -7.51455367e-01 3.78842533e-01
-8.82790208e-01 2.65986416e-02 -2.40316719e-01 8.60082686e-01
-4.47426647e-01 6.79184616e-01 -5.99705815e-01 -1.01667869e+00
-3.94211501e-01 -9.92348194e-01 7.09622860e-01 3.86318952e-01
-1.37556195e-01 -6.90529943e-01 1.18151151e-01 1.47044733e-01
1.17673419e-01 2.48588085e-01 4.41811442e-01 -9.05794024e-01
-2.20685735e-01 -5.92040420e-01 -1.33946285e-01 7.20529437e-01
4.05641615e-01 -3.13868076e-01 -1.07648480e+00 -2.06914723e-01
7.25593328e-01 -2.16123477e-01 4.63273764e-01 4.20497000e-01
1.41193295e+00 -2.97217160e-01 -3.21326256e-01 4.93452638e-01
1.22734916e+00 3.94399822e-01 8.75786722e-01 -1.29529759e-01
7.89558291e-01 6.42727673e-01 7.85458267e-01 5.63428164e-01
8.30633417e-02 6.08718932e-01 -1.28815234e-01 -1.67395622e-01
2.45493889e-01 -1.85713947e-01 1.29133493e-01 1.08089113e+00
1.61705852e-01 2.24563032e-02 -6.09649777e-01 4.27038252e-01
-1.85979211e+00 -7.98456848e-01 -9.25402269e-02 2.52357149e+00
1.14506924e+00 3.28020118e-02 3.26096676e-02 1.79904372e-01
1.16016638e+00 -3.11728984e-01 -4.18096930e-01 -6.68506548e-02
2.32313767e-01 3.48126367e-02 4.56647038e-01 1.66990981e-01
-1.01898181e+00 6.07838035e-01 5.70800209e+00 9.54870045e-01
-1.17705488e+00 6.74041063e-02 9.26697373e-01 6.20269515e-02
3.39921772e-01 -1.59299653e-03 -7.06666768e-01 1.18773651e+00
1.24011147e+00 8.77700225e-02 4.14661556e-01 7.56912351e-01
2.47312471e-01 -1.36824593e-01 -1.23534083e+00 1.49860573e+00
2.15822771e-01 -8.89946282e-01 -5.84724486e-01 -4.77644578e-02
4.74724174e-01 -3.36291850e-01 -2.75535733e-01 1.55940384e-01
-3.03886175e-01 -9.34602261e-01 2.17307970e-01 6.45744562e-01
9.48128879e-01 -5.38443863e-01 9.49177980e-01 4.76983935e-01
-8.92300010e-01 1.41851813e-01 -6.60473287e-01 1.82066455e-01
1.00765593e-01 1.03704405e+00 -8.60843658e-01 4.28698868e-01
4.78230566e-01 7.02075660e-01 -1.90389842e-01 1.12305725e+00
-1.43320099e-01 8.32718134e-01 -4.21874851e-01 1.97391778e-01
-1.40197173e-01 -3.55286837e-01 1.18214428e-01 1.01142621e+00
4.39111352e-01 4.57711905e-01 2.56320983e-01 8.59078526e-01
1.25327051e-01 1.77004665e-01 -4.81202871e-01 3.44729722e-02
5.79192400e-01 1.44541442e+00 -5.57236135e-01 -6.94112778e-01
-4.77656275e-01 1.11435008e+00 4.04354110e-02 1.84186757e-01
-9.02028441e-01 -6.24930263e-01 -8.93446151e-03 -3.96052003e-02
1.06407031e-01 2.64112204e-01 -2.74488479e-01 -1.21840501e+00
1.51280731e-01 -7.28315473e-01 5.60836971e-01 -6.88220441e-01
-1.32838559e+00 5.74061811e-01 -2.45787278e-01 -1.80137885e+00
-2.78083056e-01 -2.60116130e-01 -5.45219362e-01 9.56938386e-01
-1.28736007e+00 -5.94625652e-01 -4.50014293e-01 6.63538933e-01
2.33979866e-01 -1.33414313e-01 9.80099618e-01 6.38450503e-01
-6.79099917e-01 6.67745233e-01 9.99107882e-02 2.17508644e-01
1.05669641e+00 -1.12098193e+00 -2.45814666e-01 5.26241124e-01
-2.34717086e-01 5.35588443e-01 4.23257113e-01 -7.43172288e-01
-8.57572138e-01 -9.68533933e-01 9.25203443e-01 -4.15566117e-01
4.09441829e-01 -4.67967391e-01 -1.00597239e+00 4.18666691e-01
-1.50435120e-01 4.05099928e-01 9.20678914e-01 -1.98095247e-01
-1.02297291e-01 -3.94679457e-01 -1.26152003e+00 1.62906155e-01
4.59798068e-01 -3.17715734e-01 -5.46353996e-01 3.51165742e-01
4.67302769e-01 -4.99565661e-01 -1.09765291e+00 3.19908082e-01
3.94584298e-01 -7.90145934e-01 4.81613517e-01 -3.53891462e-01
3.24075937e-01 -4.76641983e-01 1.70346662e-01 -1.19560361e+00
-3.03856224e-01 -6.84166074e-01 -3.70820537e-02 1.25395787e+00
1.93947464e-01 -7.26347446e-01 9.36730385e-01 8.66055608e-01
-2.15176150e-01 -5.92564166e-01 -1.05206013e+00 -6.45153880e-01
-4.73186761e-01 -2.71329731e-02 2.33329415e-01 1.19251013e+00
6.03292286e-01 3.50704610e-01 -7.60412276e-01 1.41030267e-01
6.99094236e-01 2.20899940e-01 3.99338126e-01 -1.22971439e+00
-4.98715550e-01 3.04235548e-01 -4.58533257e-01 -7.48434067e-01
-1.94278985e-01 -6.04612768e-01 3.35693330e-01 -1.08490419e+00
3.13454241e-01 -5.66188157e-01 -6.26978099e-01 5.24302542e-01
-3.48194361e-01 3.17942530e-01 -9.54391956e-02 4.84146118e-01
-7.70609796e-01 3.26197416e-01 1.08147156e+00 3.37360620e-01
-4.19892013e-01 1.04121439e-01 -5.31713843e-01 6.24941111e-01
6.36071086e-01 -7.45694280e-01 -3.16716403e-01 2.18851924e-01
-1.37989819e-01 3.32704335e-01 -7.54039809e-02 -1.05549407e+00
2.29846194e-01 1.57129660e-01 5.65855622e-01 -5.11491120e-01
1.19812302e-01 -1.08078051e+00 2.28900403e-01 5.65981269e-01
-2.42218181e-01 -3.26761574e-01 -1.89631265e-02 4.70117360e-01
-2.86270887e-01 -3.07718724e-01 9.77093875e-01 -2.42937490e-01
-2.80517220e-01 3.12081724e-01 -1.42547593e-01 -8.79679620e-02
1.12568152e+00 -1.91128537e-01 -9.39899608e-02 -2.94218391e-01
-7.62624145e-01 1.82263166e-01 1.68335140e-01 1.16499528e-01
3.98745775e-01 -1.37972224e+00 -5.75963438e-01 4.14874971e-01
2.67139524e-01 3.28777693e-02 4.24779147e-01 1.11236858e+00
-3.42933118e-01 2.99273819e-01 1.92557741e-02 -6.92336619e-01
-1.07441592e+00 6.50296986e-01 3.32397163e-01 -1.89415827e-01
-6.83424234e-01 5.85154653e-01 1.84112981e-01 -1.10000007e-01
2.54024714e-01 -4.37719487e-02 -3.25812966e-01 3.11922468e-02
6.46073341e-01 2.69813329e-01 1.15845710e-01 -2.03751266e-01
-3.18922222e-01 4.81442720e-01 -2.49588594e-01 7.16111958e-02
1.02945209e+00 -1.12340264e-01 -1.29344061e-01 4.97037470e-01
1.38309383e+00 -1.63648680e-01 -1.11260235e+00 -8.71721804e-02
-8.81047025e-02 -5.08441091e-01 -1.69004008e-01 -7.49839962e-01
-9.92539585e-01 9.26312625e-01 1.19095063e+00 1.71110645e-01
1.13498139e+00 -9.69751105e-02 5.99684179e-01 9.05940905e-02
2.25665957e-01 -1.42668200e+00 3.18028815e-02 -7.24610463e-02
2.35568225e-01 -1.19251156e+00 -8.12002867e-02 -3.65205944e-01
-1.12364364e+00 1.00451589e+00 5.75066686e-01 -2.32030421e-01
5.60499310e-01 3.33545841e-02 2.81417549e-01 -1.99691594e-01
-6.23138309e-01 2.30147153e-01 3.07530537e-02 4.00502145e-01
2.65223712e-01 9.26012695e-02 -6.44252658e-01 7.92783320e-01
2.10457459e-01 3.78907084e-01 5.33181846e-01 5.12041867e-01
-3.63870770e-01 -9.93344009e-01 -2.24226564e-01 6.70878768e-01
-6.34665728e-01 1.48447342e-02 1.82205960e-01 4.19243693e-01
2.57498950e-01 8.93574297e-01 1.42940670e-01 -3.12917352e-01
1.46195799e-01 3.74307483e-01 2.03322291e-01 -7.04635859e-01
-1.44154593e-01 5.93836784e-01 -1.28256068e-01 -6.21917844e-01
-3.76483768e-01 -6.14870667e-01 -1.30778897e+00 3.52639824e-01
-5.51459134e-01 4.98900056e-01 5.24701774e-01 6.94521368e-01
3.56737405e-01 3.02481025e-01 1.05881083e+00 -3.18986535e-01
-7.56700099e-01 -1.19929349e+00 -9.58369911e-01 7.65480936e-01
1.04473256e-01 -4.20959294e-01 -7.10482657e-01 2.42937490e-01] | [14.031670570373535, 2.705820083618164] |
27fecde8-2811-479d-b4f3-c1bcd91ac424 | cross-level-cross-scale-cross-attention | 2104.13053 | null | https://arxiv.org/abs/2104.13053v1 | https://arxiv.org/pdf/2104.13053v1.pdf | Cross-Level Cross-Scale Cross-Attention Network for Point Cloud Representation | Self-attention mechanism recently achieves impressive advancement in Natural Language Processing (NLP) and Image Processing domains. And its permutation invariance property makes it ideally suitable for point cloud processing. Inspired by this remarkable success, we propose an end-to-end architecture, dubbed Cross-Level Cross-Scale Cross-Attention Network (CLCSCANet), for point cloud representation learning. First, a point-wise feature pyramid module is introduced to hierarchically extract features from different scales or resolutions. Then a cross-level cross-attention is designed to model long-range inter-level and intra-level dependencies. Finally, we develop a cross-scale cross-attention module to capture interactions between-and-within scales for representation enhancement. Compared with state-of-the-art approaches, our network can obtain competitive performance on challenging 3D object classification, point cloud segmentation tasks via comprehensive experimental evaluation. | ['Guo-Qiang Xiao', 'Jia Chen', 'Zhang-Yue He', 'Xian-Feng Han'] | 2021-04-27 | null | null | null | null | ['3d-object-classification'] | ['computer-vision'] | [ 3.00154258e-02 -3.61146241e-01 -7.78372958e-02 -5.36707342e-01
-9.48845148e-01 -3.03883761e-01 3.90084326e-01 2.51318544e-01
-1.62018970e-01 2.01417934e-02 -1.28039017e-01 -1.73487306e-01
-1.34262025e-01 -8.78485024e-01 -1.17773652e+00 -3.19672555e-01
-2.34744936e-01 1.20905049e-01 3.53197664e-01 -6.27942150e-03
4.35845107e-01 1.21898699e+00 -1.48226595e+00 4.40284491e-01
8.95406008e-01 1.07771552e+00 4.55349416e-01 5.44023931e-01
-5.52664042e-01 5.44000447e-01 -2.21696541e-01 -1.90505609e-01
2.78520137e-01 3.70780826e-01 -7.83405602e-01 2.06221312e-01
7.08303034e-01 -1.59142658e-01 -4.23912376e-01 1.15729463e+00
4.98960882e-01 2.64610667e-02 5.33655405e-01 -9.87585723e-01
-1.26879668e+00 1.70697346e-01 -1.25419319e+00 3.29319298e-01
1.26743540e-02 2.03143045e-01 1.04348934e+00 -1.16754925e+00
4.17114884e-01 1.58253849e+00 5.36628902e-01 2.60020971e-01
-1.10539830e+00 -7.79128551e-01 5.86906075e-01 1.37568206e-01
-1.42158997e+00 1.71749771e-01 1.16926229e+00 -3.55174571e-01
1.31163394e+00 -1.16866827e-01 5.68699479e-01 6.76270127e-01
2.29203656e-01 1.16101253e+00 7.68962383e-01 -1.20200567e-01
-9.60767940e-02 -3.32594991e-01 1.07546777e-01 6.59124196e-01
4.11225706e-02 -6.98074475e-02 -2.88987190e-01 -2.11941414e-02
1.32225704e+00 2.29347602e-01 7.76294544e-02 -4.92677510e-01
-1.11263251e+00 6.65307462e-01 1.31890583e+00 5.08389413e-01
-7.75310636e-01 3.19841266e-01 2.92166263e-01 2.07983796e-02
7.24674165e-01 3.57365698e-01 -7.43805170e-01 3.16502780e-01
-8.18038762e-01 2.02013329e-01 -4.11536060e-02 1.23866510e+00
8.26245904e-01 8.25880691e-02 -2.85813987e-01 1.01856792e+00
4.64101732e-01 4.24171865e-01 1.70332327e-01 -6.55281663e-01
5.54104149e-01 1.04531407e+00 -2.60784298e-01 -9.92870629e-01
-3.31157565e-01 -6.80278063e-01 -1.07579255e+00 3.78039420e-01
-2.17055142e-01 4.09944564e-01 -1.14894891e+00 1.43298042e+00
4.08814013e-01 6.71530664e-01 -1.36093736e-01 1.03445518e+00
9.40365791e-01 7.89025843e-01 4.03602570e-01 1.18662573e-01
1.66676342e+00 -1.02735770e+00 -1.79341823e-01 -1.86102420e-01
-3.82962427e-03 -5.78774273e-01 1.11817729e+00 -4.16024104e-02
-1.42464066e+00 -1.05988491e+00 -9.05448139e-01 -5.34639537e-01
-5.40898681e-01 -3.65421362e-02 7.45570779e-01 1.22538293e-02
-8.45463514e-01 6.09883249e-01 -1.00029194e+00 -1.66684493e-01
1.03231180e+00 6.01055264e-01 -2.45367199e-01 -9.28147063e-02
-7.00511873e-01 3.56841117e-01 3.25319842e-02 8.41644108e-02
-4.58124757e-01 -1.02221024e+00 -8.52105260e-01 2.90826976e-01
-9.08204839e-02 -1.11906445e+00 9.62693691e-01 -8.44800830e-01
-1.37133801e+00 1.17877579e+00 -3.61285359e-01 -3.00344527e-01
7.78403580e-02 -2.97840238e-01 -9.36874449e-02 2.89479136e-01
3.21920335e-01 1.10948217e+00 1.09922338e+00 -1.44975853e+00
-8.03367376e-01 -7.98006177e-01 -2.04380788e-02 2.94578046e-01
7.88211543e-03 2.50147402e-01 -1.07768035e+00 -7.86338389e-01
5.14916480e-01 -6.54485106e-01 -4.37625647e-01 -7.00172782e-02
-2.67229438e-01 -7.19251871e-01 8.94317806e-01 -2.25191474e-01
6.92226350e-01 -2.19136024e+00 1.26825899e-01 9.92751718e-02
2.50525057e-01 1.80917293e-01 -4.75194067e-01 -2.44350228e-02
-4.18399811e-01 2.68115669e-01 -3.14120233e-01 -5.93549013e-01
8.11112449e-02 1.56978350e-02 -3.91840249e-01 3.16600770e-01
1.10400367e+00 1.22511327e+00 -6.61450744e-01 -4.87433553e-01
6.60277843e-01 7.41674662e-01 -6.81947589e-01 1.74320772e-01
-1.37960166e-01 4.89185452e-01 -7.64177322e-01 1.05408525e+00
1.11251044e+00 -5.65859318e-01 -7.89089084e-01 -3.75910342e-01
-3.17309022e-01 2.70737968e-02 -7.12585807e-01 1.99698257e+00
-6.38994515e-01 3.50711793e-01 -1.13319121e-02 -7.94109643e-01
9.29733932e-01 -1.90593258e-01 6.96682453e-01 -4.55322862e-01
2.39987716e-01 -2.65680309e-02 -1.81502014e-01 -3.70105773e-01
3.33687961e-01 4.51326258e-02 -6.65482879e-02 -2.42355227e-01
1.29671678e-01 -4.15102541e-01 -1.72772214e-01 4.43797112e-02
8.71600628e-01 -4.22068834e-02 8.84274393e-02 -2.71350890e-01
8.77320409e-01 -2.69035012e-01 4.61642057e-01 6.84096575e-01
-3.34059626e-01 9.20469284e-01 2.77122021e-01 -6.49679899e-01
-9.89611268e-01 -1.03884220e+00 -3.61809790e-01 1.17569005e+00
4.19685751e-01 -1.53845951e-01 -4.78946894e-01 -6.32771492e-01
4.83929843e-01 2.85267055e-01 -6.70111179e-01 4.59504388e-02
-5.84580362e-01 -3.72945607e-01 2.82613158e-01 9.52060580e-01
7.72728562e-01 -1.26486409e+00 -3.25129330e-01 2.54599094e-01
3.11611563e-01 -1.40604270e+00 -3.82655889e-01 1.82071999e-02
-1.00094926e+00 -7.72602320e-01 -8.29168320e-01 -1.09939075e+00
5.51951945e-01 4.61599797e-01 1.35710549e+00 -1.30797625e-02
-3.69819880e-01 3.64530802e-01 -2.37624243e-01 -6.68198705e-01
4.59430695e-01 2.98151523e-01 -1.51043028e-01 8.99859816e-02
6.34152114e-01 -7.90646732e-01 -6.18396223e-01 -3.83628421e-02
-8.31739128e-01 -1.84199572e-01 9.44100082e-01 6.82312250e-01
1.18695164e+00 -2.33331159e-01 4.60080147e-01 -5.21928251e-01
3.27744901e-01 -3.02192450e-01 -7.25692689e-01 1.56054586e-01
1.35456929e-02 -1.38518929e-01 4.97566909e-01 -9.74613801e-03
-7.05663383e-01 2.37607956e-01 -4.15153384e-01 -1.15712202e+00
-4.82350111e-01 2.34444916e-01 -3.06724459e-01 -5.89647233e-01
4.03320193e-02 2.53431469e-01 -5.09047687e-01 -4.50464129e-01
4.38256145e-01 5.73139966e-01 5.65947056e-01 -6.14684582e-01
9.54000950e-01 6.16573393e-01 -4.63132709e-02 -1.04371572e+00
-8.32023561e-01 -5.66423297e-01 -1.04066801e+00 3.07560023e-02
1.38523483e+00 -1.17520308e+00 -8.24054956e-01 5.07342935e-01
-1.54821622e+00 3.47396806e-02 -3.61519486e-01 2.42622495e-01
-6.24932051e-01 1.80785924e-01 -6.06303513e-01 -8.03003907e-01
-5.83873928e-01 -1.23080599e+00 1.88146102e+00 4.16563094e-01
5.22363782e-01 -6.82995796e-01 -2.92358667e-01 2.94754088e-01
2.07251534e-01 1.91622809e-01 9.29083109e-01 -2.90060997e-01
-1.12309086e+00 -7.95051977e-02 -9.56907690e-01 2.83135623e-01
-6.57729059e-02 9.37115997e-02 -7.75897026e-01 -2.70996302e-01
-2.21804351e-01 -1.69687957e-01 9.91950333e-01 5.91122389e-01
1.95403028e+00 3.29802155e-01 -3.22845221e-01 1.08743465e+00
1.40872014e+00 -2.08380640e-01 5.68328500e-01 1.78972229e-01
1.14681554e+00 2.26423174e-01 2.27346361e-01 1.83176488e-01
5.95867932e-01 5.07601082e-01 4.99310642e-01 -4.35354471e-01
-4.05839346e-02 -2.19274580e-01 -1.98476106e-01 8.15976262e-01
-1.66392103e-01 6.01555370e-02 -9.72593129e-01 5.53112686e-01
-1.62610984e+00 -7.17670083e-01 -5.17901108e-02 1.61609066e+00
1.56744018e-01 2.54427344e-01 -1.75547361e-01 -3.16484779e-01
6.87307596e-01 2.89769262e-01 -9.44356203e-01 -2.91857541e-01
-1.84490204e-01 4.36916858e-01 4.70375657e-01 1.56633958e-01
-1.52147830e+00 1.32927835e+00 5.55831623e+00 9.47031736e-01
-1.23321450e+00 -3.81631963e-02 6.08971059e-01 2.39105120e-01
-1.83900427e-02 -4.51544136e-01 -5.91605663e-01 3.33099961e-01
2.50369310e-01 1.25194594e-01 3.61956656e-02 1.01750708e+00
-1.06331162e-01 5.13456523e-01 -8.13623309e-01 1.29957235e+00
1.97790172e-02 -1.35058045e+00 2.85214603e-01 -6.67085424e-02
8.22120070e-01 7.63090432e-01 1.36491269e-01 4.88438964e-01
6.09725267e-02 -7.54925072e-01 5.18827975e-01 3.54669243e-01
6.67251945e-01 -1.11649871e+00 6.44211709e-01 1.71119422e-01
-1.74845219e+00 -1.13536425e-01 -7.21636176e-01 3.19226891e-01
5.37234806e-02 3.77971947e-01 -1.05859108e-01 8.03238153e-01
1.10037410e+00 1.14219534e+00 -5.78040898e-01 1.09317195e+00
-4.12191153e-02 1.80482417e-01 -3.99960577e-01 2.54919946e-01
6.91749632e-01 -2.22287312e-01 6.63412094e-01 1.31178451e+00
2.30997652e-01 4.99921948e-01 1.54116556e-01 1.11010361e+00
-3.51798177e-01 7.78278857e-02 -5.71642637e-01 1.62148401e-01
3.08895200e-01 1.24295938e+00 -7.37888992e-01 -5.10762632e-01
-6.85281217e-01 1.01054335e+00 6.35857761e-01 3.45358789e-01
-8.01841557e-01 -4.40782905e-01 1.17667568e+00 -9.22334790e-02
8.51590991e-01 -4.72989678e-01 -5.88271439e-01 -1.19420767e+00
-3.76755223e-02 -3.74207854e-01 2.42667630e-01 -8.94365430e-01
-1.89770043e+00 8.14343810e-01 -1.93525568e-01 -1.41754687e+00
5.56345582e-01 -7.50606179e-01 -7.50669360e-01 8.91260147e-01
-2.03024411e+00 -1.75133026e+00 -3.25633049e-01 6.86806083e-01
8.77833188e-01 -1.40706554e-01 3.45991701e-01 3.98567736e-01
-4.44109082e-01 5.49875379e-01 -2.26556674e-01 2.15553701e-01
2.32195213e-01 -1.15549207e+00 8.50381851e-01 6.87834799e-01
2.69512802e-01 6.80130661e-01 -3.60677987e-02 -5.58690786e-01
-1.38023090e+00 -1.52550805e+00 5.96963286e-01 -4.40854549e-01
5.29882133e-01 -4.25271749e-01 -1.24220395e+00 5.04700005e-01
7.12866858e-02 6.42812848e-01 3.16823006e-01 3.70852910e-02
-4.83347476e-01 -7.51510262e-02 -9.78901446e-01 2.99808741e-01
1.37598634e+00 -6.22932017e-01 -6.93898320e-01 4.32462841e-01
1.10295177e+00 -5.45223355e-01 -8.90700936e-01 9.07237232e-01
1.85536265e-01 -7.53368020e-01 1.44843805e+00 -7.89661288e-01
6.70358598e-01 -3.81014735e-01 -1.58103570e-01 -9.42868233e-01
-9.17650580e-01 -2.36451030e-01 1.35732427e-01 1.22390676e+00
1.71038300e-01 -3.59880716e-01 8.86502922e-01 3.49918127e-01
-6.00325048e-01 -9.53421950e-01 -9.48732674e-01 -6.28894985e-01
4.24206167e-01 -6.22323453e-01 9.42881167e-01 9.24787164e-01
-4.95339751e-01 3.45434874e-01 7.36381859e-02 7.73766458e-01
8.02638292e-01 7.50041008e-01 6.90050006e-01 -1.23905301e+00
1.13756351e-01 -8.10388744e-01 -8.05654347e-01 -1.48213911e+00
4.00943309e-01 -9.85126734e-01 -9.07852501e-02 -1.46340811e+00
4.85491753e-02 -3.09246868e-01 -6.71465218e-01 1.97498485e-01
-3.77628207e-01 4.08719063e-01 5.30693829e-01 2.57403910e-01
-8.13918471e-01 9.43373024e-01 1.41578662e+00 -2.52203375e-01
-3.45785260e-01 -9.65271220e-02 -5.00277996e-01 9.47380006e-01
4.78186458e-01 -2.80644834e-01 7.62457922e-02 -8.88841093e-01
-2.81770647e-01 -2.39209652e-01 3.83494437e-01 -1.25278485e+00
3.72709930e-01 -2.49589072e-03 6.39342606e-01 -1.34928942e+00
3.06747198e-01 -9.96643424e-01 -5.47525942e-01 -7.08429050e-03
-2.53005445e-01 1.33295432e-01 4.76604253e-01 6.36768460e-01
-4.91979748e-01 3.23945999e-01 8.79276514e-01 -2.26593465e-01
-7.86281705e-01 1.03135240e+00 1.57070845e-01 -2.26008326e-01
1.00945032e+00 -1.08151384e-01 8.28056633e-02 2.26898983e-01
-6.33472502e-01 5.16991019e-01 3.72774452e-01 7.45160460e-01
8.50500643e-01 -1.48358893e+00 -7.88210630e-01 6.18014812e-01
3.55422974e-01 6.00679994e-01 6.85953021e-01 4.70629930e-01
-5.17205775e-01 4.19118404e-01 -2.59909898e-01 -1.07728982e+00
-1.04520655e+00 6.45507872e-01 1.77386984e-01 -2.12711453e-01
-6.22702122e-01 1.35523081e+00 5.87371588e-01 -6.27541423e-01
5.43611236e-02 -7.12345660e-01 -4.27987762e-02 -3.75743657e-01
4.01229352e-01 -1.16488375e-01 6.29069358e-02 -8.10839117e-01
-5.64422965e-01 1.46117353e+00 -1.65793762e-01 4.17839795e-01
1.64518607e+00 -1.21483225e-02 -2.25722656e-01 3.32559347e-01
1.42230940e+00 -2.86820769e-01 -1.47772038e+00 -4.54865932e-01
-1.43023834e-01 -6.82522178e-01 2.46235564e-01 -4.08446848e-01
-1.23664403e+00 1.22500169e+00 5.84885299e-01 1.62176475e-01
1.28579295e+00 3.36707890e-01 8.45387578e-01 2.51004308e-01
3.16163749e-01 -7.08640277e-01 -6.53535426e-02 7.53158629e-01
1.20118725e+00 -1.61951017e+00 -5.94682321e-02 -6.62730813e-01
-4.76154536e-01 8.24652195e-01 8.74744713e-01 -8.18753004e-01
7.95469999e-01 7.62119368e-02 -2.41275981e-01 -5.15071511e-01
-7.01840520e-01 -5.07827282e-01 6.05585277e-01 4.99136358e-01
2.76259273e-01 -2.37891488e-02 3.79971176e-01 5.52937210e-01
-2.12219134e-02 -7.61841834e-02 -2.59276509e-01 6.70355380e-01
-2.92565554e-01 -8.07154715e-01 -2.90048987e-01 3.03062886e-01
-2.85643518e-01 2.64468510e-02 -4.32096034e-01 7.94023693e-01
3.36040288e-01 5.03400207e-01 5.96784472e-01 -3.05268884e-01
6.90301716e-01 -1.77192777e-01 4.74217772e-01 -5.51415384e-01
-8.02852392e-01 1.34761363e-01 -7.53135622e-01 -7.64272809e-01
-4.61403817e-01 -5.97314537e-01 -1.35585999e+00 -9.47236642e-03
-2.02710539e-01 -1.44300193e-01 6.70092642e-01 6.19949996e-01
9.85627413e-01 8.10900152e-01 7.55968809e-01 -1.40895367e+00
-1.16393730e-01 -8.68212342e-01 -3.87965202e-01 5.31793594e-01
5.52953124e-01 -6.06335521e-01 1.79299358e-02 -2.12951109e-01] | [7.956623077392578, -3.5137083530426025] |
003ceffd-0890-410e-8f6b-fe156cd78902 | principled-multi-aspect-evaluation-measures | 2212.00492 | null | https://arxiv.org/abs/2212.00492v1 | https://arxiv.org/pdf/2212.00492v1.pdf | Principled Multi-Aspect Evaluation Measures of Rankings | Information Retrieval evaluation has traditionally focused on defining principled ways of assessing the relevance of a ranked list of documents with respect to a query. Several methods extend this type of evaluation beyond relevance, making it possible to evaluate different aspects of a document ranking (e.g., relevance, usefulness, or credibility) using a single measure (multi-aspect evaluation). However, these methods either are (i) tailor-made for specific aspects and do not extend to other types or numbers of aspects, or (ii) have theoretical anomalies, e.g. assign maximum score to a ranking where all documents are labelled with the lowest grade with respect to all aspects (e.g., not relevant, not credible, etc.). We present a theoretically principled multi-aspect evaluation method that can be used for any number, and any type, of aspects. A thorough empirical evaluation using up to 5 aspects and a total of 425 runs officially submitted to 10 TREC tracks shows that our method is more discriminative than the state-of-the-art and overcomes theoretical limitations of the state-of-the-art. | ['Christina Lioma', 'Jakob Grue Simonsen', 'Lucas Chaves Lima', 'Maria Maistro'] | 2022-12-01 | null | null | null | null | ['document-ranking'] | ['natural-language-processing'] | [-2.03792140e-01 8.26458633e-02 -2.86980540e-01 -4.56919700e-01
-1.19136882e+00 -9.72778857e-01 9.86427367e-01 7.45435834e-01
-6.78598166e-01 7.82753408e-01 3.96316856e-01 -4.09776896e-01
-7.76217759e-01 -6.39754176e-01 -2.73832858e-01 -5.25367200e-01
3.33089828e-02 9.58415508e-01 7.44598806e-01 -3.83300185e-01
7.62322724e-01 3.88356179e-01 -1.67910254e+00 2.87636966e-01
4.76386368e-01 8.02363276e-01 -3.23043883e-01 4.48716253e-01
-4.96554226e-02 2.33067587e-01 -9.56873357e-01 -5.56917787e-01
-2.32297257e-01 -3.01767915e-01 -1.00388992e+00 -2.94285744e-01
1.60476685e-01 -1.04350381e-01 2.10849434e-01 1.02701938e+00
4.05144602e-01 5.80570661e-02 1.09448147e+00 -9.37589705e-01
-5.06691337e-01 3.53388518e-01 -3.35739732e-01 1.71284914e-01
6.50957167e-01 -5.07210314e-01 1.37834740e+00 -7.41687357e-01
4.70801711e-01 1.16854703e+00 3.31172258e-01 2.09446490e-01
-9.17588532e-01 -2.00228870e-01 2.69493103e-01 5.91955818e-02
-1.09709251e+00 -5.58591709e-02 2.71157980e-01 -5.26645899e-01
6.44716740e-01 6.11638367e-01 5.57565331e-01 5.52417517e-01
2.41786718e-01 4.42257553e-01 1.33011734e+00 -6.17217600e-01
5.81044257e-01 4.43005323e-01 4.99273777e-01 -4.69393097e-03
7.36240923e-01 -2.50024833e-02 -2.92620510e-01 -5.86529434e-01
-5.12873605e-02 -4.25904036e-01 -2.47023746e-01 -2.87082762e-01
-1.04290187e+00 7.42891014e-01 3.06438990e-02 5.66452801e-01
-4.13908511e-01 -1.70165837e-01 4.07964081e-01 3.37573439e-01
5.56047261e-01 6.60569668e-01 -3.82517368e-01 -9.06591266e-02
-9.91009176e-01 4.63511318e-01 9.10294235e-01 5.35866857e-01
6.71314299e-01 -5.41331887e-01 -6.76124752e-01 6.21621549e-01
5.20289004e-01 4.25871819e-01 4.59953696e-01 -8.23676646e-01
1.49045303e-01 7.37880826e-01 2.84809798e-01 -8.10787439e-01
-4.09747660e-01 -5.75350225e-01 -3.46071512e-01 4.61202741e-01
1.26135960e-01 4.33461927e-02 -6.69883132e-01 1.33485854e+00
1.85044676e-01 -9.11124647e-01 -1.17645261e-03 9.06488180e-01
9.34516966e-01 5.89999616e-01 -2.90758386e-02 -3.45720977e-01
1.52184033e+00 -5.04179180e-01 -6.39761806e-01 1.48026079e-01
6.02277935e-01 -1.03853154e+00 1.07020366e+00 6.62330508e-01
-1.01593840e+00 -2.62509733e-01 -9.76848066e-01 2.59502739e-01
-5.88637829e-01 -1.68358624e-01 5.93542874e-01 9.62464988e-01
-1.35204208e+00 4.74796474e-01 -1.52613670e-01 -1.81843936e-01
-3.90690029e-01 1.94557890e-01 -3.25898409e-01 5.98422587e-02
-1.52381921e+00 1.08345234e+00 5.94567358e-01 -2.77354270e-01
-6.52634442e-01 -4.33040783e-02 -2.45617449e-01 2.60619730e-01
4.88170385e-01 -5.49846768e-01 1.26315701e+00 -6.08390212e-01
-1.08723342e+00 6.68161929e-01 -7.56307468e-02 -2.66856458e-02
3.25489551e-01 -3.08069319e-01 -5.80580592e-01 4.09311771e-01
2.99729258e-01 1.50778681e-01 2.66745120e-01 -1.44537354e+00
-8.05655777e-01 -2.64526039e-01 6.65129483e-01 4.73677874e-01
-5.17886460e-01 5.23543298e-01 -6.36588871e-01 -2.90563226e-01
1.42132640e-01 -8.89119267e-01 -5.11493273e-02 -2.63754874e-01
-2.43482500e-01 -4.72855896e-01 5.71218073e-01 -4.08408076e-01
1.63088906e+00 -1.62741590e+00 -1.42345369e-01 4.97372091e-01
-6.49802238e-02 4.22768354e-01 9.45229754e-02 8.02630544e-01
2.29772151e-01 5.38710296e-01 4.07439768e-02 1.54759223e-02
1.81818753e-01 2.92841252e-02 -7.15885386e-02 2.66064972e-01
-2.79857397e-01 6.62356853e-01 -1.09530044e+00 -6.91518009e-01
-1.54411495e-01 5.45373559e-01 -2.74457008e-01 -2.54924536e-01
-1.09566309e-01 2.70875469e-02 -8.20744455e-01 5.54026663e-01
2.59899348e-01 -2.42460355e-01 9.18632150e-02 -6.56583952e-03
-2.12533772e-01 7.00710714e-01 -1.15826046e+00 1.00252569e+00
-3.23972732e-01 5.49183726e-01 -4.59479958e-01 -6.45998776e-01
8.32377315e-01 6.83473468e-01 3.19758266e-01 -5.02368271e-01
-1.21116355e-01 5.44336021e-01 -7.56635219e-02 -2.19980225e-01
9.88173306e-01 -1.04730375e-01 -1.23932831e-01 8.66055429e-01
-8.40736702e-02 -1.59858167e-01 6.06154680e-01 3.65916431e-01
9.20163214e-01 2.79680211e-02 5.75916588e-01 -4.65862304e-01
7.39726901e-01 -1.05860569e-01 1.87757105e-01 8.42586040e-01
-4.29717936e-02 6.18665457e-01 7.92857587e-01 -6.91163242e-02
-9.69119132e-01 -8.50761831e-01 -2.69573033e-01 1.17543101e+00
1.45990163e-01 -7.11115122e-01 -6.03510201e-01 -8.04624319e-01
-2.35078782e-01 9.09501433e-01 -5.85671246e-01 -3.52455191e-02
1.10882120e-02 -7.22365260e-01 4.08504069e-01 8.61189812e-02
1.46748722e-01 -8.84427190e-01 -6.12875938e-01 7.40962178e-02
-3.51174533e-01 -6.41437173e-01 -2.31853962e-01 7.24470243e-02
-8.50016475e-01 -1.06678891e+00 -9.03105736e-01 -2.76011169e-01
5.17777145e-01 2.93606848e-01 1.33200586e+00 1.22179195e-01
4.89725143e-01 6.31285787e-01 -7.11242437e-01 -2.66663522e-01
-3.89134496e-01 -7.64372339e-03 -5.73305562e-02 -2.13836223e-01
4.51295137e-01 -8.86894614e-02 -6.65106475e-01 6.75172329e-01
-1.16797817e+00 -7.65982151e-01 8.19800317e-01 5.12031019e-01
6.36004031e-01 3.63767534e-01 4.95682061e-01 -9.55560148e-01
1.17787218e+00 -2.31769159e-01 -4.14306641e-01 7.23795593e-01
-1.24752009e+00 1.93927377e-01 1.66125253e-01 -2.38904029e-01
-9.41808820e-01 -5.73398888e-01 -6.09601550e-02 4.09099609e-01
1.54654952e-02 1.02806115e+00 -4.08168845e-02 9.83904228e-02
8.12021613e-01 -1.80524904e-02 -5.21035373e-01 -3.18487406e-01
2.08423227e-01 7.19477534e-01 1.13375492e-01 -5.98098218e-01
6.42665982e-01 2.07988471e-01 3.16414148e-01 -5.72266400e-01
-1.04375660e+00 -9.51725721e-01 -3.89877409e-01 -5.29233456e-01
4.22176331e-01 -5.56372821e-01 -3.66131902e-01 7.43527263e-02
-1.03215873e+00 4.33935940e-01 -2.12158680e-01 7.22236693e-01
-2.62018442e-01 6.96497619e-01 -2.54967660e-01 -9.82798100e-01
-3.97666514e-01 -1.04989934e+00 1.03901601e+00 1.40610069e-01
-4.14684802e-01 -9.80035245e-01 2.57931799e-01 2.56312639e-01
6.64880514e-01 9.74379778e-02 8.59521806e-01 -9.12219048e-01
-1.46847665e-01 -8.75860035e-01 -3.84448878e-02 3.08878928e-01
8.93688947e-02 2.53780186e-01 -8.56838346e-01 -3.05978239e-01
1.36947066e-01 -1.78756744e-01 8.41861904e-01 2.71735877e-01
5.08214176e-01 -5.56073189e-01 -3.21414590e-01 -3.64846766e-01
1.40142000e+00 2.60381579e-01 9.12036777e-01 8.99479151e-01
-1.57939076e-01 9.16071832e-01 1.04382157e+00 1.63237035e-01
1.12577163e-01 1.00729907e+00 3.88829529e-01 1.77172095e-01
2.34200135e-01 1.22913621e-01 1.52105778e-01 6.20515525e-01
-3.88039440e-01 -5.10214686e-01 -7.55979180e-01 5.31756341e-01
-1.61809123e+00 -1.02106333e+00 -2.42242664e-01 2.81308746e+00
7.51540780e-01 5.92956185e-01 1.44507587e-01 4.24371600e-01
8.75790417e-01 5.88615052e-02 2.62907683e-03 -5.09236455e-01
-1.90171450e-01 -1.06944881e-01 9.54120904e-02 5.84281445e-01
-9.39115465e-01 4.51319277e-01 6.38409042e+00 9.22790706e-01
-6.08608127e-01 7.13987425e-02 5.12148380e-01 1.58541605e-01
-8.56648207e-01 4.66737688e-01 -8.76229465e-01 1.97458476e-01
8.44073892e-01 -6.12374723e-01 -2.43357643e-01 7.90045679e-01
-1.67503953e-01 -4.53061968e-01 -8.96879494e-01 2.87941933e-01
4.06846181e-02 -6.96566641e-01 3.25800478e-01 2.74789661e-01
6.69822216e-01 -3.70888352e-01 -7.72192935e-03 3.26857328e-01
6.50048107e-02 -7.69304693e-01 7.35556841e-01 5.28425515e-01
6.34229958e-01 -7.30660141e-01 1.25099337e+00 3.35469931e-01
-7.41684616e-01 3.45536500e-01 -4.66461599e-01 2.05353871e-01
2.17735872e-01 9.88734841e-01 -3.30237418e-01 9.29591954e-01
7.51000464e-01 2.92406790e-02 -7.56565034e-01 1.41855800e+00
-5.08302033e-01 5.25218010e-01 -2.59217829e-01 -5.27036905e-01
3.42118889e-01 7.46309683e-02 7.16251314e-01 1.18445075e+00
4.51266408e-01 -7.37971067e-02 -1.45738363e-01 1.94437072e-01
3.32476616e-01 3.85628164e-01 -3.68706465e-01 1.00977123e-01
4.57939088e-01 1.10975492e+00 -8.24640810e-01 -4.72440690e-01
-1.99723080e-01 3.35807711e-01 -1.41638160e-01 2.87009537e-01
-1.42893597e-01 -5.48266232e-01 -8.41323510e-02 2.33328804e-01
4.97891940e-02 9.02090520e-02 -2.27265712e-02 -6.90843225e-01
4.01305079e-01 -6.30133271e-01 6.07448876e-01 -6.56623363e-01
-1.07278669e+00 1.04663587e+00 4.25152719e-01 -1.58715701e+00
-5.73440015e-01 -5.18894076e-01 -3.52808923e-01 8.95884752e-01
-1.67152870e+00 -5.42662382e-01 -8.53771716e-02 2.32499942e-01
-1.34714127e-01 -1.36081830e-01 8.95563781e-01 3.59067991e-02
3.92758012e-01 2.88408786e-01 2.46000811e-01 -3.55742544e-01
9.10512924e-01 -1.37171137e+00 -1.29087538e-01 5.94702363e-01
8.78563598e-02 8.16121519e-01 9.37126160e-01 -4.47943240e-01
-6.03347003e-01 -5.23795485e-01 1.60929251e+00 -4.74984199e-01
5.21423280e-01 2.49841213e-01 -8.30279768e-01 1.28032520e-01
1.29232958e-01 -5.85141361e-01 8.53001595e-01 5.71105242e-01
-4.76965338e-01 -1.26920640e-01 -1.03366828e+00 5.28900504e-01
3.46270949e-01 -4.34963733e-01 -9.08506632e-01 3.47535104e-01
4.65445518e-01 1.00550381e-02 -8.43202293e-01 6.65249527e-01
8.38509023e-01 -1.38952923e+00 8.38242352e-01 -5.08349657e-01
3.74296248e-01 -5.19421995e-01 -2.78501570e-01 -1.20446336e+00
-3.82799059e-01 -1.94999799e-01 4.32445901e-03 1.35470319e+00
7.78050959e-01 -7.13027179e-01 3.58959198e-01 6.42578125e-01
-5.50260618e-02 -8.04877102e-01 -1.00508308e+00 -9.86035883e-01
1.50631204e-01 -2.56114960e-01 5.67482829e-01 6.63159132e-01
2.87433714e-01 4.20585036e-01 -1.14199579e-01 -1.44462779e-01
3.73539746e-01 9.06210840e-02 3.22136790e-01 -1.66201210e+00
-3.74206424e-01 -7.09569633e-01 -4.90357548e-01 -6.50335848e-01
-2.78131515e-01 -5.73875487e-01 -1.57155432e-02 -2.04703879e+00
4.19345289e-01 -3.39862257e-01 -7.44953930e-01 3.90477479e-01
-2.87252486e-01 2.54576802e-01 -3.25643197e-02 6.42950296e-01
-8.54822457e-01 3.08970839e-01 1.07828474e+00 -1.11080565e-01
5.37559274e-04 4.17679876e-01 -1.07207990e+00 5.45284510e-01
7.31337428e-01 -6.24188900e-01 -3.58564705e-01 -1.28596574e-02
9.63095486e-01 7.76630566e-02 1.86422810e-01 -9.61656630e-01
2.67793030e-01 -1.50474548e-01 9.58713666e-02 -5.93188524e-01
2.23603725e-01 -7.16183066e-01 8.44361335e-02 3.73119950e-01
-6.99474573e-01 -2.08705408e-03 -2.00252444e-01 4.11369115e-01
-5.88582933e-01 -1.02161634e+00 2.38590941e-01 -1.13570891e-01
-4.95604873e-01 -1.44861415e-01 -3.06181073e-01 -7.19472207e-03
6.17409170e-01 -1.36548311e-01 -6.45514190e-01 -7.35560119e-01
-4.12980974e-01 1.66054219e-01 4.46329206e-01 3.27946275e-01
4.12898660e-01 -1.27150047e+00 -8.51002812e-01 -7.79088438e-01
7.06288040e-01 -4.54238027e-01 -5.55586293e-02 6.99426651e-01
-2.04720765e-01 9.16830063e-01 1.71032727e-01 -3.59715164e-01
-1.29489565e+00 5.09285748e-01 -4.35582213e-02 -8.46553385e-01
-2.14724571e-01 2.85042286e-01 2.40932658e-01 -3.01793575e-01
1.71309352e-01 1.64959103e-01 -8.94327164e-01 3.10335517e-01
6.55106962e-01 3.64799291e-01 4.78575677e-01 -8.46068799e-01
-4.99801397e-01 8.74671578e-01 -1.65931970e-01 -6.46139801e-01
9.62301075e-01 -6.41336143e-02 -2.80129641e-01 5.68414092e-01
8.55856895e-01 1.98508486e-01 -4.57903922e-01 -2.66119778e-01
3.18996131e-01 -3.78849596e-01 1.76541671e-01 -1.18182242e+00
-6.25069022e-01 6.28048778e-01 2.28231817e-01 6.56684339e-01
1.13413537e+00 1.12491623e-01 -6.56783395e-03 7.01335192e-01
5.13390124e-01 -1.26740801e+00 1.29102290e-01 4.31647927e-01
1.03022540e+00 -9.64586914e-01 4.11756575e-01 -1.45476431e-01
-7.23487139e-01 1.17486334e+00 1.69989783e-02 2.42381647e-01
7.40706205e-01 -2.61462510e-01 -1.30292894e-02 -4.98463362e-01
-8.34065318e-01 -2.20685869e-01 9.24360991e-01 4.32853818e-01
8.85434687e-01 -6.21730834e-02 -1.32007158e+00 2.28848100e-01
-1.35104850e-01 -5.80498390e-02 5.69398940e-01 9.82103527e-01
-7.54700899e-01 -1.13099682e+00 -5.53938448e-01 5.27536750e-01
-8.74708474e-01 -8.75553265e-02 -6.36852086e-01 8.18912208e-01
-3.68870944e-01 1.39932883e+00 -5.77722728e-01 -1.95170864e-01
4.96957630e-01 -6.49023354e-02 8.25468674e-02 -5.31266987e-01
-6.69101179e-01 2.69008577e-01 5.02482414e-01 -7.56232366e-02
-8.08595836e-01 -6.75880253e-01 -6.56808794e-01 8.77981167e-03
-7.00926661e-01 9.08520520e-01 9.00526941e-01 9.91219759e-01
3.88720594e-02 2.31218621e-01 5.57237864e-01 -5.00354290e-01
-7.49680877e-01 -1.08277977e+00 -7.90058672e-01 1.78670466e-01
9.29292738e-02 -6.81349993e-01 -6.97045743e-01 -4.54547703e-01] | [10.322091102600098, 7.694666862487793] |
b5c9a90c-27d1-4865-bad0-1ba22df91ba3 | learning-to-kindle-the-starlight | 2211.09206 | null | https://arxiv.org/abs/2211.09206v1 | https://arxiv.org/pdf/2211.09206v1.pdf | Learning to Kindle the Starlight | Capturing highly appreciated star field images is extremely challenging due to light pollution, the requirements of specialized hardware, and the high level of photographic skills needed. Deep learning-based techniques have achieved remarkable results in low-light image enhancement (LLIE) but have not been widely applied to star field image enhancement due to the lack of training data. To address this problem, we construct the first Star Field Image Enhancement Benchmark (SFIEB) that contains 355 real-shot and 854 semi-synthetic star field images, all having the corresponding reference images. Using the presented dataset, we propose the first star field image enhancement approach, namely StarDiffusion, based on conditional denoising diffusion probabilistic models (DDPM). We introduce dynamic stochastic corruptions to the inputs of conditional DDPM to improve the performance and generalization of the network on our small-scale dataset. Experiments show promising results of our method, which outperforms state-of-the-art low-light image enhancement algorithms. The dataset and codes will be open-sourced. | ['Han Pan', 'Shuyuan Zhu', 'Henry Leung', 'Zhongliang Jing', 'Lindong Wang', 'Jiaqi Wu', 'Yu Yuan'] | 2022-11-16 | null | null | null | null | ['low-light-image-enhancement'] | ['computer-vision'] | [ 4.04853642e-01 -3.95131171e-01 4.07651424e-01 -2.34772101e-01
-5.25634706e-01 -2.42130771e-01 5.60088873e-01 -2.76987106e-01
-6.07882500e-01 6.64233685e-01 -1.11544706e-01 -6.68697804e-02
-2.19058007e-01 -8.38168681e-01 -7.49343514e-01 -8.21099997e-01
2.99968213e-01 -3.18807550e-02 7.15031266e-01 -4.27182406e-01
1.43130153e-01 4.54508543e-01 -1.62412083e+00 2.56934650e-02
1.13440979e+00 1.06101406e+00 6.66149199e-01 5.47060013e-01
1.13757625e-01 1.08587694e+00 -4.10510659e-01 -5.31752050e-01
4.21491504e-01 -2.44750798e-01 -4.83205110e-01 4.11814749e-01
7.35173464e-01 -5.45943201e-01 -6.40801907e-01 1.46523154e+00
6.84522450e-01 5.05697131e-01 3.44390661e-01 -1.03337467e+00
-8.81718934e-01 6.46383315e-02 -5.93632340e-01 5.08673787e-01
-3.68103355e-01 4.25734371e-01 4.82072890e-01 -8.49579692e-01
6.03228450e-01 9.11024988e-01 6.77864254e-01 4.88404363e-01
-1.02207983e+00 -4.52070087e-01 -3.40524375e-01 4.49864179e-01
-9.16036189e-01 -3.58631313e-01 9.18118536e-01 -2.44876787e-01
6.64427221e-01 -1.61228016e-01 6.02664649e-01 9.67241824e-01
1.66802630e-01 5.52517235e-01 1.70805931e+00 -3.26822579e-01
3.44053626e-01 1.06848635e-01 -1.95394047e-02 6.92839682e-01
2.48577133e-01 6.85193121e-01 -5.38072586e-01 2.14070857e-01
8.17580044e-01 7.65655190e-02 -3.13085467e-01 -1.44779816e-01
-9.01225507e-01 7.36852765e-01 5.70656955e-01 2.32968107e-01
-5.33074915e-01 2.01405920e-02 -1.91378638e-01 1.13453776e-01
5.60636640e-01 1.26560152e-01 -2.14184910e-01 1.27249286e-01
-9.99882519e-01 1.87356263e-01 3.50922555e-01 5.70665359e-01
8.24266315e-01 2.88383216e-01 -2.75005788e-01 8.58949661e-01
1.09112747e-01 8.55499029e-01 3.04495934e-02 -1.12919116e+00
1.37774706e-01 2.87483543e-01 2.28948370e-01 -8.25326622e-01
-1.04828440e-01 -6.24928474e-01 -9.67737317e-01 7.92572439e-01
4.87826437e-01 5.15305884e-02 -9.14769053e-01 1.47783422e+00
1.78104535e-01 2.23846018e-01 3.35648566e-01 1.26515043e+00
1.01487637e+00 8.33977461e-01 1.13555208e-01 -1.44596875e-01
1.20344090e+00 -1.08771682e+00 -7.13561773e-01 -3.08386058e-01
2.80307047e-02 -7.45650589e-01 9.87808466e-01 6.02402687e-01
-1.29838824e+00 -5.94013155e-01 -9.66183007e-01 -1.68431371e-01
-3.63326520e-01 1.85176626e-01 5.97777784e-01 5.84118426e-01
-1.22430432e+00 6.34858370e-01 -5.31887174e-01 -1.89748809e-01
4.99171227e-01 -4.27685790e-02 -1.95246860e-01 -3.29384625e-01
-1.21787453e+00 8.84348989e-01 2.25445047e-01 4.87985760e-02
-1.24181390e+00 -7.70001709e-01 -6.11261427e-01 8.11595470e-02
2.84291208e-01 -5.52232504e-01 9.77017522e-01 -9.30198789e-01
-1.54244232e+00 7.59999812e-01 2.72035033e-01 -4.77580339e-01
2.15892509e-01 6.37996867e-02 -4.41229045e-01 4.13965404e-01
-3.31751883e-01 7.24577308e-01 1.26339030e+00 -1.32086933e+00
-4.94297624e-01 -2.81745732e-01 -3.27089205e-02 1.23177379e-01
-4.13161725e-01 3.44422132e-01 -2.88983762e-01 -6.43302977e-01
-4.07506198e-01 -7.59684384e-01 -2.57038087e-01 2.12738961e-01
1.24956354e-01 -5.20937778e-02 9.81637239e-01 -8.94171476e-01
7.98578739e-01 -2.28565454e+00 -1.70229256e-01 -2.75668144e-01
2.36459225e-01 6.89741790e-01 -1.94725916e-01 -1.08959831e-01
-6.60977364e-02 -5.37191749e-01 -6.40374660e-01 -3.34324688e-01
-2.83889413e-01 2.17164263e-01 -1.59950987e-01 5.80552101e-01
1.24952786e-01 8.21474850e-01 -9.14036632e-01 -5.59631705e-01
6.26959682e-01 8.01525593e-01 -2.58616537e-01 2.33107805e-01
-2.16623023e-01 2.72482365e-01 -5.07910829e-03 5.70560277e-01
1.08666277e+00 -2.60231197e-01 -3.49583864e-01 -4.19426918e-01
-2.74262697e-01 -3.69967759e-01 -9.76883173e-01 1.64436138e+00
-5.61506093e-01 7.20201135e-01 1.16848119e-01 -6.95289552e-01
1.12620008e+00 -1.97643369e-01 3.87047946e-01 -8.34053397e-01
2.31290162e-01 5.69456816e-02 -2.71910489e-01 -5.04257202e-01
6.24451041e-01 -3.44953597e-01 5.16486883e-01 2.81851590e-01
2.86105126e-01 -5.02124190e-01 3.51302117e-01 3.00387263e-01
9.16792214e-01 -6.15455094e-04 -2.95448571e-01 -2.26770625e-01
5.15080392e-01 -2.32334197e-01 4.43096548e-01 6.11733377e-01
-4.21078712e-01 7.28098392e-01 -1.47557795e-01 -2.82503515e-01
-1.22678220e+00 -1.25840926e+00 -7.14589059e-02 8.12339723e-01
3.91072065e-01 -4.14977893e-02 -7.85466909e-01 -5.28772593e-01
-4.27714020e-01 7.11407483e-01 -3.89290124e-01 -1.78715870e-01
-2.21597701e-01 -1.16844249e+00 2.21731842e-01 3.02281231e-01
1.09322810e+00 -1.13366950e+00 -6.89825356e-01 7.36037120e-02
-2.90930361e-01 -1.47728908e+00 -2.18322352e-01 1.61010012e-01
-7.54051566e-01 -8.81031573e-01 -1.06010520e+00 -6.68243527e-01
4.11309421e-01 5.13847768e-01 1.12070882e+00 -1.83801249e-01
-3.48492891e-01 3.05474609e-01 -3.06960940e-01 -3.84325117e-01
-3.87204111e-01 -4.71526057e-01 -1.29950240e-01 2.42904961e-01
2.82293916e-01 -5.07443726e-01 -8.16755950e-01 2.27090806e-01
-1.36509991e+00 -1.34134457e-01 8.90229583e-01 9.68607903e-01
5.30746877e-01 4.14000064e-01 1.05083361e-02 -2.95653313e-01
4.52116966e-01 -2.04677489e-02 -1.01495039e+00 3.42963904e-01
-8.59424829e-01 1.56518951e-01 4.01206762e-01 -3.98618013e-01
-1.56136334e+00 4.03668657e-02 -2.71877229e-01 -5.13553619e-01
-1.76094785e-01 -1.13869064e-01 -3.97804193e-02 -7.13904977e-01
7.43896902e-01 3.61519635e-01 -4.76760119e-02 -5.11291921e-01
2.50401467e-01 6.38533533e-01 1.04366672e+00 -1.41219527e-01
9.71376657e-01 8.86410892e-01 1.83680654e-01 -9.66273606e-01
-9.39681292e-01 -4.50889707e-01 -3.03973585e-01 -5.69862664e-01
1.04182398e+00 -9.21588719e-01 -4.54611897e-01 8.66207540e-01
-1.18853402e+00 -4.83915806e-01 -3.98310870e-01 4.56637055e-01
-5.60010850e-01 7.74853408e-01 -6.04910254e-01 -7.47405887e-01
-5.31410456e-01 -1.16476870e+00 8.85867059e-01 6.22725725e-01
7.42731869e-01 -9.47791994e-01 3.41025810e-03 4.64448214e-01
6.78533852e-01 8.10148716e-02 4.85883921e-01 2.85762429e-01
-7.98586786e-01 2.09282413e-01 -5.86985409e-01 1.03370202e+00
-2.71930903e-01 -2.06133530e-01 -1.02619672e+00 -2.64949441e-01
2.15105325e-01 -3.78889233e-01 1.21430540e+00 6.73634112e-01
1.14883089e+00 2.65424609e-01 1.81945249e-01 9.79650676e-01
1.62399554e+00 -1.03548057e-01 1.05622637e+00 5.02868533e-01
4.25056815e-01 6.03239536e-01 5.75680137e-01 4.62439716e-01
1.11532867e-01 5.81547379e-01 7.43774831e-01 -5.82693040e-01
-7.59225368e-01 -9.56546590e-02 5.92893183e-01 3.96257162e-01
-2.01529905e-01 -2.67194569e-01 -4.02167141e-01 6.66021347e-01
-1.37241471e+00 -1.12123883e+00 -5.62997878e-01 1.89343238e+00
7.14609146e-01 -1.04155354e-01 -1.73596293e-01 2.33558580e-01
5.27025282e-01 3.92989337e-01 -4.26053852e-01 -1.02187462e-01
-6.50062740e-01 3.10225546e-01 7.85892665e-01 4.77170259e-01
-1.15947235e+00 9.39181149e-01 5.88246346e+00 1.05475092e+00
-1.08963597e+00 5.23973227e-01 6.10467792e-01 1.47033155e-01
-8.47634971e-02 -7.04922378e-02 -6.87433660e-01 5.81694543e-01
7.99585164e-01 2.82125175e-01 7.09393620e-01 6.81406736e-01
3.43993336e-01 -2.92616725e-01 -1.64593428e-01 1.30771685e+00
4.00193453e-01 -1.19191062e+00 -2.75193155e-01 -1.43687561e-01
9.99866307e-01 3.30653936e-01 3.52068812e-01 -9.11994427e-02
4.10321236e-01 -6.61339819e-01 3.99951279e-01 7.36169934e-01
9.40010130e-01 -5.48486412e-01 6.63718104e-01 1.62175253e-01
-7.05660760e-01 -2.43438333e-01 -7.76876867e-01 2.31378347e-01
3.48212391e-01 9.90507841e-01 -1.24039747e-01 4.19908524e-01
1.29159009e+00 6.74115121e-01 -8.90377939e-01 1.11904323e+00
-3.59894902e-01 5.88193715e-01 -3.64695817e-01 2.54189581e-01
2.14941680e-01 -5.23776114e-01 5.06184161e-01 1.03376365e+00
3.76023293e-01 1.81429744e-01 -3.13021541e-01 1.10305488e+00
1.61057264e-02 -2.00965077e-01 -5.09055078e-01 3.47634740e-02
-3.90295267e-01 1.58744299e+00 -4.81163293e-01 -3.10957879e-01
-6.23502374e-01 1.07829154e+00 -8.67798105e-02 4.73778278e-01
-6.84203565e-01 -4.15704638e-01 3.47002566e-01 2.59547651e-01
3.57012689e-01 -2.41914958e-01 -1.81287438e-01 -1.20439279e+00
-1.72783628e-01 -8.20725620e-01 2.95749307e-01 -1.31239641e+00
-1.34067953e+00 7.81804800e-01 -2.15348881e-02 -1.07278705e+00
3.16131145e-01 -7.85817623e-01 -4.91215587e-01 7.72001386e-01
-2.04082179e+00 -1.20426154e+00 -8.73032868e-01 8.70994210e-01
4.00883377e-01 -2.98421234e-01 1.71820134e-01 5.00332057e-01
-1.69601336e-01 1.07733354e-01 4.36984450e-01 -1.07335493e-01
8.43302846e-01 -9.96103346e-01 1.67843774e-01 1.42355216e+00
8.69991779e-02 -1.17819518e-01 6.69007957e-01 -3.93874913e-01
-1.12715721e+00 -1.18268561e+00 7.63780653e-01 -3.56285453e-01
6.08007252e-01 -8.14495385e-02 -9.21560705e-01 2.22143978e-01
5.18019795e-01 3.58803719e-01 1.55364692e-01 -6.22753024e-01
-1.17676072e-02 -3.97948563e-01 -1.21471369e+00 3.21104497e-01
9.59075689e-01 -2.80893028e-01 -5.02899468e-01 3.44611704e-01
5.59049189e-01 -7.67001212e-02 -7.20787168e-01 6.48840427e-01
-1.16593540e-01 -1.31721675e+00 1.19225585e+00 -9.92220733e-03
6.26738787e-01 -4.66597050e-01 -7.19297454e-02 -1.39479578e+00
-3.57849598e-01 -5.38618684e-01 -1.04015902e-01 1.20920491e+00
7.16610439e-03 -3.81701380e-01 7.10796237e-01 2.89886653e-01
-1.18635498e-01 -4.14151847e-01 -5.62563002e-01 -7.48822272e-01
-1.01170793e-01 -2.65208244e-01 3.94375920e-01 6.76918089e-01
-7.58244216e-01 1.47193223e-01 -6.88823402e-01 1.68234006e-01
1.13854337e+00 2.05429807e-01 5.84751248e-01 -1.12443376e+00
-5.55072248e-01 -1.01984650e-01 -1.42623842e-01 -9.25903916e-01
1.17331482e-01 -8.24825883e-01 2.64286458e-01 -1.40719295e+00
3.22647452e-01 -2.37310618e-01 -2.19936818e-01 1.33541048e-01
-3.04851651e-01 6.49925947e-01 2.32651353e-01 -5.50562032e-02
-6.73195004e-01 9.21337783e-01 1.32564437e+00 -4.08794910e-01
9.38640013e-02 -1.22083910e-01 -4.33471531e-01 6.16980910e-01
6.92661285e-01 -3.95379186e-01 -2.51941293e-01 -5.45108795e-01
3.59914526e-02 -1.44354910e-01 7.27671802e-01 -1.34201741e+00
3.20132494e-01 -9.35706571e-02 4.42288756e-01 -5.46305537e-01
5.81986129e-01 -7.79125690e-01 -7.94950724e-02 4.07442868e-01
-5.23813777e-02 -1.02407530e-01 8.49613324e-02 5.34133255e-01
-3.51445049e-01 -2.95580119e-01 1.51368809e+00 -6.29962906e-02
-1.17006218e+00 5.34120619e-01 -2.77812034e-01 7.82707110e-02
9.15783048e-01 1.01791456e-01 -5.35018682e-01 -4.07474011e-01
-3.21765631e-01 -3.09729487e-01 6.02555573e-01 1.19272977e-01
9.43962336e-01 -8.90184402e-01 -8.68303180e-01 2.76503891e-01
1.28738776e-01 -1.79600477e-01 5.89198709e-01 8.45036626e-01
-6.58451021e-01 -6.02478087e-02 -4.97490436e-01 -3.97716165e-01
-1.23720980e+00 9.53899682e-01 5.20664871e-01 -2.27289677e-01
-8.27036262e-01 7.90335834e-01 2.38029242e-01 -2.22490877e-01
8.55077729e-02 -1.05463900e-01 -1.41553998e-01 -6.99478209e-01
6.78507209e-01 4.08283114e-01 7.50353560e-02 -5.46253324e-01
5.67815565e-02 6.03903115e-01 2.40331784e-01 -4.31506485e-02
1.46929991e+00 -2.49508888e-01 -9.64815840e-02 -2.54774362e-01
7.39055336e-01 -2.80613989e-01 -1.63289428e+00 -4.26640898e-01
-3.36139560e-01 -8.47897470e-01 7.63875484e-01 -9.40009534e-01
-1.46411955e+00 1.14644897e+00 1.21216619e+00 -3.71168166e-01
1.62577927e+00 -3.79294902e-02 1.10596645e+00 2.00042799e-01
2.00583607e-01 -1.12993717e+00 2.37364724e-01 3.73460144e-01
5.93711197e-01 -1.60320818e+00 -5.23774140e-02 -3.44148010e-01
-7.39404678e-01 8.57277393e-01 5.41631341e-01 -9.13614482e-02
6.40732169e-01 4.66823816e-01 1.69483945e-01 -1.14982791e-01
-5.79796731e-01 -5.69356561e-01 4.29908782e-02 9.52339590e-01
-9.49035212e-02 -2.67879725e-01 -2.72322148e-01 2.47793674e-01
2.60930926e-01 2.13190272e-01 5.54575086e-01 3.96728575e-01
-4.65879351e-01 -8.60247612e-01 -6.38757229e-01 1.06046572e-01
-4.39104050e-01 -3.91414762e-01 -1.87130466e-01 3.46655577e-01
2.97177315e-01 1.16162837e+00 -1.55223697e-01 -1.81745082e-01
1.44440547e-01 -4.09958392e-01 6.06383443e-01 1.54834494e-01
-4.37052161e-01 -1.96777895e-01 -3.19384336e-01 -3.88097018e-01
-6.61834419e-01 -4.25810426e-01 -6.42055988e-01 -3.83838922e-01
-1.11258693e-01 -2.56846976e-02 8.51415515e-01 6.76299334e-01
1.72166839e-01 4.18421745e-01 5.58683038e-01 -8.93146455e-01
-5.83822966e-01 -9.78113532e-01 -8.84082735e-01 7.17036009e-01
1.25620246e-01 -7.80391872e-01 -3.96082968e-01 3.09967995e-01] | [10.741560935974121, -2.482313394546509] |
056ab5a7-5242-4f41-8929-8dd4927c1b28 | locating-and-editing-factual-knowledge-in-gpt | 2202.05262 | null | https://arxiv.org/abs/2202.05262v5 | https://arxiv.org/pdf/2202.05262v5.pdf | Locating and Editing Factual Associations in GPT | We analyze the storage and recall of factual associations in autoregressive transformer language models, finding evidence that these associations correspond to localized, directly-editable computations. We first develop a causal intervention for identifying neuron activations that are decisive in a model's factual predictions. This reveals a distinct set of steps in middle-layer feed-forward modules that mediate factual predictions while processing subject tokens. To test our hypothesis that these computations correspond to factual association recall, we modify feed-forward weights to update specific factual associations using Rank-One Model Editing (ROME). We find that ROME is effective on a standard zero-shot relation extraction (zsRE) model-editing task, comparable to existing methods. To perform a more sensitive evaluation, we also evaluate ROME on a new dataset of counterfactual assertions, on which it simultaneously maintains both specificity and generalization, whereas other methods sacrifice one or another. Our results confirm an important role for mid-layer feed-forward modules in storing factual associations and suggest that direct manipulation of computational mechanisms may be a feasible approach for model editing. The code, dataset, visualizations, and an interactive demo notebook are available at https://rome.baulab.info/ | ['Yonatan Belinkov', 'Alex Andonian', 'David Bau', 'Kevin Meng'] | 2022-02-10 | null | null | null | null | ['model-editing'] | ['natural-language-processing'] | [ 3.12218964e-01 3.79709363e-01 -1.27086148e-01 -4.29400891e-01
-4.44065779e-01 -5.33177376e-01 1.24126732e+00 4.94880825e-01
-4.27354544e-01 7.63306379e-01 6.74348235e-01 -5.08157372e-01
-1.51525185e-01 -1.01250780e+00 -8.70593071e-01 -2.14118376e-01
-2.13119417e-01 2.68385202e-01 -1.16000414e-01 -3.64998847e-01
7.11824000e-01 4.38833237e-01 -1.31642807e+00 9.55899775e-01
4.91662979e-01 5.69825709e-01 1.73271839e-02 3.22707534e-01
-7.50987008e-02 1.11973059e+00 -5.37980974e-01 -8.33120167e-01
-9.54271927e-02 -3.80573034e-01 -8.03275406e-01 -9.97927070e-01
5.31945825e-01 -2.81886071e-01 -2.89511383e-01 1.00456190e+00
3.83639991e-01 3.40904444e-01 7.54189670e-01 -1.07633674e+00
-1.13257504e+00 1.25606954e+00 -8.36871937e-02 8.18757832e-01
3.97551417e-01 2.46434882e-01 1.09812236e+00 -1.26205933e+00
8.48757744e-01 1.33241475e+00 6.86419249e-01 4.63581592e-01
-1.60547483e+00 -9.57581401e-01 4.37013417e-01 7.55089596e-02
-1.24104989e+00 -8.45409155e-01 8.60538065e-01 -6.24856055e-01
1.71614051e+00 3.24576586e-01 8.84104371e-01 1.25018203e+00
7.57308185e-01 2.48966128e-01 1.14247477e+00 -2.64080346e-01
2.70468920e-01 7.72064328e-02 1.70892015e-01 7.63286591e-01
4.50361878e-01 6.86478198e-01 -1.06236899e+00 -3.84195149e-01
7.29762197e-01 -6.37906417e-02 -1.62389472e-01 -8.16514045e-02
-1.26766539e+00 6.49354160e-01 3.53032380e-01 5.06874025e-01
-5.91803312e-01 5.98627210e-01 3.80478382e-01 4.29085672e-01
5.84154367e-01 1.08798981e+00 -5.26318133e-01 1.46402374e-01
-1.02759683e+00 5.20203769e-01 4.52945203e-01 4.37884986e-01
4.11420137e-01 3.18951488e-01 -4.71090198e-01 5.85943282e-01
9.24340785e-02 1.75094143e-01 5.43728948e-01 -7.10423827e-01
3.45967382e-01 4.88938361e-01 6.34649321e-02 -1.40691936e+00
-4.31724787e-01 -4.83437687e-01 -6.27394140e-01 1.42409317e-02
6.06810972e-02 2.20286041e-01 -4.80991989e-01 1.98662388e+00
-1.11239664e-01 6.03406094e-02 -9.12613869e-02 7.10077763e-01
6.70421004e-01 3.83499324e-01 8.39135170e-01 -1.09871544e-01
1.34508824e+00 -1.01506025e-01 -6.66627288e-01 -3.61345321e-01
7.10100055e-01 -4.92885262e-01 1.17836559e+00 1.56648055e-01
-1.30892253e+00 -2.39470735e-01 -1.03894579e+00 -4.14242625e-01
-8.39242041e-01 -1.28719434e-02 1.25324845e+00 3.17268461e-01
-9.73357737e-01 7.78331876e-01 -5.91554046e-01 -1.47727147e-01
7.30058193e-01 1.58627018e-01 -2.98028409e-01 5.29885113e-01
-1.70753348e+00 1.38782239e+00 3.35524797e-01 4.76893261e-02
-1.09555626e+00 -1.23694599e+00 -9.50313687e-01 4.18305606e-01
2.74249315e-01 -9.18779612e-01 1.16587925e+00 -6.96737945e-01
-1.00874484e+00 1.05882192e+00 -3.20200950e-01 -6.77373648e-01
-2.97039654e-02 -1.17983380e-02 -5.57503879e-01 -2.32245684e-01
-3.80443409e-02 4.89512295e-01 7.19784975e-01 -1.17499840e+00
-1.27975553e-01 -3.60154241e-01 2.43642554e-01 6.01175167e-02
-1.16141573e-01 2.28789181e-01 4.64157462e-01 -1.06960130e+00
-8.46552774e-02 -2.40896598e-01 6.01872010e-03 -9.14406106e-02
-2.17079371e-01 -5.50694242e-02 4.54106741e-02 -6.78663075e-01
1.48522246e+00 -1.72187817e+00 -1.65428579e-01 1.57194853e-01
4.41296875e-01 -2.14723195e-03 -6.11895658e-02 3.65263671e-01
-7.59825408e-01 6.82155669e-01 -3.85825098e-01 -1.46420375e-02
2.02267412e-02 -2.44517833e-01 -1.00942516e+00 2.55590975e-01
4.19245392e-01 1.34849417e+00 -9.62866783e-01 -2.82274157e-01
4.60278988e-02 2.98039645e-01 -7.26827741e-01 -1.57581478e-01
-3.95590723e-01 -1.15647770e-01 -3.71630751e-02 2.66059667e-01
3.75574291e-01 3.77182476e-02 4.07053500e-01 -4.04857934e-01
-2.31891006e-01 1.18352067e+00 -7.54406214e-01 1.58888865e+00
-5.06736696e-01 6.51621163e-01 -4.16017115e-01 -9.29731667e-01
7.41020024e-01 3.47511500e-01 -4.52966057e-03 -8.19819748e-01
4.94445637e-02 -2.36077204e-01 1.24428496e-01 -1.80611491e-01
8.44608486e-01 -6.65094852e-01 -1.44568577e-01 8.02329183e-01
2.24702477e-01 -1.94990560e-01 -9.49723553e-03 6.63634121e-01
9.74237978e-01 1.62114173e-01 6.79484606e-01 -5.27659774e-01
-1.27030745e-01 -1.10069640e-01 4.20355618e-01 7.93106675e-01
1.09617911e-01 1.28003821e-01 8.16976964e-01 -3.21797103e-01
-9.43746567e-01 -1.29072750e+00 -5.26597258e-03 1.04713905e+00
-5.51113427e-01 -5.51342726e-01 -3.35971653e-01 -3.59130710e-01
2.63643377e-02 1.70326626e+00 -1.07281959e+00 -7.46312857e-01
-4.73213345e-01 -8.44142973e-01 8.21589589e-01 5.92197895e-01
1.86362922e-01 -1.20947087e+00 -8.87935042e-01 2.07367852e-01
-1.60056695e-01 -3.33346575e-01 -2.14093983e-01 1.74341246e-01
-9.88982081e-01 -9.19687450e-01 1.08490475e-02 -2.35668957e-01
6.84199095e-01 -1.07396767e-01 1.30912340e+00 1.77465960e-01
-1.42589226e-01 1.71027273e-01 4.32884097e-01 -4.70983624e-01
-2.74578512e-01 -1.97848961e-01 4.66829948e-02 -6.22651756e-01
4.87863481e-01 -8.03478479e-01 -2.64946371e-01 -2.49715000e-01
-8.64801347e-01 4.39854145e-01 3.06535691e-01 9.01586294e-01
4.93135422e-01 -2.05063060e-01 7.14678884e-01 -1.28325951e+00
1.08671641e+00 -6.96439087e-01 -5.80462337e-01 2.71872193e-01
-9.65492189e-01 2.95278907e-01 6.34471655e-01 -1.88178524e-01
-1.58686543e+00 -3.12519580e-01 4.07637730e-02 -1.89088002e-01
8.26956034e-02 8.16869259e-01 -2.58445852e-02 4.08506423e-01
1.03775299e+00 1.13175094e-01 -3.38903546e-01 -1.18226230e-01
6.66307688e-01 -1.16665825e-01 5.14880061e-01 -6.32337034e-01
6.79129660e-01 4.18839037e-01 -3.27223718e-01 -1.73087344e-01
-7.05291688e-01 2.18513399e-01 -5.36344588e-01 -4.00051884e-02
5.05681455e-01 -8.21482122e-01 -8.98493230e-01 2.03199498e-03
-1.43781984e+00 -5.77388346e-01 -4.33305681e-01 3.33882749e-01
-4.43490446e-01 -2.54423946e-01 -6.52519703e-01 -6.73366845e-01
-3.55769545e-01 -4.24876004e-01 5.62700748e-01 -1.31486595e-01
-8.03844571e-01 -1.06979465e+00 1.14968635e-01 -2.52740532e-01
3.98282677e-01 -3.48200239e-02 1.41625381e+00 -8.04041743e-01
-3.50058496e-01 8.68849009e-02 -1.63783342e-01 -3.17333072e-01
-2.17980593e-01 -8.62792658e-04 -1.08334601e+00 2.83938229e-01
2.35023573e-02 -7.58295506e-02 1.25602901e+00 1.59458056e-01
1.32895803e+00 -7.91512430e-01 -5.36724687e-01 3.56026888e-01
1.13092363e+00 7.52266422e-02 7.73639083e-01 -1.24060594e-01
3.41155589e-01 5.95706820e-01 4.02100593e-01 1.52641103e-01
4.08204198e-01 3.86456251e-01 2.94864899e-03 4.18951027e-02
-8.75335485e-02 -7.49765694e-01 4.10925567e-01 3.75724286e-01
-8.23894963e-02 7.05507472e-02 -9.20373440e-01 8.21865141e-01
-1.57020378e+00 -1.65836382e+00 1.86288744e-01 2.19913507e+00
1.34829044e+00 4.22711253e-01 -4.60822970e-01 -2.68661100e-02
2.78670371e-01 3.56655031e-01 -3.57182473e-01 -9.00548398e-01
-1.71692789e-01 4.00120497e-01 2.81425864e-01 7.99714029e-01
-7.05839336e-01 1.30952907e+00 6.93737745e+00 5.84420383e-01
-1.01551151e+00 3.43206167e-01 7.30154991e-01 -5.78634083e-01
-1.09177160e+00 2.19736129e-01 -4.11869526e-01 2.31349871e-01
1.23745906e+00 -5.63744903e-01 7.07994401e-01 5.26494205e-01
4.44354653e-01 -3.55816007e-01 -1.41202843e+00 4.88115281e-01
2.35473048e-02 -2.03806281e+00 4.32115048e-01 -2.68665284e-01
4.71210659e-01 -3.14402997e-01 8.17990080e-02 3.64926636e-01
6.68890357e-01 -1.25002086e+00 9.46663857e-01 1.10020053e+00
7.80762553e-01 -4.61816609e-01 1.54017016e-01 -4.42926325e-02
-8.22863102e-01 4.72725108e-02 -4.73008305e-01 -4.56843436e-01
1.03931405e-01 7.99507916e-01 -9.01187420e-01 1.06358536e-01
4.72871482e-01 5.82219005e-01 -6.68148041e-01 4.00284857e-01
-4.76362854e-01 6.89994097e-01 -1.07410066e-01 -5.85654452e-02
-5.18294156e-01 4.05007392e-01 4.05450672e-01 1.39404762e+00
-4.46439832e-02 5.35129845e-01 -7.28189230e-01 1.71277595e+00
-2.97564507e-01 -2.12665379e-01 -9.52394009e-01 -2.27734745e-01
8.03656340e-01 8.53555441e-01 -4.81455714e-01 -6.43953502e-01
5.84147824e-03 7.63085485e-01 7.44080186e-01 5.65636218e-01
-8.01078081e-01 -9.96699110e-02 6.73378110e-01 1.24848105e-01
-8.38412121e-02 -2.57365644e-01 -9.00379300e-01 -1.16145349e+00
-1.93706527e-01 -7.54059672e-01 3.59016955e-01 -1.21833813e+00
-9.64904547e-01 1.65700972e-01 2.16599777e-01 -4.30688649e-01
-3.51891756e-01 -4.50613528e-01 -7.20719397e-01 1.02422321e+00
-1.00119567e+00 -9.87293124e-01 3.20232838e-01 4.15003687e-01
1.24755949e-01 1.13296993e-01 1.25521302e+00 1.14839032e-01
-4.44992542e-01 5.95914543e-01 -5.56705415e-01 5.96398488e-02
5.69886327e-01 -9.63916481e-01 7.07275867e-01 9.93234277e-01
4.05334353e-01 1.67413437e+00 7.81723201e-01 -1.09731627e+00
-1.10227847e+00 -1.01094913e+00 1.39387882e+00 -8.55735123e-01
5.86067617e-01 -6.03913486e-01 -1.00764847e+00 1.25832069e+00
1.94208011e-01 -3.90327722e-01 7.73399413e-01 4.69062179e-01
-7.47468352e-01 9.02164653e-02 -1.07620299e+00 8.86703074e-01
1.39565694e+00 -8.69820416e-01 -1.20315778e+00 -1.60537660e-02
7.18485653e-01 -1.29754633e-01 -7.57651269e-01 -8.55858903e-03
5.59787929e-01 -6.10722721e-01 9.84442711e-01 -1.26144814e+00
8.27149451e-01 -2.52316177e-01 1.02824997e-03 -1.60583127e+00
-5.74544013e-01 -3.04845244e-01 -1.32893369e-01 1.22195256e+00
9.06414151e-01 -6.92218482e-01 4.44271564e-02 1.00126314e+00
-2.64901817e-02 -5.60353458e-01 -9.64159608e-01 -3.08932126e-01
2.52196789e-01 -8.91231000e-01 7.69360662e-01 1.36379921e+00
5.48814237e-01 3.77559036e-01 -6.76885024e-02 -2.01991543e-01
2.66956031e-01 9.96353477e-02 1.00056164e-01 -9.47401047e-01
-3.73042300e-02 -5.60523212e-01 1.03590868e-01 -2.16837913e-01
3.62481892e-01 -1.49053323e+00 -2.47290447e-01 -1.53688109e+00
2.79933572e-01 -3.86419773e-01 -4.42505807e-01 1.12296820e+00
-1.92625195e-01 -1.14042692e-01 6.86170310e-02 4.30867728e-03
-1.26134187e-01 6.30935431e-01 6.91720426e-01 -8.47240463e-02
2.41596475e-02 -5.27998388e-01 -1.18593001e+00 7.39294887e-01
9.37685430e-01 -6.72305465e-01 -3.57470542e-01 -4.12967026e-01
7.61394382e-01 -2.24259719e-01 9.37685072e-01 -4.66138333e-01
2.40998566e-01 -3.50792140e-01 7.79341340e-01 -2.11910591e-01
3.38606447e-01 -4.41018760e-01 1.10664137e-01 4.69838470e-01
-9.84959781e-01 2.39975631e-01 7.20137894e-01 3.37426841e-01
1.32329270e-01 1.15326703e-01 4.36542898e-01 -3.41166854e-01
-7.49458373e-01 -2.29307145e-01 -7.31932163e-01 4.27072532e-02
4.95462358e-01 9.99312401e-02 -8.55035901e-01 -2.62116820e-01
-6.27341449e-01 -1.73591122e-01 2.83965588e-01 4.29535449e-01
8.57507586e-01 -1.19622970e+00 -6.06692612e-01 1.06023304e-01
-5.76712890e-03 -7.94150054e-01 1.16852872e-01 7.99726665e-01
7.71285892e-02 6.76383615e-01 -3.57235461e-01 3.26781154e-01
-6.32079363e-01 8.38827074e-01 5.04292488e-01 -3.32383543e-01
-2.79960871e-01 9.40661550e-01 2.92676151e-01 -3.16092074e-01
-4.04128224e-01 -6.09969676e-01 -1.72802329e-01 2.17124119e-01
5.69016755e-01 1.74089059e-01 1.50679603e-01 -1.91017509e-01
-5.00889242e-01 -2.28232682e-01 -1.56722978e-01 -4.32193547e-01
1.25248647e+00 9.28755477e-02 -4.32968140e-01 7.56700456e-01
6.53165996e-01 2.71357030e-01 -8.41876388e-01 -3.98886316e-02
7.32938200e-02 -3.61785948e-01 -2.52111022e-05 -1.34957790e+00
-7.99090981e-01 7.91997552e-01 2.36535206e-01 -1.68030694e-01
8.89626682e-01 -1.09723307e-01 -3.19920015e-03 3.59876096e-01
3.96524429e-01 -8.71494710e-01 -3.46301377e-01 4.91136909e-01
1.48876536e+00 -7.13658571e-01 2.71700263e-01 -2.02976674e-01
-5.24205863e-01 6.73393965e-01 7.54489601e-01 -1.38679534e-01
6.21660531e-01 4.66864854e-01 -4.00431663e-01 -5.04417062e-01
-1.42003667e+00 2.95977294e-01 7.50887468e-02 4.20996904e-01
1.01918745e+00 2.24442229e-01 -6.60873771e-01 1.27414894e+00
-5.99871099e-01 2.29592487e-01 4.80280846e-01 3.96818638e-01
-7.32903033e-02 -6.02590561e-01 -1.93194583e-01 1.01007140e+00
-3.19891512e-01 -1.00457847e+00 -4.95293677e-01 5.72394252e-01
-1.19924285e-01 7.22425878e-01 1.60654843e-01 -3.62728804e-01
2.96450496e-01 5.95873058e-01 5.27376354e-01 -7.75974691e-01
-1.03115785e+00 -5.39605856e-01 4.91647422e-01 -8.56483221e-01
-1.58425923e-02 -6.22224212e-01 -1.49524045e+00 -5.04236460e-01
3.12866867e-02 -6.59054518e-02 4.03500795e-01 7.77248681e-01
6.39212728e-01 7.50845611e-01 -1.92423359e-01 -3.70779037e-01
-2.78320819e-01 -1.00622392e+00 -2.84889758e-01 2.60955811e-01
7.23882914e-02 -7.77343035e-01 -1.57736242e-01 2.91758865e-01] | [9.996282577514648, 7.938222885131836] |
f650e1e7-4dc5-41f4-8cba-2a707eceabd8 | genetic-multi-armed-bandits-a-reinforcement | 2302.07695 | null | https://arxiv.org/abs/2302.07695v1 | https://arxiv.org/pdf/2302.07695v1.pdf | Genetic multi-armed bandits: a reinforcement learning approach for discrete optimization via simulation | This paper proposes a new algorithm, referred to as GMAB, that combines concepts from the reinforcement learning domain of multi-armed bandits and random search strategies from the domain of genetic algorithms to solve discrete stochastic optimization problems via simulation. In particular, the focus is on noisy large-scale problems, which often involve a multitude of dimensions as well as multiple local optima. Our aim is to combine the property of multi-armed bandits to cope with volatile simulation observations with the ability of genetic algorithms to handle high-dimensional solution spaces accompanied by an enormous number of feasible solutions. For this purpose, a multi-armed bandit framework serves as a foundation, where each observed simulation is incorporated into the memory of GMAB. Based on this memory, genetic operators guide the search, as they provide powerful tools for exploration as well as exploitation. The empirical results demonstrate that GMAB achieves superior performance compared to benchmark algorithms from the literature in a large variety of test problems. In all experiments, GMAB required considerably fewer simulations to achieve similar or (far) better solutions than those generated by existing methods. At the same time, GMAB's overhead with regard to the required runtime is extremely small due to the suggested tree-based implementation of its memory. Furthermore, we prove its convergence to the set of global optima as the simulation effort goes to infinity. | ['Michael Krapp', 'Deniz Preil'] | 2023-02-15 | null | null | null | null | ['stochastic-optimization', 'multi-armed-bandits'] | ['methodology', 'miscellaneous'] | [-1.61481187e-01 -3.89784098e-01 -4.17648077e-01 1.00917198e-01
-1.02586389e+00 -6.19236887e-01 2.83560932e-01 -7.74914548e-02
-3.65505487e-01 1.40029442e+00 -2.67279863e-01 -5.90338290e-01
-7.08070278e-01 -1.03539169e+00 -6.96264625e-01 -1.08671665e+00
-3.23013306e-01 7.06551909e-01 -1.36685491e-01 -2.50455707e-01
3.94323260e-01 3.94840866e-01 -1.39943850e+00 -1.85730696e-01
1.09405315e+00 1.17165244e+00 2.53992766e-01 2.99429715e-01
-1.76819682e-01 1.24128819e-01 -5.47525764e-01 -3.14904600e-01
3.57312471e-01 -4.20603663e-01 -5.65339923e-01 1.72221631e-01
-3.57714593e-01 -1.25684008e-01 5.36149889e-02 1.16498899e+00
4.78417575e-01 4.61314082e-01 2.83602715e-01 -1.06936848e+00
-2.44065925e-01 7.10815728e-01 -9.28999901e-01 3.30550432e-01
1.63499713e-01 3.66984069e-01 1.06395471e+00 -2.68124968e-01
3.52484941e-01 1.25561833e+00 3.82262230e-01 1.68145850e-01
-1.19580865e+00 -6.21158779e-01 3.67241561e-01 1.77003264e-01
-1.16047347e+00 -1.04778372e-01 8.09566498e-01 1.67614147e-01
6.96684301e-01 5.81503391e-01 9.90638256e-01 8.57814074e-01
-8.90080929e-02 9.25517321e-01 1.19576478e+00 -5.34203529e-01
8.78869474e-01 -1.91907398e-02 -7.25234672e-02 6.77411020e-01
5.42086542e-01 4.23776984e-01 -2.03121215e-01 -6.29791796e-01
6.06714725e-01 -2.42373183e-01 -1.58191413e-01 -5.82096398e-01
-8.42443347e-01 1.31676030e+00 3.88777941e-01 5.70399128e-02
-7.47547686e-01 2.97930896e-01 1.99725747e-01 1.96413428e-01
3.59789163e-01 6.20894909e-01 -2.34178558e-01 -2.56996661e-01
-7.52966106e-01 4.49448526e-01 7.27549255e-01 5.33686161e-01
6.05801165e-01 2.42117375e-01 -9.78109688e-02 7.80021310e-01
2.86536664e-01 5.43544054e-01 6.71361804e-01 -6.73112035e-01
8.55113924e-01 3.98794293e-01 5.55395067e-01 -1.14729691e+00
-3.62748027e-01 -1.01783144e+00 -7.71732211e-01 1.76225320e-01
3.40302795e-01 -3.55157524e-01 -7.31872320e-01 1.61896741e+00
7.52514362e-01 1.64062738e-01 1.08576283e-01 8.96038711e-01
-4.33170274e-02 9.09680665e-01 -2.39135295e-01 -8.52486789e-01
1.01174366e+00 -1.14598227e+00 -6.26345456e-01 -4.07319307e-01
4.71831948e-01 -4.35559422e-01 8.78382146e-01 5.61324835e-01
-1.31825745e+00 -1.66038543e-01 -9.89563882e-01 1.02516222e+00
-2.01149702e-01 -2.11841986e-01 1.09670186e+00 1.15094233e+00
-8.70811820e-01 3.47838670e-01 -7.62996733e-01 6.05439022e-02
3.70693892e-01 3.23826134e-01 4.11500931e-01 5.41572757e-02
-1.06454670e+00 6.36541247e-01 7.91072130e-01 2.81683415e-01
-5.73130488e-01 -2.20374852e-01 -5.71934283e-01 2.10948899e-01
7.55759299e-01 -6.24617279e-01 1.22001731e+00 -9.19253230e-01
-1.65582740e+00 6.70039728e-02 -1.75789654e-01 -6.97900116e-01
7.43180096e-01 8.57172683e-02 -2.67318100e-01 4.09810469e-02
-4.27783653e-02 8.41889977e-02 7.63053775e-01 -1.21619081e+00
-8.57739270e-01 -3.87360811e-01 1.93955079e-02 3.61859143e-01
-3.88484180e-01 -2.42654681e-01 -2.70092517e-01 -7.89247990e-01
1.47632167e-01 -9.45632815e-01 -7.32947350e-01 -7.22986817e-01
-2.28198409e-01 -1.79791003e-02 4.71137762e-01 -1.29656881e-01
1.57025135e+00 -1.87364197e+00 2.44114712e-01 8.15002561e-01
-3.77534956e-01 2.24700451e-01 -1.80879027e-01 5.87031841e-01
6.31339699e-02 7.86678866e-02 -1.25027090e-01 -2.81220246e-02
9.73834321e-02 2.66763926e-01 -3.84982616e-01 4.94115561e-01
-2.66986668e-01 9.37051833e-01 -9.12210107e-01 -3.13537568e-01
1.02377772e-01 -3.92747045e-01 -6.08459175e-01 -1.02690756e-01
-5.54262698e-01 1.89820126e-01 -1.05793869e+00 8.00696850e-01
6.15518212e-01 -4.71407175e-01 3.87217849e-01 4.04511064e-01
-7.97183067e-02 -1.81990251e-01 -1.62732220e+00 1.39929783e+00
-3.72590601e-01 -2.95897245e-01 9.79603305e-02 -1.51320553e+00
9.12488937e-01 1.25135586e-01 2.81874090e-01 -7.51415491e-01
1.95097670e-01 4.33169544e-01 -8.99677444e-03 -2.03098744e-01
3.03176254e-01 3.62069495e-02 -6.44815266e-02 7.87494779e-01
-5.10479033e-01 -1.30149409e-01 6.04371905e-01 -3.59866679e-01
7.92260230e-01 -1.36245489e-01 4.39997107e-01 -1.52086034e-01
5.74158370e-01 1.68437645e-01 5.70653796e-01 1.13202953e+00
1.10176623e-01 -1.82882532e-01 2.90199488e-01 -6.55660391e-01
-7.25576878e-01 -7.54278958e-01 8.70008469e-02 9.32624996e-01
2.82175899e-01 9.43178609e-02 -5.11003733e-01 -4.12122250e-01
1.79517075e-01 9.42664623e-01 -4.80297148e-01 1.59488078e-02
-4.64894384e-01 -1.33346677e+00 1.17271870e-01 1.56627625e-01
6.02685630e-01 -9.82682526e-01 -8.18240166e-01 6.36037469e-01
9.11373869e-02 -6.46936715e-01 -1.27123678e-02 2.81115919e-01
-1.08920860e+00 -9.29273307e-01 -7.43452311e-01 -5.20185888e-01
4.39807326e-01 2.09722906e-01 8.25573146e-01 2.79086977e-02
-1.39704257e-01 -6.50257571e-03 -3.82167518e-01 -3.38662863e-01
-3.56156796e-01 6.90858066e-02 -4.74116281e-02 1.00120589e-01
-4.44173440e-02 -6.83678925e-01 -5.32354236e-01 4.86196280e-01
-8.90569031e-01 -1.42749399e-01 5.10346115e-01 1.24676621e+00
5.73399067e-01 4.59775597e-01 8.51779759e-01 -4.78775024e-01
9.64777529e-01 -6.51836514e-01 -1.29784977e+00 4.61249769e-01
-5.75990200e-01 2.24951178e-01 6.98504448e-01 -5.11758983e-01
-9.33438838e-01 -2.29232386e-01 2.61346430e-01 -2.10709408e-01
2.47452617e-01 8.58961940e-01 1.77382410e-01 -8.81991982e-02
5.17567217e-01 5.26040554e-01 8.20489898e-02 -2.97349393e-01
2.63571680e-01 5.90287387e-01 6.17520697e-03 -8.81314158e-01
6.91556156e-01 2.13949621e-01 1.90391511e-01 -3.60193878e-01
-8.17514479e-01 -1.53585151e-01 1.69757426e-01 -6.83365464e-02
1.16600573e-01 -5.57078958e-01 -8.55956018e-01 2.47424573e-01
-6.05974674e-01 -2.38282561e-01 -6.84348941e-02 5.31956255e-01
-8.72130811e-01 3.16341728e-01 -2.59817809e-01 -1.33315504e+00
-3.37919921e-01 -1.28815937e+00 4.89484221e-01 3.61079246e-01
3.54304560e-03 -8.83852482e-01 4.34263647e-02 4.17951345e-01
2.46940479e-01 3.38038683e-01 1.04890907e+00 -5.97530842e-01
-6.68618381e-01 -1.58864602e-01 1.75560951e-01 -1.11436687e-01
1.06778145e-01 -1.74879164e-01 -3.34619582e-01 -7.10566401e-01
1.45010725e-01 -4.52771425e-01 5.85611820e-01 6.59465849e-01
1.23227310e+00 -4.59910810e-01 -4.07043695e-01 5.54068863e-01
1.49167430e+00 8.32614422e-01 2.87451506e-01 1.02031755e+00
-2.42744252e-01 2.50196218e-01 1.09568405e+00 9.69779968e-01
-9.52757820e-02 6.63621426e-01 7.57547855e-01 1.33003309e-01
7.68892765e-01 -1.51176006e-01 -7.58575723e-02 3.83530617e-01
1.10086396e-01 -5.19025862e-01 -8.81878734e-01 5.50490260e-01
-2.21863818e+00 -9.34905529e-01 3.27064365e-01 2.28227377e+00
6.95917010e-01 2.02407077e-01 3.92369032e-01 3.19071829e-01
9.67866778e-01 9.63313356e-02 -8.32052767e-01 -6.72803879e-01
-9.21930000e-02 6.80839494e-02 5.67023635e-01 2.35582411e-01
-9.35474217e-01 6.95747375e-01 6.26745224e+00 1.32724690e+00
-9.51988459e-01 -3.29341054e-01 1.03208232e+00 -2.49281883e-01
-2.10453421e-01 -2.35546052e-01 -6.29169464e-01 6.61263406e-01
9.33480144e-01 -3.67356569e-01 9.17841077e-01 8.40821743e-01
3.39735985e-01 -3.81694376e-01 -5.74426234e-01 1.01701868e+00
-4.04278636e-01 -1.50168478e+00 -5.94377592e-02 2.46969447e-01
1.18647289e+00 -3.17224234e-01 3.91727626e-01 9.58568156e-02
7.38590717e-01 -1.01068521e+00 6.31858230e-01 7.05025196e-02
3.32424760e-01 -1.42260742e+00 7.49626160e-01 7.15022802e-01
-7.73667455e-01 -7.53073394e-01 -3.53931993e-01 -1.82655051e-01
2.39285558e-01 4.48135465e-01 -7.11422861e-01 8.45990777e-01
4.81527954e-01 3.11909206e-02 -6.97628129e-03 1.54275572e+00
-5.97454645e-02 6.07601941e-01 -7.36855388e-01 -6.77187443e-01
1.08691502e+00 -6.56654000e-01 5.39130390e-01 7.16835022e-01
6.76716030e-01 5.57666719e-02 3.12836409e-01 6.59238338e-01
3.29828382e-01 3.36361825e-01 -3.19352835e-01 -8.76892731e-02
7.44661689e-01 8.70741308e-01 -8.50302160e-01 -2.81044036e-01
-6.34669662e-02 3.82189304e-01 3.48744571e-01 4.45901632e-01
-1.07643151e+00 -4.85017598e-02 3.85247409e-01 -5.06833911e-01
7.47313023e-01 -2.27645766e-02 -3.09515208e-01 -7.01429904e-01
-4.08469588e-02 -1.16693521e+00 7.00000644e-01 -4.22343463e-01
-1.01904714e+00 5.36182225e-01 7.53493863e-04 -1.01462150e+00
-7.11600423e-01 -2.53343761e-01 -3.61710101e-01 7.60543585e-01
-1.35134149e+00 -5.43548226e-01 1.69370160e-01 5.87614954e-01
5.66583574e-01 -4.27325785e-01 7.73088515e-01 -6.16503172e-02
-7.48822808e-01 6.02384448e-01 9.05426502e-01 -4.04479146e-01
-1.93039104e-01 -9.68148410e-01 1.82458520e-01 6.40563309e-01
2.38313928e-01 3.97228152e-01 8.69562685e-01 -5.22661209e-01
-1.53371048e+00 -7.18610287e-01 5.94340563e-02 3.37702632e-01
1.01388097e+00 9.93852541e-02 -4.25553143e-01 2.60749578e-01
-1.57049790e-01 -3.07762418e-02 4.36055064e-01 2.24989578e-01
2.74610698e-01 -1.27424076e-01 -1.17616534e+00 7.00751007e-01
8.35623682e-01 2.44345218e-01 -3.36983621e-01 4.23348874e-01
3.52498919e-01 -5.52448630e-01 -4.30116385e-01 3.22819769e-01
3.04348171e-01 -9.75269556e-01 9.36657429e-01 -7.45494723e-01
1.19891070e-01 1.63318515e-02 -7.37313777e-02 -1.57783091e+00
-3.20625395e-01 -9.49128747e-01 -3.79081964e-01 8.89778018e-01
4.49853063e-01 -9.71056521e-01 9.58466709e-01 1.94924161e-01
3.69916409e-01 -1.12186992e+00 -1.32460797e+00 -1.00736737e+00
-6.42506257e-02 -1.94726750e-01 9.20966744e-01 5.64920306e-01
-2.74051219e-01 -1.10365093e-01 -4.44064915e-01 1.17419451e-01
7.28947163e-01 8.54727566e-01 6.44498229e-01 -8.37035716e-01
-7.55792201e-01 -5.51069379e-01 5.35791218e-02 -1.02209973e+00
-6.80671260e-02 -5.00600517e-01 -2.24336788e-01 -1.07170320e+00
2.70251781e-02 -8.24984074e-01 -5.62430859e-01 1.13497689e-01
-7.38817379e-02 -1.59134101e-02 1.47717595e-01 -3.85708399e-02
-5.74065566e-01 6.83650911e-01 1.29454315e+00 -9.81279239e-02
-5.38678586e-01 6.47551358e-01 -7.03020513e-01 7.20220566e-01
9.35138464e-01 -4.00270075e-01 -4.96359438e-01 -2.15600461e-01
3.08460355e-01 6.37034833e-01 -2.94024609e-02 -7.48345673e-01
7.47200055e-03 -5.69891512e-01 1.44693539e-01 -6.32621765e-01
3.79545063e-01 -8.27608049e-01 3.64310324e-01 7.37319648e-01
-1.64419904e-01 1.75739437e-01 2.60168642e-01 8.04533780e-01
-2.29361773e-01 -5.29341102e-01 6.01244092e-01 -2.61903375e-01
-4.79692698e-01 1.38481017e-02 -3.44626427e-01 -1.90336853e-02
1.31051910e+00 -3.97517592e-01 -3.56461369e-02 -4.90832835e-01
-7.26186872e-01 5.78426600e-01 1.11632742e-01 7.35515133e-02
3.55846047e-01 -1.24226451e+00 -4.07640189e-01 4.08716835e-02
-3.84781957e-01 -1.44581452e-01 1.15684211e-01 5.80728471e-01
-4.74987030e-01 6.40787601e-01 -7.67557602e-03 -4.82877910e-01
-1.07015324e+00 8.12887728e-01 2.06409261e-01 -7.61232793e-01
-3.09553206e-01 7.58544326e-01 -3.07626069e-01 -1.52765781e-01
3.72534931e-01 -1.06054224e-01 1.29985914e-01 2.48882711e-01
4.19206440e-01 4.75447416e-01 -5.83949983e-02 6.77463189e-02
-8.56519341e-02 3.62553924e-01 3.48205380e-02 -1.64543092e-01
1.69758129e+00 -1.84721276e-01 1.12715242e-02 2.01947853e-01
6.99836731e-01 -9.06121731e-02 -1.06651390e+00 -3.00976902e-01
7.92467296e-02 -7.23251462e-01 1.82806000e-01 -9.70674396e-01
-9.86964166e-01 2.57473767e-01 5.08751929e-01 5.89697897e-01
1.17464471e+00 -3.69512677e-01 6.46393716e-01 6.57985270e-01
8.55230093e-01 -1.25089848e+00 -1.11874446e-01 2.43634924e-01
7.07531452e-01 -1.05459166e+00 8.46724808e-02 -1.60918266e-01
-4.78097677e-01 9.49979901e-01 2.47118667e-01 -1.00509472e-01
1.30754113e-01 7.12595508e-02 -2.61774719e-01 9.65952650e-02
-8.86437237e-01 -1.30399823e-01 -1.03971019e-01 2.23952413e-01
-4.24739003e-01 7.23846769e-03 -6.62319660e-01 4.16079611e-01
-2.38713324e-01 -7.74477124e-02 1.58069193e-01 9.18500960e-01
-6.56463623e-01 -1.06070662e+00 -8.10581863e-01 3.86161923e-01
-4.34332550e-01 5.08684106e-03 1.75312504e-01 8.74347270e-01
-4.22211409e-01 1.04826021e+00 -1.62307367e-01 1.86684415e-01
-1.29497293e-02 -2.09441751e-01 3.65994185e-01 -9.00174752e-02
-5.12637317e-01 3.58209252e-01 1.96604788e-01 -6.20365679e-01
-3.29472065e-01 -6.29481614e-01 -8.39499533e-01 -2.46672034e-01
-5.12614727e-01 8.03433478e-01 5.89904845e-01 8.91219616e-01
9.83141214e-02 2.96308130e-01 9.75327432e-01 -6.25273526e-01
-1.32316816e+00 -5.97549796e-01 -7.50874102e-01 -6.04617596e-03
1.28240883e-01 -7.97218263e-01 -2.53176898e-01 -8.91608238e-01] | [4.596925258636475, 3.2274274826049805] |
256a08f9-a8ed-4d0b-a252-e2a27d8dcfc6 | exploring-instance-level-uncertainty-for | 2012.12880 | null | https://arxiv.org/abs/2012.12880v3 | https://arxiv.org/pdf/2012.12880v3.pdf | Exploring Instance-Level Uncertainty for Medical Detection | The ability of deep learning to predict with uncertainty is recognized as key for its adoption in clinical routines. Moreover, performance gain has been enabled by modelling uncertainty according to empirical evidence. While previous work has widely discussed the uncertainty estimation in segmentation and classification tasks, its application on bounding-box-based detection has been limited, mainly due to the challenge of bounding box aligning. In this work, we explore to augment a 2.5D detection CNN with two different bounding-box-level (or instance-level) uncertainty estimates, i.e., predictive variance and Monte Carlo (MC) sample variance. Experiments are conducted for lung nodule detection on LUNA16 dataset, a task where significant semantic ambiguities can exist between nodules and non-nodules. Results show that our method improves the evaluating score from 84.57% to 88.86% by utilizing a combination of both types of variances. Moreover, we show the generated uncertainty enables superior operating points compared to using the probability threshold only, and can further boost the performance to 89.52%. Example nodule detections are visualized to further illustrate the advantages of our method. | ['Lei He', 'Kun Wang', 'Weinan Song', 'Yao Zhang', 'Yuan Liang', 'Jiawei Yang'] | 2020-12-23 | null | null | null | null | ['lung-nodule-detection'] | ['medical'] | [ 1.01966739e-01 6.13906503e-01 -1.28581092e-01 -5.33386707e-01
-1.22871244e+00 -2.31930286e-01 4.18639988e-01 2.62778819e-01
-3.08874100e-01 6.46640539e-01 -6.07305989e-02 -4.75915998e-01
-2.55909264e-01 -5.67698002e-01 -5.78763127e-01 -7.11773932e-01
-6.05909079e-02 5.39168537e-01 4.02764350e-01 4.28056657e-01
-1.17866307e-01 3.49918723e-01 -1.15064967e+00 4.22289401e-01
1.07570767e+00 1.33781970e+00 1.27962530e-01 3.83650631e-01
-8.93957689e-02 4.08544660e-01 -5.98154068e-01 -4.42578822e-01
2.18335867e-01 -2.17922091e-01 -5.50567508e-01 -1.51385263e-01
2.49698892e-01 -4.67074335e-01 1.17731601e-01 1.07474744e+00
4.53283042e-01 -2.74269372e-01 9.98489738e-01 -8.05163383e-01
-1.63374379e-01 9.18579102e-01 -6.63861811e-01 4.42171507e-02
-8.80962685e-02 7.34631792e-02 7.14596927e-01 -4.89782333e-01
3.54708076e-01 9.03351784e-01 9.71449733e-01 4.31683928e-01
-1.00906050e+00 -4.63332146e-01 -2.62082547e-01 -6.40458018e-02
-1.39022470e+00 8.38654786e-02 2.99141377e-01 -6.51132345e-01
8.93375039e-01 2.69084483e-01 5.70409715e-01 8.96602809e-01
8.57689142e-01 5.40430605e-01 1.18365145e+00 -3.36725891e-01
2.52110422e-01 3.95541131e-01 -1.07624091e-01 5.52521110e-01
7.62413740e-01 2.97722012e-01 6.73372746e-02 -9.55460891e-02
7.80922890e-01 -2.53974855e-01 -1.32366329e-01 -4.92204785e-01
-9.94193614e-01 7.94744253e-01 8.32773149e-01 2.60572582e-01
-2.38957673e-01 4.97490674e-01 3.67918640e-01 -4.88504857e-01
4.07922238e-01 6.57600045e-01 -5.93219101e-02 -6.62525510e-03
-1.25626040e+00 1.11325234e-01 8.96614373e-01 7.90016830e-01
1.33452015e-02 -1.63123682e-01 -5.50363779e-01 6.18897021e-01
5.99915564e-01 4.39752012e-01 3.52865517e-01 -6.56126320e-01
5.41575179e-02 3.39561701e-01 9.75772366e-02 -7.51305878e-01
-8.32201719e-01 -6.88749731e-01 -8.40561032e-01 3.55453044e-01
6.63364291e-01 -2.73136012e-02 -1.31161702e+00 1.40753794e+00
3.09363574e-01 1.45813331e-01 -1.55692577e-01 9.10527170e-01
9.49748218e-01 3.30571681e-02 1.53568029e-01 3.21228802e-02
1.71872520e+00 -7.46912301e-01 -6.75850391e-01 -1.44559413e-01
6.26799107e-01 -6.77183032e-01 5.79361975e-01 3.75566304e-01
-9.53471005e-01 -3.79449368e-01 -1.23636103e+00 3.08413804e-01
-4.17306535e-02 3.62100095e-01 5.30912578e-01 1.14009035e+00
-7.37708569e-01 6.29586220e-01 -1.34435785e+00 -6.93791285e-02
8.45554411e-01 3.69813502e-01 6.79766983e-02 1.59423292e-01
-1.23572695e+00 1.15864944e+00 5.61009049e-01 7.41606355e-02
-7.40442097e-01 -1.03902113e+00 -8.21533263e-01 2.35581487e-01
4.68311489e-01 -7.53630757e-01 1.59499586e+00 -3.75081807e-01
-1.37654984e+00 5.33851027e-01 4.13725853e-01 -9.07207310e-01
1.14181018e+00 -1.20165080e-01 -3.87471244e-02 7.46397451e-02
-3.05383578e-02 7.71633148e-01 7.46191561e-01 -1.15080118e+00
-5.84493160e-01 -2.28741840e-01 -1.97774678e-01 -9.48681161e-02
2.91590858e-02 -2.74402380e-01 -4.37700450e-01 -3.30711216e-01
3.41515899e-01 -9.86167431e-01 -4.96346861e-01 1.20536841e-01
-4.98845577e-01 8.60742554e-02 3.20464492e-01 -6.91631973e-01
1.27584589e+00 -1.88397598e+00 -5.05730331e-01 4.78302062e-01
4.81426567e-01 8.69147182e-02 5.14784098e-01 -2.07765788e-01
3.35262045e-02 3.55796903e-01 -5.42300284e-01 -1.14564128e-01
-7.71261472e-03 8.14893469e-02 1.37637660e-01 3.59634459e-01
4.90103424e-01 9.73438978e-01 -7.92193174e-01 -7.03873515e-01
5.68569243e-01 6.63738906e-01 -4.31958705e-01 -2.63899177e-01
-2.06563175e-01 3.65166217e-01 -4.41553116e-01 6.23862207e-01
8.94855380e-01 -4.98384714e-01 2.27935493e-01 -4.13645595e-01
2.22783983e-01 1.21412754e-01 -1.18907750e+00 1.37518430e+00
-6.51717305e-01 4.01695698e-01 -2.05522209e-01 -3.68819863e-01
7.27630973e-01 1.65594399e-01 4.30583715e-01 -2.15611145e-01
4.23189104e-01 5.59469342e-01 6.73695862e-01 -3.07705104e-01
2.45453820e-01 -3.19209397e-01 -3.16539183e-02 3.48749361e-03
-1.10147230e-01 -6.26641750e-01 -6.72878027e-02 2.81155272e-03
9.65038538e-01 1.55250117e-01 7.88432240e-01 -5.55487216e-01
3.33922923e-01 8.51598084e-02 2.67738223e-01 9.94127870e-01
-5.61831295e-01 8.50726724e-01 7.32988775e-01 -7.64540723e-03
-9.38732266e-01 -1.08144629e+00 -9.11644161e-01 1.63598046e-01
1.17311977e-01 -2.40380205e-02 -8.44861627e-01 -9.62611794e-01
2.58475780e-01 9.01265800e-01 -7.75628507e-01 -1.59896344e-01
-2.74426907e-01 -9.34371054e-01 5.27485549e-01 9.47941720e-01
3.81371289e-01 -5.52649796e-01 -1.10619462e+00 2.13923961e-01
1.54581014e-02 -1.05866003e+00 -1.46234343e-02 4.16608691e-01
-9.66511786e-01 -1.09855211e+00 -7.27557540e-01 -1.28975540e-01
5.00151575e-01 -3.06526810e-01 1.24917436e+00 -2.16429487e-01
-7.69327760e-01 2.74951398e-01 -1.24596573e-01 -6.27187312e-01
-7.55577385e-01 2.27479767e-02 -1.95086852e-01 -5.85665405e-01
2.31628716e-01 3.69828306e-02 -7.73068190e-01 4.43971545e-01
-9.01815712e-01 1.05966181e-01 9.60563183e-01 9.81262326e-01
7.33887255e-01 2.94128191e-02 4.72722203e-01 -9.86098051e-01
3.66572827e-01 -3.91627192e-01 -6.13832593e-01 9.86678302e-02
-7.70909786e-01 2.36677334e-01 -1.15082882e-01 -2.43516490e-01
-1.32097530e+00 1.11519530e-01 -1.33061513e-01 -2.78200120e-01
-1.35457978e-01 4.82684284e-01 2.21837893e-01 -1.76419001e-02
9.52285051e-01 -5.19907355e-01 1.83854803e-01 3.78617346e-02
3.43427181e-01 7.43065774e-01 3.37630808e-01 -4.78540838e-01
2.36040935e-01 5.72331309e-01 2.42063865e-01 -3.49549323e-01
-9.33216810e-01 -3.81782234e-01 -4.28723365e-01 -2.65158594e-01
1.07613087e+00 -7.43999302e-01 -4.61022019e-01 4.55959067e-02
-8.73072565e-01 -8.88993070e-02 -4.92157817e-01 9.30023491e-01
-5.33568084e-01 2.71935523e-01 -5.29020369e-01 -8.60444546e-01
-2.89360046e-01 -1.56097162e+00 1.29275811e+00 2.89252490e-01
-2.65401900e-01 -7.79560804e-01 -3.42400223e-01 1.70686483e-01
5.70306599e-01 4.93624508e-01 7.05030382e-01 -1.03067005e+00
-6.12792552e-01 -4.28532481e-01 -5.15262544e-01 2.85995156e-01
4.34512012e-02 1.23828419e-01 -1.15369809e+00 -1.24764424e-02
2.73330092e-01 -9.61132646e-02 1.01450253e+00 1.05086863e+00
1.33428729e+00 4.07072365e-01 -6.81790948e-01 2.95951366e-01
1.33858347e+00 1.13661312e-01 4.94362652e-01 2.50358790e-01
2.79565990e-01 4.84983861e-01 6.77626550e-01 5.02149224e-01
-2.42251113e-01 5.26698232e-01 8.67703259e-01 6.83949366e-02
-3.86227578e-01 1.41012937e-01 -2.77041495e-01 1.99120626e-01
8.15000311e-02 -1.31836414e-01 -1.35342300e+00 2.52183169e-01
-1.67248034e+00 -3.54675651e-01 -2.87362218e-01 2.15244198e+00
6.73823774e-01 5.77762544e-01 -3.51718575e-01 -2.35614255e-01
7.40640640e-01 -1.58916578e-01 -6.49556220e-01 -1.33046806e-01
4.02254015e-01 6.61822185e-02 7.64715910e-01 3.86977166e-01
-1.37244642e+00 4.63597506e-01 5.94622564e+00 1.24696541e+00
-9.13157046e-01 4.30241935e-02 1.01672173e+00 -8.06061029e-02
-1.70643240e-01 -3.30091596e-01 -8.39711607e-01 2.90250212e-01
8.34491432e-01 1.06658697e-01 -2.84103930e-01 1.09244275e+00
-8.64130333e-02 -4.71493632e-01 -1.30580330e+00 8.18961263e-01
-1.32657021e-01 -1.17384493e+00 -3.36528957e-01 1.43819332e-01
8.52509201e-01 -1.34955391e-01 2.84680098e-01 3.99709821e-01
3.10766157e-02 -1.28582621e+00 5.04529893e-01 4.68033791e-01
7.67810106e-01 -6.47760332e-01 1.45064044e+00 2.24888682e-01
-8.51682961e-01 2.80037880e-01 -2.63871223e-01 5.89574873e-01
1.48578063e-01 1.10880232e+00 -1.68345666e+00 7.58467972e-01
7.69876063e-01 1.44974709e-01 -5.70880890e-01 1.30585361e+00
-1.72518924e-01 4.57642198e-01 -7.05793202e-01 -1.97290376e-01
1.98500514e-01 8.70025083e-02 3.99063170e-01 1.30973375e+00
5.50990045e-01 -2.54439890e-01 -1.95461527e-01 1.31318676e+00
1.95257992e-01 -1.03987157e-01 -2.81774461e-01 1.83146656e-01
3.40046883e-01 1.46108699e+00 -1.11532032e+00 -2.20014036e-01
-1.62616357e-01 5.43831050e-01 -1.07358128e-01 -2.00610533e-01
-1.33198392e+00 -1.60230413e-01 1.72416955e-01 1.03209421e-01
2.97571987e-01 1.11832790e-01 -6.42612398e-01 -6.01163566e-01
-9.97665226e-02 -4.04376060e-01 3.34314972e-01 -6.51587665e-01
-1.35341561e+00 7.25797296e-01 3.16864341e-01 -1.25426042e+00
-4.36145812e-01 -1.07353342e+00 -2.57738888e-01 8.94335389e-01
-1.29252803e+00 -1.04211450e+00 -4.77265686e-01 -2.01835677e-01
3.40693414e-01 2.81207472e-01 6.13274813e-01 7.92628899e-02
-1.99212134e-01 5.98474383e-01 2.09088400e-01 -1.21884704e-01
7.18025863e-01 -1.59926736e+00 3.26656885e-02 5.20883143e-01
-1.24188393e-01 2.87737101e-01 7.12523580e-01 -7.54390478e-01
-7.73287714e-01 -1.04374111e+00 2.23920345e-01 -6.69602990e-01
7.11269557e-01 -9.73288044e-02 -8.15320373e-01 3.53729278e-01
-1.90670669e-01 2.56531775e-01 5.92843711e-01 -1.80316612e-01
-9.21098515e-02 2.25009203e-01 -1.56891537e+00 5.45730114e-01
5.64504921e-01 -3.84040251e-02 -4.06793445e-01 1.65366575e-01
7.27017999e-01 -1.05786073e+00 -1.19975317e+00 1.04697323e+00
7.48194754e-01 -1.20146036e+00 8.99202883e-01 -1.88296214e-01
5.90082705e-01 -1.98709160e-01 -4.12017256e-02 -1.07907224e+00
4.87591885e-02 5.47872074e-02 -1.37910143e-01 8.36379409e-01
6.64767504e-01 -5.99238276e-01 9.11350131e-01 8.03223133e-01
-3.83529216e-01 -1.13315558e+00 -1.18396747e+00 -7.46477008e-01
2.28568390e-01 -8.77499759e-01 4.17591304e-01 4.08062041e-01
-2.61792064e-01 -3.97508204e-01 1.66907430e-01 3.23050946e-01
3.97669584e-01 -5.83071448e-02 2.09280580e-01 -1.12791145e+00
-4.63918239e-01 -7.77126312e-01 -5.61229646e-01 -6.92813694e-01
-3.75631303e-01 -8.09478462e-01 4.14620221e-01 -1.60248733e+00
2.88697541e-01 -4.75598514e-01 -3.48177791e-01 3.02517209e-02
-2.57971734e-01 1.32835418e-01 -1.17179556e-02 5.04027307e-03
-2.73666829e-01 2.20494136e-01 1.35561633e+00 -3.99954207e-02
-3.44654694e-02 3.65895420e-01 -4.94437546e-01 8.64255548e-01
8.96717966e-01 -4.71278757e-01 -2.52298981e-01 1.49466321e-01
5.48383817e-02 8.87796432e-02 3.87146026e-01 -1.24658608e+00
-1.13365926e-01 2.84027308e-01 6.89223766e-01 -8.18788290e-01
1.46381795e-01 -9.99826312e-01 2.01784223e-01 9.01349962e-01
-2.91894525e-01 -6.90058827e-01 5.95808506e-01 8.57214093e-01
-8.11815187e-02 -6.68467760e-01 9.96634185e-01 -1.36766031e-01
-3.64933193e-01 -5.21566533e-02 -1.48285463e-01 -1.82066694e-01
1.25228977e+00 -1.72946885e-01 -1.78860873e-01 -7.52393603e-02
-1.00638640e+00 2.02534959e-01 -5.08308485e-02 3.79761867e-02
3.42092484e-01 -1.22246909e+00 -5.88352680e-01 -3.79560143e-02
3.06749791e-01 3.40627193e-01 4.69488770e-01 1.25183403e+00
-7.12911844e-01 7.57643163e-01 1.49319887e-01 -1.29529619e+00
-1.05682278e+00 2.92116106e-01 7.58608222e-01 -6.33945286e-01
-4.78059381e-01 1.10928011e+00 1.83225796e-01 -2.54819930e-01
3.64869505e-01 -1.06318092e+00 -5.75666763e-02 -1.72758698e-01
2.36539796e-01 4.14619505e-01 3.81097227e-01 4.23482060e-02
-4.95344460e-01 2.55241394e-01 -2.15476021e-01 -7.67939240e-02
7.14350522e-01 1.85943395e-01 1.80245861e-01 3.52338612e-01
5.95218897e-01 -1.43458307e-01 -1.26366639e+00 1.08646788e-02
1.69722870e-01 -3.37197721e-01 1.73696414e-01 -1.16814315e+00
-7.91502178e-01 1.01640713e+00 1.13220656e+00 2.99793869e-01
7.71432877e-01 2.49245584e-01 2.91630208e-01 1.74434781e-01
2.46564388e-01 -7.84346938e-01 -6.66552410e-02 -5.48274070e-03
7.43968427e-01 -1.59437752e+00 1.65881798e-01 -7.21577048e-01
-7.28407681e-01 1.17162645e+00 7.59473681e-01 6.42232411e-03
7.89477468e-01 5.54757357e-01 2.58937366e-02 -8.48868638e-02
-3.55949342e-01 -1.61751226e-01 7.05175340e-01 5.03320754e-01
6.19669795e-01 3.30810815e-01 -3.87440711e-01 9.05912399e-01
-9.76844430e-02 9.94776003e-03 4.18250442e-01 7.93939710e-01
-5.41431189e-01 -7.57929623e-01 -4.93572742e-01 1.00942147e+00
-7.21691668e-01 -3.77359092e-02 1.52918443e-01 1.24155462e+00
1.73961103e-01 6.20114028e-01 2.24978656e-01 -3.19549032e-02
2.03957900e-01 -1.43376822e-02 5.56779623e-01 -7.51866698e-01
-4.87013012e-01 3.56710076e-01 8.86383727e-02 -5.80741763e-01
-1.98523968e-01 -5.22793531e-01 -1.47633922e+00 2.64353603e-01
-8.93366635e-01 1.98640954e-02 9.75920141e-01 6.88430071e-01
7.54285231e-02 1.15791810e+00 -2.67259646e-02 -6.88334167e-01
-1.19727588e+00 -1.12317467e+00 -5.53691626e-01 -6.68992475e-02
1.10728897e-01 -8.91914010e-01 -4.22536492e-01 -3.33822995e-01] | [14.411467552185059, -2.1201834678649902] |
f0230c48-82cc-42d1-8513-740d9d93bc47 | ifcnet-a-benchmark-dataset-for-ifc-entity | 2106.09712 | null | https://arxiv.org/abs/2106.09712v1 | https://arxiv.org/pdf/2106.09712v1.pdf | IFCNet: A Benchmark Dataset for IFC Entity Classification | Enhancing interoperability and information exchange between domain-specific software products for BIM is an important aspect in the Architecture, Engineering, Construction and Operations industry. Recent research started investigating methods from the areas of machine and deep learning for semantic enrichment of BIM models. However, training and evaluation of these machine learning algorithms requires sufficiently large and comprehensive datasets. This work presents IFCNet, a dataset of single-entity IFC files spanning a broad range of IFC classes containing both geometric and semantic information. Using only the geometric information of objects, the experiments show that three different deep learning models are able to achieve good classification performance. | ['Christoph van Treeck', 'Jérôme Frisch', 'Veronika Richter', 'Nicolas Pauen', 'Christoph Emunds'] | 2021-06-17 | null | null | null | null | ['ifc-entity-classification'] | ['computer-vision'] | [-1.76248506e-01 -3.12517248e-02 1.07969224e-01 -6.92622781e-01
-6.40242159e-01 -2.18651891e-01 5.07705033e-01 3.27677339e-01
1.18792370e-01 6.95433021e-01 5.40324636e-02 -2.91589171e-01
-6.08641148e-01 -1.36118793e+00 -8.71683419e-01 -3.29562843e-01
-8.01026076e-02 8.54828954e-01 9.12752599e-02 -4.68098819e-01
4.34511974e-02 7.49784172e-01 -1.93897188e+00 8.16744506e-01
6.30989492e-01 1.48154461e+00 2.14739367e-01 1.31323233e-01
-5.77673614e-01 7.22371817e-01 -3.84398729e-01 -3.33939940e-01
3.97896618e-01 2.53081024e-01 -1.14089596e+00 -3.40577543e-01
5.52696764e-01 -1.37841880e-01 -4.33674343e-02 9.10869420e-01
4.29614276e-01 -2.68170744e-01 4.79160428e-01 -1.50179136e+00
-5.92970550e-01 7.64042616e-01 -1.57315791e-01 -5.30819952e-01
2.04797626e-01 -2.71164626e-01 9.30588782e-01 -7.75385618e-01
7.51165569e-01 9.70481277e-01 1.03056538e+00 -1.32472128e-01
-7.71729112e-01 -7.85257399e-01 -2.64544934e-01 7.06636131e-01
-1.39929664e+00 1.02545261e-01 1.10293674e+00 -6.44725442e-01
1.16931355e+00 1.22361556e-01 7.93748915e-01 5.08558214e-01
4.94568720e-02 5.68660259e-01 8.29634368e-01 -4.58507687e-01
2.40971252e-01 3.25002857e-02 4.55855966e-01 2.08551899e-01
5.36108553e-01 1.00384884e-01 -2.39435062e-01 1.43279538e-01
5.85660577e-01 -7.95345306e-02 2.70602435e-01 -5.35834014e-01
-7.45217383e-01 7.44927943e-01 5.08892417e-01 6.03668272e-01
-7.12016165e-01 5.55287719e-01 5.07233858e-01 4.53613043e-01
2.64987826e-01 4.50413108e-01 -1.03122675e+00 -3.62603754e-01
-3.71968865e-01 3.35580796e-01 7.32388139e-01 1.46257460e+00
9.01414692e-01 -1.17270641e-01 4.99705046e-01 8.43895972e-01
2.96367198e-01 3.07805985e-01 -1.27044067e-01 -1.20663333e+00
5.66187978e-01 1.03257036e+00 8.78535658e-02 -1.01730561e+00
-6.68718278e-01 -3.97545159e-01 -5.71466029e-01 4.49434280e-01
3.98882572e-03 2.16282785e-01 -6.72695756e-01 1.11122251e+00
3.28379005e-01 -4.52549905e-01 -1.26057595e-01 5.39920866e-01
1.18092942e+00 2.11312607e-01 3.35942239e-01 5.07534802e-01
1.28400052e+00 -5.78531802e-01 -8.13762188e-01 -3.18340026e-02
1.11946058e+00 -7.65567303e-01 6.96284413e-01 3.10862333e-01
-9.95531201e-01 -7.39728332e-01 -1.16672373e+00 -2.05457062e-01
-1.18211436e+00 3.66491303e-02 9.89457488e-01 4.02828783e-01
-6.57088399e-01 7.90914834e-01 -4.30223197e-01 -2.22329691e-01
8.05942059e-01 4.21264589e-01 -6.35953784e-01 -4.07368988e-01
-1.25867951e+00 1.19852865e+00 8.65267992e-01 7.37947673e-02
-5.63009799e-01 -1.03196466e+00 -8.74546647e-01 3.66782509e-02
1.06225677e-01 -5.44928968e-01 1.17015910e+00 -7.86952257e-01
-8.67263794e-01 7.08892345e-01 8.12999606e-01 -3.48170370e-01
2.91022897e-01 -5.20438969e-01 -6.79338872e-01 -3.87575567e-01
6.42518923e-02 6.51015520e-01 1.98728621e-01 -1.43599701e+00
-8.06727111e-01 -4.79784757e-01 3.37081790e-01 -3.07456613e-01
-1.21748865e-01 -1.08040705e-01 5.24488315e-02 -5.29142499e-01
2.74635535e-02 -7.53245175e-01 1.94867216e-02 -1.45024449e-01
8.09340253e-02 -3.81365776e-01 9.75784481e-01 -9.00785625e-01
7.95632541e-01 -2.04767275e+00 -3.35692942e-01 1.91552415e-01
-9.72383991e-02 1.12942159e-01 -2.27030873e-01 4.96459931e-01
-3.92503113e-01 -1.20019756e-01 7.67664313e-02 2.16478959e-01
3.81660998e-01 2.15358406e-01 4.94147927e-01 2.29777634e-01
-4.59874049e-03 1.05480504e+00 -5.89247942e-01 -4.21636850e-01
5.30962765e-01 3.40387404e-01 -3.27650517e-01 -1.86782435e-01
-4.44349647e-01 2.24760145e-01 -3.98842305e-01 9.17067349e-01
1.05147016e+00 -7.50466511e-02 1.72471017e-01 -7.27441609e-01
1.18067034e-01 2.66457945e-01 -1.24760044e+00 2.02324724e+00
-7.78816938e-01 7.20542431e-01 -9.20473188e-02 -1.31767368e+00
9.94990587e-01 2.60098070e-01 9.75205958e-01 -1.07813120e+00
3.85365993e-01 5.53841174e-01 -2.93051060e-02 -5.77247322e-01
5.68590164e-01 -2.88501799e-01 -2.82683134e-01 3.44991624e-01
1.73954263e-01 -1.37006640e-01 1.96372673e-01 -3.43321472e-01
1.03334975e+00 2.83067197e-01 1.19273011e-02 -3.43677580e-01
5.39808512e-01 3.22862506e-01 2.46229514e-01 3.73490363e-01
1.94358081e-01 4.52805579e-01 -2.67597139e-01 -8.60765457e-01
-1.16253018e+00 -9.29341972e-01 -3.74787331e-01 1.04798293e+00
1.63044453e-01 -2.31867030e-01 -6.21853590e-01 -7.44306684e-01
5.41307151e-01 9.45292175e-01 -5.14681041e-01 -2.76407361e-01
-7.75962949e-01 -3.06708127e-01 2.52466500e-01 1.22542620e+00
6.10387862e-01 -9.97919917e-01 -4.95262861e-01 4.59748149e-01
-2.94970479e-02 -1.21919310e+00 4.07992691e-01 1.81661740e-01
-9.50919092e-01 -1.50771177e+00 -8.95536169e-02 -9.43020344e-01
3.21076989e-01 -6.08450361e-02 1.52181697e+00 1.07107610e-01
-5.43190241e-01 4.02524531e-01 -5.45378566e-01 -7.01879680e-01
-5.83760917e-01 1.95722431e-01 -6.43577456e-01 -5.60601771e-01
7.09664881e-01 -5.70230484e-01 -3.46990168e-01 3.51411015e-01
-7.84292102e-01 2.78839797e-01 6.16354525e-01 3.73365402e-01
3.97759646e-01 5.17094195e-01 7.12961376e-01 -6.48396373e-01
1.62332028e-01 -3.57542098e-01 -4.76562858e-01 1.92520112e-01
-5.65073550e-01 -6.21442311e-02 1.36179745e-01 1.06427141e-01
-1.31840694e+00 -3.51062194e-02 -4.80152756e-01 1.28573820e-01
-6.94786608e-01 8.00197899e-01 -8.00586104e-01 -4.63890843e-03
3.34712118e-01 -2.87733823e-01 -6.04367405e-02 -9.27791238e-01
6.05716705e-01 7.75354028e-01 5.71396470e-01 -9.74887252e-01
4.40383613e-01 4.51064706e-01 1.47545263e-01 -5.22856057e-01
-4.48732018e-01 -4.15973961e-01 -1.01888442e+00 -4.75623369e-01
6.65356755e-01 -7.41236806e-01 -4.53383893e-01 5.92598677e-01
-1.35964322e+00 -1.52008221e-01 -6.75894976e-01 5.73926568e-01
-7.16220140e-01 -1.17604651e-01 -4.73135740e-01 -3.07829767e-01
-2.80573756e-01 -9.57630932e-01 1.19396806e+00 -2.99603671e-01
-1.68484569e-01 -1.08202064e+00 -1.38434991e-01 4.45551425e-01
3.81658792e-01 5.60902834e-01 1.34916198e+00 -6.55226946e-01
-7.58365035e-01 -8.15476596e-01 -3.83713663e-01 7.10233808e-01
2.89503157e-01 -3.41020346e-01 -7.90763676e-01 6.52271584e-02
-3.04205537e-01 -3.01204324e-02 2.62714624e-01 2.14169100e-01
1.37275052e+00 2.19493166e-01 -4.40744013e-01 1.76793396e-01
1.74203217e+00 5.72944045e-01 9.58757877e-01 7.81829178e-01
8.38657618e-01 9.25188839e-01 9.72726464e-01 4.07809585e-01
5.27071774e-01 7.01202273e-01 7.58854926e-01 -3.41389179e-02
-3.65590721e-01 2.36376271e-01 -3.40409845e-01 6.05570853e-01
-3.83603960e-01 1.79619536e-01 -1.22873437e+00 7.20809102e-01
-1.86211050e+00 -6.96144104e-01 -6.13918185e-01 1.66180146e+00
6.13975465e-01 -2.96271518e-02 -3.76705974e-01 5.38067758e-01
5.95548689e-01 -3.73179823e-01 -9.51517671e-02 -1.69521526e-01
-1.32580400e-01 5.27450740e-01 5.72359741e-01 -1.99833587e-01
-1.20516026e+00 5.78160167e-01 6.58048582e+00 6.50923312e-01
-5.36388576e-01 3.55205894e-01 8.56149420e-02 1.68518335e-01
-2.34823078e-01 -2.06095144e-01 -4.35968667e-01 2.34363720e-01
9.43917930e-01 2.79160850e-02 -3.26611139e-02 1.11255538e+00
-1.39309019e-01 1.10023499e-01 -1.33464837e+00 9.02788699e-01
-2.55038679e-01 -1.82681119e+00 -6.87313378e-02 1.25909656e-01
8.31771970e-01 -5.71761057e-02 -3.48670930e-01 4.06420738e-01
6.07946992e-01 -9.68147993e-01 1.07539165e+00 6.12733066e-01
5.54530084e-01 -7.93769240e-01 1.43585372e+00 -1.21023767e-01
-1.09431946e+00 -1.83554649e-01 -1.13209814e-01 -1.22320332e-01
2.21198857e-01 5.09027541e-01 -3.03836137e-01 1.17267597e+00
1.28348100e+00 8.94798756e-01 -6.36382580e-01 9.54174399e-01
1.85945764e-01 7.75294602e-02 -3.47956002e-01 3.28341395e-01
5.45542091e-02 -2.60956828e-02 -9.61848795e-02 1.06891966e+00
5.03447056e-01 -2.64239162e-01 -9.73715633e-02 8.04455698e-01
8.78492445e-02 1.89717300e-02 -7.28051364e-01 7.35632405e-02
3.76621991e-01 9.53361988e-01 -4.22265470e-01 -2.41602853e-01
-6.34375155e-01 2.05526352e-01 -8.08811858e-02 7.20012039e-02
-1.14609361e+00 -3.47098321e-01 7.72223651e-01 4.44441438e-01
4.64612812e-01 -1.62465960e-01 -9.24546957e-01 -1.03777707e-01
2.08498791e-01 -3.78649831e-01 4.28132147e-01 -8.80409479e-01
-1.37648010e+00 1.49022415e-01 3.36917877e-01 -1.35311103e+00
1.08056009e-01 -9.73996639e-01 -2.10715428e-01 3.92467260e-01
-1.37962162e+00 -1.85866940e+00 -8.31277788e-01 4.24194694e-01
2.35635653e-01 -3.67222577e-01 5.56673706e-01 9.94361281e-01
-4.87494841e-02 2.61985153e-01 2.37940118e-01 2.26023495e-01
2.61402130e-01 -1.03857446e+00 1.45598412e-01 3.84273604e-02
9.46078543e-03 2.39016801e-01 3.62649232e-01 -5.74507058e-01
-1.33917511e+00 -1.53763723e+00 5.36093056e-01 -3.34541947e-01
5.82718194e-01 -9.57190320e-02 -7.58063257e-01 8.11375022e-01
1.43560037e-01 -1.57711059e-01 8.53528380e-01 4.29644212e-02
-2.31264576e-01 -3.66074145e-01 -1.12796915e+00 -4.04588319e-02
1.32959580e+00 -4.69301164e-01 -6.53625846e-01 2.85598576e-01
5.84107161e-01 -2.64472872e-01 -1.68765152e+00 8.68048787e-01
6.09581470e-01 -7.87200570e-01 1.23336625e+00 -7.92868972e-01
5.98674238e-01 -3.06055307e-01 -6.98781133e-01 -9.34474170e-01
-3.79099578e-01 3.27789128e-01 -1.59478813e-01 1.41696000e+00
2.65485078e-01 -4.48516428e-01 7.78149903e-01 5.22724748e-01
-7.99356401e-01 -3.19040537e-01 -8.20794284e-01 -1.03146279e+00
4.94613886e-01 -1.15853965e+00 1.51519239e+00 9.21295881e-01
-3.70995969e-01 -1.20976167e-02 3.40237945e-01 -5.23535684e-02
3.73682141e-01 4.45079863e-01 8.61769915e-01 -1.92961836e+00
3.86961967e-01 -3.86849731e-01 -9.49725747e-01 -1.96589410e-01
3.07910234e-01 -1.05308461e+00 -5.24776839e-02 -2.17458081e+00
-2.73133293e-02 -8.59293044e-01 -4.46049988e-01 5.17219365e-01
6.26538932e-01 8.17106515e-02 2.44805310e-02 1.72506412e-03
-5.49918890e-01 5.63484490e-01 9.24096942e-01 -6.68780684e-01
3.06524962e-01 -2.12263808e-01 -4.97050285e-01 7.66629100e-01
8.10902476e-01 -2.97572166e-01 1.91039424e-02 -5.83549619e-01
3.19156140e-01 -5.64410090e-01 6.92234635e-01 -1.13972485e+00
-1.29644182e-02 -2.21751705e-02 2.41553798e-01 -1.23102844e+00
1.02096088e-01 -1.59621370e+00 9.81156111e-01 3.65638673e-01
-1.28795244e-02 -2.82925606e-01 4.52542007e-01 2.47716263e-01
-4.74205822e-01 -3.56237531e-01 6.87090635e-01 -1.72366947e-01
-1.29903781e+00 1.15346342e-01 -3.38896327e-02 -3.84936899e-01
1.19510162e+00 -4.67886329e-01 -3.37372959e-01 1.65670037e-01
-8.31894636e-01 -4.31188792e-02 4.17725533e-01 8.91006827e-01
4.14766520e-01 -1.94288278e+00 -5.58145761e-01 8.61503407e-02
5.54592252e-01 2.31694803e-01 2.12453499e-01 5.61787009e-01
-9.66262102e-01 5.90396643e-01 -5.50366163e-01 -6.13290250e-01
-1.05428362e+00 4.48030382e-01 4.07245934e-01 -6.94879740e-02
-6.46554351e-01 2.79693991e-01 -3.15037258e-02 -8.07012498e-01
5.56181557e-02 -2.54834145e-01 -7.59132430e-02 2.42490042e-02
9.43047181e-03 7.61478484e-01 7.09785998e-01 -6.29691720e-01
-2.95669794e-01 4.75803256e-01 1.26576200e-01 5.03258288e-01
1.66252482e+00 1.33522972e-01 -3.63194495e-01 1.51585013e-01
1.47915232e+00 -7.60012925e-01 -7.02025354e-01 -1.62033997e-02
6.86576545e-01 -5.36054909e-01 2.20699951e-01 -9.37723935e-01
-1.45611596e+00 7.71322429e-01 9.99375224e-01 1.10423028e-01
9.38706994e-01 4.00556266e-01 9.43975389e-01 3.36881816e-01
1.00867629e+00 -1.57072568e+00 -1.84998084e-02 3.87751043e-01
1.15858364e+00 -1.16229916e+00 -4.63952571e-02 -6.06265008e-01
-2.84393672e-02 1.17344701e+00 7.36577928e-01 1.27847567e-01
1.19858491e+00 5.07416308e-01 -3.13588500e-01 -8.31724226e-01
-1.50750116e-01 -2.99029678e-01 1.98863521e-01 1.11597788e+00
5.02572298e-01 1.85003281e-01 -2.71805823e-01 9.34614301e-01
-3.18565875e-01 2.95918822e-01 7.82961324e-02 1.42394257e+00
-1.85851678e-01 -1.14025211e+00 -2.47946680e-01 4.57262784e-01
-9.45452079e-02 3.03644210e-01 -1.71828151e-01 1.43417680e+00
6.42458618e-01 6.76418304e-01 3.69118243e-01 -4.50735033e-01
7.94957340e-01 -6.48883060e-02 6.47445798e-01 -3.62100899e-01
-4.57810581e-01 -4.19668257e-01 5.59213281e-01 -6.56860471e-01
-5.54021835e-01 -6.86396658e-01 -1.43881476e+00 -3.29245359e-01
-5.65540552e-01 -1.64821461e-01 1.06141007e+00 9.10978019e-01
4.67492938e-01 8.21923971e-01 3.02209169e-01 -5.18294096e-01
-8.53590369e-02 -1.06619263e+00 -9.03014958e-01 8.74109447e-01
-4.29348320e-01 -1.07784760e+00 2.93489635e-01 2.44322449e-01] | [7.381726264953613, 1.796416163444519] |
d62c2752-0415-49dd-bcc2-1bef9d99704c | model-based-safe-deep-reinforcement-learning | 2210.07573 | null | https://arxiv.org/abs/2210.07573v1 | https://arxiv.org/pdf/2210.07573v1.pdf | Model-based Safe Deep Reinforcement Learning via a Constrained Proximal Policy Optimization Algorithm | During initial iterations of training in most Reinforcement Learning (RL) algorithms, agents perform a significant number of random exploratory steps. In the real world, this can limit the practicality of these algorithms as it can lead to potentially dangerous behavior. Hence safe exploration is a critical issue in applying RL algorithms in the real world. This problem has been recently well studied under the Constrained Markov Decision Process (CMDP) Framework, where in addition to single-stage rewards, an agent receives single-stage costs or penalties as well depending on the state transitions. The prescribed cost functions are responsible for mapping undesirable behavior at any given time-step to a scalar value. The goal then is to find a feasible policy that maximizes reward returns while constraining the cost returns to be below a prescribed threshold during training as well as deployment. We propose an On-policy Model-based Safe Deep RL algorithm in which we learn the transition dynamics of the environment in an online manner as well as find a feasible optimal policy using the Lagrangian Relaxation-based Proximal Policy Optimization. We use an ensemble of neural networks with different initializations to tackle epistemic and aleatoric uncertainty issues faced during environment model learning. We compare our approach with relevant model-free and model-based approaches in Constrained RL using the challenging Safe Reinforcement Learning benchmark - the Open AI Safety Gym. We demonstrate that our algorithm is more sample efficient and results in lower cumulative hazard violations as compared to constrained model-free approaches. Further, our approach shows better reward performance than other constrained model-based approaches in the literature. | ['Shalabh Bhatnagar', 'Ashish Kumar Jayant'] | 2022-10-14 | null | null | null | null | ['safe-exploration'] | ['robots'] | [ 1.49032354e-01 3.88782799e-01 -4.81479347e-01 6.34323061e-02
-8.49472106e-01 -5.40175378e-01 5.94541788e-01 4.19644445e-01
-9.89295721e-01 1.21002316e+00 -1.72108576e-01 -3.79449815e-01
-5.77757776e-01 -8.57011616e-01 -8.81672084e-01 -9.05273914e-01
-5.98003328e-01 8.35898757e-01 6.24905191e-02 -8.31405073e-02
1.35226175e-01 3.63375604e-01 -1.30286455e+00 -5.30097127e-01
9.25638437e-01 9.37003255e-01 1.48242816e-01 5.09576976e-01
5.50778806e-01 8.48281622e-01 -4.24349487e-01 1.13444649e-01
4.25583541e-01 -1.47029623e-01 -6.68926537e-01 -1.23566516e-01
-5.70134699e-01 -5.96673369e-01 9.52003673e-02 1.18918550e+00
5.59812605e-01 6.76227927e-01 5.26079595e-01 -1.16394234e+00
1.26041234e-01 8.20656598e-01 -4.10825014e-01 -1.89851731e-01
1.24403268e-01 5.20492673e-01 8.80449533e-01 -8.04220513e-02
4.42232788e-01 1.24122739e+00 1.25697985e-01 6.89301014e-01
-1.40486670e+00 -4.64531809e-01 4.92060781e-01 4.92862388e-02
-9.80956554e-01 1.36501184e-02 5.22562683e-01 -2.54074037e-01
9.89396751e-01 -2.98547894e-02 9.17298377e-01 1.28458285e+00
4.55655873e-01 6.59712493e-01 1.24826479e+00 -3.73852342e-01
1.07808232e+00 -1.84409127e-01 -4.21458364e-01 5.69071591e-01
2.06115037e-01 8.74852240e-01 -2.31429264e-01 -2.58534729e-01
6.26456439e-01 -1.66801780e-01 1.32389545e-01 -6.52447581e-01
-9.27871943e-01 1.13237095e+00 2.15617701e-01 -2.16546565e-01
-7.43915856e-01 7.19110310e-01 3.78069729e-01 2.83373207e-01
1.47874102e-01 6.70558691e-01 -2.91882336e-01 -2.96107829e-01
-7.11013675e-01 8.40337574e-01 8.10770273e-01 4.63038862e-01
4.23735768e-01 2.56919056e-01 -4.22200769e-01 3.48032594e-01
5.52155614e-01 3.16732079e-01 1.61669170e-03 -1.25647962e+00
5.12701333e-01 2.33551368e-01 7.81907976e-01 -4.30883765e-01
-4.00878966e-01 -6.10753894e-01 -2.99493968e-01 8.41483891e-01
4.39530551e-01 -6.12641752e-01 -8.66877973e-01 2.09712386e+00
6.24069393e-01 2.74486303e-01 2.67240733e-01 7.12473929e-01
-2.20257729e-01 7.22140908e-01 3.31743807e-01 -6.06863141e-01
8.48891973e-01 -5.57731032e-01 -7.27446139e-01 -4.07398373e-01
5.53232312e-01 -4.39641662e-02 9.65617239e-01 7.21514285e-01
-1.11953270e+00 1.54339090e-01 -1.09062123e+00 7.77345598e-01
-4.03926037e-02 -4.06058341e-01 5.83965480e-01 6.20971203e-01
-6.57620251e-01 8.72244537e-01 -1.21460295e+00 5.03533445e-02
5.55381179e-01 7.08951414e-01 1.28706008e-01 2.98715591e-01
-1.21139264e+00 1.17399025e+00 7.75684536e-01 1.01867884e-01
-1.88564420e+00 -4.02747273e-01 -7.00457036e-01 9.41284820e-02
1.25951612e+00 -3.30706120e-01 1.66171432e+00 -6.93398118e-01
-1.99494243e+00 7.86562264e-02 5.39978206e-01 -1.02599466e+00
9.14533973e-01 -2.97070891e-01 1.86799988e-01 -5.53165227e-02
-1.70690402e-01 5.99512577e-01 8.41107130e-01 -1.24153566e+00
-6.30220354e-01 3.42687704e-02 4.95827228e-01 6.37763679e-01
-9.82871652e-02 -1.15277618e-01 7.77062774e-02 -2.52556056e-01
-6.22170508e-01 -1.29099131e+00 -8.93009245e-01 -3.81040782e-01
-2.26919726e-01 -2.19637483e-01 3.01676273e-01 -2.97166944e-01
1.21004593e+00 -1.67375207e+00 4.53664929e-01 3.88613790e-01
-2.76289970e-01 6.56809732e-02 4.31467146e-02 5.11515856e-01
2.93812633e-01 -2.35453863e-02 -5.14800489e-01 -3.45678538e-01
3.19016784e-01 6.18959010e-01 -4.30128783e-01 5.58155954e-01
-7.51629770e-02 5.98723769e-01 -1.21439552e+00 -2.89629161e-01
2.73916900e-01 1.13646030e-01 -7.34670103e-01 4.62607116e-01
-8.12278152e-01 6.02331638e-01 -7.03805566e-01 3.72480065e-01
2.10775226e-01 3.73138070e-01 2.33365282e-01 6.84815645e-01
-2.53490090e-01 8.28140751e-02 -1.31924272e+00 1.40754032e+00
-6.44193769e-01 3.44322957e-02 1.03886105e-01 -1.04825783e+00
5.33977389e-01 3.59914929e-01 4.58299458e-01 -4.59707797e-01
2.95162916e-01 7.20485747e-02 1.29298627e-01 -1.82028264e-01
3.37027311e-01 -3.85995746e-01 -3.89608383e-01 5.66812634e-01
-1.88668206e-01 -3.27877253e-01 3.98582891e-02 -1.28670454e-01
1.10867834e+00 7.26743281e-01 5.23563325e-01 -3.93402576e-01
2.43957520e-01 -3.01127359e-02 7.94727325e-01 1.02706361e+00
-2.46487021e-01 -1.54809594e-01 8.86403501e-01 -1.72519892e-01
-6.73317850e-01 -7.90080369e-01 1.38157725e-01 8.05535197e-01
3.89933363e-02 -7.72857144e-02 -8.28580379e-01 -8.72292340e-01
-3.36475819e-02 1.21202004e+00 -7.80171812e-01 -4.54727679e-01
-5.95547438e-01 -6.72662020e-01 2.64353544e-01 2.50155658e-01
3.29033077e-01 -1.32021451e+00 -1.53138471e+00 4.43071783e-01
3.31738234e-01 -5.91721594e-01 -1.61381945e-01 7.92255580e-01
-6.94155276e-01 -8.06823015e-01 -4.00524050e-01 -1.41806543e-01
6.99387431e-01 -5.91189086e-01 6.90010667e-01 -2.74772316e-01
9.43663493e-02 3.87621641e-01 -5.02309576e-02 -5.33451378e-01
-4.75915253e-01 -7.30485395e-02 3.15478861e-01 -1.14057347e-01
-2.11358055e-01 -3.07484388e-01 -4.44062471e-01 5.07949218e-02
-9.05567944e-01 -1.06401928e-01 3.65391344e-01 9.63166177e-01
6.65454268e-01 4.24254566e-01 6.66058600e-01 -6.02173746e-01
9.77900505e-01 -5.95018983e-01 -1.44728589e+00 2.05623567e-01
-8.18568051e-01 5.12194872e-01 7.94396877e-01 -7.61138499e-01
-1.11220157e+00 2.10081309e-01 1.35590658e-01 -4.02599245e-01
2.18137298e-02 5.44669688e-01 -1.50006816e-01 2.18634214e-02
5.12420475e-01 7.28517100e-02 -1.52292037e-02 -1.30557552e-01
5.32792546e-02 2.08710179e-01 -9.09196958e-03 -1.19661081e+00
7.63100803e-01 1.24913424e-01 4.67941135e-01 -2.89204776e-01
-7.60335147e-01 2.07631379e-01 -1.04695000e-01 -5.21623671e-01
6.67676508e-01 -6.82946861e-01 -1.26678014e+00 2.14067921e-01
-6.67619109e-01 -9.85611677e-01 -6.72379851e-01 4.61038202e-01
-1.13676834e+00 8.45568478e-02 -1.10094063e-01 -1.47521806e+00
-1.47297904e-02 -1.39952743e+00 5.52446306e-01 3.49678159e-01
-3.90867814e-02 -9.73958433e-01 2.77685612e-01 -1.14371873e-01
2.10208908e-01 6.82196736e-01 7.21221149e-01 -5.16052485e-01
-6.21274292e-01 1.26075432e-01 6.24260366e-01 8.36667642e-02
-2.66209126e-01 -2.61082083e-01 -5.09813845e-01 -5.58574855e-01
-1.20777944e-02 -6.91438615e-01 6.10680819e-01 5.30427217e-01
8.89607668e-01 -8.15523148e-01 -1.34184599e-01 3.41528386e-01
1.50195348e+00 7.09579468e-01 2.13073924e-01 7.99648046e-01
1.34029791e-01 6.82936370e-01 1.13278174e+00 8.84079695e-01
1.84129015e-01 7.26177156e-01 1.10521281e+00 4.00552809e-01
7.38753140e-01 -4.48228151e-01 7.13293552e-01 -3.44016075e-01
-9.53146443e-02 -1.59801453e-01 -8.11183989e-01 4.86734778e-01
-2.30025101e+00 -9.85379219e-01 5.87278128e-01 2.81277609e+00
1.03827858e+00 4.21452343e-01 3.08785409e-01 -9.19911563e-02
3.48391712e-01 -9.33295563e-02 -1.06510246e+00 -7.43825257e-01
3.77736598e-01 1.40324265e-01 8.25985610e-01 7.83115029e-01
-1.00351906e+00 9.16841269e-01 5.63723803e+00 7.90953279e-01
-7.91993856e-01 8.67957845e-02 6.36299968e-01 -5.41646719e-01
-2.02570841e-01 1.80061713e-01 -7.62521863e-01 3.88789594e-01
1.32880211e+00 -2.21814170e-01 9.17141140e-01 9.97398794e-01
6.04867101e-01 -6.20820165e-01 -1.12209225e+00 4.15827274e-01
-7.10320950e-01 -1.02310920e+00 -4.59948778e-01 3.41271967e-01
7.53960252e-01 -6.86925277e-02 8.05575401e-02 6.10384226e-01
1.00832939e+00 -1.15538251e+00 1.19241834e+00 4.25145864e-01
3.82925481e-01 -1.51333117e+00 5.01353920e-01 7.62909830e-01
-7.43647337e-01 -5.66163182e-01 -1.02879174e-01 -2.27817774e-01
2.27600202e-01 1.88038945e-01 -8.43568504e-01 3.25266123e-01
4.43993121e-01 2.95165420e-01 1.18728183e-01 1.05242944e+00
-5.74144065e-01 6.49880767e-01 -5.58568001e-01 -3.24596941e-01
8.02277803e-01 -3.82522434e-01 7.16233134e-01 6.45546436e-01
1.52767614e-01 -2.34799325e-01 5.21346211e-01 9.98764694e-01
3.66492867e-01 -1.84378609e-01 -6.27115190e-01 4.18815343e-03
4.45933312e-01 9.07353878e-01 -8.50285590e-01 8.74968525e-03
1.36132404e-01 4.81389076e-01 4.02237028e-01 3.73602271e-01
-9.68297482e-01 -8.26179758e-02 5.35684586e-01 -3.25314760e-01
1.37223795e-01 -2.62470484e-01 -3.91297638e-02 -7.18888342e-01
-2.38432854e-01 -8.59564006e-01 4.68031913e-01 -7.60033131e-02
-7.49550521e-01 3.06950271e-01 4.77562070e-01 -1.07562041e+00
-7.98398614e-01 -2.41942495e-01 -5.10599971e-01 6.86103702e-01
-1.59049928e+00 -6.76938057e-01 4.68737662e-01 4.49452758e-01
5.25860846e-01 -6.90186024e-02 7.26131797e-01 -3.27132106e-01
-8.57660890e-01 1.63122669e-01 1.70868799e-01 -4.90535587e-01
2.06215754e-01 -1.43943477e+00 -1.56903908e-01 8.62797320e-01
-4.63788569e-01 4.09039587e-01 1.11170411e+00 -8.00354600e-01
-1.32364202e+00 -9.65010285e-01 5.20830117e-02 -1.20187774e-01
7.29129434e-01 -3.47272545e-01 -5.22477210e-01 6.66692257e-01
1.25174299e-01 -5.87256849e-02 1.11249559e-01 -1.81096882e-01
2.81181842e-01 -1.32062018e-03 -1.31421888e+00 9.19536531e-01
7.00745404e-01 6.77242875e-02 -3.99764091e-01 2.21883491e-01
5.63826501e-01 -5.45944571e-01 -7.18065262e-01 3.12722296e-01
3.26383054e-01 -7.49558568e-01 7.43970275e-01 -6.40402317e-01
6.21158034e-02 -2.70098388e-01 6.63802624e-02 -1.58804607e+00
7.53265154e-03 -1.32722414e+00 -5.28029919e-01 8.19400549e-01
3.67748648e-01 -6.18790209e-01 7.01178074e-01 7.70181537e-01
8.54409486e-03 -1.09246790e+00 -1.34993649e+00 -1.04014504e+00
2.79958695e-01 -4.32231605e-01 4.31982636e-01 3.17248195e-01
-5.56274503e-03 -2.48986974e-01 -6.64757371e-01 2.83674717e-01
1.03087664e+00 -1.02494322e-02 3.80730957e-01 -9.33930218e-01
-6.05308771e-01 -3.41459662e-01 3.21354836e-01 -6.03147268e-01
5.56572914e-01 -4.59650218e-01 4.43270296e-01 -1.37113917e+00
-2.03463305e-02 -7.14606941e-01 -3.46505046e-01 6.94625616e-01
3.73478867e-02 -6.23281956e-01 2.68789649e-01 -1.31669670e-01
-6.44332170e-01 8.41213703e-01 1.11187255e+00 3.85990404e-02
-5.69949031e-01 2.90509284e-01 -4.15912211e-01 5.71499646e-01
9.10406828e-01 -7.11758316e-01 -7.36287773e-01 4.55995835e-02
5.40106595e-01 5.33023298e-01 1.52490795e-01 -8.31234336e-01
-1.28396200e-02 -1.03817761e+00 -9.92721990e-02 -2.57773429e-01
3.60592842e-01 -1.01447463e+00 1.46031573e-01 9.98146117e-01
-7.39142179e-01 -9.07983482e-02 1.07114628e-01 8.88346672e-01
2.79896766e-01 -5.52166581e-01 9.43298519e-01 -3.15549791e-01
-4.68781710e-01 2.66302943e-01 -6.92466736e-01 1.98622290e-02
1.52059436e+00 9.88241285e-02 8.76322836e-02 -5.46419084e-01
-6.27532482e-01 6.98575377e-01 3.25115412e-01 1.91838756e-01
3.90487045e-01 -8.42238963e-01 -3.95525783e-01 -1.93717808e-01
-1.52001232e-01 2.64540404e-01 6.25213385e-02 6.91174567e-01
-2.04187423e-01 3.26578885e-01 -2.45923460e-01 -1.43987894e-01
-7.16554224e-01 6.35833204e-01 7.65591443e-01 -7.56457984e-01
-4.86872464e-01 5.87249160e-01 -1.32409453e-01 -3.57573420e-01
6.24179006e-01 -3.19594532e-01 -2.07323492e-01 -3.32031064e-02
3.11422020e-01 3.56213391e-01 -2.60738075e-01 -1.39780343e-01
-2.01635703e-01 1.18373856e-01 3.58331650e-02 -6.51780963e-01
1.47242808e+00 1.13063872e-01 3.35668474e-01 3.94424945e-01
4.27190781e-01 -3.25981110e-01 -2.07420206e+00 7.98313692e-02
2.92819291e-01 -2.67825723e-01 4.97707635e-01 -1.00153649e+00
-6.62363529e-01 5.06123483e-01 5.40268362e-01 4.96993549e-02
8.90177786e-01 -4.36557353e-01 2.85724968e-01 5.19860327e-01
6.69035375e-01 -1.58810985e+00 9.08583477e-02 4.28094029e-01
6.78821564e-01 -1.06552672e+00 2.28716619e-02 4.42750454e-01
-9.39205825e-01 7.94179261e-01 6.97170317e-01 -2.49009296e-01
2.61721522e-01 3.78350645e-01 -3.43651533e-01 2.32639939e-01
-1.16382086e+00 -2.67226666e-01 -2.58958846e-01 5.24529696e-01
-3.65680277e-01 1.53834492e-01 -2.30179027e-01 2.76727200e-01
1.46741569e-01 3.88716608e-02 6.79525256e-01 1.24660277e+00
-5.18692255e-01 -1.29953814e+00 -4.36920553e-01 2.27976486e-01
-5.44297457e-01 2.01459631e-01 1.38927624e-01 7.21761107e-01
-9.59233120e-02 9.62145448e-01 -2.89667159e-01 1.13960661e-01
8.37835521e-02 -6.23868369e-02 4.38528419e-01 -6.62270665e-01
-7.66137481e-01 2.45796621e-01 2.89778411e-01 -8.74070823e-01
-1.97402060e-01 -8.43029499e-01 -1.62428880e+00 1.16854250e-01
-1.90558106e-01 2.91373909e-01 6.87776566e-01 1.12940896e+00
4.28394899e-02 5.42251647e-01 7.22181678e-01 -9.62153077e-01
-1.30542481e+00 -5.75145423e-01 -5.03184497e-01 -1.12137198e-01
5.81167698e-01 -1.04836893e+00 -3.97302091e-01 -5.73880017e-01] | [4.365190505981445, 2.2343578338623047] |
8411c316-6fe5-418f-b8a1-d728cca821f6 | optimize-the-co-expansion-of-generation-and | 2101.04048 | null | https://arxiv.org/abs/2101.04048v1 | https://arxiv.org/pdf/2101.04048v1.pdf | Optimize the Co-expansion of Generation and Transmission Considering Wind Power in the US Eastern Interconnection | This paper studies the generation and transmission expansion co-optimization problem with a high wind power penetration rate in large-scale power grids. In this paper, generation and transmission expansion co-optimization is modeled as a mixed-integer programming (MIP) problem. A scenario creation method is proposed to capture the variation and correlation of both load and wind power across regions for large-scale power grids. Obtained scenarios that represent load and wind uncertainties can be easily introduced into the MIP problem and then solved to obtain the co-optimized generation and transmission expansion plan. Simulation results show that the proposed planning model and the scenario creation method can improve the expansion result significantly through modeling more detailed information of wind and load variation among regions in the US EI system. The improved expansion plan that combines generation and transmission will aid system planners and policy makers to maximize the social welfare in large-scale power grids. | ['Shutang You'] | 2021-01-11 | null | null | null | null | ['paper-generation'] | ['natural-language-processing'] | [-5.45428336e-01 -3.34135324e-01 -3.36436719e-01 -1.22003118e-03
-2.98007410e-02 -8.64881158e-01 8.80659297e-02 8.36228356e-02
1.56358704e-01 1.48328125e+00 3.56478244e-01 -3.71770769e-01
-5.75698197e-01 -1.42125201e+00 2.10188299e-01 -9.05054152e-01
-4.67922896e-01 4.85600084e-01 -4.43240076e-01 -3.45706433e-01
-1.85895219e-01 6.29571736e-01 -7.85508215e-01 -3.92418265e-01
1.40886223e+00 5.12959540e-01 6.66747808e-01 2.85598040e-01
1.69739068e-01 1.75158873e-01 -8.78484547e-01 3.05286676e-01
4.65169460e-01 -1.28517136e-01 -4.26297635e-01 -2.39882693e-02
-1.12477505e+00 -4.29236442e-01 3.36481065e-01 1.22081494e+00
8.60988796e-01 3.68016154e-01 4.52492326e-01 -1.70116186e+00
-4.76792157e-01 1.25733650e+00 -1.23885715e+00 5.10432661e-01
1.61119968e-01 1.29718417e-02 8.21548700e-01 -4.35225397e-01
2.67274827e-01 1.16170180e+00 1.56293854e-01 -3.38865995e-01
-1.16372252e+00 -6.56671047e-01 1.27763525e-01 2.49921028e-02
-1.40350997e+00 5.39782047e-01 8.19055378e-01 -3.21630985e-01
1.54760778e+00 3.35774302e-01 1.35737002e+00 -3.99588138e-01
2.59414852e-01 3.41587424e-01 8.13999355e-01 -3.89996737e-01
2.59063482e-01 -6.20546974e-02 -3.29355091e-01 -2.44356856e-01
6.78278983e-01 1.51212677e-01 2.86690384e-01 8.30059424e-02
5.19667566e-01 -1.50250241e-01 -3.74778301e-01 3.82540435e-01
-8.59775007e-01 9.69689727e-01 7.00266302e-01 6.13032699e-01
-5.97409785e-01 -1.68798134e-01 1.67653695e-01 -1.76357422e-02
5.64075053e-01 6.27662301e-01 -6.16877913e-01 1.12900369e-01
-1.29460716e+00 3.84147555e-01 4.16390866e-01 9.30206656e-01
3.73185426e-01 1.14649475e+00 -5.16785607e-02 6.33426070e-01
2.71701276e-01 1.20818973e+00 2.71165848e-01 -4.23421741e-01
6.43003047e-01 3.75776082e-01 5.70468485e-01 -8.30632985e-01
-7.05200374e-01 -7.62189329e-01 -1.06703651e+00 4.38068300e-01
1.91966780e-02 -1.19594657e+00 -3.86647403e-01 1.46222651e+00
2.13396307e-02 -1.53338179e-01 2.42419299e-02 5.46613932e-01
-6.90739453e-02 1.28743577e+00 -1.67530686e-01 -1.18533564e+00
1.27766442e+00 -8.65387321e-01 -1.16850233e+00 1.88079014e-01
4.10374641e-01 -8.05423021e-01 3.57446410e-02 5.30475378e-02
-1.41994035e+00 -3.01696092e-01 -1.00050652e+00 8.12108636e-01
-6.15660846e-01 2.55120188e-01 4.42958534e-01 6.08750582e-01
-9.52016234e-01 4.67947692e-01 -7.37570703e-01 -3.02012321e-02
1.89334169e-01 3.01697880e-01 2.54839063e-01 4.95671362e-01
-1.35354149e+00 1.19066930e+00 1.02038968e+00 5.24443984e-01
-3.84288967e-01 -8.91266882e-01 -8.44362080e-01 6.81762159e-01
2.50671118e-01 -7.10514009e-01 9.47134435e-01 -2.41489843e-01
-1.27994931e+00 -5.02389193e-01 -1.16994031e-01 -3.78616154e-01
7.07636401e-02 3.81981850e-01 -7.19630241e-01 -4.21856195e-02
1.98257059e-01 -4.05722484e-02 3.85464542e-02 -1.10830939e+00
-1.02654910e+00 -2.04940692e-01 -1.47598252e-01 5.88353157e-01
1.66713241e-02 7.28600770e-02 5.47328651e-01 -9.90491509e-01
-4.47734833e-01 -4.93576735e-01 -9.45892692e-01 -1.26817691e+00
-5.01468658e-01 -9.14664716e-02 9.22783494e-01 -9.77762401e-01
1.67397511e+00 -1.46027684e+00 1.59485564e-01 6.10215187e-01
-6.08787119e-01 2.44925022e-01 1.00757293e-01 1.04084826e+00
-3.83496881e-01 6.48091674e-01 1.76901277e-02 8.84649754e-02
2.84622520e-01 5.69752932e-01 -3.02665651e-01 -1.03419237e-01
-1.60996288e-01 8.11339736e-01 -1.11049390e+00 2.24370480e-01
7.69673944e-01 -5.40031530e-02 -2.74928093e-01 -3.66706401e-02
-1.30960405e-01 2.24242777e-01 -5.29886186e-01 4.57531929e-01
1.32198751e+00 1.20453194e-01 2.77586877e-01 -1.99553579e-01
-4.92059052e-01 -5.20855308e-01 -1.39483023e+00 1.08604813e+00
-8.20395648e-01 6.30456358e-02 1.68011948e-01 -1.13811600e+00
7.33327627e-01 4.17148948e-01 9.75990295e-01 -4.46298897e-01
-2.23210916e-01 4.82405759e-02 3.08654606e-01 -3.76455754e-01
6.14929914e-01 -2.28093520e-01 2.05947697e-01 9.10267830e-01
-1.59012780e-01 -6.71544313e-01 9.59968388e-01 -6.27525747e-02
1.11072369e-01 -4.04840976e-01 6.59735322e-01 -9.02282536e-01
4.12535667e-01 -1.11040659e-01 1.02752638e+00 -9.70191881e-03
2.92069077e-01 -2.10444495e-01 4.43482906e-01 -6.48375079e-02
-1.11154091e+00 -1.11854446e+00 -2.65078902e-01 3.44884843e-01
1.61679760e-02 -6.34268522e-02 -3.49942088e-01 -2.63085932e-01
2.18912348e-01 1.28467226e+00 -1.23624615e-01 5.48832893e-01
-3.17552060e-01 -1.80605841e+00 -4.17360753e-01 6.47180319e-01
6.64656103e-01 -6.23331964e-01 -3.54656905e-01 5.14018834e-01
3.79040763e-02 -7.14224219e-01 -1.70833588e-01 2.17359737e-01
-5.16507030e-01 -6.46701038e-01 -7.47433424e-01 -6.64215684e-01
1.06085432e+00 -5.11837974e-02 9.73778009e-01 1.08933216e-02
-2.02412263e-01 -1.50878936e-01 -3.58151287e-01 -5.99753320e-01
2.16871724e-01 -9.25476551e-02 2.20640481e-01 -6.74284697e-01
-4.92528051e-01 -8.17122698e-01 -6.24949455e-01 2.58570462e-01
-7.82316148e-01 8.68978575e-02 -1.09582298e-01 9.42784846e-01
3.39794040e-01 1.65472937e+00 1.48716307e+00 -4.08705324e-01
1.00529170e+00 -6.74309433e-01 -1.28498197e+00 7.03016520e-01
-1.01667440e+00 -3.61099333e-01 7.95535326e-01 9.82633084e-02
-1.48275006e+00 -2.00560585e-01 -4.12540883e-03 2.77549267e-01
2.30429709e-01 1.11331522e+00 -3.16691756e-01 8.45399871e-02
-4.41141903e-01 9.05648712e-03 -6.10036075e-01 -3.87494445e-01
2.56714195e-01 2.31521741e-01 -2.64047869e-02 -4.61666375e-01
1.41350663e+00 4.07844335e-02 3.37600201e-01 -4.95297343e-01
-1.19971998e-01 -1.78900659e-01 -4.68194216e-01 -1.93708807e-01
7.04926670e-01 -1.05426478e+00 -5.51306307e-01 5.42760074e-01
-8.39363515e-01 -2.99086690e-01 -5.37458241e-01 7.71394134e-01
-1.58970118e-01 1.27782598e-01 -3.13286126e-01 -1.22442222e+00
-4.28219885e-01 -1.42347074e+00 2.29885191e-01 6.16206110e-01
5.13164103e-01 -1.65244782e+00 8.42715576e-02 -3.71316969e-02
8.49572599e-01 6.24398172e-01 9.20359015e-01 7.77757308e-03
-4.41676348e-01 2.53356844e-01 -1.86461225e-01 2.48424158e-01
5.26541114e-01 4.11469012e-01 -2.79265016e-01 -6.71349108e-01
-1.97282404e-01 2.54726678e-01 -1.42121315e-01 8.97478163e-01
7.90246248e-01 -7.31886804e-01 -2.31105149e-01 5.42140603e-01
2.39564610e+00 9.19840753e-01 3.94669265e-01 2.11866960e-01
1.13963082e-01 1.41119257e-01 7.25504458e-01 1.20232213e+00
4.24174011e-01 4.72076774e-01 4.56866950e-01 -3.09687167e-01
6.04822040e-01 2.09766757e-02 -8.18743557e-02 8.93362045e-01
-5.44089735e-01 -7.23214507e-01 -7.09027946e-01 7.18073666e-01
-1.91039610e+00 -1.48315239e+00 -6.53849542e-02 1.83621979e+00
3.47305685e-01 -3.65563959e-01 7.78909177e-02 1.97891608e-01
7.36920774e-01 1.74212158e-01 -1.26328945e-01 -7.32522428e-01
-4.20193791e-01 1.93684012e-01 8.55049729e-01 8.08357060e-01
-6.57622278e-01 2.50556439e-01 6.49687338e+00 1.24428546e+00
-8.92695844e-01 -2.50807405e-01 8.20439100e-01 -8.24997425e-02
-7.63247430e-01 -3.10702212e-02 -7.65173495e-01 8.34736049e-01
4.80177611e-01 -1.55758524e+00 9.51970637e-01 3.17253232e-01
1.41995311e+00 -6.36392236e-01 -3.32638830e-01 4.28073376e-01
-6.64851606e-01 -1.24562490e+00 -1.06156200e-01 2.89525807e-01
1.71437299e+00 1.11236520e-01 -3.41891974e-01 1.06125429e-01
8.71607542e-01 -8.89904737e-01 2.13430166e-01 3.48844409e-01
2.36072406e-01 -1.29446113e+00 1.26340771e+00 3.38677675e-01
-1.82331395e+00 -5.80902278e-01 -2.74998069e-01 -4.79328841e-01
1.38734305e+00 1.10780787e+00 -7.31403172e-01 1.53526759e+00
5.74893296e-01 6.78685844e-01 1.18776731e-01 1.17357457e+00
-5.77177107e-01 5.62108755e-01 -7.09729135e-01 2.58558482e-01
1.58865318e-01 -8.56620967e-01 4.98612225e-01 5.79886913e-01
9.56619620e-01 2.81003535e-01 6.09852076e-01 6.72418237e-01
2.96296805e-01 1.46259323e-01 -4.86472428e-01 1.11992791e-01
9.14334714e-01 1.38534725e+00 -7.53126740e-01 -1.77366629e-01
-2.25350961e-01 -7.64049739e-02 -3.88930678e-01 9.28008139e-01
-7.64983773e-01 -4.45282280e-01 4.59912807e-01 -3.77262473e-01
1.20931298e-01 -2.22200856e-01 -7.16820359e-01 -1.16869557e+00
-9.66313109e-02 -6.16308600e-02 5.36661148e-01 -1.03954995e+00
-1.30566645e+00 3.20135146e-01 6.05984330e-01 -1.09347844e+00
-8.37853253e-01 -1.39478430e-01 -1.50057650e+00 1.54825020e+00
-2.03452015e+00 -1.05338931e+00 3.19874793e-01 3.33832860e-01
6.01062179e-01 -3.34347099e-01 6.76276505e-01 2.21742570e-01
-1.10313725e+00 -3.53256543e-03 8.85344326e-01 -4.24997322e-02
-5.18677175e-01 -1.60099864e+00 -3.10512543e-01 1.37986290e+00
-5.18675387e-01 2.14273930e-01 5.98449707e-01 -7.62511313e-01
-9.98065591e-01 -1.01846683e+00 7.43699610e-01 6.81102753e-01
1.01921606e+00 2.64346093e-01 -3.13535959e-01 5.52766740e-01
1.31368279e+00 -2.82617122e-01 5.40135622e-01 -8.48184302e-02
7.29158223e-01 -3.35265368e-01 -1.65721476e+00 4.26333785e-01
1.37224212e-01 1.47900686e-01 -3.18129152e-01 5.10577023e-01
3.26754868e-01 -7.00114528e-03 -1.54241371e+00 6.41495049e-01
1.13932461e-01 -1.20095640e-01 9.59855318e-01 -2.80693471e-02
-2.19334736e-01 -6.35607600e-01 1.65247805e-02 -2.60593271e+00
-4.95367438e-01 -8.08206916e-01 3.32423121e-01 1.55538702e+00
7.04959869e-01 -9.85065103e-01 2.11953029e-01 5.35101652e-01
1.81937993e-01 -6.66974664e-01 -1.06906843e+00 -1.01893735e+00
3.15953821e-01 -1.73814539e-02 1.45562780e+00 1.10999501e+00
8.26771379e-01 -6.13305792e-02 -2.19667688e-01 7.35038757e-01
6.30295157e-01 5.03180802e-01 1.73947196e-02 -3.82572532e-01
4.75341082e-02 -4.76169139e-01 1.54464915e-01 -9.94903073e-02
-9.63486657e-02 -6.06761873e-01 -4.17045027e-01 -2.12185693e+00
-1.14454508e-01 -4.85229313e-01 -1.13126568e-01 6.34446681e-01
-1.53844818e-01 -1.40006259e-01 5.51331460e-01 -2.44583040e-01
3.84954721e-01 1.01468730e+00 1.36784422e+00 -1.06406480e-01
-2.83709884e-01 4.20016378e-01 -1.67937070e-01 6.14249647e-01
1.30500674e+00 6.69678599e-02 -8.76729965e-01 -5.89241683e-01
4.58188891e-01 7.62538195e-01 -4.68526334e-01 -8.01605046e-01
3.00851390e-02 -8.44369173e-01 2.45236531e-01 -1.16833615e+00
-2.08021685e-01 -1.37581563e+00 8.71828496e-01 6.60755217e-01
1.84458390e-01 8.24069440e-01 3.57153535e-01 2.21523061e-01
-3.67141426e-01 -5.18365502e-01 4.71930772e-01 -4.18235399e-02
-3.59733254e-01 3.66341680e-01 -5.16005635e-01 -3.05623412e-01
1.58867371e+00 2.41097942e-01 -7.35451102e-01 -5.14139056e-01
-1.06932259e+00 1.34547591e+00 7.49059990e-02 -4.32884088e-03
1.52416795e-01 -1.36312604e+00 -9.01276767e-01 -3.37412432e-02
-8.60446513e-01 -8.14824253e-02 3.90461266e-01 3.83721530e-01
-7.45730996e-01 5.02535999e-01 -3.05232882e-01 -7.31917694e-02
-7.22388685e-01 5.16227782e-01 4.73915696e-01 -1.06471419e+00
1.61054373e-01 1.92719683e-01 -3.64885688e-01 -1.87511683e-01
-4.98158962e-01 -4.20324057e-01 -5.60541511e-01 4.97704774e-01
4.28257436e-01 8.76105845e-01 -2.54311353e-01 -4.81639475e-01
-2.59132311e-02 4.86989290e-01 7.18774319e-01 -1.55111730e-01
1.59920192e+00 -7.42870688e-01 -3.85286063e-01 -1.37883320e-01
6.09035254e-01 1.57724649e-01 -7.00062931e-01 3.43874395e-01
-4.41747546e-01 -5.28379679e-01 2.99069762e-01 -1.34281075e+00
-1.89670610e+00 3.51432353e-01 1.81245387e-01 8.15731406e-01
1.63511074e+00 -8.90296698e-01 5.26873112e-01 -1.93631992e-01
6.08540714e-01 -1.20257366e+00 -8.53032291e-01 3.12958598e-01
9.25296545e-01 -8.26536894e-01 4.84060377e-01 -4.22196031e-01
-3.75107735e-01 1.13058937e+00 4.65007782e-01 -3.02880630e-02
1.15460896e+00 1.01889944e+00 -4.44017053e-01 3.80161196e-01
-7.61073291e-01 -2.73388952e-01 -1.83436215e-01 7.03782201e-01
2.93469448e-02 7.54242301e-01 -7.93091536e-01 6.39491856e-01
-3.03597122e-01 1.87732890e-01 8.09794366e-01 7.36741483e-01
-2.57912248e-01 -1.18339574e+00 -4.70431894e-01 5.53542852e-01
-4.48533356e-01 -2.76280999e-01 8.96666944e-01 5.08343935e-01
5.80679178e-01 1.31843734e+00 2.18685925e-01 1.48427978e-01
1.25018418e-01 -5.25106072e-01 -3.66076007e-02 -4.01412755e-01
-6.14020288e-01 4.49518204e-01 9.82160345e-02 1.84332982e-01
-3.43359172e-01 -6.71578705e-01 -1.35477984e+00 -3.49006534e-01
-8.18766594e-01 9.84678090e-01 6.06070220e-01 8.93682480e-01
-1.29760802e-01 7.46425271e-01 1.40164757e+00 -1.06722498e+00
-3.96752000e-01 -8.90160918e-01 -1.36601341e+00 -3.29565316e-01
-2.96919703e-01 -5.59412479e-01 -4.60747898e-01 -3.75636011e-01] | [5.6439666748046875, 2.5525319576263428] |
8c9e1e26-9c9b-471a-929a-c2de1f023004 | block-coordinate-plug-and-play-methods-for | 2305.12672 | null | https://arxiv.org/abs/2305.12672v1 | https://arxiv.org/pdf/2305.12672v1.pdf | Block Coordinate Plug-and-Play Methods for Blind Inverse Problems | Plug-and-play (PnP) prior is a well-known class of methods for solving imaging inverse problems by computing fixed-points of operators combining physical measurement models and learned image denoisers. While PnP methods have been extensively used for image recovery with known measurement operators, there is little work on PnP for solving blind inverse problems. We address this gap by presenting a new block-coordinate PnP (BC-PnP) method that efficiently solves this joint estimation problem by introducing learned denoisers as priors on both the unknown image and the unknown measurement operator. We present a new convergence theory for BC-PnP compatible with blind inverse problems by considering nonconvex data-fidelity terms and expansive denoisers. Our theory analyzes the convergence of BC-PnP to a stationary point of an implicit function associated with an approximate minimum mean-squared error (MMSE) denoiser. We numerically validate our method on two blind inverse problems: automatic coil sensitivity estimation in magnetic resonance imaging (MRI) and blind image deblurring. Our results show that BC-PnP provides an efficient and principled framework for using denoisers as PnP priors for jointly estimating measurement operators and images. | ['Ulugbek S. Kamilov', 'Hongyu An', 'Jiaming Liu', 'Yuyang Hu', 'Shirin Shoushtari', 'Weijie Gan'] | 2023-05-22 | null | null | null | null | ['deblurring', 'blind-image-deblurring'] | ['computer-vision', 'computer-vision'] | [ 5.17641962e-01 1.15451440e-01 2.89104313e-01 -1.35111675e-01
-1.14409161e+00 -1.93545356e-01 2.36554846e-01 -7.10074067e-01
-4.34537739e-01 8.08877409e-01 6.26373351e-01 -7.81758726e-02
-5.87093711e-01 -2.41465017e-01 -8.29871893e-01 -1.11760676e+00
-2.43553184e-02 4.96461689e-01 -2.08215326e-01 -1.43168628e-01
8.95953923e-02 2.29319185e-01 -8.60789716e-01 -8.55978206e-02
1.14239430e+00 6.73362732e-01 4.87148106e-01 6.95000827e-01
5.63201666e-01 6.90351963e-01 -2.37898245e-01 -1.91394597e-01
6.32264555e-01 -8.95374000e-01 -7.63749897e-01 3.98511849e-02
-8.21215659e-02 -4.59162652e-01 -4.72611934e-01 1.48800647e+00
9.07073081e-01 3.25000584e-01 9.17885423e-01 -7.27790356e-01
-8.77374113e-01 4.36372757e-01 -8.75771582e-01 3.30109686e-01
2.79344082e-01 4.73194011e-02 2.08550856e-01 -8.94567788e-01
4.52949107e-01 9.39031482e-01 1.10836697e+00 5.48545599e-01
-1.32167101e+00 -1.65292606e-01 -7.07247972e-01 4.69616532e-01
-1.39910173e+00 -7.20209539e-01 6.38025641e-01 -3.82768363e-01
5.33051133e-01 6.78825736e-01 4.30820614e-01 6.39506459e-01
4.09191072e-01 5.79254687e-01 1.64723802e+00 -6.09690547e-01
2.85370946e-01 -1.95143089e-01 -1.57675743e-01 5.27238607e-01
-4.22197096e-02 2.73173571e-01 -3.19546372e-01 -2.92514354e-01
1.25429654e+00 -2.51786917e-01 -9.44292426e-01 -3.40678990e-01
-1.38442183e+00 6.18345141e-01 4.61833358e-01 3.78846347e-01
-1.03120863e+00 1.77843943e-01 3.21984431e-03 2.50869989e-01
5.43824375e-01 4.12276328e-01 9.24723893e-02 1.24050163e-01
-1.14609253e+00 8.20420608e-02 7.73503065e-01 5.85177183e-01
4.31180954e-01 8.18697959e-02 -3.34742367e-01 1.10043609e+00
3.69833767e-01 1.04035604e+00 5.81206262e-01 -1.51839113e+00
1.89608466e-02 -6.19255245e-01 3.40489179e-01 -9.52439249e-01
-2.05826581e-01 -6.11582279e-01 -1.29291975e+00 1.02272376e-01
3.37537587e-01 -3.99741121e-02 -9.10614491e-01 1.79358661e+00
3.85130465e-01 4.98442084e-01 -8.82546008e-02 1.54314482e+00
3.86898637e-01 6.82618499e-01 -2.13423744e-01 -7.15998888e-01
1.37960041e+00 -7.08239257e-01 -1.08192396e+00 -2.65461475e-01
-4.72123884e-02 -8.42971444e-01 4.77233231e-01 4.10645962e-01
-1.65814686e+00 -2.00313553e-01 -9.55349803e-01 -2.61219759e-02
6.28675222e-01 -1.88908547e-01 1.65407479e-01 6.44558311e-01
-1.41534102e+00 6.27791882e-01 -9.02824700e-01 -6.02731220e-02
2.45081767e-01 2.44848013e-01 -3.23202699e-01 -3.67167056e-01
-1.13875794e+00 1.40695298e+00 -1.90345615e-01 5.97115040e-01
-1.20687175e+00 -9.80270565e-01 -5.93722820e-01 -2.85845518e-01
3.56142409e-02 -1.07753789e+00 1.07655561e+00 -9.94652212e-01
-1.72451210e+00 7.65727937e-01 -2.72138089e-01 -4.18724149e-01
6.65523469e-01 -1.72886282e-01 -3.38509470e-01 4.61542189e-01
3.88565600e-01 2.66163737e-01 1.55771828e+00 -1.35716975e+00
4.49141771e-01 -3.40816557e-01 -5.14688075e-01 4.91168916e-01
1.34707898e-01 8.43069050e-03 -1.84596360e-01 -7.48788893e-01
7.26649284e-01 -6.46789670e-01 -4.93032873e-01 5.00984164e-03
-3.54055285e-01 5.90595841e-01 1.15823336e-01 -1.53922462e+00
8.18692029e-01 -1.99147308e+00 6.59872115e-01 4.24975842e-01
3.13432276e-01 -3.75124440e-02 -1.30647317e-01 -2.64651421e-02
-4.47287977e-01 -3.46740603e-01 -8.79752696e-01 -1.13015197e-01
-1.03984475e-01 3.64402235e-01 -8.29263031e-02 1.21959150e+00
-4.95459646e-01 9.30976510e-01 -9.74993765e-01 -2.89191216e-01
2.59633482e-01 7.95416117e-01 -5.75114131e-01 2.39610672e-01
4.12474900e-01 1.16718805e+00 -1.96476966e-01 2.44579285e-01
9.88140941e-01 -1.85303748e-01 1.16868116e-01 -5.37164390e-01
2.57882606e-02 -2.75333703e-01 -1.20563030e+00 1.74759424e+00
-4.85052764e-01 4.43837672e-01 9.96147513e-01 -1.48755872e+00
3.72891784e-01 7.61431277e-01 7.93818474e-01 -7.11779475e-01
1.48704529e-01 5.82026482e-01 -2.92508185e-01 -9.01965141e-01
9.16125998e-02 -1.04284847e+00 4.69185323e-01 4.28343475e-01
1.28187224e-01 -3.77360731e-01 -2.45505854e-01 8.43076929e-02
1.27374303e+00 -1.58873931e-01 4.29606587e-01 -7.46510386e-01
5.26342630e-01 -3.86226207e-01 3.48823011e-01 1.05311954e+00
-2.85361230e-01 1.10575616e+00 -5.03454208e-02 3.58664915e-02
-1.17254567e+00 -1.00715160e+00 -5.81819713e-01 4.97004777e-01
1.36951521e-01 3.23050648e-01 -1.11716700e+00 6.28361339e-03
-4.05418545e-01 6.72963262e-01 -4.75681752e-01 1.81289732e-01
-6.62842810e-01 -1.22775292e+00 2.21221209e-01 4.96242195e-02
6.96323335e-01 -7.68899143e-01 -3.17232907e-01 3.23981583e-01
-9.07190919e-01 -9.10090983e-01 -7.26108968e-01 -7.39323944e-02
-9.33814764e-01 -9.27275062e-01 -1.58480334e+00 -6.88692570e-01
9.41346705e-01 2.08894968e-01 8.89181197e-01 -3.61433297e-01
-1.21483095e-01 8.00001204e-01 3.15746926e-02 1.12806864e-01
-7.09497571e-01 -7.46211231e-01 1.96339086e-01 3.57677668e-01
-3.74560535e-01 -9.20513630e-01 -9.29553568e-01 5.38862884e-01
-9.35003221e-01 8.05954412e-02 6.65273309e-01 9.39345002e-01
7.33310759e-01 -8.90878588e-02 3.42950672e-01 -3.88778001e-01
8.81424427e-01 -6.14058971e-01 -3.95712614e-01 2.17703104e-01
-4.52110857e-01 2.30010390e-01 1.45876676e-01 -4.24146056e-01
-1.22025383e+00 -1.02793304e-02 -3.20425183e-01 -5.52738667e-01
2.79095501e-01 4.62444663e-01 -5.24594821e-03 -4.55080926e-01
1.01534224e+00 4.98793691e-01 3.46309692e-01 -5.95929384e-01
5.71617484e-01 5.03814220e-01 1.18477809e+00 -6.18926108e-01
7.28429496e-01 6.60857916e-01 1.84937328e-01 -8.02796900e-01
-7.00031102e-01 -6.56913996e-01 -1.75543323e-01 -3.11962992e-01
1.08211410e+00 -7.24992037e-01 -6.66866779e-01 5.70530653e-01
-1.13044095e+00 -3.19541186e-01 -5.70997775e-01 7.85850406e-01
-1.02681148e+00 9.32130337e-01 -7.25500286e-01 -6.94092274e-01
-4.81436282e-01 -1.39851725e+00 7.88086951e-01 -1.94633782e-01
-5.00991009e-03 -1.16781652e+00 2.45707139e-01 4.79095221e-01
8.64039183e-01 1.67609751e-01 4.55415756e-01 1.97909400e-01
-3.53019834e-01 -5.67929298e-02 -1.82205200e-01 6.97630584e-01
-2.14895025e-01 -1.26323676e+00 -7.77246296e-01 -3.93602908e-01
1.36627269e+00 2.53794581e-01 6.47133231e-01 1.25798202e+00
9.79862034e-01 -6.28981292e-01 -1.41178951e-01 1.05940175e+00
1.57680428e+00 -2.25099772e-01 1.07519102e+00 1.35513350e-01
4.70682919e-01 1.62545443e-01 -1.12822175e-01 2.11390123e-01
-6.45834282e-02 7.08198845e-01 2.55391479e-01 -6.97105303e-02
-4.61257309e-01 5.60199767e-02 2.96009630e-01 1.17534530e+00
-4.48982716e-01 1.34530321e-01 -7.54778922e-01 7.43359983e-01
-1.71074843e+00 -8.35730135e-01 -4.76687521e-01 2.28676867e+00
9.62104976e-01 -7.51405120e-01 -4.20400798e-01 1.41733393e-01
8.96416306e-01 -2.12640792e-01 -5.03053367e-01 2.27022424e-01
-1.69382989e-01 4.22291428e-01 9.41108108e-01 1.04049015e+00
-7.98333108e-01 1.11419363e-02 6.91231060e+00 9.36208248e-01
-6.23518288e-01 1.02380025e+00 2.35918730e-01 3.26560587e-01
-3.35795105e-01 9.19142962e-02 1.32558316e-01 4.47665036e-01
7.17969418e-01 -2.55025718e-02 1.06061637e+00 2.30442196e-01
3.36380631e-01 -8.25686231e-02 -8.09161425e-01 1.44887650e+00
2.08532482e-01 -1.11642480e+00 -3.93835843e-01 1.62270460e-02
1.06420350e+00 1.79141313e-01 5.07603064e-02 -3.70292544e-01
1.81733429e-01 -1.01666248e+00 6.76904380e-01 9.24522936e-01
1.00396490e+00 -1.40076399e-01 6.03051722e-01 2.85684228e-01
-4.60082918e-01 1.50998950e-01 -3.26925129e-01 1.47433639e-01
7.44412780e-01 1.12198281e+00 -3.56071889e-01 6.01251721e-01
4.63496387e-01 7.51347899e-01 2.83910930e-01 1.27000928e+00
-3.94093007e-01 4.39488262e-01 -3.99681687e-01 6.82823062e-01
-1.32007390e-01 -6.68887019e-01 1.15853512e+00 8.17936242e-01
6.16302252e-01 6.63472056e-01 -4.09016103e-01 1.08963406e+00
8.16716552e-02 -1.45867601e-01 -1.65529214e-02 5.45109034e-01
6.40157834e-02 1.00425184e+00 -6.13562822e-01 -1.93337366e-01
4.14132588e-02 1.19384015e+00 -3.30634981e-01 8.79186034e-01
-6.87173069e-01 1.23065405e-01 2.35346124e-01 2.86383957e-01
1.44261882e-01 -4.57363427e-02 -4.32383180e-01 -1.44704282e+00
7.60176107e-02 -9.83327806e-01 7.89737776e-02 -1.10257113e+00
-1.53380501e+00 3.25804830e-01 1.45514861e-01 -1.13969862e+00
-2.26827487e-02 -3.25364590e-01 -3.07496905e-01 1.24925709e+00
-1.33823311e+00 -7.06043661e-01 -1.94031358e-01 8.02417338e-01
1.01110093e-01 2.04958022e-01 7.10157812e-01 4.65122432e-01
9.53749046e-02 7.97562674e-02 6.86856985e-01 -2.10312858e-01
5.06767392e-01 -1.09257662e+00 -2.29569018e-01 1.03507316e+00
-3.99244010e-01 6.33774817e-01 1.13526046e+00 -5.45317352e-01
-1.56545317e+00 -4.18707162e-01 3.50436360e-01 -2.54699707e-01
5.14335990e-01 2.28941426e-01 -8.81579459e-01 6.94372535e-01
3.61680388e-01 1.37718260e-01 2.05976248e-01 -4.57266331e-01
1.97397575e-01 -1.09102456e-02 -1.56992924e+00 2.55212694e-01
9.02974725e-01 -4.54403192e-01 -5.75323522e-01 7.12251365e-01
1.96977202e-02 -8.10354412e-01 -1.02936363e+00 4.51678216e-01
2.51055777e-01 -8.77561450e-01 1.36947548e+00 1.93905950e-01
3.59114528e-01 -3.56405109e-01 -1.37552813e-01 -1.66856384e+00
-3.23891759e-01 -1.10484171e+00 -1.62049696e-01 4.87974405e-01
-1.37014851e-01 -8.94829690e-01 2.36817807e-01 4.95189816e-01
-4.44272608e-01 -3.03455472e-01 -1.31311357e+00 -6.57056510e-01
9.81708765e-02 -5.38383722e-01 1.27618074e-01 9.81266022e-01
-2.29105577e-02 -4.57289703e-02 -7.69719481e-01 3.35696936e-01
1.38239753e+00 -5.41583180e-01 5.51609695e-02 -4.79933202e-01
-9.16476607e-01 -2.51123101e-01 -2.52259940e-01 -1.42928040e+00
-2.24330425e-02 -1.13784182e+00 6.42256498e-01 -1.72191858e+00
4.24257547e-01 -4.11189467e-01 -3.39041762e-02 2.28284374e-02
-9.44551006e-02 3.52173984e-01 -9.78298634e-02 6.82504296e-01
-1.69231653e-01 4.26057160e-01 1.71694434e+00 -9.94771868e-02
9.25822183e-02 -1.62817724e-02 -5.43585122e-01 6.39251471e-01
2.09491685e-01 -5.66789448e-01 -3.37552786e-01 -6.90609455e-01
2.04429567e-01 6.18471980e-01 6.93916559e-01 -9.50395286e-01
4.81290460e-01 2.26382360e-01 2.33654514e-01 1.12629659e-01
2.89018035e-01 -7.32666373e-01 7.91227102e-01 5.23527026e-01
-1.62857786e-01 -5.16439676e-01 -1.77017525e-01 6.19376540e-01
-6.59524128e-02 -6.02516055e-01 1.00539172e+00 -3.29515517e-01
-2.79223889e-01 -4.22940254e-02 -3.73121470e-01 6.83879778e-02
4.06713873e-01 -5.41830435e-03 1.54659594e-03 -7.33538866e-01
-1.30662680e+00 -2.15789303e-01 1.26235828e-01 -3.58862936e-01
8.34285855e-01 -1.28414524e+00 -1.17826867e+00 3.77199143e-01
-6.60631120e-01 -2.24714026e-01 6.30454600e-01 1.60695088e+00
-8.11795652e-01 -3.21236476e-02 -3.86948586e-02 -6.99714541e-01
-7.06894159e-01 6.35806561e-01 9.22943592e-01 -2.14677900e-01
-8.28416407e-01 9.45648372e-01 3.10574919e-01 -4.43828076e-01
-1.11899488e-01 9.46597755e-02 3.21938396e-01 -5.61341584e-01
7.30141521e-01 6.55888736e-01 1.04884401e-01 -8.01157832e-01
-2.26586699e-01 6.51115656e-01 4.00810480e-01 -7.20091581e-01
1.42469132e+00 -4.54526275e-01 -7.93991506e-01 -4.63074930e-02
1.32085562e+00 -1.15846731e-01 -1.31729996e+00 -2.28651196e-01
-3.95020187e-01 -4.16260391e-01 5.18227100e-01 -8.74329388e-01
-1.26988423e+00 4.44087327e-01 9.22363400e-01 2.72267386e-02
1.46914053e+00 -1.58205966e-03 7.42461681e-01 -1.96050465e-01
3.79482657e-01 -7.35732138e-01 -2.17501879e-01 8.65179524e-02
1.45472658e+00 -9.43868697e-01 2.55110621e-01 -4.27922219e-01
-2.04086736e-01 6.78141534e-01 -3.92141402e-01 -3.33133072e-01
8.98019671e-01 1.62379771e-01 3.74400169e-02 -1.64913327e-01
3.42406720e-01 1.63591444e-01 4.85853821e-01 7.07766950e-01
4.83985804e-02 2.53447384e-01 -7.69849539e-01 2.72978038e-01
4.76852506e-02 2.63577014e-01 4.91537571e-01 6.06899858e-01
-2.07337454e-01 -8.18392456e-01 -1.06785917e+00 2.50225157e-01
-4.95197564e-01 -3.63999724e-01 4.99847412e-01 9.15896967e-02
-1.46727547e-01 8.60574424e-01 -4.64040935e-01 2.26651937e-01
7.05943331e-02 -1.65543258e-01 1.03385007e+00 -1.86871201e-01
-1.81983635e-01 4.12229657e-01 -3.16189408e-01 -5.18622100e-01
-8.04453969e-01 -7.04229295e-01 -9.74238276e-01 -2.64154315e-01
-1.76510260e-01 2.91973740e-01 8.85938168e-01 1.08416951e+00
7.10271820e-02 4.58427429e-01 3.87643903e-01 -1.17779768e+00
-7.88931787e-01 -1.03225684e+00 -8.41220677e-01 4.82649833e-01
5.56370854e-01 -3.28320920e-01 -6.15474522e-01 1.47993237e-01] | [11.77672290802002, -2.4311134815216064] |
d8a97c95-de81-4df3-99ac-84d9e474391b | beyond-ann-exploiting-structural-knowledge | 2103.08366 | null | https://arxiv.org/abs/2103.08366v1 | https://arxiv.org/pdf/2103.08366v1.pdf | Beyond ANN: Exploiting Structural Knowledge for Efficient Place Recognition | Visual place recognition is the task of recognizing same places of query images in a set of database images, despite potential condition changes due to time of day, weather or seasons. It is important for loop closure detection in SLAM and candidate selection for global localization. Many approaches in the literature perform computationally inefficient full image comparisons between queries and all database images. There is still a lack of suited methods for efficient place recognition that allow a fast, sparse comparison of only the most promising image pairs without any loss in performance. While this is partially given by ANN-based methods, they trade speed for precision and additional memory consumption, and many cannot find arbitrary numbers of matching database images in case of loops in the database. In this paper, we propose a novel fast sequence-based method for efficient place recognition that can be applied online. It uses relocalization to recover from sequence losses, and exploits usually available but often unused intra-database similarities for a potential detection of all matching database images for each query in case of loops or stops in the database. We performed extensive experimental evaluations over five datasets and 21 sequence combinations, and show that our method outperforms two state-of-the-art approaches and even full image comparisons in many cases, while providing a good tradeoff between performance and percentage of evaluated image pairs. Source code for Matlab will be provided with publication of this paper. | ['Peter Protzel', 'Peer Neubert', 'Stefan Schubert'] | 2021-03-15 | null | null | null | null | ['loop-closure-detection', 'visual-place-recognition'] | ['computer-vision', 'computer-vision'] | [ 2.14966297e-01 -8.16015482e-01 -9.87039134e-02 -2.44127259e-01
-7.38941967e-01 -8.00104141e-01 4.72023070e-01 6.57915354e-01
-8.58188689e-01 7.00168252e-01 -4.26394731e-01 -1.34805560e-01
-2.78412938e-01 -6.51581347e-01 -8.32686067e-01 -5.87188065e-01
-3.86557400e-01 4.34122622e-01 8.78540099e-01 -1.88809171e-01
6.36292815e-01 1.02669466e+00 -1.96278656e+00 -3.69504094e-02
6.21042252e-01 1.04469156e+00 7.00471222e-01 6.43292606e-01
-4.25793715e-02 6.10620320e-01 -4.80853677e-01 6.54611960e-02
4.52323973e-01 -3.05628359e-01 -5.27158380e-01 6.66252822e-02
6.25731230e-01 -5.63383810e-02 -3.19837272e-01 1.09354448e+00
6.44590080e-01 3.16453516e-01 3.37114394e-01 -1.33303523e+00
-2.00351141e-02 -1.69873074e-01 -4.89695430e-01 3.92378688e-01
8.17721307e-01 1.49866372e-01 6.35231316e-01 -1.10534847e+00
9.33099210e-01 8.16322923e-01 6.68519139e-01 -1.29151613e-01
-1.20922768e+00 -5.53772986e-01 -3.30570012e-01 6.86061800e-01
-2.02099323e+00 -5.74349403e-01 5.75810909e-01 -2.65224546e-01
1.08775485e+00 5.36220253e-01 6.97125733e-01 3.30982774e-01
1.07858710e-01 4.96744096e-01 9.87248421e-01 -6.06418312e-01
3.20121258e-01 9.89236459e-02 -3.31776142e-01 7.79933393e-01
1.46330982e-01 1.62492067e-01 -9.81349945e-01 -3.28911006e-01
6.57097936e-01 1.86118305e-01 -3.18029165e-01 -8.94098401e-01
-1.67410469e+00 5.39322019e-01 7.17094719e-01 4.01546627e-01
-6.00540102e-01 -2.40681996e-03 1.51781097e-01 4.59517300e-01
-1.56976461e-01 3.55846852e-01 -8.38190168e-02 -1.09810382e-01
-1.27525115e+00 3.24675530e-01 6.38534665e-01 1.05894852e+00
1.34443164e+00 -4.18891281e-01 1.34145081e-01 6.50499880e-01
8.59054774e-02 6.82382286e-01 4.22115237e-01 -6.06239140e-01
3.22925091e-01 6.86415374e-01 2.55172163e-01 -1.53992593e+00
-4.21837986e-01 -1.29160390e-03 -6.24552846e-01 2.09218577e-01
2.81272382e-01 4.11071450e-01 -6.97905242e-01 1.26940274e+00
4.74108398e-01 1.01597376e-01 2.35529020e-02 9.69030678e-01
4.46685344e-01 6.16800368e-01 -4.83752578e-01 -2.32463941e-01
1.16869140e+00 -7.77532935e-01 -3.78932387e-01 -5.56528151e-01
4.45057571e-01 -1.23261476e+00 5.68135560e-01 5.86320795e-02
-6.19552732e-01 -4.96554464e-01 -1.15745962e+00 2.65903622e-01
-5.78564823e-01 1.41875744e-01 3.76693219e-01 2.96809018e-01
-1.11981893e+00 5.31311452e-01 -5.24160326e-01 -9.75977540e-01
-1.76771909e-01 6.51169360e-01 -7.25731492e-01 -1.29921660e-01
-7.82898962e-01 9.42549229e-01 5.79572260e-01 3.02091926e-01
-6.58879399e-01 -1.78640544e-01 -9.43252444e-01 -3.45497310e-01
2.97608882e-01 -2.80869186e-01 8.60945761e-01 -9.18585896e-01
-8.51092815e-01 1.01743484e+00 -4.98200476e-01 -7.04822659e-01
6.69717073e-01 2.97134984e-02 -4.12537694e-01 1.82033956e-01
2.58248448e-01 8.52029324e-01 6.80804491e-01 -1.06550848e+00
-9.13180053e-01 -4.02316421e-01 -3.07322830e-01 3.87983769e-01
2.69247562e-01 -4.16262150e-02 -7.85969913e-01 -3.05120736e-01
6.95644379e-01 -1.14046323e+00 -3.55394572e-01 3.51526707e-01
-9.71055478e-02 2.62569636e-01 8.00010920e-01 -4.23135698e-01
8.94502342e-01 -2.09594560e+00 -3.31640601e-01 5.16230106e-01
-4.35653925e-01 2.04542533e-01 -1.36970162e-01 8.95547330e-01
2.24221662e-01 -3.89145970e-01 -2.57815689e-01 -1.15665428e-01
-3.56707782e-01 3.95383954e-01 -3.61408055e-01 1.00213873e+00
-1.50959760e-01 5.64723730e-01 -8.89886141e-01 -6.01903081e-01
6.81719005e-01 3.01180869e-01 -1.61420897e-01 1.02119550e-01
2.22695395e-01 3.94907027e-01 -2.18739390e-01 8.13033938e-01
7.86254764e-01 1.01243496e-01 4.45293561e-02 -1.22165345e-01
-3.87639552e-01 1.18732505e-01 -1.67276657e+00 1.81278050e+00
-2.92947948e-01 8.26834679e-01 -1.61185965e-01 -8.44960809e-01
1.14853799e+00 2.41908198e-03 4.26417798e-01 -1.09666538e+00
-2.99716264e-01 7.05087245e-01 -1.92028552e-01 -2.15298474e-01
8.98393750e-01 4.31219339e-01 8.71215463e-02 1.00047134e-01
-1.72479942e-01 -1.43629864e-01 3.45840693e-01 -1.59683198e-01
1.02933478e+00 4.28423136e-02 5.85794449e-01 -2.15498880e-01
7.98259854e-01 3.97946000e-01 3.92010778e-01 1.01297641e+00
-4.14512545e-01 6.78203940e-01 -1.41663969e-01 -7.07018971e-01
-1.02072763e+00 -7.34658539e-01 -8.29805434e-03 6.86028421e-01
9.20514882e-01 -2.46273071e-01 -2.02356979e-01 -2.19185829e-01
1.04401007e-01 2.11496145e-01 -2.99741924e-01 8.26100633e-02
-5.15668213e-01 -3.34302336e-01 3.53284121e-01 -4.85365950e-02
6.93788826e-01 -1.09312713e+00 -1.33114648e+00 2.70872295e-01
-1.83134601e-01 -1.00860345e+00 -2.67440528e-01 2.50175148e-01
-9.00556266e-01 -1.17823088e+00 -6.83081985e-01 -9.99803782e-01
7.74193287e-01 7.12406993e-01 8.74718904e-01 2.49260679e-01
-3.26335698e-01 3.18703383e-01 -2.48958111e-01 -4.88236360e-02
-1.48773298e-01 -1.05445266e-01 1.42906725e-01 1.00359008e-01
1.62174225e-01 -4.23862219e-01 -7.01915443e-01 7.05004394e-01
-7.51306117e-01 -2.49171063e-01 6.43563509e-01 8.36786211e-01
1.02265680e+00 -1.48879439e-01 -8.83857831e-02 -2.18055785e-01
1.90501302e-01 -2.31919080e-01 -1.02609146e+00 5.00398636e-01
-5.69493651e-01 2.59970367e-01 1.82827473e-01 -2.57898629e-01
-3.60472858e-01 7.84203291e-01 1.31114483e-01 -5.36151946e-01
-1.69289634e-01 4.43633348e-01 2.67420232e-01 -7.62063324e-01
7.89022446e-01 7.56195128e-01 7.81390220e-02 -2.53642410e-01
9.54617560e-02 4.73876834e-01 7.99586356e-01 -1.12486809e-01
8.27800691e-01 6.38349950e-01 1.78354830e-01 -9.72738206e-01
5.99439442e-02 -1.17256534e+00 -7.68213511e-01 -2.61269182e-01
3.88512611e-01 -8.58420372e-01 -5.31981766e-01 3.27152401e-01
-1.13692951e+00 2.20709115e-01 2.23210528e-02 4.89971757e-01
-5.60086191e-01 4.82886821e-01 1.34893522e-01 -9.26302493e-01
-1.65356904e-01 -1.20871150e+00 1.00744784e+00 2.87921041e-01
1.82318836e-02 -6.11830354e-01 1.50241047e-01 -1.55527547e-01
3.09954852e-01 1.68435946e-01 1.10230245e-01 -5.87914228e-01
-1.08338678e+00 -5.91414273e-01 -7.74998963e-02 -2.66085833e-01
1.10377580e-01 -2.02058747e-01 -6.74928665e-01 -4.59515959e-01
-4.05759662e-01 3.38663198e-02 6.43203616e-01 1.22515604e-01
5.33651292e-01 -1.55505627e-01 -6.61648691e-01 6.22114301e-01
1.81046271e+00 4.89151329e-01 7.03932464e-01 7.74228811e-01
3.20250332e-01 5.17952919e-01 1.15178668e+00 5.13438106e-01
1.59901857e-01 1.02211130e+00 5.13371110e-01 -1.19857453e-01
2.68261045e-01 -3.70247543e-01 2.13126019e-01 1.50555193e-01
1.81161538e-01 -3.75112146e-02 -1.11563385e+00 9.95185435e-01
-2.05667615e+00 -9.79276776e-01 -6.92137778e-02 2.86690187e+00
2.33963266e-01 -1.64479569e-01 -2.04148050e-02 1.39165804e-01
9.22951996e-01 1.94047347e-01 -3.38787824e-01 -5.07968627e-02
-3.61411422e-01 -1.29212797e-01 1.07275903e+00 5.22285819e-01
-1.22902191e+00 8.30894887e-01 5.75647783e+00 6.89210176e-01
-1.21796083e+00 -1.03029005e-01 1.71955869e-01 4.31756824e-02
1.74116671e-01 3.59390914e-01 -6.53633118e-01 1.91259846e-01
5.63791275e-01 -1.19463235e-01 3.37287277e-01 8.44413877e-01
-3.70051526e-03 -9.22678947e-01 -9.81774867e-01 1.61358893e+00
1.61753714e-01 -1.37856889e+00 -2.70831525e-01 1.93255730e-02
6.43701017e-01 3.64276528e-01 -3.62790346e-01 -2.31309950e-01
-4.17698264e-01 -6.42642915e-01 1.00447547e+00 4.02457088e-01
3.86614829e-01 -8.41212034e-01 7.06433833e-01 5.49768269e-01
-1.48455465e+00 -1.88311078e-02 -4.23925728e-01 -7.77738541e-02
-6.61812499e-02 2.61464298e-01 -1.15314102e+00 6.91994965e-01
8.59812200e-01 4.97553468e-01 -8.52887332e-01 1.72086060e+00
9.32909697e-02 -7.88066834e-02 -8.17715466e-01 -2.15485349e-01
3.40941817e-01 -8.43058228e-02 7.21581638e-01 1.12154961e+00
6.54774725e-01 -2.85938352e-01 3.59414071e-01 4.94753927e-01
4.71337080e-01 3.79749924e-01 -1.08357096e+00 3.59208494e-01
8.11494887e-01 9.50448573e-01 -1.07997549e+00 -1.42639473e-01
-1.32195041e-01 1.34875965e+00 6.89294934e-02 2.17444882e-01
-5.18647850e-01 -5.84447265e-01 5.23489058e-01 7.48483017e-02
3.62029910e-01 -5.50208211e-01 1.89543366e-01 -8.67203414e-01
3.49889159e-01 -7.05845416e-01 4.91335839e-01 -7.02468216e-01
-7.01626599e-01 5.89570701e-01 -1.24543771e-01 -1.72273242e+00
-4.99364257e-01 -1.88731000e-01 -2.00386360e-01 7.79019594e-01
-1.43189836e+00 -1.00268865e+00 -4.17052776e-01 8.59809697e-01
3.37444216e-01 -2.54648663e-02 6.46166980e-01 4.02756363e-01
9.19686928e-02 3.46093059e-01 3.90210181e-01 -1.03095978e-01
8.35803807e-01 -8.10054600e-01 2.58209020e-01 1.19554245e+00
6.80762708e-01 5.29511690e-01 8.97278786e-01 -5.98474264e-01
-1.68757534e+00 -6.89247549e-01 1.35890448e+00 -8.22953209e-02
2.51684606e-01 -3.42368811e-01 -7.34061480e-01 3.72490436e-01
6.70075417e-02 2.57014036e-01 1.40709847e-01 -4.26140159e-01
-6.11610264e-02 -3.38040650e-01 -1.08266115e+00 4.94760543e-01
8.04308176e-01 -7.57651091e-01 -2.78674662e-01 3.64972860e-01
9.79624987e-02 -6.90308690e-01 -4.39042777e-01 3.02765787e-01
5.19102752e-01 -1.37489462e+00 1.10184669e+00 4.08128440e-01
-4.46735799e-01 -1.00545478e+00 -3.99599850e-01 -8.17372561e-01
1.13754839e-01 -5.83499908e-01 4.28867251e-01 9.85871613e-01
2.30260611e-01 -5.98828435e-01 7.00707376e-01 3.13175410e-01
2.87106931e-01 -4.00862426e-01 -1.25528204e+00 -9.72716391e-01
-9.32180166e-01 -4.19967383e-01 4.30815250e-01 7.09825695e-01
-3.04362744e-01 -1.52320117e-01 -4.41529214e-01 5.55129647e-01
6.00224495e-01 5.15934944e-01 1.03071845e+00 -8.76444995e-01
5.76362759e-02 -1.98566169e-01 -1.18445385e+00 -9.17219281e-01
-1.72038496e-01 -6.27798259e-01 3.57885331e-01 -1.45873845e+00
1.15553299e-02 -4.84838277e-01 -1.50135204e-01 3.97115499e-01
2.54758835e-01 4.73563701e-01 2.66933501e-01 6.93290532e-01
-7.37294555e-01 2.70159513e-01 5.75231493e-01 -1.39024466e-01
-4.79030758e-01 -2.92803496e-02 2.59824336e-01 3.68795246e-01
4.11518931e-01 -6.30984724e-01 -2.54396200e-01 -1.83397353e-01
7.87887573e-02 1.09548971e-01 7.31392384e-01 -1.70743215e+00
7.65787125e-01 -1.19151704e-01 3.95790309e-01 -1.11066949e+00
3.61065567e-01 -1.15526068e+00 6.47940457e-01 7.74481416e-01
1.68044884e-02 4.93049115e-01 2.85273463e-01 6.48945689e-01
-6.03960812e-01 -1.91648856e-01 6.84129953e-01 -1.45465448e-01
-1.66580820e+00 2.32275844e-01 -2.15187937e-01 -5.22911251e-01
1.20864141e+00 -7.35750616e-01 1.07556149e-01 -4.38541323e-01
-3.50800335e-01 1.66685671e-01 9.63548541e-01 5.03619134e-01
8.13988149e-01 -1.30128372e+00 -3.21708292e-01 3.61184239e-01
5.61664164e-01 -2.11769223e-01 1.45476624e-01 9.52749193e-01
-1.08818197e+00 4.86997068e-01 -4.43923771e-01 -1.00619769e+00
-1.50928330e+00 5.79550862e-01 5.65835953e-01 8.07593390e-02
-4.41829592e-01 5.85117102e-01 -1.35385185e-01 -3.65718484e-01
2.67051369e-01 -1.71911359e-01 1.14818051e-01 1.27347395e-01
5.32732904e-01 1.53626606e-01 3.49054754e-01 -1.10707033e+00
-1.00987542e+00 8.97092938e-01 4.14622337e-01 -2.12111816e-01
9.26871300e-01 -3.89129788e-01 -1.95988670e-01 4.20162439e-01
1.21085799e+00 2.32136156e-02 -1.01865315e+00 -2.85962015e-01
3.23112220e-01 -8.81690621e-01 -2.68698394e-01 -4.80389714e-01
-8.18497360e-01 5.42676091e-01 1.24206853e+00 -1.79169402e-01
1.37041879e+00 -1.91661775e-01 4.01251644e-01 7.23096609e-01
9.39948082e-01 -8.60330820e-01 -3.31653893e-01 5.11481106e-01
8.41595054e-01 -1.32777107e+00 7.27930963e-02 -1.04930714e-01
-4.75909144e-01 1.09465933e+00 3.25303227e-01 -1.27552569e-01
3.56023341e-01 -5.49195632e-02 9.05265957e-02 -8.29432160e-02
-2.95900613e-01 -4.97125655e-01 2.43203655e-01 6.53226972e-01
3.50757800e-02 -2.18566686e-01 -3.20409805e-01 -5.61137140e-01
3.27841155e-02 -1.53442889e-01 1.71854421e-01 1.41442645e+00
-5.94189703e-01 -1.04292202e+00 -7.64127493e-01 -7.66647309e-02
-1.76550020e-02 -3.87299806e-02 -3.83523941e-01 7.89699435e-01
9.42621529e-02 8.40488195e-01 1.22340471e-01 -3.42825264e-01
4.18051273e-01 -5.87685034e-02 4.59712505e-01 -5.81827201e-02
-5.68451405e-01 -1.66262053e-02 -9.75660905e-02 -9.00141597e-01
-6.32102132e-01 -7.15389490e-01 -1.16299748e+00 -2.30642200e-01
-3.69270176e-01 2.19191968e-01 8.41866255e-01 7.06476867e-01
6.34220302e-01 -3.77769917e-01 5.62573731e-01 -1.10954607e+00
-1.72540963e-01 -4.93751258e-01 -6.40547276e-01 3.65403116e-01
6.35885894e-01 -6.98733449e-01 -2.44919751e-02 -1.87351611e-02] | [7.5980377197265625, -1.9684937000274658] |
6d1a3d99-6d54-43ae-a675-dd48f9d6dd21 | aggression-and-misogyny-detection-using-bert | null | null | https://aclanthology.org/2020.trac-1.20 | https://aclanthology.org/2020.trac-1.20.pdf | Aggression and Misogyny Detection using BERT: A Multi-Task Approach | In recent times, the focus of the NLP community has increased towards offensive language, aggression, and hate-speech detection.This paper presents our system for TRAC-2 shared task on {``}Aggression Identification{''} (sub-task A) and {``}Misogynistic Aggression Identification{''} (sub-task B). The data for this shared task is provided in three different languages - English, Hindi, and Bengali. Each data instance is annotated into one of the three aggression classes - Not Aggressive, Covertly Aggressive, Overtly Aggressive, as well as one of the two misogyny classes - Gendered and Non-Gendered. We propose an end-to-end neural model using attention on top of BERT that incorporates a multi-task learning paradigm to address both the sub-tasks simultaneously. Our team, {``}na14{''}, scored 0.8579 weighted F1-measure on the English sub-task B and secured 3rd rank out of 15 teams for the task. The code and the model weights are publicly available at https://github.com/NiloofarSafi/TRAC-2. Keywords: Aggression, Misogyny, Abusive Language, Hate-Speech Detection, BERT, NLP, Neural Networks, Social Media | ['Prerana Mukherjee', 'Srinivas Pykl', 'Amitava Das', 'Parth Patwa', 'Niloofar Safi Samghabadi', 'Thamar Solorio'] | 2020-05-01 | null | null | null | lrec-2020-5 | ['misogynistic-aggression-identification', 'aggression-identification'] | ['natural-language-processing', 'natural-language-processing'] | [-5.00003695e-01 -7.35937580e-02 3.01054031e-01 -2.84719527e-01
-8.66880119e-01 -4.75256383e-01 4.66544062e-01 3.47222760e-02
-8.51230383e-01 8.01965177e-01 2.72562146e-01 7.70675344e-03
-2.22703442e-01 -1.21336810e-01 4.28716354e-02 -6.38930023e-01
-1.14894314e-02 7.58694410e-01 -6.03939220e-02 -5.53690791e-01
3.84217799e-01 3.68634731e-01 -9.15474951e-01 4.42096323e-01
5.06071270e-01 5.46124220e-01 -4.73643273e-01 1.12920737e+00
2.74339318e-01 1.26909101e+00 -9.74338055e-01 -1.24340987e+00
-7.35696852e-02 -2.51090914e-01 -1.18130124e+00 -7.62788653e-01
6.64784670e-01 -5.96141934e-01 -5.58506966e-01 1.25661933e+00
7.97848344e-01 2.61465132e-01 7.79378176e-01 -1.51743114e+00
-7.84710407e-01 8.48316312e-01 -8.74098718e-01 7.50378132e-01
2.54433185e-01 4.40522224e-01 1.05888486e+00 -6.82389617e-01
2.51496911e-01 1.46136999e+00 7.43361115e-01 1.18415451e+00
-8.95719826e-01 -1.28123069e+00 -2.93255389e-01 2.87212849e-01
-1.37479615e+00 -6.10124588e-01 6.63947344e-01 -8.87176335e-01
1.12140179e+00 3.17706496e-01 2.62456000e-01 2.06993771e+00
-2.04255179e-01 7.73233235e-01 9.44141507e-01 1.80555373e-01
-1.52760476e-01 1.30047381e-01 5.82865357e-01 6.35511339e-01
-1.53344467e-01 3.26187909e-01 -6.09104633e-01 -5.13876438e-01
2.49042258e-01 -5.30074060e-01 3.26477855e-01 7.01392949e-01
-5.65298617e-01 1.13173306e+00 8.88560563e-02 4.21671391e-01
-1.50140315e-01 1.97982341e-01 1.04764771e+00 3.06919497e-02
6.22527003e-01 6.39510274e-01 -2.38427877e-01 -9.51356292e-01
-5.45540988e-01 7.25175381e-01 5.57263434e-01 4.94296193e-01
5.09246252e-03 3.33630413e-01 -3.82051229e-01 1.62670732e+00
1.39910877e-01 5.36895931e-01 5.17409384e-01 -9.30129111e-01
7.59130836e-01 1.41251639e-01 -1.54403448e-01 -8.80528092e-01
-9.46545899e-01 -2.25113675e-01 -6.76980257e-01 2.71193236e-01
4.54814047e-01 -3.54636222e-01 -6.56984389e-01 2.06792974e+00
1.95160568e-01 -1.96599007e-01 -2.56844521e-01 8.51036549e-01
9.73828197e-01 5.63132048e-01 4.71193254e-01 4.67444025e-02
1.34792471e+00 -1.04624915e+00 -5.52720666e-01 -9.96517837e-02
8.06199491e-01 -5.59508502e-01 1.11152589e+00 7.36820817e-01
-1.27942991e+00 -7.62954801e-02 -5.34677625e-01 -3.03067505e-01
-5.33837557e-01 -2.55968273e-01 2.63888419e-01 1.00863135e+00
-7.86563456e-01 1.77574232e-01 -5.62961221e-01 -2.66381443e-01
6.54618144e-01 2.77087957e-01 -6.51735663e-01 4.59441215e-01
-1.24750078e+00 1.21674323e+00 2.04676226e-01 -8.57534781e-02
-1.19174707e+00 -5.46770275e-01 -5.41090548e-01 -1.52759165e-01
9.56573859e-02 -1.63459092e-01 1.38359261e+00 -6.88279986e-01
-1.27475071e+00 1.41868138e+00 4.18381214e-01 -1.18957624e-01
5.87935686e-01 -5.87499261e-01 -5.84213614e-01 -1.53999537e-01
3.17159444e-01 2.09669396e-01 6.85489953e-01 -9.75472927e-01
-6.94781363e-01 -6.60386443e-01 6.29950613e-02 1.49234697e-01
-5.00031352e-01 1.28908730e+00 3.07425112e-01 -7.17002332e-01
-9.40768719e-01 -9.05058503e-01 2.95011878e-01 -4.39323753e-01
-1.03049850e+00 -5.68479061e-01 7.70553648e-01 -1.08701217e+00
1.61758268e+00 -2.33714271e+00 5.75298965e-01 -7.85197169e-02
6.98067844e-01 6.71538532e-01 2.06151098e-01 4.43879515e-01
-3.04772794e-01 3.77517223e-01 -3.15295160e-01 -7.91614771e-01
1.32135585e-01 5.13500497e-02 2.53566541e-02 6.77160144e-01
-1.11936457e-01 4.91413176e-01 -1.04581392e+00 -2.35907450e-01
-1.40839770e-01 2.02230528e-01 -6.34045064e-01 4.24272031e-01
4.81679648e-01 6.51457548e-01 -1.58197865e-01 4.59569395e-01
3.90323490e-01 5.19439995e-01 -6.60046816e-01 3.66603196e-01
-3.76157522e-01 2.15584770e-01 -4.80112523e-01 1.01687396e+00
-3.47196162e-01 5.43020487e-01 5.39406955e-01 -5.18633187e-01
7.35026956e-01 2.65453190e-01 2.07880810e-01 -5.52503347e-01
5.35801053e-01 5.26965380e-01 5.84474385e-01 -9.06352639e-01
2.96831608e-01 -3.27136219e-01 -5.85457265e-01 5.25114357e-01
1.59929380e-01 -5.96688222e-03 3.74236941e-01 3.96310806e-01
1.54640007e+00 -3.41870487e-01 1.87261492e-01 -8.39897171e-02
5.46989202e-01 -3.61574739e-01 4.09239024e-01 7.56091952e-01
-9.29908574e-01 6.06529117e-01 8.27831149e-01 -4.12701726e-01
-1.12648010e+00 -1.10475898e+00 -1.19075761e-04 1.94895625e+00
-6.35581195e-01 -4.92427558e-01 -7.48063564e-01 -8.33288610e-01
-1.16187083e-02 1.17677295e+00 -8.62370729e-01 -6.23157203e-01
-6.70212805e-01 -6.60630465e-01 1.70683241e+00 4.39374715e-01
3.18673611e-01 -1.36380124e+00 -4.40062046e-01 -1.39751852e-01
-2.16569424e-01 -1.06694746e+00 -3.78566921e-01 4.57690023e-02
3.44061524e-01 -8.40881228e-01 -5.56419849e-01 -2.99561203e-01
-8.13954100e-02 -3.79864901e-01 7.04394042e-01 2.29574472e-01
-6.15729749e-01 9.40005407e-02 -5.47578275e-01 -4.22656924e-01
-6.12216294e-01 5.99474423e-02 3.59947741e-01 2.49931011e-02
5.99464536e-01 -5.98739922e-01 -1.46875586e-02 -1.80786010e-02
-5.11462629e-01 -1.48891494e-01 3.86681296e-02 6.63164198e-01
-4.78831798e-01 -3.96079987e-01 3.83749902e-01 -9.09521639e-01
9.31587160e-01 -9.34673309e-01 -1.10034971e-03 -4.11308020e-01
3.15639943e-01 -8.97660196e-01 8.80336344e-01 -3.48873138e-01
-7.62032270e-01 -5.58557093e-01 -8.00482571e-01 -6.18025184e-01
-6.47805691e-01 2.77057648e-01 1.04705274e-01 2.75609910e-01
8.49003136e-01 -1.96227729e-01 -1.73954815e-01 -6.19411051e-01
-4.48920904e-03 1.09762061e+00 6.87127888e-01 -6.07683241e-01
1.00226057e+00 -5.89620182e-03 -2.61906564e-01 -1.19279456e+00
-1.19736743e+00 -6.58354282e-01 -6.64657533e-01 -2.84692287e-01
1.30998671e+00 -7.14349508e-01 -1.05764258e+00 1.08707762e+00
-1.27975512e+00 -5.17922282e-01 3.08501482e-01 2.99083710e-01
-4.36803013e-01 2.46145993e-01 -1.07483065e+00 -1.24881184e+00
-7.49479890e-01 -9.18954611e-01 1.05506492e+00 -9.57118049e-02
-8.29245985e-01 -7.37349451e-01 5.19728482e-01 9.80446219e-01
3.36856544e-01 1.96622729e-01 1.09087586e+00 -1.45215118e+00
8.31880927e-01 -1.95817575e-01 -2.34553009e-01 5.16283274e-01
-1.96317375e-01 4.15714651e-01 -9.91783321e-01 -1.86338261e-01
-2.88752794e-01 -1.09460056e+00 5.77876806e-01 -1.12578109e-01
1.17368150e+00 -5.82783759e-01 2.97271907e-01 5.03716290e-01
5.49957275e-01 2.62890756e-01 4.68384534e-01 3.45058709e-01
1.19568837e+00 6.85273409e-01 1.83429465e-01 8.33176672e-01
4.08510417e-01 1.06531692e+00 4.04288530e-01 2.24446490e-01
-6.09655008e-02 -8.06739852e-02 5.71030498e-01 6.05987132e-01
-3.15457135e-01 -2.01767847e-01 -1.42377722e+00 5.30105472e-01
-1.66996348e+00 -1.48848009e+00 -3.50531906e-01 1.75927114e+00
6.57430351e-01 -1.18191183e-01 1.18045402e+00 1.71529070e-01
7.54217029e-01 1.48926288e-01 -2.77806103e-01 -1.39339590e+00
-1.06505174e-02 5.99008650e-02 5.86477742e-02 7.75260985e-01
-1.41957319e+00 1.30223787e+00 4.58433771e+00 1.00110936e+00
-9.22332525e-01 5.58362007e-01 4.89811450e-01 -7.19830096e-01
4.70085770e-01 -7.21037388e-01 -9.12518501e-01 8.31933439e-01
1.15366232e+00 -2.30380669e-01 5.79788148e-01 7.33278453e-01
8.29469562e-02 2.13249847e-01 -7.64526784e-01 1.05130470e+00
4.82194275e-01 -4.62408990e-01 -1.58083826e-01 1.81021303e-01
1.19293094e-01 2.40087613e-01 4.37184215e-01 8.95439148e-01
6.87315881e-01 -1.40418744e+00 1.00934446e+00 -9.87729058e-03
7.11744368e-01 -9.13273036e-01 7.19933391e-01 6.32517755e-01
-4.48804736e-01 -6.82600498e-01 -8.41697976e-02 -1.76575974e-01
1.75599784e-01 -1.11694038e-01 -4.03086901e-01 1.05510920e-01
1.19568670e+00 3.33877981e-01 -5.78552902e-01 7.62067735e-01
-3.27829838e-01 8.78694057e-01 -2.87010759e-01 -1.72242835e-01
5.72983801e-01 -9.42755789e-02 1.08840704e+00 1.56012332e+00
-1.64510339e-01 2.72203654e-01 3.92758846e-02 5.20846069e-01
7.19640031e-02 1.48985729e-01 -6.59304678e-01 -3.35411489e-01
2.91636169e-01 1.38095260e+00 -1.04200669e-01 2.65396871e-02
-1.71530813e-01 8.65048409e-01 1.04018283e+00 -4.63341214e-02
-1.13756847e+00 -4.79608446e-01 8.83427620e-01 -5.07101640e-02
-2.55781353e-01 -1.31038085e-01 -2.06618935e-01 -7.52006829e-01
-4.23482984e-01 -1.10769367e+00 8.09465587e-01 -6.03989244e-01
-1.81272173e+00 6.75958514e-01 2.66663134e-01 -4.69488949e-01
-2.36044243e-01 -7.87326217e-01 -7.11485982e-01 8.05135906e-01
-5.87857783e-01 -1.38074923e+00 1.04318626e-01 6.00169897e-01
6.16692126e-01 -6.15525663e-01 8.11775863e-01 7.37918913e-01
-1.39852154e+00 8.93953502e-01 -3.27405691e-01 6.64651573e-01
5.87377846e-01 -1.25159252e+00 1.42356411e-01 5.24676859e-01
-3.04682821e-01 4.84967440e-01 7.85953045e-01 -6.33219123e-01
-4.92231071e-01 -8.60178828e-01 8.80431056e-01 -8.73314500e-01
1.47681248e+00 -7.57217169e-01 -8.04895163e-01 7.24554479e-01
1.09880954e-01 -4.61519331e-01 1.14219165e+00 4.87033576e-01
-7.49109149e-01 4.73011076e-01 -1.18822479e+00 7.35784471e-01
1.24919558e+00 -4.95256066e-01 -5.67103386e-01 6.97907269e-01
2.87267089e-01 -5.05038738e-01 -6.21004999e-01 -8.90716836e-02
5.54756701e-01 -1.17491603e+00 6.99717343e-01 -1.31543946e+00
1.19335437e+00 5.04129291e-01 -2.71645375e-02 -1.32932520e+00
-8.15785766e-01 -5.99092543e-01 -5.28817400e-02 1.23588943e+00
5.01481712e-01 -4.00308281e-01 4.30630028e-01 5.94459891e-01
-4.51617241e-01 -5.47249973e-01 -1.62531006e+00 -7.65266001e-01
7.71127820e-01 -4.58430320e-01 3.79369259e-02 1.26384580e+00
3.76371086e-01 7.92013526e-01 -8.75223458e-01 9.34387743e-02
5.45238495e-01 -8.00195098e-01 9.70065951e-01 -1.07482231e+00
-2.45517343e-01 -1.05519652e+00 -5.73499799e-01 -1.25825733e-01
5.64539075e-01 -9.80294585e-01 -1.97205827e-01 -1.03502584e+00
4.14752692e-01 -8.64161178e-02 -3.20870546e-04 9.01353538e-01
-1.84217051e-01 5.37036896e-01 3.40650052e-01 -3.92099805e-02
-5.98966539e-01 5.45702517e-01 5.89242756e-01 1.05261706e-01
1.44502506e-01 -8.93642679e-02 -6.81501567e-01 9.06038404e-01
9.25860405e-01 -6.84903085e-01 3.60877998e-02 -5.37010491e-01
2.37404108e-01 -2.77389318e-01 4.44185048e-01 -8.35635126e-01
3.45212519e-02 -1.90987274e-01 -6.16216892e-03 -3.36970389e-02
8.88795435e-01 -3.63457710e-01 -4.05814826e-01 4.15194571e-01
-4.55999166e-01 2.40389764e-01 1.75838679e-01 -2.64237467e-02
2.57245719e-01 -4.91788715e-01 1.14481044e+00 -1.02002859e-01
-8.59057680e-02 4.73886341e-01 -6.21510506e-01 5.51924765e-01
1.06606770e+00 1.11909091e-01 -8.47139060e-01 -4.80554909e-01
-8.03669214e-01 3.16176623e-01 -9.77845043e-02 6.12375677e-01
4.24202263e-01 -8.98279607e-01 -1.41314042e+00 -7.19436109e-02
2.24712938e-01 -6.79463208e-01 4.73726690e-01 1.03920925e+00
-5.46292245e-01 7.41540343e-02 -4.49014783e-01 1.00222998e-03
-1.59330320e+00 2.81495750e-01 3.71631861e-01 -1.44881174e-01
-3.28658313e-01 1.35164535e+00 1.57851115e-01 -7.65425265e-01
1.57979265e-01 9.46788371e-01 -4.00391549e-01 7.88675770e-02
6.01336598e-01 8.57079625e-01 -2.97499955e-01 -1.65213227e+00
-4.38763469e-01 1.16854168e-01 -7.20891535e-01 -2.98402160e-01
1.25892115e+00 2.02101260e-01 -3.50332022e-01 7.05718398e-01
1.41175389e+00 2.00677007e-01 -3.78671914e-01 2.47462571e-01
2.45541930e-02 -7.34380484e-01 -1.47845179e-01 -1.13440418e+00
-8.24361444e-01 9.69054222e-01 6.71034157e-02 4.23468173e-01
4.97742832e-01 1.85764104e-01 8.20778966e-01 1.43471092e-01
6.84436113e-02 -1.18376434e+00 2.26066068e-01 1.09471560e+00
1.35314941e+00 -9.42559421e-01 -2.12889344e-01 -1.61214352e-01
-1.02975631e+00 8.15256476e-01 1.24399972e+00 -1.30770385e-01
2.87045866e-01 1.99224725e-01 2.63792604e-01 -6.13449931e-01
-6.79141521e-01 -1.04939344e-03 -1.26774803e-01 7.29597390e-01
4.37630564e-01 2.13087067e-01 -2.89801717e-01 8.93125832e-01
-9.18913841e-01 -7.32279539e-01 3.57508093e-01 4.87810880e-01
-1.48778245e-01 -3.86976451e-01 -3.75782073e-01 5.60802281e-01
-9.07348394e-01 -7.63307065e-02 -1.21364832e+00 7.76479542e-01
9.87987742e-02 9.62666094e-01 1.18388712e-01 -6.36022925e-01
5.46476305e-01 8.46316863e-04 3.17618698e-01 -7.66246676e-01
-1.38768828e+00 -3.17514151e-01 8.07794273e-01 -3.64369303e-01
3.97336453e-01 -6.58818543e-01 -8.57048571e-01 -7.91162074e-01
8.23985692e-03 -2.15238389e-02 3.47669840e-01 8.76080871e-01
-1.66978434e-01 1.55350372e-01 4.90770102e-01 -8.25664759e-01
-6.84441864e-01 -1.36120510e+00 -6.96163356e-01 8.62962782e-01
2.68623382e-01 -9.34755981e-01 -9.22453403e-01 -7.28933692e-01] | [8.800849914550781, 10.751529693603516] |
fc37ade0-b7a2-4322-b9b2-e80ca9474c4c | a-geometric-model-for-polarization-imaging-on | 2211.16986 | null | https://arxiv.org/abs/2211.16986v1 | https://arxiv.org/pdf/2211.16986v1.pdf | A Geometric Model for Polarization Imaging on Projective Cameras | The vast majority of Shape-from-Polarization (SfP) methods work under the oversimplified assumption of using orthographic cameras. Indeed, it is still not well understood how to project the Stokes vectors when the incoming rays are not orthogonal to the image plane. We try to answer this question presenting a geometric model describing how a general projective camera captures the light polarization state. Based on the optical properties of a tilted polarizer, our model is implemented as a pre-processing operation acting on raw images, followed by a per-pixel rotation of the reconstructed normal field. In this way, all the existing SfP methods assuming orthographic cameras can behave like they were designed for projective ones. Moreover, our model is consistent with state-of-the-art forward and inverse renderers (like Mitsuba3 and ART), intrinsically enforces physical constraints among the captured channels, and handles demosaicing of DoFP sensors. Experiments on existing and new datasets demonstrate the accuracy of the model when applied to commercially available polarimetric cameras. | ['Filippo Bergamasco', 'Mara Pistellato'] | 2022-11-29 | null | null | null | null | ['demosaicking'] | ['computer-vision'] | [ 3.41172338e-01 1.95043549e-01 2.60761499e-01 -2.51474857e-01
7.00149536e-02 -5.43820620e-01 7.63163149e-01 -7.11128831e-01
-4.78503287e-01 5.80704391e-01 6.51807636e-02 -1.06624737e-01
-2.29342766e-02 -9.47485566e-01 -7.19180107e-01 -9.58967686e-01
5.92975914e-01 8.50625873e-01 1.47182852e-01 -3.30258459e-01
2.22468138e-01 6.66783690e-01 -1.56585777e+00 -8.09812695e-02
5.79406500e-01 6.62202358e-01 -1.89176779e-02 5.16466081e-01
1.77399352e-01 3.88653696e-01 -1.24827065e-01 -6.02089643e-01
4.78266984e-01 -2.24383891e-01 -5.96288800e-01 3.23836297e-01
5.75698137e-01 -5.85858524e-01 -4.37713772e-01 1.03086925e+00
8.25421810e-02 -3.65030318e-01 4.81799424e-01 -7.17590392e-01
-3.98682773e-01 -8.78814459e-02 -3.05316329e-01 -3.89952630e-01
5.20360351e-01 2.19874844e-01 7.25123465e-01 -8.34215820e-01
9.95674253e-01 8.63674104e-01 5.66115320e-01 3.54022384e-01
-1.34428644e+00 -2.76379380e-02 -3.38222623e-01 -9.46209839e-05
-1.26709688e+00 -5.43232560e-01 8.64306271e-01 -5.31212449e-01
5.70673227e-01 2.88874865e-01 9.55078125e-01 1.01865375e+00
-1.82851236e-02 6.26880378e-02 1.45347226e+00 -7.14610279e-01
2.86807299e-01 3.09585720e-01 -3.45885642e-02 3.21088403e-01
4.97613460e-01 1.96798876e-01 -7.89175868e-01 6.41785264e-02
7.15052664e-01 -2.79302686e-01 -6.80037498e-01 -8.92382681e-01
-1.20327330e+00 3.18185389e-01 7.45616928e-02 -1.29009053e-01
-3.08353096e-01 -2.78513789e-01 -1.35917455e-01 -2.79421937e-02
3.46651465e-01 5.65654874e-01 -7.02612698e-02 3.35054211e-02
-8.34580123e-01 1.31877810e-01 9.11081910e-01 9.71035421e-01
1.02040732e+00 2.69153677e-02 4.77273166e-01 4.02033448e-01
4.90609497e-01 1.14876890e+00 -6.43327162e-02 -1.21497738e+00
3.11439127e-01 3.00021082e-01 3.54996949e-01 -8.83616507e-01
-4.35964257e-01 -3.50114435e-01 -5.34889877e-01 1.74362794e-01
4.28793937e-01 -1.49463683e-01 -5.57265997e-01 1.15772235e+00
2.80133903e-01 -1.15578644e-01 3.98427099e-01 1.12789011e+00
2.51834452e-01 5.55408180e-01 -7.42618859e-01 -2.95466661e-01
1.38923597e+00 -3.40789109e-01 -6.34436131e-01 -2.98326939e-01
3.44663739e-01 -8.71278763e-01 6.58998787e-01 7.95730889e-01
-9.27880824e-01 -1.41428888e-01 -9.86927211e-01 -9.13921967e-02
1.48390055e-01 4.80234116e-01 4.44817722e-01 9.33679879e-01
-9.15599763e-01 2.75827795e-01 -8.76540542e-01 -4.61551130e-01
-5.52191883e-02 1.15866691e-01 -6.00613892e-01 -1.26549691e-01
-6.88744366e-01 8.36887121e-01 2.00697571e-01 3.64971697e-01
-4.73960161e-01 -5.38531423e-01 -4.25315559e-01 -9.77104381e-02
2.52395689e-01 -9.36526120e-01 8.45139027e-01 -8.31861973e-01
-2.16756082e+00 1.05671203e+00 -7.32462704e-02 -3.43572766e-01
5.73046267e-01 -3.33665639e-01 -3.05046231e-01 5.37637353e-01
-2.98635781e-01 2.36523420e-01 6.60930753e-01 -1.55749536e+00
-1.72155440e-01 -5.38378179e-01 1.78579748e-01 4.56398666e-01
-3.49222645e-02 -2.46515751e-01 -4.81627256e-01 1.18158124e-01
6.81892455e-01 -1.20929658e+00 -1.60798058e-02 4.84968089e-02
-5.85932136e-01 8.98470163e-01 8.80916893e-01 -3.23707283e-01
5.72838366e-01 -2.18078065e+00 2.78659344e-01 1.61429062e-01
-1.36987284e-01 3.18143457e-01 2.44198769e-01 6.94311678e-01
1.83395501e-02 -3.94490957e-01 -2.02283591e-01 -3.90301496e-01
-1.89101785e-01 5.82209766e-01 -5.13663054e-01 1.02533424e+00
-2.19052210e-01 2.37661600e-01 -5.48817098e-01 -5.50577976e-02
4.57463920e-01 7.23851383e-01 -8.16124737e-01 1.48257956e-01
-5.06407022e-02 8.69348466e-01 -2.21154913e-01 3.65527928e-01
1.31294298e+00 2.77358890e-01 5.77266574e-01 -2.80283779e-01
-8.23600709e-01 4.06967878e-01 -1.41669500e+00 1.41271102e+00
-3.08431387e-01 6.85298026e-01 5.08162498e-01 -8.68487716e-01
8.88927996e-01 2.27609456e-01 3.77988160e-01 -6.36180937e-01
9.53401253e-02 3.60624224e-01 -3.03444803e-01 -5.80059290e-01
8.07416677e-01 -4.00329798e-01 3.85025054e-01 -8.93718656e-03
6.82614893e-02 -5.28434217e-01 2.67617684e-02 -1.27472848e-01
5.75292528e-01 5.81093609e-01 1.33436486e-01 -4.03289109e-01
6.02239370e-01 1.51553869e-01 5.89886427e-01 4.23702747e-01
5.00695288e-01 1.08324099e+00 6.37021244e-01 -5.55198610e-01
-1.23698092e+00 -8.57607543e-01 -5.59332371e-01 2.22455766e-02
2.32800275e-01 -3.46108198e-01 -8.86982083e-01 1.39605731e-01
-2.62005389e-01 5.55877924e-01 -1.45165637e-01 3.56883466e-01
-4.62086976e-01 -1.02334332e+00 2.70801157e-01 -1.89795479e-01
6.55592561e-01 -4.22781616e-01 -9.25288677e-01 -8.70864689e-02
-1.38952076e-01 -1.47347736e+00 5.29753506e-01 -1.84840873e-01
-8.69169831e-01 -1.32324648e+00 -4.73776549e-01 -5.16401455e-02
7.51461148e-01 5.13683081e-01 9.77587402e-01 -4.65302169e-01
1.54199049e-01 5.76064825e-01 -2.82717288e-01 -9.65829268e-02
-2.28993744e-01 -3.58247072e-01 1.73525155e-01 6.07864082e-01
1.26749054e-01 -8.29647005e-01 -5.69992721e-01 4.80776489e-01
-9.04920399e-01 3.70939195e-01 3.36632192e-01 4.88167018e-01
6.05901659e-01 -3.14743817e-01 -7.04183996e-01 -1.06022704e+00
-1.68207228e-01 -1.28145099e-01 -1.15967047e+00 -3.77628654e-02
-4.52602923e-01 1.41158089e-01 7.21380770e-01 2.75106400e-01
-1.45429528e+00 3.91136736e-01 -2.28182748e-01 -2.29095519e-01
-3.91264588e-01 1.68632179e-01 -3.45073730e-01 -3.20604473e-01
4.22057599e-01 3.71140867e-01 -1.19416885e-01 -5.64155877e-01
2.82585293e-01 5.25880277e-01 5.89155555e-01 -5.20318568e-01
9.32303250e-01 1.43861508e+00 7.12917387e-01 -1.41637373e+00
-6.02659822e-01 -3.93687963e-01 -8.41963589e-01 -3.46906215e-01
8.92034233e-01 -9.05692816e-01 -8.02071571e-01 7.74861276e-01
-1.29219854e+00 1.20812103e-01 -3.37630898e-01 8.87174070e-01
-6.09253943e-01 6.85049474e-01 -1.28834337e-01 -8.46774936e-01
5.49906977e-02 -1.24569356e+00 1.30407524e+00 1.21867426e-01
2.73339957e-01 -7.47692525e-01 4.45424795e-01 5.29123306e-01
2.16924101e-01 1.49448123e-02 4.79987562e-01 2.77641058e-01
-9.45418179e-01 -4.95205782e-02 2.52972431e-02 5.02785802e-01
-5.49649119e-01 4.81686324e-01 -1.23061764e+00 1.27625510e-01
3.61457497e-01 1.88239023e-01 7.20965505e-01 3.64890277e-01
4.81700391e-01 -2.13928685e-01 -8.91853794e-02 1.37042749e+00
1.79733658e+00 -1.18313439e-01 1.08866155e+00 4.17519331e-01
7.45824575e-01 7.58249581e-01 2.96647847e-01 5.40638864e-01
5.31838059e-01 1.03339136e+00 1.00211120e+00 1.11026771e-01
1.05759017e-01 -1.10735297e-01 3.24545324e-01 7.75706708e-01
-7.00986683e-01 -2.24068224e-01 -9.69357669e-01 1.98356390e-01
-1.54372203e+00 -9.10629392e-01 -9.50825036e-01 2.57330441e+00
2.17225328e-01 -1.78174853e-01 -5.14832735e-01 4.17398885e-02
2.06118107e-01 4.14483547e-01 1.54570460e-01 -2.47902021e-01
-5.56432843e-01 1.05987102e-01 9.96420026e-01 7.98752248e-01
-7.71414280e-01 6.11162126e-01 6.00355482e+00 4.42523044e-03
-1.51379025e+00 8.57813731e-02 -9.38289389e-02 1.58545543e-02
-4.82031137e-01 7.70424306e-01 -8.83408606e-01 2.98971295e-01
6.10854626e-01 3.99235815e-01 5.33679426e-01 6.58112168e-01
2.86070496e-01 -5.69763720e-01 -8.71039569e-01 1.08982909e+00
2.56089419e-01 -1.13715124e+00 -7.00596347e-02 3.08117956e-01
7.01889634e-01 2.35625863e-01 4.58283387e-02 -3.76362175e-01
-2.40512878e-01 -3.54500741e-01 8.22720230e-01 9.17111695e-01
5.93601644e-01 -1.65463850e-01 4.34839100e-01 4.56705540e-01
-5.64803720e-01 8.60944167e-02 -3.76625359e-01 -3.36682051e-01
6.17499828e-01 9.66920435e-01 -6.75820351e-01 9.76657450e-01
5.87134123e-01 5.38859248e-01 -1.75189540e-01 6.59887850e-01
-6.11518741e-01 5.54082870e-01 -6.39018595e-01 5.84552646e-01
9.77693871e-02 -1.06037533e+00 8.41274679e-01 8.22990358e-01
6.05237067e-01 1.23178899e-01 -3.67726326e-01 7.18877256e-01
2.40705416e-01 -1.05388835e-01 -6.81694031e-01 4.30666238e-01
-7.50376955e-02 1.33991706e+00 -4.87887174e-01 -6.76751211e-02
-5.49214840e-01 8.02527189e-01 -1.75213650e-01 5.83852351e-01
-5.38874269e-01 2.02401817e-01 6.09279871e-01 5.96467197e-01
2.33086094e-01 -5.39161921e-01 -2.52484351e-01 -1.72104514e+00
4.50723529e-01 -4.53639567e-01 -2.86107808e-01 -1.30293787e+00
-6.24632776e-01 3.76899868e-01 -2.51936018e-02 -1.45616734e+00
1.84641361e-01 -1.17820942e+00 -3.44164550e-01 8.72317314e-01
-1.75133276e+00 -1.22445655e+00 -4.07358676e-01 4.01708990e-01
-2.68921375e-01 3.65347952e-01 7.78459549e-01 1.35764748e-01
-3.36484671e-01 -5.69078088e-01 4.82780606e-01 -1.64309487e-01
6.67927384e-01 -1.08042848e+00 -1.73292886e-02 1.26120055e+00
-1.54713979e-02 6.79938138e-01 1.02709115e+00 -2.26443246e-01
-2.04008818e+00 -5.89431047e-01 7.82289505e-01 -3.17166209e-01
3.60680699e-01 -4.63719755e-01 -5.52334726e-01 7.72868693e-01
3.92699838e-01 3.44819397e-01 3.12927365e-01 -8.02554935e-03
-2.77958840e-01 -4.62406993e-01 -7.73358762e-01 1.95742577e-01
9.27727044e-01 -5.69392145e-01 -4.55849588e-01 5.14160156e-01
-1.13195136e-01 -7.72631645e-01 -6.05713010e-01 2.24063605e-01
7.79780447e-01 -1.61379540e+00 9.09625411e-01 -1.00864008e-01
4.08073783e-01 -5.18674910e-01 -2.62558967e-01 -1.27543771e+00
9.40101743e-02 -8.29969108e-01 3.78680944e-01 1.04834557e+00
-4.11414430e-02 -1.04902375e+00 6.60969794e-01 3.90127242e-01
-4.13181335e-01 -2.25761622e-01 -9.78543878e-01 -6.16595566e-01
-2.97626823e-01 -2.73630530e-01 3.34736377e-01 9.19809341e-01
-2.29864821e-01 9.73159075e-02 -5.01595914e-01 8.15360546e-01
7.87280917e-01 7.20876455e-02 1.04292214e+00 -1.40091431e+00
-4.76941645e-01 -6.43818006e-02 -3.69255364e-01 -1.24032450e+00
1.73831396e-02 -4.19582337e-01 -2.03299820e-01 -1.11969972e+00
-1.74616843e-01 -3.98441523e-01 5.58436930e-01 -1.38232946e-01
4.59469348e-01 1.64580569e-01 5.35234474e-02 5.91919899e-01
-6.82701766e-02 5.83441317e-01 1.09055614e+00 3.45386803e-01
3.76930870e-02 -7.56794736e-02 -1.81018561e-01 1.07714784e+00
3.46910775e-01 -4.63218987e-01 -9.53049883e-02 -7.84019530e-01
1.10173690e+00 -2.19014753e-03 6.47363663e-01 -1.08278465e+00
3.40284735e-01 -1.06257968e-01 -1.18806265e-01 -5.35736501e-01
7.39108801e-01 -1.05780673e+00 7.82534719e-01 1.96068957e-01
4.01021004e-01 -2.98193514e-01 -3.63189995e-01 1.54348075e-01
-3.09053570e-01 -5.36263168e-01 9.29772496e-01 -2.59461045e-01
-4.81319666e-01 -5.82466274e-02 -2.03418121e-01 -1.22635506e-01
8.21241021e-01 -3.47003549e-01 -6.39566779e-01 -1.45565465e-01
-2.64136165e-01 -3.60181004e-01 1.16140485e+00 -4.13731843e-01
3.30740452e-01 -8.51880133e-01 -4.13319468e-01 5.03087461e-01
1.88683227e-01 2.02702940e-01 3.78678471e-01 1.23201358e+00
-1.41419530e+00 6.18669093e-01 -2.36913130e-01 -8.01433980e-01
-1.22409630e+00 2.74996489e-01 5.73777080e-01 1.14046969e-01
-7.44870365e-01 4.13881630e-01 2.82949567e-01 -6.08111858e-01
-5.59709966e-01 -2.68548131e-01 -5.50258905e-02 -6.51362538e-02
2.10726112e-01 3.56740892e-01 3.17007571e-01 -1.03674269e+00
-3.88625681e-01 1.12713253e+00 5.14380574e-01 -4.09129500e-01
1.55635941e+00 -2.06534967e-01 -4.10503358e-01 4.66935337e-02
8.23973179e-01 5.46921074e-01 -1.39204931e+00 -2.83239204e-02
-6.20514691e-01 -6.10104859e-01 9.66175199e-02 -1.93517357e-01
-8.97622645e-01 1.16280782e+00 1.08483903e-01 9.25229862e-02
1.14531231e+00 -2.92243689e-01 4.06731032e-02 5.75642109e-01
6.39454365e-01 -1.08868647e+00 -7.95274556e-01 5.95683396e-01
6.18945897e-01 -7.02590883e-01 2.58704185e-01 -8.04191947e-01
-5.65866530e-01 1.51006651e+00 1.23927392e-01 -1.09607846e-01
4.19934034e-01 1.94398612e-01 1.83743834e-02 -1.85143307e-01
-3.21626544e-01 6.49129897e-02 -1.14756323e-01 6.12180769e-01
2.67171919e-01 6.63551614e-02 -3.47376794e-01 -1.80548951e-01
-2.87797987e-01 -1.62265461e-03 1.18729687e+00 8.06867599e-01
-3.26252520e-01 -1.23731709e+00 -7.81458259e-01 -3.12392533e-01
-2.33379707e-01 5.75019754e-02 -1.76424146e-01 8.21470976e-01
2.26262346e-01 3.96074980e-01 1.75402448e-01 6.19747043e-02
6.02105975e-01 -6.22145943e-02 5.53107977e-01 -5.13747036e-01
-8.48505199e-02 -2.23161187e-02 2.88920790e-01 -6.09932363e-01
-1.02558315e+00 -9.77281630e-01 -7.58028984e-01 -1.83587223e-01
-1.64356619e-01 -4.42285091e-02 1.07801354e+00 9.03792739e-01
3.58091414e-01 -5.37339598e-02 5.85374594e-01 -1.03730726e+00
-4.11773473e-01 -6.12234652e-01 -9.71731603e-01 3.33224863e-01
2.19101280e-01 -7.34761834e-01 -4.67310399e-01 1.96330667e-01] | [9.9786958694458, -2.7933340072631836] |
66142419-bc38-44a1-af57-540812af4ad4 | regularizing-end-to-end-speech-translation | 2112.10991 | null | https://arxiv.org/abs/2112.10991v2 | https://arxiv.org/pdf/2112.10991v2.pdf | Regularizing End-to-End Speech Translation with Triangular Decomposition Agreement | End-to-end speech-to-text translation (E2E-ST) is becoming increasingly popular due to the potential of its less error propagation, lower latency, and fewer parameters. Given the triplet training corpus $\langle speech, transcription, translation\rangle$, the conventional high-quality E2E-ST system leverages the $\langle speech, transcription\rangle$ pair to pre-train the model and then utilizes the $\langle speech, translation\rangle$ pair to optimize it further. However, this process only involves two-tuple data at each stage, and this loose coupling fails to fully exploit the association between triplet data. In this paper, we attempt to model the joint probability of transcription and translation based on the speech input to directly leverage such triplet data. Based on that, we propose a novel regularization method for model training to improve the agreement of dual-path decomposition within triplet data, which should be equal in theory. To achieve this goal, we introduce two Kullback-Leibler divergence regularization terms into the model training objective to reduce the mismatch between output probabilities of dual-path. Then the well-trained model can be naturally transformed as the E2E-ST models by the pre-defined early stop tag. Experiments on the MuST-C benchmark demonstrate that our proposed approach significantly outperforms state-of-the-art E2E-ST baselines on all 8 language pairs, while achieving better performance in the automatic speech recognition task. Our code is open-sourced at https://github.com/duyichao/E2E-ST-TDA. | ['Tong Xu', 'Jun Xie', 'Boxing Chen', 'Weizhi Wang', 'Zhirui Zhang', 'Yichao Du'] | 2021-12-21 | null | null | null | null | ['speech-to-text-translation'] | ['natural-language-processing'] | [ 1.52767196e-01 -1.38038294e-02 -2.55566239e-01 -4.89593685e-01
-1.51269639e+00 -5.60813606e-01 4.02434230e-01 -3.16597492e-01
-2.47637734e-01 5.56582868e-01 3.16460997e-01 -8.92201424e-01
2.42783964e-01 -1.91518441e-01 -7.11113513e-01 -6.30689561e-01
3.37750167e-01 5.32802165e-01 -1.07419537e-02 -2.55949199e-01
-1.71582595e-01 -1.06524609e-01 -8.99931610e-01 3.94141018e-01
1.03305578e+00 9.98401642e-01 4.95760709e-01 6.12247288e-01
-1.82928726e-01 3.98460567e-01 -2.42862582e-01 -4.95768577e-01
2.62414575e-01 -6.84051394e-01 -5.67577183e-01 -1.12268038e-01
2.99730450e-02 -8.74039456e-02 -2.86223501e-01 9.33943331e-01
6.28596008e-01 2.79224999e-02 4.61894095e-01 -9.90895212e-01
-4.60068017e-01 8.42599154e-01 -5.10366440e-01 1.42905086e-01
1.69336572e-01 1.89898565e-01 1.25687575e+00 -1.34572947e+00
2.83463687e-01 1.24743378e+00 4.80104655e-01 5.29565811e-01
-1.24714875e+00 -8.97659481e-01 1.21827595e-01 4.35988139e-03
-1.42556036e+00 -9.10187483e-01 6.01969779e-01 -2.30345592e-01
1.19594800e+00 3.18906367e-01 1.75639987e-01 1.31301105e+00
8.39582235e-02 1.01348174e+00 1.11567616e+00 -5.77992260e-01
-1.11071728e-01 1.87985212e-01 1.78924855e-02 7.19984233e-01
-4.39516068e-01 1.68618485e-01 -7.91140676e-01 -7.46962726e-02
3.91609132e-01 -3.01729083e-01 -3.26368183e-01 1.99872345e-01
-1.22985172e+00 4.59326267e-01 -5.10773808e-02 3.97118658e-01
-2.26453319e-01 2.22987048e-02 4.35357034e-01 5.19973874e-01
5.75805187e-01 5.96990660e-02 -8.04323792e-01 -5.41034877e-01
-7.88820982e-01 -2.85345852e-01 6.57449424e-01 1.05803084e+00
6.76782191e-01 8.81575644e-02 -3.32953542e-01 1.15562010e+00
5.83198488e-01 7.33559132e-01 4.74096894e-01 -6.72167003e-01
1.15643764e+00 2.21388578e-01 -1.78798333e-01 -3.85691792e-01
1.04190148e-01 -6.51930809e-01 -9.03743684e-01 -5.30544877e-01
1.65178869e-02 -1.97957441e-01 -9.00420606e-01 1.86144853e+00
1.32547542e-01 2.80275583e-01 2.84440845e-01 6.61290228e-01
3.92448574e-01 1.04192555e+00 -2.62133986e-01 -4.29617494e-01
1.08531356e+00 -1.30976331e+00 -7.89025366e-01 -5.41369379e-01
9.00709510e-01 -1.14205360e+00 1.29660177e+00 1.35178044e-01
-1.10203171e+00 -4.89783198e-01 -8.43499362e-01 -3.42347398e-02
1.23901278e-01 4.97736156e-01 1.20011799e-01 5.48633039e-01
-9.64941621e-01 2.73438036e-01 -1.11169493e+00 -1.52353212e-01
-3.14189568e-02 3.74193549e-01 -1.81956410e-01 -4.56738397e-02
-1.26400626e+00 8.38003099e-01 3.22209954e-01 2.72844315e-01
-7.00174570e-01 -4.75674003e-01 -7.57907033e-01 -2.15539318e-02
3.52026939e-01 -6.56105876e-01 1.50987828e+00 -6.28349006e-01
-1.92150939e+00 5.78891754e-01 -6.57289743e-01 -3.18137318e-01
2.80300945e-01 -2.11536288e-01 -5.38018286e-01 -2.27958336e-01
1.42369131e-02 4.23148245e-01 6.50596023e-01 -1.10329330e+00
-5.28862774e-01 -2.27856874e-01 -4.10146594e-01 3.20235729e-01
-4.10989016e-01 2.29479834e-01 -7.91187882e-01 -7.35442400e-01
2.70026296e-01 -1.02145457e+00 -6.34815991e-02 -6.13686025e-01
-5.26645958e-01 -2.26826429e-01 6.06255949e-01 -7.97263145e-01
1.62964988e+00 -2.17996097e+00 1.68815732e-01 1.42909095e-01
-1.66198805e-01 5.40385723e-01 -3.98972005e-01 6.72998309e-01
1.62068576e-01 2.89015770e-01 -3.41623396e-01 -1.00485170e+00
4.94518466e-02 3.60705495e-01 -3.94684374e-01 1.37949675e-01
1.20593637e-01 7.82001138e-01 -7.49158800e-01 -4.64913100e-01
6.33960813e-02 4.49600488e-01 -4.38327551e-01 3.62662792e-01
-1.48354098e-01 5.81530392e-01 -4.37680304e-01 3.57381463e-01
4.01681840e-01 -1.07810229e-01 3.33364159e-01 2.24513784e-02
-1.61140263e-01 1.09097683e+00 -7.37249255e-01 1.80405688e+00
-7.08157957e-01 3.93952549e-01 2.01291531e-01 -9.23586249e-01
1.06130898e+00 6.31327271e-01 3.31671536e-01 -6.72198296e-01
1.06593184e-01 7.12900460e-01 1.16479546e-01 -2.40337402e-01
1.94755733e-01 -1.90173984e-01 6.76375404e-02 4.68129635e-01
3.88789140e-02 -1.00673012e-01 -1.77541394e-02 -1.19784288e-02
9.11786497e-01 1.67443782e-01 1.21228367e-01 5.29701859e-02
5.92772782e-01 -4.03026879e-01 8.41115892e-01 4.95406687e-01
-1.13350838e-01 5.99527478e-01 2.59401679e-01 1.59182355e-01
-9.63085830e-01 -1.09930670e+00 1.49805039e-01 1.01861322e+00
-1.45872802e-01 -6.73021913e-01 -8.47795367e-01 -7.31407046e-01
-4.56739783e-01 8.59359741e-01 -3.23872268e-02 -1.82261974e-01
-8.28349352e-01 -5.47483981e-01 7.61687160e-01 3.28671515e-01
4.15609509e-01 -6.27252698e-01 2.36561686e-01 3.66316587e-01
-6.79838121e-01 -1.30060804e+00 -1.08311403e+00 3.75162601e-01
-8.74577641e-01 -3.75239521e-01 -5.09149313e-01 -9.52979624e-01
4.15301532e-01 2.98285365e-01 9.25522149e-01 -1.10166654e-01
4.75067943e-01 -1.19120352e-01 -4.59209770e-01 1.90497171e-02
-7.35406935e-01 3.03883672e-01 2.11065918e-01 8.87197629e-02
4.17555898e-01 -6.63167179e-01 -3.76423836e-01 5.55275798e-01
-5.88986933e-01 3.78287286e-01 8.02487552e-01 1.11993206e+00
7.42801905e-01 -3.39165598e-01 4.26452965e-01 -5.25489151e-01
4.49145168e-01 -3.40925545e-01 -3.34301144e-01 4.90083218e-01
-7.86715150e-01 2.33047366e-01 8.12309027e-01 -3.23172212e-01
-8.82262290e-01 1.50536851e-03 -5.32018363e-01 -5.70061505e-01
1.76113963e-01 7.88943291e-01 -3.92517447e-01 3.50098640e-01
3.15087616e-01 5.39206445e-01 -1.73098028e-01 -6.40629053e-01
2.90737718e-01 1.19405770e+00 3.37626934e-01 -6.59979939e-01
6.79717302e-01 -1.43852144e-01 -4.49610323e-01 -5.78601837e-01
-8.09568167e-01 -5.90141416e-01 -3.08005124e-01 1.13681205e-01
5.82211375e-01 -9.39892471e-01 -3.86408418e-01 3.93100262e-01
-1.44796836e+00 -4.72808152e-01 2.69143820e-01 7.82185733e-01
-5.96430540e-01 4.22759473e-01 -6.97485685e-01 -9.12327230e-01
-5.33188879e-01 -1.37273836e+00 1.15339518e+00 -3.12864870e-01
8.27254578e-02 -8.04352164e-01 7.41718113e-02 6.21118486e-01
2.69909561e-01 -5.47034085e-01 9.46893990e-01 -6.56711221e-01
-5.60075045e-01 7.19362348e-02 -1.04437314e-01 7.04331458e-01
1.38825566e-01 -1.49040520e-01 -8.90539765e-01 -3.59700680e-01
1.43660009e-01 -1.40450150e-01 7.06632316e-01 1.60622045e-01
8.40944171e-01 -5.48575699e-01 -2.20701009e-01 6.38017178e-01
1.09832966e+00 3.07938010e-01 4.84136641e-01 2.19776947e-02
6.94952011e-01 3.93879235e-01 6.20001018e-01 2.24043012e-01
6.10767901e-01 9.79382515e-01 3.77445482e-02 3.44019420e-02
-2.02219218e-01 -4.54436630e-01 9.73785579e-01 1.94913697e+00
2.74337709e-01 -5.62120318e-01 -1.06888413e+00 4.30972397e-01
-1.82275176e+00 -7.14699090e-01 -1.01429401e-02 2.22512269e+00
1.10569215e+00 1.82271376e-01 -1.07341357e-01 -1.09573752e-01
7.35128939e-01 2.13727549e-01 -2.75730371e-01 -4.02377844e-01
-7.48136863e-02 8.79113078e-02 4.03942227e-01 8.56489480e-01
-6.77927971e-01 1.16413713e+00 5.22864437e+00 1.23040342e+00
-1.22995651e+00 4.55290943e-01 6.74140036e-01 -1.67555772e-02
-5.52806258e-01 2.20607400e-01 -1.18160570e+00 5.51140428e-01
1.36774397e+00 -2.00799163e-02 6.99529588e-01 3.46778780e-01
5.75866282e-01 3.95187467e-01 -1.21058428e+00 1.05272532e+00
-9.40931439e-02 -1.03016114e+00 -2.40065213e-02 8.15820415e-03
6.00716770e-01 3.60515058e-01 2.00812384e-01 4.79335755e-01
2.89863288e-01 -6.91158772e-01 7.83868730e-01 9.38685089e-02
1.01344609e+00 -4.55799580e-01 4.99663413e-01 7.78369963e-01
-1.35860634e+00 5.22989593e-02 5.75142242e-02 2.02725783e-01
4.24427927e-01 6.46634161e-01 -1.08955550e+00 7.72207975e-01
3.05424899e-01 5.21292269e-01 -4.84290607e-02 4.03372049e-01
-3.39266479e-01 1.08097744e+00 -4.56017852e-01 1.06340930e-01
4.30434138e-01 -3.90231788e-01 6.46589577e-01 1.38271475e+00
7.05604553e-01 2.00329870e-02 3.07772249e-01 5.10741293e-01
-1.49752036e-01 1.92504272e-01 -4.02223110e-01 -2.57962644e-01
7.18039155e-01 7.80327439e-01 -1.87995568e-01 -2.33645961e-01
-5.85162282e-01 1.13872647e+00 4.80275989e-01 4.84009951e-01
-8.09458435e-01 -1.53899446e-01 7.13763893e-01 -1.14614837e-01
2.97326207e-01 -5.02906442e-01 -3.19204956e-01 -1.30959129e+00
4.24798727e-01 -1.18191707e+00 1.49081589e-03 -5.10372400e-01
-1.07162154e+00 9.84891832e-01 -2.99225032e-01 -1.32061040e+00
-6.15145862e-01 -3.64030480e-01 -3.24092686e-01 1.13469172e+00
-1.57797551e+00 -1.22478092e+00 3.77047807e-01 4.24203098e-01
8.21875036e-01 -1.96330279e-01 8.15577567e-01 5.33220351e-01
-8.24637711e-01 1.06655717e+00 5.07663727e-01 1.01741657e-01
8.51171732e-01 -9.08719599e-01 7.15764165e-01 1.11284506e+00
3.45363617e-01 6.70652926e-01 5.14829814e-01 -5.21750212e-01
-1.39175284e+00 -1.05348051e+00 1.51546717e+00 -3.32245380e-01
7.15366006e-01 -6.80517793e-01 -7.95072496e-01 7.22064257e-01
2.44489819e-01 -9.48556885e-02 6.78973436e-01 7.87124261e-02
-4.37284470e-01 -3.07348162e-01 -5.58049917e-01 7.43740141e-01
1.09934604e+00 -9.87263262e-01 -4.42415237e-01 2.02872559e-01
1.05062616e+00 -3.59831691e-01 -8.04269552e-01 4.52007502e-01
4.28385526e-01 -7.13988841e-01 6.80147946e-01 -3.65313768e-01
3.15587163e-01 -2.17572585e-01 -5.88309944e-01 -1.51044357e+00
-3.73638012e-02 -1.18997693e+00 -1.19218044e-01 1.35217130e+00
9.96865094e-01 -7.41329730e-01 4.93678123e-01 2.78576225e-01
-7.53617167e-01 -1.04190409e+00 -1.22238123e+00 -1.07082593e+00
3.23314629e-02 -6.26537323e-01 5.94685137e-01 6.48286700e-01
6.37098178e-02 6.54177547e-01 -5.63898206e-01 2.10114405e-01
2.70847559e-01 -7.12783858e-02 7.26090550e-01 -6.12662315e-01
-6.98823810e-01 -3.30722481e-01 1.76527739e-01 -1.78966820e+00
2.78068453e-01 -1.08608353e+00 3.67451310e-01 -1.39601338e+00
3.74587327e-02 -7.13521183e-01 -3.58568192e-01 5.31632602e-01
-3.15536678e-01 -1.16305947e-02 3.14050138e-01 4.46950078e-01
-3.83475751e-01 9.16536570e-01 1.24194503e+00 3.38507220e-02
-3.53307158e-01 9.12970230e-02 -4.30419207e-01 4.23604578e-01
8.00716519e-01 -6.36710882e-01 -3.14911425e-01 -9.10462141e-01
-2.66829617e-02 4.17661279e-01 -2.20837742e-01 -6.68529630e-01
3.33018661e-01 -7.60771260e-02 -4.10245955e-01 -6.29579842e-01
6.00901783e-01 -6.78852618e-01 1.30599529e-01 1.10671639e-01
-3.53034467e-01 1.07196771e-01 2.02014036e-02 3.63117158e-01
-4.71185803e-01 -6.58247173e-02 5.83351612e-01 2.11375400e-01
-1.25883445e-01 4.10541862e-01 -3.01786780e-01 2.94856187e-02
4.74597991e-01 -9.36814118e-03 -1.97880179e-01 -3.61704499e-01
-4.71150130e-01 2.30055019e-01 1.57003954e-01 4.82564181e-01
4.14524317e-01 -1.37303066e+00 -8.66518974e-01 3.59160751e-01
6.46265447e-02 -1.41364306e-01 1.54054388e-02 1.19335938e+00
4.50489074e-02 6.11822069e-01 3.88462752e-01 -6.76498294e-01
-1.23049819e+00 1.56830668e-01 2.23277301e-01 -6.14595115e-01
-2.39266351e-01 9.21637714e-01 1.45290852e-01 -7.69464076e-01
2.91114748e-01 -2.79434919e-01 3.84354264e-01 -3.93323004e-01
2.38483950e-01 1.65588751e-01 2.74815500e-01 -7.65218258e-01
-4.13064718e-01 4.23440605e-01 -2.81068504e-01 -6.59446597e-01
1.03093827e+00 -5.10223925e-01 3.92569229e-03 5.16687810e-01
1.48602712e+00 1.76072657e-01 -1.08671749e+00 -5.26463628e-01
2.98399087e-02 -3.40615392e-01 5.16985171e-02 -8.24715853e-01
-7.93397069e-01 1.05785441e+00 2.98982382e-01 -1.56229645e-01
1.13698506e+00 -2.82667652e-02 1.31107867e+00 3.55330616e-01
2.35002801e-01 -1.05317771e+00 -4.82587293e-02 1.00673008e+00
7.74970353e-01 -1.19372940e+00 -5.96473277e-01 -5.43869376e-01
-6.60158932e-01 7.92944491e-01 4.07281369e-01 4.41290051e-01
4.94090527e-01 3.42792720e-01 3.63137454e-01 3.73684585e-01
-1.03098071e+00 -1.05017394e-01 3.02793473e-01 2.23627552e-01
6.82633698e-01 1.92319244e-01 -3.90567750e-01 6.07245386e-01
-2.53919214e-01 -1.69885397e-01 2.96213082e-03 5.81642032e-01
-3.07151228e-01 -1.68299925e+00 -2.14458987e-01 2.63796747e-01
-5.75347781e-01 -5.66905975e-01 -2.73125947e-01 2.38757119e-01
-1.47979811e-01 1.32196498e+00 -2.64185041e-01 -6.90561056e-01
2.52126962e-01 2.73941398e-01 2.04399779e-01 -6.72438323e-01
-7.19623148e-01 5.96827328e-01 3.37809622e-01 -5.27019858e-01
-8.83897319e-02 -6.97135329e-01 -1.24915814e+00 -2.61429638e-01
-4.74496752e-01 3.70818079e-01 8.61097872e-01 9.80244458e-01
6.70184374e-01 2.56485403e-01 1.06941235e+00 -4.93463099e-01
-9.17899430e-01 -1.15953410e+00 -1.66096881e-01 7.38613084e-02
4.13166732e-01 -2.92223155e-01 -4.61346328e-01 4.62473072e-02] | [14.449492454528809, 7.144014358520508] |
0d142271-5c91-47fd-87d8-e859f55443d2 | applying-intelligent-reflector-surfaces-for | 2108.07834 | null | https://arxiv.org/abs/2108.07834v2 | https://arxiv.org/pdf/2108.07834v2.pdf | Applying Intelligent Reflector Surfaces for Detecting Violent Expiratory Aerosol Cloud using Terahertz Signals | The recent COVID-19 pandemic has driven researchers from different spectrum to develop novel solutions that can improve detection and understanding of SARS-CoV-2 virus. In this article we propose the use of Intelligent Reflector Surface (IRS) emitting terahertz signals to detect airborne respiratory aerosol cloud that are secreted from people. Our proposed approach makes use of future IRS infrastructure to extend beyond communication functionality by adding environmental scanning for aerosol clouds. Simulations have also been conducted to analyze the accuracy of aerosol cloud detection based on a signal scanning and path optimization algorithm. Utilizing IRS for detecting respiratory aerosol cloud can lead to new added value of telecommunication infrastructures for sensor monitoring data that can be used for public health. | ['Sasitharan Balasubramaniam', 'Nicola Marchetti', 'Daniel Perez Martins', "Nathan D'Arcy", 'Michael Taynnan Barros', 'Harun Šiljak'] | 2021-08-17 | null | null | null | null | ['cloud-detection'] | ['computer-vision'] | [ 4.33998555e-01 -3.88183624e-01 6.32284045e-01 -3.03848803e-01
-3.63168210e-01 -6.33480668e-01 3.13764900e-01 -1.52298331e-01
-9.04709697e-02 6.96312249e-01 -3.21106054e-02 -4.87954974e-01
-1.60736945e-02 -1.18463540e+00 -6.89192638e-02 -7.30651557e-01
-4.11277562e-02 7.96785474e-01 3.02701265e-01 -4.57036346e-01
-4.15911153e-02 8.85984302e-01 -1.08491075e+00 3.91742885e-01
6.98153436e-01 6.55344963e-01 5.49959958e-01 8.98652196e-01
-2.06522155e-03 -2.37562224e-01 -5.91656208e-01 6.86066508e-01
6.66434526e-01 -2.04069257e-01 -4.31266390e-02 -3.52889955e-01
-1.13985524e-01 -3.47924769e-01 7.06105232e-01 6.46369815e-01
5.62759399e-01 -1.79805979e-01 7.58966446e-01 -8.58557820e-01
3.19024525e-03 -1.62958369e-01 -5.70372581e-01 5.97397745e-01
3.66001278e-01 4.53119934e-01 9.77286994e-02 -5.85693479e-01
5.41081667e-01 8.80632758e-01 8.52515280e-01 1.42586514e-01
-3.07238668e-01 -9.20257986e-01 -4.12708193e-01 5.48958704e-02
-1.22835553e+00 8.72239321e-02 3.31400275e-01 -5.38529932e-01
1.26810277e+00 8.04340720e-01 8.34102869e-01 3.84831578e-01
4.77412879e-01 -3.73955220e-01 1.41927898e+00 -2.82859266e-01
1.15671426e-01 5.27967334e-01 1.00564025e-01 6.39178634e-01
1.09347188e+00 2.43225798e-01 2.21654177e-01 -9.29892212e-02
1.96957082e-01 4.35569555e-01 -9.01819170e-02 7.60046005e-01
-5.28562605e-01 7.24083185e-01 2.88684100e-01 7.37949729e-01
-8.44368517e-01 8.48223865e-02 -2.81548262e-01 3.85643333e-01
8.52154136e-01 5.39165497e-01 -4.01008964e-01 2.50409693e-01
-8.29515815e-01 2.39794731e-01 1.24535881e-01 3.96549523e-01
6.40303254e-01 2.69986510e-01 -1.55932948e-01 2.70872086e-01
1.00453770e+00 1.75914824e+00 -3.39910418e-01 -3.53251398e-01
1.94758505e-01 7.49971569e-01 4.54088151e-01 -8.32318425e-01
-6.73977375e-01 -6.70787811e-01 -1.35825619e-01 1.43086135e-01
-2.79333770e-01 -7.29945719e-01 -9.17034149e-01 6.26382411e-01
7.26733565e-01 4.88671660e-01 2.43387640e-01 9.46737826e-01
7.86609113e-01 1.20515490e+00 -4.40209448e-01 -1.75813362e-01
1.67619336e+00 -1.76130220e-01 -4.96485561e-01 2.98013479e-01
5.94473243e-01 -8.69606197e-01 4.31265652e-01 3.97828341e-01
-4.04654682e-01 -1.52616411e-01 -9.67348814e-01 8.76855373e-01
-5.13886690e-01 -1.56735256e-01 4.45304774e-02 1.46342289e+00
-8.47844899e-01 3.16566937e-02 -7.50333250e-01 -8.46884310e-01
2.94163167e-01 4.06394571e-01 5.50845325e-01 -7.95088932e-02
-8.49309623e-01 7.90094614e-01 -3.29935849e-01 -1.44078583e-02
-6.75575614e-01 -8.71160924e-01 2.95573063e-02 -2.98724443e-01
-2.76728630e-01 -8.45535457e-01 7.59418249e-01 -4.22484130e-01
-1.39152718e+00 5.33833623e-01 -3.34917635e-01 -5.12185037e-01
-2.75031719e-02 -2.50273496e-02 -8.67494762e-01 4.51984584e-01
-3.44109237e-02 -1.50932014e-01 5.36120474e-01 -1.37048042e+00
-6.91022754e-01 -8.21199715e-01 -5.43357670e-01 -1.73985004e-01
4.26434427e-02 6.43855453e-01 6.99617505e-01 -1.21369056e-01
-1.95689827e-01 -1.06480825e+00 -2.74405122e-01 -7.80710280e-01
-2.71610349e-01 3.55734453e-02 1.67293513e+00 -5.13346910e-01
6.67046726e-01 -1.38432336e+00 -1.00962627e+00 7.18392432e-01
-1.33733273e-01 7.38978148e-01 -2.57218480e-01 8.39578807e-01
3.45022380e-01 1.71578825e-01 -6.31605983e-02 1.74452111e-01
-5.38935900e-01 -2.32205600e-01 -5.20237386e-01 5.75723708e-01
1.38810307e-01 5.89554608e-01 -5.94182968e-01 2.18726784e-01
2.08561167e-01 8.99729729e-01 -2.92790264e-01 6.40097335e-02
-4.75678086e-01 1.16652739e+00 -1.03235626e+00 6.54313564e-01
1.38503349e+00 -6.53343089e-03 -2.64034510e-01 1.86275110e-01
-8.46969664e-01 1.72302619e-01 -7.64922261e-01 3.38035971e-01
-2.46718228e-01 3.57432157e-01 1.55957773e-01 -4.99454588e-01
9.68854666e-01 6.47487640e-01 8.49780023e-01 -5.61718225e-01
3.63240510e-01 1.02446489e-01 -1.84262767e-01 -1.05108023e+00
3.23571503e-01 -6.42326355e-01 6.62111759e-01 6.71597540e-01
-8.94326627e-01 -2.21273646e-01 -3.90673548e-01 -4.07767117e-01
9.15192962e-01 -3.34354043e-01 -1.56374231e-01 -1.95018247e-01
6.12501740e-01 5.98682225e-01 4.49305540e-03 3.95295322e-01
1.25797629e-01 1.75523102e-01 -6.61619186e-01 -3.07750553e-01
-6.83074653e-01 -1.07180834e+00 -1.55312970e-01 5.01843631e-01
-1.97338268e-01 3.77866477e-01 -6.17802918e-01 -1.48233935e-01
-1.27689824e-01 7.51665354e-01 -9.12784711e-02 3.97979468e-01
-5.92039466e-01 -8.67976189e-01 5.20716548e-01 -1.17472462e-01
4.28539246e-01 -8.32493126e-01 -1.11533475e+00 2.52950519e-01
-5.19515276e-02 -1.11374438e+00 1.81135669e-01 -5.09720862e-01
-1.24521005e+00 -1.18680334e+00 -3.86189729e-01 -1.14988268e-03
5.46657979e-01 1.04310358e+00 3.95492584e-01 3.84224266e-01
-7.73797572e-01 9.89821613e-01 -6.75544441e-01 -1.51184750e+00
-4.43311006e-01 -3.62490296e-01 -7.49789085e-03 1.59889348e-02
6.24024332e-01 -2.64846951e-01 -1.00451362e+00 2.32640371e-01
-5.91197193e-01 -5.40502787e-01 1.80258349e-01 -6.66834831e-01
3.17447037e-01 1.08732045e-01 8.38095725e-01 -8.33679199e-01
5.31908631e-01 -1.01074505e+00 -1.02389205e+00 -1.40221089e-01
-6.64409041e-01 -3.66470218e-01 3.13348323e-01 3.65708411e-01
-1.12698328e+00 -5.03971994e-01 -2.45317712e-01 -2.64079254e-02
-3.13344419e-01 1.57039583e-01 6.63367331e-01 -1.56336784e-01
7.00250089e-01 5.50490692e-02 -7.73331374e-02 -2.95269847e-01
-3.20989132e-01 1.10979140e+00 -3.94318998e-01 1.09461043e-02
1.15860820e+00 1.25392616e+00 2.40759075e-01 -1.66791296e+00
-4.55895782e-01 -1.19028246e+00 2.19855711e-01 -6.72720730e-01
1.23477888e+00 -8.17412853e-01 -6.49998963e-01 1.07581273e-01
-1.30861366e+00 9.22846496e-02 5.73079400e-02 9.72963393e-01
1.38437226e-01 -3.44447158e-02 -6.13278002e-02 -1.43118703e+00
-8.55514109e-01 -6.61183596e-01 9.03442502e-01 1.56034604e-01
1.49748370e-01 -8.09781969e-01 9.25037801e-01 7.71320820e-01
1.11952150e+00 4.42458570e-01 3.18369001e-01 -3.21709156e-01
-1.14981365e+00 -1.89348012e-01 -1.65276036e-01 6.10469058e-02
4.21885341e-01 1.42538741e-01 -1.01977766e+00 -5.16173124e-01
3.97266358e-01 4.86975312e-01 8.38792503e-01 6.71376765e-01
3.16979140e-01 -1.77877977e-01 -7.39812434e-01 4.53284621e-01
1.68013740e+00 5.11015773e-01 7.23112524e-01 7.21271560e-02
2.46843711e-01 6.80164635e-01 7.52452374e-01 5.80499053e-01
1.59781680e-01 2.67835677e-01 5.59020519e-01 -2.61780322e-01
1.13873243e-01 5.82427859e-01 3.88175726e-01 5.46555340e-01
-7.66873956e-01 -3.77580196e-01 -9.24818456e-01 1.41232923e-01
-1.00779617e+00 -1.38620317e+00 -7.98405290e-01 2.15190792e+00
-4.24231552e-02 -6.10981524e-01 2.14960426e-01 -3.14863622e-01
5.38812339e-01 -2.77233481e-01 4.09477726e-02 -5.43684065e-01
4.10020322e-01 8.38388562e-01 8.39718044e-01 6.58937156e-01
-4.67544764e-01 4.81532425e-01 5.99984169e+00 -1.63904265e-01
-1.61841285e+00 5.87827384e-01 -4.14013118e-01 3.29459719e-02
-6.49607003e-01 -1.69778526e-01 -9.90325689e-01 1.97637007e-01
1.08551896e+00 1.33292943e-01 2.25917652e-01 4.18725610e-03
1.31607699e+00 -9.62818936e-02 1.55669311e-02 2.27535754e-01
-3.33352119e-01 -1.27995849e+00 7.77524039e-02 3.41705650e-01
7.90277064e-01 8.53266656e-01 1.10581309e-01 -3.15218508e-01
-9.62564200e-02 -7.62809992e-01 -2.51037389e-01 7.45032430e-01
5.01662135e-01 -5.52872062e-01 6.41023517e-01 4.00575221e-01
-1.34891343e+00 8.80827382e-02 -3.77583772e-01 -8.63565952e-02
2.44769990e-01 7.35886812e-01 -1.87698364e+00 3.44584227e-01
4.11165774e-01 -8.97488743e-02 1.02465734e-01 1.03674018e+00
2.24005759e-01 9.99426246e-01 -7.34771073e-01 -4.45706695e-01
3.76008064e-01 -6.00360990e-01 1.13217723e+00 1.46679509e+00
1.08128464e+00 5.99220634e-01 -1.89189896e-01 9.43366528e-01
4.71439511e-01 2.64718950e-01 -1.04824245e+00 -7.05914274e-02
4.58397925e-01 1.06970215e+00 -7.28451550e-01 -2.78481275e-01
-1.71117380e-01 3.65514457e-01 -9.94045079e-01 4.36343193e-01
-7.29135871e-01 -1.52017429e-01 8.24454665e-01 1.17467570e+00
3.73620898e-01 -4.43523258e-01 3.82956229e-02 -3.19523543e-01
-4.76249427e-01 -4.18917328e-01 2.28822857e-01 -6.30551159e-01
-6.08811319e-01 4.20619726e-01 1.15293920e-01 -1.12395263e+00
2.11232588e-01 -5.43036222e-01 -1.02880383e+00 7.43973494e-01
-1.84593582e+00 -1.13729680e+00 -4.90048885e-01 6.71162546e-01
1.61164314e-01 -2.01885149e-01 7.71630824e-01 1.26390770e-01
1.08729318e-01 -3.66476893e-01 3.65464777e-01 -5.59667349e-01
2.41376951e-01 -3.64595950e-01 -7.44619034e-03 5.56451678e-01
-2.88023889e-01 6.68977976e-01 9.88238633e-01 -1.07080054e+00
-1.45436883e+00 -1.69703341e+00 5.18438697e-01 -2.39255428e-01
-6.55307397e-02 8.18788633e-02 -2.14369342e-01 7.04658449e-01
5.89381576e-01 -4.36299592e-01 7.81258464e-01 -3.89896989e-01
-1.83536317e-02 -3.86441648e-01 -1.82424128e+00 1.54247314e-01
3.04990292e-01 5.65417111e-02 -3.42043459e-01 8.48579705e-01
6.70468330e-01 3.40707630e-01 -3.37315977e-01 5.41817427e-01
4.60124791e-01 -9.95914102e-01 8.85174930e-01 -5.59061766e-01
-5.30172288e-01 -5.07622480e-01 -8.57880488e-02 -1.19798183e+00
5.72467372e-02 -3.45069945e-01 8.34517479e-01 4.69170004e-01
3.67777288e-01 -1.38280249e+00 7.37713099e-01 -2.94239610e-01
-2.18145445e-01 -2.91391313e-01 -8.51207316e-01 -8.50778759e-01
-3.56663316e-01 -6.35042131e-01 5.45916975e-01 7.88092852e-01
-4.95701671e-01 -1.48904405e-03 -1.36945948e-01 1.04266334e+00
6.39945328e-01 -5.61162755e-02 5.66321731e-01 -1.52961242e+00
-1.15528785e-01 1.68675095e-01 -9.41496864e-02 -1.40059203e-01
-6.62012815e-01 -7.04916418e-01 -2.42423207e-01 -1.80881858e+00
-9.64043140e-02 -6.38952494e-01 -8.11107606e-02 -9.43813995e-02
4.56076860e-01 5.19549549e-01 1.95448399e-02 -6.35930747e-02
5.37745766e-02 -1.39346477e-02 1.09833539e+00 8.21522102e-02
-4.55294132e-01 4.98619795e-01 -2.30059177e-01 6.35726810e-01
1.37040889e+00 -1.11434567e+00 -4.62130219e-01 -2.32176438e-01
6.89730823e-01 -2.19368979e-01 4.74115670e-01 -1.34595919e+00
-8.54464546e-02 -4.26253229e-01 -3.39648314e-02 -9.92111325e-01
3.83132875e-01 -1.24704361e+00 6.34240687e-01 1.29291832e+00
6.64657533e-01 8.40892196e-02 1.71337739e-01 4.91722703e-01
3.79110038e-01 -3.34161699e-01 8.54293346e-01 -2.66152710e-01
2.95083016e-01 3.43948156e-01 -1.09789908e+00 -3.20871562e-01
1.50923455e+00 -6.23677492e-01 -7.24873066e-01 -1.55965254e-01
-1.63056850e-01 -9.92593840e-02 2.19393879e-01 -9.33353379e-02
8.00762534e-01 -5.59622228e-01 -1.13144004e+00 1.77774206e-01
-2.86412276e-02 -5.59675694e-01 1.15975423e-03 9.62217748e-01
-9.16362226e-01 9.22832727e-01 -1.55669644e-01 -6.39292836e-01
-1.77412641e+00 7.92991444e-02 4.87559438e-01 -1.24899991e-01
-3.15730840e-01 6.04950309e-01 -3.96435082e-01 -3.64219397e-01
-6.07067525e-01 -5.31468153e-01 -2.48146564e-01 -6.87415898e-02
5.38893998e-01 8.02158535e-01 2.23223343e-01 -3.29229653e-01
-7.17903554e-01 1.00273645e+00 4.51263577e-01 -1.23692438e-01
1.52128339e+00 -9.45601314e-02 -3.12868118e-01 2.92380214e-01
7.50934064e-01 7.85575569e-01 -3.93201530e-01 2.93940544e-01
-3.75792891e-01 -3.50540042e-01 4.27175164e-02 -9.14650977e-01
-6.83335006e-01 6.85034215e-01 1.17970228e+00 3.21051836e-01
1.06234241e+00 -1.20208822e-02 9.67026293e-01 7.28602946e-01
4.12561715e-01 -7.49609530e-01 -1.03379644e-01 3.03652942e-01
7.21729338e-01 -1.10246634e+00 2.01006860e-01 -5.78454494e-01
-8.02612975e-02 1.12368143e+00 -1.52244642e-01 -3.22244793e-01
8.13216925e-01 5.65634668e-01 4.56209928e-01 -9.95638371e-01
-5.07324100e-01 -4.29942518e-01 -1.99756175e-01 1.06057239e+00
3.38289648e-01 4.99304473e-01 -4.58597541e-01 4.57823649e-02
2.78322041e-01 3.33613098e-01 5.72257638e-01 7.92652905e-01
-1.41015470e+00 -1.09046149e+00 -1.05868363e+00 8.21392417e-01
-6.05277002e-01 -4.52982783e-02 -3.25232774e-01 5.34660280e-01
4.42577511e-01 1.21833169e+00 -6.01644032e-02 -2.40239218e-01
2.30815455e-01 -2.12000832e-01 6.16124213e-01 -8.41945887e-01
-7.62615323e-01 -1.82143942e-01 1.83911905e-01 -1.19216591e-01
-6.82420790e-01 -5.70729017e-01 -1.70242178e+00 -7.65918195e-02
-2.41727844e-01 2.83756614e-01 1.32974386e+00 9.31927085e-01
5.35114586e-01 2.73925006e-01 9.27819014e-01 -2.23287553e-01
-2.82234639e-01 -9.66486394e-01 -4.90122795e-01 -4.81118828e-01
4.32528973e-01 -4.10747588e-01 -2.80758411e-01 -3.95301372e-01] | [14.40982723236084, 3.8301804065704346] |
a4df9a94-7a8d-4c7b-83a7-1b1370ff511e | pre-emptive-learning-to-defer-for-sequential | 2109.06312 | null | https://arxiv.org/abs/2109.06312v2 | https://arxiv.org/pdf/2109.06312v2.pdf | Learning-to-defer for sequential medical decision-making under uncertainty | Learning-to-defer is a framework to automatically defer decision-making to a human expert when ML-based decisions are deemed unreliable. Existing learning-to-defer frameworks are not designed for sequential settings. That is, they defer at every instance independently, based on immediate predictions, while ignoring the potential long-term impact of these interventions. As a result, existing frameworks are myopic. Further, they do not defer adaptively, which is crucial when human interventions are costly. In this work, we propose Sequential Learning-to-Defer (SLTD), a framework for learning-to-defer to a domain expert in sequential decision-making settings. Contrary to existing literature, we pose the problem of learning-to-defer as model-based reinforcement learning (RL) to i) account for long-term consequences of ML-based actions using RL and ii) adaptively defer based on the dynamics (model-based). Our proposed framework determines whether to defer (at each time step) by quantifying whether a deferral now will improve the value compared to delaying deferral to the next time step. To quantify the improvement, we account for potential future deferrals. As a result, we learn a pre-emptive deferral policy (i.e. a policy that defers early if using the ML-based policy could worsen long-term outcomes). Our deferral policy is adaptive to the non-stationarity in the dynamics. We demonstrate that adaptive deferral via SLTD provides an improved trade-off between long-term outcomes and deferral frequency on synthetic, semi-synthetic, and real-world data with non-stationary dynamics. Finally, we interpret the deferral decision by decomposing the propagated (long-term) uncertainty around the outcome, to justify the deferral decision. | ['Finale Doshi-Velez', 'Sonali Parbhoo', 'Shalmali Joshi'] | 2021-09-13 | null | null | null | null | ['decision-making-under-uncertainty', 'decision-making-under-uncertainty'] | ['medical', 'reasoning'] | [ 1.01027004e-01 1.07420638e-01 -4.23721403e-01 -3.83616000e-01
-5.94277084e-01 -6.82458878e-01 4.95362073e-01 4.30083215e-01
-6.39710546e-01 1.11887395e+00 -4.87719662e-02 -9.34370875e-01
-3.72587949e-01 -6.74774528e-01 -8.19417477e-01 -5.64353347e-01
-2.49110445e-01 5.00900328e-01 2.89277643e-01 9.45888758e-02
2.82965779e-01 2.28523999e-01 -1.12568986e+00 -4.71554557e-03
1.02125013e+00 9.51206207e-01 2.11158291e-01 5.55292368e-01
2.92016715e-01 1.07370996e+00 -4.11166489e-01 -1.55438427e-02
2.78779536e-01 -2.05698684e-01 -4.93895978e-01 -1.36910468e-01
-2.99482703e-01 -9.35989797e-01 5.99926896e-02 7.75789201e-01
3.65721583e-01 4.16688882e-02 6.37585104e-01 -1.12782860e+00
-3.34767699e-01 7.07506716e-01 -4.00893420e-01 3.50618213e-01
2.78086662e-01 6.32412076e-01 7.85597146e-01 -3.35180283e-01
1.05555989e-01 1.48473537e+00 5.00085294e-01 4.37849522e-01
-1.37777805e+00 -3.99872631e-01 9.66889203e-01 2.84087472e-02
-8.96880925e-01 -2.05389515e-01 5.33734858e-01 -6.61893487e-01
7.68980622e-01 -1.31755576e-01 5.05716264e-01 1.02271819e+00
9.26276982e-01 6.60121977e-01 1.37436867e+00 -3.52386147e-01
8.20556641e-01 -2.85423607e-01 -2.01760918e-01 2.95593023e-01
1.64098457e-01 7.69928634e-01 -3.28061044e-01 -2.47327819e-01
4.82561618e-01 2.97930151e-01 1.72165424e-01 1.37524888e-01
-9.35429811e-01 5.94216704e-01 8.39824229e-02 -4.03712362e-01
-8.99471939e-01 2.18576893e-01 3.79384726e-01 7.31688261e-01
3.29774171e-01 2.70091712e-01 -6.38358057e-01 -1.33144468e-01
-7.28518963e-01 4.16255385e-01 6.56052113e-01 4.97423589e-01
7.22598553e-01 1.35759562e-01 -6.53015971e-01 1.63358569e-01
1.22654788e-01 3.45924526e-01 4.03227389e-01 -1.13230300e+00
4.54767436e-01 1.00130238e-01 8.99871588e-01 -5.53988695e-01
-3.51446092e-01 -3.82755667e-01 -5.00115216e-01 5.20784676e-01
3.47072780e-01 -8.18390369e-01 -7.28925884e-01 2.07756090e+00
3.58657330e-01 3.55112463e-01 2.15072650e-02 8.28110993e-01
-5.01527071e-01 5.44272482e-01 1.52432486e-01 -7.80268252e-01
8.02861094e-01 -4.67245162e-01 -5.56420147e-01 -2.67520964e-01
5.56331575e-01 -4.34448153e-01 1.01271486e+00 6.56952202e-01
-8.69790137e-01 -4.13522780e-01 -6.96397960e-01 8.73352945e-01
2.05032036e-01 -1.06667504e-01 3.42388749e-01 1.57089457e-01
-9.02644515e-01 8.11445713e-01 -1.12502313e+00 7.00024366e-02
-5.43939099e-02 1.75816849e-01 3.49367082e-01 2.49805227e-01
-1.42182064e+00 7.34054208e-01 2.46692553e-01 6.57206997e-02
-1.50131774e+00 -7.07148254e-01 -2.86041617e-01 -1.21702820e-01
1.29507220e+00 -6.32820845e-01 1.82589269e+00 -1.16175306e+00
-1.61095858e+00 1.54609248e-01 -1.03477001e-01 -9.93959367e-01
1.03049636e+00 -3.69467437e-01 -2.01523617e-01 -1.94247141e-01
2.80789912e-01 2.34012172e-01 1.06310284e+00 -1.06033611e+00
-1.09303892e+00 -1.90314457e-01 4.81530935e-01 4.22995031e-01
-8.23350847e-02 -4.36578721e-01 2.28098676e-01 -6.21585667e-01
-4.65728164e-01 -1.19172430e+00 -4.14009899e-01 -2.21813098e-01
-3.62711847e-01 -4.52037305e-01 6.34541035e-01 -3.49374920e-01
1.62628472e+00 -1.87841010e+00 -2.17898801e-01 -1.69923931e-01
-1.43182084e-01 -4.48916815e-02 2.55941954e-02 6.44495785e-01
3.05737883e-01 2.14834705e-01 -1.60370752e-01 -2.83905327e-01
-1.30665824e-01 4.84828770e-01 -5.70545673e-01 2.08494633e-01
9.13776159e-02 2.90021807e-01 -9.60768759e-01 -7.93771073e-02
1.84077591e-01 -2.81615525e-01 -5.67698181e-01 4.31434661e-01
-5.56509554e-01 5.33327043e-01 -9.00814474e-01 5.74495256e-01
4.09731835e-01 -7.62756094e-02 2.94255286e-01 3.71804982e-01
-4.35842633e-01 1.82359368e-01 -9.72380340e-01 8.70278597e-01
-8.30102980e-01 4.89426963e-02 1.87715709e-01 -1.20461845e+00
5.63314497e-01 2.93496817e-01 3.70155305e-01 -6.31055176e-01
-1.11681215e-01 2.11186334e-01 1.70314640e-01 -2.31728241e-01
1.34120405e-01 -2.34744698e-01 -1.24597371e-01 6.33906543e-01
-5.54323137e-01 5.45938946e-02 4.41622846e-02 1.07504919e-01
1.16496921e+00 3.31338376e-01 5.46244621e-01 -3.07123721e-01
3.49463791e-01 -3.21093172e-01 1.19341958e+00 1.07036483e+00
-3.96285832e-01 -1.88652426e-02 8.48747194e-01 -4.35958534e-01
-5.31473219e-01 -1.03440511e+00 2.31809184e-01 1.06067514e+00
1.77960888e-01 9.83149707e-02 -4.31440204e-01 -9.02273893e-01
5.14625490e-01 1.20740342e+00 -6.64802074e-01 -3.56938690e-01
-2.12369993e-01 -3.92742187e-01 -1.39953911e-01 3.90981525e-01
2.72338361e-01 -9.07473624e-01 -1.05441272e+00 6.98372006e-01
2.24018946e-01 -6.23880029e-01 -6.88718557e-01 3.52918267e-01
-8.39624465e-01 -9.29487109e-01 -5.22744656e-01 7.64160305e-02
5.24165809e-01 4.42061871e-02 8.83300424e-01 4.65810075e-02
4.76393223e-01 6.89287126e-01 -5.48895001e-02 -4.65250224e-01
-4.63975340e-01 -2.42587015e-01 4.92290050e-01 1.43711597e-01
-1.29780024e-01 -4.10735875e-01 -8.53525400e-01 4.44118649e-01
-7.78867781e-01 -8.40266943e-02 4.73314017e-01 9.78499293e-01
8.35442066e-01 6.77437410e-02 1.32279778e+00 -9.67319667e-01
1.10192120e+00 -6.67781651e-01 -1.06596816e+00 5.67218244e-01
-1.34650660e+00 2.54171431e-01 1.04760623e+00 -7.11385489e-01
-1.36878288e+00 -1.51959628e-01 3.23095322e-01 -4.18984503e-01
-1.19187951e-01 8.33020926e-01 1.58036754e-01 5.49172163e-01
3.47245902e-01 -4.44747731e-02 -2.53593475e-01 -2.72587508e-01
-3.63304131e-02 6.55677617e-01 1.64361015e-01 -1.11943042e+00
3.94630969e-01 7.23546147e-02 -1.78976136e-03 2.60276422e-02
-1.00838816e+00 1.27396464e-01 -2.54545778e-01 -4.53549325e-01
3.98118496e-01 -9.99778211e-01 -1.12322164e+00 3.94428104e-01
-7.69663155e-01 -8.52236450e-01 -1.89108923e-01 4.45551306e-01
-7.98632443e-01 8.38364959e-02 -4.02289480e-01 -1.35328388e+00
-3.51919048e-02 -1.12901413e+00 4.80792642e-01 4.51400936e-01
-9.22800675e-02 -1.19463682e+00 -3.10883243e-02 -1.37749389e-01
1.78647771e-01 3.27374548e-01 8.56915772e-01 -4.97506768e-01
-4.27352726e-01 1.52907059e-01 3.10891211e-01 2.63293773e-01
2.78908104e-01 4.38151089e-03 -5.40572762e-01 -6.17710233e-01
-8.64065588e-02 -3.56196374e-01 6.19277596e-01 6.34987414e-01
1.03556895e+00 -7.59263992e-01 -2.27224380e-01 2.73220092e-02
1.18885481e+00 8.13323796e-01 1.55356722e-02 5.49631476e-01
5.56128509e-02 4.47691649e-01 1.25319862e+00 1.14093065e+00
6.56247139e-01 4.83840108e-01 6.59220576e-01 3.26756388e-01
3.87412727e-01 -7.69314349e-01 7.47192740e-01 6.41012117e-02
8.19560215e-02 -3.49251240e-01 -8.14771712e-01 6.10965431e-01
-2.03434777e+00 -7.37946153e-01 4.32858199e-01 2.96053267e+00
1.14018798e+00 7.97462761e-01 4.34416354e-01 -2.66515911e-01
5.45792401e-01 -1.95188537e-01 -1.35168445e+00 -6.41557276e-01
3.39185476e-01 -4.56579477e-01 5.40122569e-01 8.25357080e-01
-7.78742492e-01 7.78798282e-01 5.26376247e+00 6.81120872e-01
-1.45754421e+00 -1.10863566e-01 1.15636384e+00 2.21008845e-02
-3.48014265e-01 4.07319754e-01 -8.38035762e-01 8.63216400e-01
1.15007532e+00 -5.59238195e-01 5.66638589e-01 6.95119262e-01
8.47913086e-01 -5.74514568e-01 -1.45844185e+00 1.94048434e-01
-6.76199436e-01 -9.00566220e-01 -2.55793154e-01 -5.60764372e-02
7.32777417e-01 -3.22521180e-01 7.93398470e-02 6.27764523e-01
8.19004893e-01 -5.98853111e-01 1.18512094e+00 8.26813757e-01
6.35496616e-01 -1.01325095e+00 6.37042701e-01 9.38081503e-01
-9.84010637e-01 -6.56271577e-01 -4.85840626e-02 -4.47013825e-01
2.52215147e-01 8.40534985e-01 -8.78711283e-01 4.15603966e-01
3.17967862e-01 5.50887942e-01 -2.21171483e-01 5.12541473e-01
-6.10712528e-01 9.95047152e-01 -1.64625168e-01 -4.18079421e-02
2.51805842e-01 -1.52456239e-01 5.93753815e-01 7.47667313e-01
3.73012096e-01 -6.34886622e-02 8.60227346e-01 8.42701137e-01
4.14070636e-01 -5.36129475e-01 -3.98159504e-01 1.19447997e-02
9.39948082e-01 6.73852205e-01 -5.95461547e-01 -3.55754614e-01
-2.03912139e-01 6.08822405e-01 1.99871063e-01 7.02510297e-01
-6.37991905e-01 -1.26157641e-01 3.44928354e-01 1.74441472e-01
1.49858862e-01 -1.77973092e-01 -2.60482341e-01 -7.76839733e-01
-7.33359009e-02 -7.90301740e-01 5.53146541e-01 -4.04895812e-01
-1.28401947e+00 1.42597079e-01 9.59637314e-02 -1.22458839e+00
-7.13091433e-01 -1.27627715e-01 -7.74273634e-01 8.71053994e-01
-1.63022566e+00 -6.49657011e-01 5.53691626e-01 3.59267354e-01
7.44357407e-01 1.27293378e-01 3.37319791e-01 -3.07497352e-01
-6.09032691e-01 3.90712231e-01 1.13874987e-01 -4.73704010e-01
8.81744146e-01 -1.29991078e+00 -5.46476766e-02 8.32843244e-01
-6.59046113e-01 5.46337724e-01 8.98685992e-01 -8.81840110e-01
-1.25396991e+00 -9.85168397e-01 4.59463060e-01 -1.62804574e-01
7.32653022e-01 1.61941461e-02 -1.00122738e+00 5.81622183e-01
-1.46999195e-01 -1.90018103e-01 6.84704781e-02 7.37156123e-02
1.36672035e-01 -4.24859226e-01 -1.24091256e+00 5.44632852e-01
5.64258695e-01 -3.85388702e-01 -4.99258369e-01 -2.01361813e-03
8.48774254e-01 -1.86472490e-01 -4.88913238e-01 4.35851336e-01
5.12710571e-01 -1.03758240e+00 5.73924243e-01 -4.99461979e-01
2.93780357e-01 -2.89961696e-01 -8.12460184e-02 -1.63299394e+00
-2.09419996e-01 -1.10671902e+00 -2.90234745e-01 1.05610573e+00
2.03516826e-01 -9.66805458e-01 3.29050779e-01 7.72294998e-01
-3.65341548e-03 -9.80957687e-01 -1.14997852e+00 -1.04545915e+00
2.43323252e-01 -2.68232405e-01 5.18789291e-01 5.80357790e-01
-2.76769072e-01 6.78627566e-02 -6.72078609e-01 3.12774330e-01
5.50225616e-01 3.46748888e-01 4.74945724e-01 -9.23092544e-01
-6.01244509e-01 -2.01204807e-01 4.14575338e-01 -1.20149159e+00
1.92804247e-01 -2.48566702e-01 2.32662514e-01 -1.17245412e+00
3.09884697e-02 -6.76618338e-01 -7.48189807e-01 7.34272957e-01
-5.13362706e-01 -8.44491422e-01 2.99525261e-01 2.48738527e-01
-5.54130614e-01 7.09633350e-01 1.19048119e+00 9.02953297e-02
-4.77119684e-01 6.02536917e-01 -6.40283108e-01 6.76820099e-01
7.93058634e-01 -6.50497079e-01 -6.50983572e-01 -8.07160884e-02
2.71396399e-01 9.61720288e-01 3.97583753e-01 -6.99307382e-01
1.02220699e-01 -1.17719412e+00 -1.53666973e-01 -5.54066956e-01
-2.35385627e-01 -5.79079866e-01 -1.01216651e-01 9.64552224e-01
-7.33695447e-01 2.13364065e-01 9.53194350e-02 1.13980782e+00
9.01268125e-02 -1.63179100e-01 5.75446904e-01 3.82799618e-02
-4.65098053e-01 2.74792850e-01 -7.63935030e-01 1.40104383e-01
1.17148566e+00 3.32448572e-01 8.72219801e-02 -8.16893220e-01
-7.56617308e-01 8.48757386e-01 1.91388220e-01 1.57992274e-01
3.54434997e-01 -8.79478991e-01 -5.42097092e-01 -1.43862575e-01
-3.15199256e-01 -8.78327861e-02 3.06129485e-01 7.90301383e-01
2.92902708e-01 3.83638740e-01 1.36172593e-01 -4.59000498e-01
-7.35137463e-01 6.56421602e-01 4.64155406e-01 -6.25021160e-01
-1.35875300e-01 4.22624171e-01 3.82222503e-01 -1.40015170e-01
2.81442612e-01 -5.58233857e-01 1.52736470e-01 1.24451645e-01
3.93595845e-01 4.46692795e-01 -2.00617872e-02 2.36636430e-01
-2.60212839e-01 1.99673474e-01 -2.15227306e-01 -4.12940532e-01
1.09726644e+00 -5.49362957e-01 4.74190384e-01 8.22120965e-01
2.10333437e-01 -1.96174622e-01 -2.22954249e+00 -3.27691048e-01
2.27873668e-01 -3.28881174e-01 1.96289629e-01 -1.21034777e+00
-6.52496934e-01 7.51197517e-01 5.80671787e-01 2.78971761e-01
1.17426944e+00 -5.34990788e-01 4.65086132e-01 5.94433963e-01
7.58992851e-01 -1.45886898e+00 7.39430487e-02 5.70567727e-01
8.40832174e-01 -1.16528904e+00 1.22014262e-01 3.12317550e-01
-9.13733304e-01 9.29291129e-01 7.61116385e-01 9.39258114e-02
6.77363813e-01 1.95345104e-01 -4.96683493e-02 5.65228164e-01
-1.42725611e+00 9.84526575e-02 -1.27122819e-01 5.00645816e-01
1.08776569e-01 6.01094007e-01 -4.35081601e-01 6.40015304e-01
3.84809554e-01 4.83737379e-01 6.59893513e-01 1.14337695e+00
-4.16525006e-01 -9.59359229e-01 -3.78295571e-01 4.71135288e-01
-3.52310628e-01 2.25845009e-01 -1.39063403e-01 4.38797325e-01
1.57029480e-02 1.12133908e+00 1.38520217e-03 -2.42252156e-01
2.75619298e-01 -1.28875837e-01 -6.20023173e-04 -4.68312740e-01
-2.88095981e-01 3.64136070e-01 6.42899051e-02 -4.44927365e-01
-6.87150136e-02 -8.59317958e-01 -1.46472406e+00 -2.06706330e-01
-8.96485150e-02 -1.10130548e-01 1.21467650e-01 1.15532100e+00
5.05884826e-01 5.90313613e-01 1.25241554e+00 -6.19384289e-01
-1.55922341e+00 -6.04159772e-01 -4.40764040e-01 -3.03026475e-02
6.35463595e-01 -1.15432465e+00 -6.73316061e-01 -2.54397541e-01] | [4.141523361206055, 2.4477651119232178] |
98883a33-9625-4137-bc34-875a7e1cc64c | a-topic-based-sentence-representation-for | null | null | https://aclanthology.org/W19-8905 | https://aclanthology.org/W19-8905.pdf | A topic-based sentence representation for extractive text summarization | In this study, we examine the effect of probabilistic topic model-based word representations, on sentence-based extractive summarization. We formulate the task of summary extraction as a binary classification problem, and we test a variety of machine learning algorithms, exploring a range of different settings. An wide experimental evaluation on the MultiLing 2015 MSS dataset illustrates that topic-based representations can prove beneficial to the extractive summarization process in terms of F1, ROUGE-L and ROUGE-W scores, compared to a TF-IDF baseline, with QDA-based analysis providing the best results. | ['Nikiforos Pittaras', 'Nikolaos Gialitsis', 'Panagiotis Stamatopoulos'] | 2019-09-01 | null | null | null | ranlp-2019-9 | ['extractive-document-summarization'] | ['natural-language-processing'] | [ 2.90474385e-01 3.31353575e-01 -8.02390158e-01 -1.19947761e-01
-1.43452907e+00 -7.13491499e-01 1.10057366e+00 6.88361108e-01
-4.67969418e-01 9.32818532e-01 1.22740161e+00 -2.72086978e-01
-3.26854706e-01 -5.74810207e-01 -2.38659769e-01 -4.16652381e-01
-7.44837746e-02 4.70821381e-01 -2.05511954e-02 -3.50487024e-01
8.48648190e-01 9.06549469e-02 -1.29651129e+00 4.12356824e-01
1.10122144e+00 5.84519804e-01 7.40526756e-03 8.27720881e-01
-2.01643705e-01 4.06016320e-01 -1.11027205e+00 -5.77444971e-01
-2.24003091e-01 -2.83232421e-01 -1.01790392e+00 -8.47155452e-02
5.39807677e-01 -8.83530900e-02 -3.72794092e-01 7.92340040e-01
5.47630608e-01 3.28816324e-01 1.28755784e+00 -7.13128567e-01
-5.22383690e-01 1.01115119e+00 -5.08080542e-01 7.07041681e-01
6.84756398e-01 -1.26691431e-01 1.57669604e+00 -8.06203246e-01
8.27520013e-01 1.33760440e+00 4.91947591e-01 1.83684886e-01
-1.34099495e+00 -2.27593839e-01 -7.51586184e-02 8.55628587e-03
-9.22056139e-01 -4.19632763e-01 6.04645789e-01 -2.44887054e-01
1.39762878e+00 3.84668708e-01 3.75904649e-01 1.06188786e+00
6.60801411e-01 1.05675042e+00 7.04212606e-01 -5.39458990e-01
8.27159882e-02 -1.19648904e-01 7.24867463e-01 2.02025086e-01
6.94441080e-01 -5.80108345e-01 -7.82014787e-01 -4.98979449e-01
-8.66911188e-02 -2.44427472e-01 -3.27014714e-01 1.01270862e-01
-1.24326313e+00 1.32993460e+00 5.99545836e-02 3.50180537e-01
-7.69602418e-01 1.30134806e-01 7.93272257e-01 1.18781053e-01
1.22373462e+00 1.08070636e+00 -4.87052053e-01 -3.72795224e-01
-1.44611621e+00 6.89372897e-01 8.64971101e-01 7.17291534e-01
4.43999171e-01 8.66986886e-02 -9.63633358e-01 9.95371044e-01
-4.41935658e-02 5.21163225e-01 7.63976932e-01 -8.57337654e-01
8.46995771e-01 3.44701171e-01 4.91454080e-02 -9.18019772e-01
-3.97270322e-01 -6.46957755e-01 -5.58325231e-01 -6.82060599e-01
-2.94120044e-01 -3.78459066e-01 -1.01812887e+00 1.31668639e+00
-2.55973995e-01 -4.36195761e-01 3.50010097e-01 3.83559138e-01
1.24667513e+00 1.20463967e+00 2.57842660e-01 -5.25619149e-01
1.23119068e+00 -9.46842968e-01 -1.09138298e+00 -3.76219213e-01
8.21751595e-01 -8.58237803e-01 7.73397803e-01 3.41809750e-01
-1.25897539e+00 -1.85786128e-01 -1.15764737e+00 -1.30061284e-01
-4.23016161e-01 3.27937186e-01 6.71026886e-01 6.07766926e-01
-9.93451357e-01 7.78805733e-01 -5.91659248e-01 -4.71616775e-01
5.10604799e-01 -2.89852265e-02 -2.20083982e-01 -6.69847280e-02
-1.16199505e+00 1.04457128e+00 8.12838435e-01 -5.09276330e-01
-6.54939473e-01 -6.72758996e-01 -9.72983539e-01 3.56246322e-01
2.63467014e-01 -8.38957191e-01 1.28917027e+00 7.46826874e-03
-1.13942122e+00 5.82634091e-01 -4.02097315e-01 -9.61011529e-01
-1.16428938e-02 -5.60317636e-01 -2.22018749e-01 3.74214739e-01
4.15587306e-01 6.37213051e-01 5.12032151e-01 -9.33839262e-01
-7.09290683e-01 -2.24244036e-02 -8.63069221e-02 3.51056635e-01
-4.19878602e-01 1.83278665e-01 4.65976447e-02 -9.14136291e-01
-1.42560050e-01 -6.90694034e-01 -1.80813640e-01 -1.21387291e+00
-7.86103427e-01 -7.79247999e-01 6.20784461e-01 -9.72085059e-01
1.53955340e+00 -1.68097043e+00 3.36544752e-01 -2.24688545e-01
1.15068533e-01 2.78994262e-01 -1.93834499e-01 8.54051828e-01
2.13429511e-01 6.39306009e-01 -1.98295653e-01 -4.54276204e-01
4.76059876e-03 -1.87922493e-01 -6.44904613e-01 1.85350388e-01
5.14573753e-01 9.40287828e-01 -1.01665759e+00 -7.31111407e-01
-5.98230287e-02 8.83353055e-02 -4.34976727e-01 2.87838019e-02
-2.72734314e-01 -2.79125661e-01 -6.13984585e-01 4.30127203e-01
2.88924038e-01 3.91420685e-02 7.65558630e-02 1.35103911e-01
-1.41490936e-01 1.07044530e+00 -2.60028660e-01 1.82547259e+00
-3.78969073e-01 1.10032904e+00 -4.98398542e-01 -8.16365659e-01
8.00066710e-01 3.26480359e-01 2.18353346e-01 -2.63081282e-01
8.28947574e-02 -9.10020769e-02 -5.47177345e-02 -2.28112713e-01
1.38618326e+00 -9.71636996e-02 -5.32551050e-01 6.57800138e-01
4.57723200e-01 -4.88774389e-01 7.04775453e-01 8.85065436e-01
1.19677532e+00 -3.04980606e-01 5.44809461e-01 -6.59372032e-01
4.07909229e-02 3.91095847e-01 1.36656895e-01 8.56188893e-01
1.64911836e-01 6.41478181e-01 7.76558042e-01 9.84515697e-02
-9.97833669e-01 -1.07715869e+00 -2.35713959e-01 1.12646723e+00
-3.26770842e-01 -9.66075718e-01 -6.35906518e-01 -7.47842252e-01
-3.40228789e-02 1.48919380e+00 -5.46973586e-01 -4.98036414e-01
-4.88571048e-01 -1.03695476e+00 6.90703452e-01 3.36620480e-01
5.82393557e-02 -1.02406907e+00 -5.64629495e-01 2.10301399e-01
-6.03615999e-01 -9.48680878e-01 -4.25207317e-01 2.33003303e-01
-1.20610952e+00 -4.69667673e-01 -9.52001691e-01 -3.91086966e-01
2.07575541e-02 2.88582653e-01 1.39329803e+00 -2.71290094e-01
5.03702872e-02 3.72056127e-01 -7.11625218e-01 -7.69127667e-01
-5.18677056e-01 9.61079717e-01 -8.40418190e-02 -7.90413797e-01
2.31552750e-01 -1.31036043e-01 -3.79591703e-01 -5.74444532e-01
-8.88271332e-01 -3.45513076e-01 6.28368914e-01 8.66203606e-01
1.65667519e-01 -1.93709567e-01 1.09979701e+00 -9.89906490e-01
1.50144553e+00 -5.26098788e-01 3.19832087e-01 2.20308870e-01
-7.84203649e-01 1.37270972e-01 2.75722891e-01 -2.21605048e-01
-1.04449356e+00 -7.68586516e-01 -3.22647020e-02 1.55414730e-01
1.47757232e-01 1.05467570e+00 1.87084898e-01 7.40835428e-01
8.41726184e-01 4.66095269e-01 -2.34815195e-01 -3.87041181e-01
6.01624668e-01 6.92139149e-01 3.25177312e-01 -3.29701841e-01
5.07248998e-01 7.08729029e-02 -3.01412940e-01 -1.11334634e+00
-1.23489130e+00 -6.88829720e-01 -4.64492410e-01 7.83201382e-02
8.56995046e-01 -9.56056654e-01 2.55338371e-01 -1.62892397e-02
-1.40818465e+00 1.40180901e-01 -4.78445143e-01 4.20873761e-01
-5.78168929e-01 5.12371182e-01 -4.98680860e-01 -6.67493582e-01
-8.69539201e-01 -8.16073954e-01 1.27072430e+00 2.90376693e-01
-8.40196490e-01 -1.00799787e+00 2.98075229e-01 3.15842479e-01
3.25705141e-01 2.32485130e-01 1.21342766e+00 -1.11018825e+00
-3.93963885e-03 -3.75708669e-01 -7.86052123e-02 3.27719569e-01
2.01894611e-01 2.48575304e-02 -7.33016849e-01 -1.85952395e-01
-2.14154154e-01 -2.88223773e-01 1.54905045e+00 7.56491840e-01
6.86096311e-01 -5.67957163e-01 -3.47409993e-01 4.75751907e-02
1.10858250e+00 -1.70636699e-01 6.62671626e-01 4.49141771e-01
3.74245286e-01 6.03471100e-01 9.29423451e-01 5.75219572e-01
3.17072332e-01 3.33325088e-01 -2.98290495e-02 2.72583008e-01
-3.12361300e-01 -1.86934650e-01 3.95964265e-01 1.09112525e+00
9.39636528e-02 -7.35557318e-01 -8.86292040e-01 7.93077052e-01
-1.65818286e+00 -1.04001188e+00 -7.55805671e-02 1.82756329e+00
8.06163609e-01 3.16305518e-01 2.39058912e-01 2.41374299e-01
6.05795801e-01 8.04818690e-01 6.49367571e-02 -8.51451516e-01
-1.49187610e-01 3.18714499e-01 4.92250621e-01 1.98787346e-01
-1.13705385e+00 1.17056978e+00 7.52013397e+00 1.10730183e+00
-7.99219370e-01 7.17837512e-02 5.41739225e-01 -1.98170364e-01
-5.50657928e-01 -1.53648004e-01 -7.59013057e-01 3.18071544e-01
1.43586397e+00 -8.51040900e-01 -2.57566422e-01 6.26511455e-01
2.50310183e-01 -3.28015238e-01 -9.53106999e-01 4.08808827e-01
4.28065151e-01 -1.61807811e+00 5.55829287e-01 5.56258447e-02
8.66969883e-01 1.05527252e-01 3.66631486e-02 6.32456303e-01
1.82970583e-01 -8.85412872e-01 6.39995396e-01 3.75153214e-01
5.57465374e-01 -9.67537940e-01 9.32481587e-01 3.27063650e-01
-5.42446196e-01 2.69993216e-01 -4.72501904e-01 2.88499668e-02
2.94417202e-01 7.73379326e-01 -1.06315207e+00 1.00224495e+00
4.04077053e-01 8.88947606e-01 -6.56905115e-01 9.80177224e-01
-2.01508075e-01 1.03844798e+00 1.14036500e-01 -4.60420519e-01
4.38674718e-01 3.23766954e-02 8.93200099e-01 1.81270862e+00
2.52328157e-01 -2.61417683e-02 -3.59231904e-02 5.39917588e-01
-3.32506329e-01 3.04310292e-01 -5.98427415e-01 -4.68757629e-01
4.47300583e-01 1.00614357e+00 -7.29671955e-01 -6.89583719e-01
1.54023662e-01 6.25252664e-01 1.41327560e-01 2.14392111e-01
-5.37972510e-01 -7.73149848e-01 3.68135840e-01 -8.15963894e-02
4.82032657e-01 -3.77731472e-01 -4.85806435e-01 -1.15700245e+00
-3.67597103e-01 -8.10552478e-01 3.37054342e-01 -5.58577120e-01
-9.39396083e-01 5.03400207e-01 4.89876509e-01 -1.02003837e+00
-4.61300313e-01 8.94214120e-03 -9.12512422e-01 7.26209939e-01
-1.50062585e+00 -8.22999299e-01 3.96186262e-01 -3.78467441e-01
1.10247350e+00 -2.55557507e-01 6.68781817e-01 -1.66548908e-01
-5.16779900e-01 3.97573680e-01 3.87288272e-01 -9.85353217e-02
6.40973449e-01 -1.38301611e+00 6.16564870e-01 7.97699094e-01
3.39327663e-01 5.91811121e-01 1.25031674e+00 -9.18466985e-01
-1.17134345e+00 -1.21389687e+00 1.07605433e+00 -5.60514271e-01
5.99569857e-01 3.95579860e-02 -7.19193459e-01 5.69633126e-01
6.75576270e-01 -9.31914091e-01 7.42728591e-01 6.93773150e-01
-9.75809842e-02 4.65245277e-01 -9.54735160e-01 3.88030201e-01
7.04976320e-01 -3.68984699e-01 -1.26660705e+00 5.12422144e-01
1.09800053e+00 -1.30194034e-02 -1.04993463e+00 3.82651269e-01
3.31210583e-01 -4.87649798e-01 1.02892065e+00 -9.61357355e-01
9.64331686e-01 2.46037677e-01 -7.62111023e-02 -1.91349566e+00
-4.62425500e-01 -3.48107487e-01 -2.81681865e-01 1.66022217e+00
8.46398652e-01 -4.46695447e-01 5.85319638e-01 7.89530203e-02
-3.36360306e-01 -7.47806847e-01 -9.53636706e-01 -8.52203488e-01
6.08296752e-01 -1.42560393e-01 2.82391459e-01 4.81592000e-01
2.59938985e-01 9.52088892e-01 -2.86959499e-01 -4.22756761e-01
3.49396318e-01 1.52412236e-01 6.54196739e-01 -1.09594953e+00
2.12708399e-01 -8.20071459e-01 -2.84486204e-01 -8.42728496e-01
3.99262041e-01 -9.93653536e-01 1.34974107e-01 -2.15426350e+00
5.16052902e-01 1.99678808e-01 -8.57474357e-02 3.00770887e-04
-6.67267621e-01 -2.45462447e-01 3.22311789e-01 3.45169514e-01
-8.97352040e-01 8.22188437e-01 9.49863613e-01 -4.84891474e-01
-6.74047172e-02 -3.51278558e-02 -1.00569737e+00 4.49753195e-01
8.21286142e-01 -7.88716495e-01 -2.61940151e-01 -4.30665851e-01
2.57792603e-02 1.00770310e-01 -2.62683928e-01 -7.19845772e-01
-5.94107471e-02 1.67334452e-02 1.04803696e-01 -1.00955129e+00
3.20258528e-01 1.50350690e-01 -5.55354536e-01 4.56745535e-01
-8.89264405e-01 1.71443492e-01 2.01812595e-01 7.03655124e-01
-3.24669689e-01 -5.82000434e-01 2.27165565e-01 7.79092684e-02
-2.09651366e-01 -1.21631563e-01 -7.49166906e-01 4.08320516e-01
5.43836772e-01 -3.52249928e-02 -5.06147027e-01 -4.97936636e-01
-1.47001773e-01 6.98440522e-02 1.84459552e-01 3.37037265e-01
5.30180275e-01 -9.72578943e-01 -1.31232679e+00 -5.46106875e-01
2.18847305e-01 -2.96141446e-01 7.71418810e-02 7.25642204e-01
-2.40999550e-01 9.67012107e-01 1.06510431e-01 -4.53422993e-01
-1.05764198e+00 3.12025845e-01 -3.94924283e-01 -8.27062249e-01
-6.49232626e-01 4.98191953e-01 -5.94737791e-02 -1.04634270e-01
1.01952925e-02 -4.62551117e-01 -6.74703777e-01 5.65665781e-01
3.77101421e-01 6.83529973e-01 2.11281925e-01 -5.57085872e-01
-1.99268371e-01 1.08655803e-01 -4.64682281e-01 -3.59801322e-01
1.31206405e+00 -3.30653042e-02 -1.41966753e-02 5.96511722e-01
1.18952715e+00 7.16937631e-02 -4.51312721e-01 -6.12715669e-02
6.51225507e-01 -2.66787738e-01 4.18748885e-01 -7.91427016e-01
-4.13458735e-01 7.02200353e-01 -1.74316645e-01 6.72004461e-01
7.36083925e-01 2.15641171e-01 9.06336069e-01 5.01906455e-01
-7.02536711e-03 -1.10550725e+00 -3.14854123e-02 7.28349030e-01
1.02183199e+00 -1.13852632e+00 5.15004098e-01 -3.01864743e-01
-8.70789051e-01 1.03250241e+00 -6.55751973e-02 -3.70673895e-01
2.94277877e-01 -1.75011218e-01 -3.96259546e-01 -3.17647189e-01
-1.08238626e+00 -9.13959667e-02 6.70636177e-01 4.02044088e-01
7.36626148e-01 1.85370147e-01 -8.66020977e-01 7.22118795e-01
-6.25261009e-01 -4.72722739e-01 8.87459815e-01 7.02585280e-01
-8.00183833e-01 -7.88280725e-01 -2.46144012e-02 1.20598960e+00
-8.95487309e-01 -3.75615299e-01 -6.02280259e-01 7.66366661e-01
-6.17105842e-01 1.29532576e+00 7.40136653e-02 -1.86214164e-01
1.44906938e-01 1.98076755e-01 2.37942427e-01 -1.18374181e+00
-7.73279369e-01 9.28735435e-02 8.52549970e-01 3.27348076e-02
-4.36721236e-01 -1.07990158e+00 -9.29464221e-01 -2.09723324e-01
-6.73751831e-01 6.13052309e-01 8.05980504e-01 1.01133204e+00
4.13322151e-01 8.20875585e-01 5.39956152e-01 -7.69680083e-01
-8.20785522e-01 -1.62658668e+00 -3.60137641e-01 1.47841200e-01
3.16625178e-01 -4.55988199e-01 -4.23372686e-01 -1.55101627e-01] | [12.438705444335938, 9.513208389282227] |
3ad7fc1d-d859-4726-bb59-7a1b19bffed8 | triviaqa-a-large-scale-distantly-supervised | 1705.03551 | null | http://arxiv.org/abs/1705.03551v2 | http://arxiv.org/pdf/1705.03551v2.pdf | TriviaQA: A Large Scale Distantly Supervised Challenge Dataset for Reading Comprehension | We present TriviaQA, a challenging reading comprehension dataset containing
over 650K question-answer-evidence triples. TriviaQA includes 95K
question-answer pairs authored by trivia enthusiasts and independently gathered
evidence documents, six per question on average, that provide high quality
distant supervision for answering the questions. We show that, in comparison to
other recently introduced large-scale datasets, TriviaQA (1) has relatively
complex, compositional questions, (2) has considerable syntactic and lexical
variability between questions and corresponding answer-evidence sentences, and
(3) requires more cross sentence reasoning to find answers. We also present two
baseline algorithms: a feature-based classifier and a state-of-the-art neural
network, that performs well on SQuAD reading comprehension. Neither approach
comes close to human performance (23% and 40% vs. 80%), suggesting that
TriviaQA is a challenging testbed that is worth significant future study. Data
and code available at -- http://nlp.cs.washington.edu/triviaqa/ | ['Luke Zettlemoyer', 'Mandar Joshi', 'Eunsol Choi', 'Daniel S. Weld'] | 2017-05-09 | triviaqa-a-large-scale-distantly-supervised-1 | https://aclanthology.org/P17-1147 | https://aclanthology.org/P17-1147.pdf | acl-2017-7 | ['triviaqa'] | ['miscellaneous'] | [-1.56878699e-02 3.87789994e-01 3.33361551e-02 -5.27585328e-01
-1.60675788e+00 -9.23904777e-01 3.99709225e-01 5.88433921e-01
-4.32804197e-01 8.40212822e-01 5.67023218e-01 -7.92312384e-01
-2.85380781e-01 -6.52232051e-01 -1.00223053e+00 4.10132296e-02
2.20490143e-01 1.07466114e+00 6.42463982e-01 -6.81990206e-01
5.41745722e-01 -3.95253390e-01 -1.37427258e+00 7.42578864e-01
1.22023475e+00 1.07240021e+00 -5.46833733e-03 1.27198553e+00
-3.03162992e-01 1.58296156e+00 -8.59002948e-01 -9.10279453e-01
-1.90445229e-01 -6.17531121e-01 -1.79361916e+00 -4.85912591e-01
1.57424760e+00 -5.62143147e-01 -7.99070150e-02 6.62686229e-01
3.40579867e-01 -1.07638119e-02 5.91056705e-01 -8.92381608e-01
-1.03232586e+00 8.35900187e-01 -2.73254663e-01 9.16299641e-01
1.22637272e+00 1.65013120e-01 1.58635902e+00 -8.42074871e-01
3.65519315e-01 1.13482869e+00 6.46493733e-01 5.40020227e-01
-9.83135402e-01 -5.37624896e-01 -2.19096661e-01 6.07450128e-01
-5.42378604e-01 -5.38761139e-01 4.81538713e-01 -3.60603780e-01
1.29058492e+00 4.40888494e-01 3.88386399e-01 1.24961340e+00
3.19042683e-01 1.12444007e+00 1.26618969e+00 -4.96247292e-01
3.18146721e-02 -1.86699361e-01 1.02544653e+00 1.09379005e+00
-1.49320930e-01 -4.01797116e-01 -7.81494975e-01 -2.59491205e-01
5.26325032e-02 -7.90124238e-01 -5.58831930e-01 -5.95828379e-03
-1.31484115e+00 8.87121022e-01 4.89622831e-01 2.29515597e-01
-8.41801763e-02 8.30597952e-02 4.43922132e-01 9.86461997e-01
1.30847856e-01 8.99728954e-01 -8.74303162e-01 -4.46616381e-01
-8.51951718e-01 6.19480669e-01 1.43804252e+00 8.16391051e-01
4.04163241e-01 -5.38008809e-01 -2.33508125e-01 7.14607894e-01
1.10578768e-01 5.75218260e-01 5.25020838e-01 -1.40746558e+00
1.27465904e+00 5.70030510e-01 5.45900241e-02 -1.03960896e+00
-5.18026948e-01 -3.69534373e-01 -3.94610941e-01 -4.20576811e-01
9.34495032e-01 3.42479683e-02 -3.23772043e-01 1.63358068e+00
1.11072592e-01 -2.53689975e-01 2.18335435e-01 7.36733794e-01
1.43676460e+00 6.47625566e-01 -1.33138761e-01 9.07657892e-02
1.73123491e+00 -1.55499709e+00 -6.16214931e-01 -5.18391967e-01
5.41255534e-01 -7.15140462e-01 1.56620741e+00 4.02142197e-01
-1.80519009e+00 -5.64212143e-01 -9.96274173e-01 -5.95801890e-01
-1.90771088e-01 -1.97688058e-01 2.57601470e-01 3.13305438e-01
-1.07438672e+00 1.58544675e-01 -2.70519704e-01 -2.87660539e-01
2.78680891e-01 -8.33458360e-03 -2.81790555e-01 -2.37782717e-01
-1.40555871e+00 1.25091636e+00 1.09726079e-01 -3.29093188e-01
-9.44572866e-01 -7.58018017e-01 -8.68242860e-01 2.93888181e-01
5.20742953e-01 -9.98037279e-01 2.15412903e+00 -7.41947174e-01
-1.32811880e+00 1.08184588e+00 -3.30168873e-01 -4.72448021e-01
2.39495069e-01 -6.54344082e-01 -2.84802824e-01 6.55372679e-01
3.54026258e-01 5.58800220e-01 6.42625391e-01 -7.72340417e-01
-4.40651685e-01 -3.31183285e-01 5.03427207e-01 1.91913262e-01
-4.98419590e-02 2.66325057e-01 1.54531458e-02 -3.04579556e-01
9.78028551e-02 -2.63074011e-01 3.38770181e-01 -2.20022827e-01
-1.76316723e-01 -9.89016891e-01 4.46923040e-02 -1.01953816e+00
1.15928984e+00 -1.44039571e+00 2.43708536e-01 -2.85361409e-01
5.51697254e-01 1.99244305e-01 -4.24847275e-01 6.27833188e-01
1.29498675e-01 5.64493388e-02 -3.04507971e-01 -1.44537956e-01
1.40016347e-01 1.71645835e-01 -5.44080377e-01 2.69330498e-02
3.82216275e-01 1.32574606e+00 -1.17240179e+00 -6.29420638e-01
-3.05938303e-01 -3.85375768e-01 -5.40074706e-01 4.33540225e-01
-7.63716280e-01 3.65494825e-02 -4.27484840e-01 5.78605354e-01
3.77305388e-01 -7.55289316e-01 -3.02826822e-01 1.30056679e-01
4.19292569e-01 1.11914659e+00 -4.32696313e-01 1.69198930e+00
-3.44404310e-01 8.27210486e-01 7.90244266e-02 -7.43621707e-01
7.72795379e-01 2.63569653e-01 -4.52614158e-01 -9.48743403e-01
2.47812599e-01 4.72807556e-01 5.68743944e-02 -9.84357595e-01
6.59503996e-01 9.80540439e-02 -1.75656170e-01 6.08522356e-01
4.09424245e-01 -4.17779654e-01 5.24161458e-01 6.35057628e-01
1.51138282e+00 -2.35866383e-01 1.96957350e-01 -4.94773179e-01
9.23004448e-01 1.90082118e-01 6.19525500e-02 9.70761180e-01
-4.10386264e-01 5.02215922e-01 6.99357450e-01 -3.48461121e-01
-7.25063682e-01 -1.35825908e+00 -1.04452126e-01 1.45823574e+00
-1.97944921e-02 -4.60039139e-01 -8.91488492e-01 -9.01808202e-01
-5.59122451e-02 1.09867895e+00 -5.55477679e-01 1.29512504e-01
-7.94952095e-01 6.68999180e-02 8.44565451e-01 7.24414110e-01
7.06823945e-01 -1.10451257e+00 -6.04415953e-01 1.56315938e-01
-8.05662274e-01 -8.84779871e-01 -4.00420040e-01 2.68567741e-01
-8.16610873e-01 -1.40900314e+00 -6.07739031e-01 -1.04570973e+00
1.96319476e-01 -1.65942423e-02 2.28729582e+00 4.66530055e-01
3.05079252e-01 7.43989468e-01 -5.75997293e-01 -2.21037224e-01
-6.47343457e-01 3.64632696e-01 -4.60064054e-01 -8.64689291e-01
7.82300353e-01 -2.99107969e-01 -4.29795980e-01 4.04939801e-01
-5.30573666e-01 -2.03737155e-01 4.00391549e-01 1.03321087e+00
3.73657495e-02 -5.08438766e-01 1.03143942e+00 -7.71906137e-01
1.11487174e+00 -7.34622180e-01 -3.48086715e-01 7.33250022e-01
-2.99299777e-01 -5.41597940e-02 5.45307875e-01 -1.94101762e-02
-9.40819442e-01 -7.55903661e-01 -5.45028329e-01 3.51675570e-01
-2.23255500e-01 5.59267998e-01 1.38188630e-01 4.15676653e-01
1.18225849e+00 4.22091633e-02 -3.89457606e-02 -3.58635068e-01
5.61858714e-01 7.52206504e-01 8.94199312e-01 -8.33596647e-01
6.07417703e-01 -1.65679753e-01 -6.04368865e-01 -5.22318125e-01
-1.61122227e+00 -5.54758787e-01 -4.98408467e-01 -4.25164700e-02
9.07466590e-01 -7.10849524e-01 -7.54383147e-01 2.82544613e-01
-1.47601604e+00 -3.80079478e-01 -1.68479845e-01 1.06484383e-01
-7.33667433e-01 4.29240853e-01 -1.12895691e+00 -4.21972364e-01
-5.91600239e-01 -7.93183148e-01 8.93488109e-01 3.14498037e-01
-7.14730918e-01 -1.06547320e+00 3.43876392e-01 1.57653069e+00
3.33019018e-01 -3.02929163e-01 1.21159267e+00 -8.81994605e-01
-4.39884961e-01 -3.42488699e-02 -2.65659183e-01 4.48324025e-01
-2.71766096e-01 -3.58590662e-01 -8.36555958e-01 -1.19588435e-01
1.60126403e-01 -1.48915911e+00 9.29152548e-01 -1.31597310e-01
9.52554703e-01 -3.64686757e-01 3.26008350e-01 -1.96426779e-01
8.27002883e-01 -3.93905818e-01 4.69699889e-01 2.94429451e-01
1.67697847e-01 8.11348200e-01 3.38108778e-01 -2.86896586e-01
1.03255415e+00 2.34790891e-01 2.84245580e-01 5.21879971e-01
-2.66731054e-01 -4.86886591e-01 5.46306193e-01 1.48994482e+00
5.66703141e-01 -5.25567353e-01 -1.19695163e+00 8.99505317e-01
-1.73419034e+00 -1.00245512e+00 -6.22624576e-01 1.46636319e+00
1.18624711e+00 1.86664134e-01 9.66767743e-02 1.30355343e-01
8.05972144e-02 2.07181677e-01 -5.39486051e-01 -7.44653106e-01
-4.21932429e-01 6.76223397e-01 -2.63753414e-01 7.36226797e-01
-7.52173960e-01 6.36465430e-01 7.04580927e+00 6.78101003e-01
-9.26823393e-02 1.62254319e-01 4.41756219e-01 3.21071178e-01
-5.95279753e-01 -1.23774029e-01 -5.78639209e-01 1.97555363e-01
1.26921189e+00 1.23726251e-02 1.90775037e-01 5.33853948e-01
-4.97807980e-01 -5.10279119e-01 -1.28246629e+00 5.03679514e-01
6.74106598e-01 -1.28202987e+00 2.78034788e-02 -6.63515806e-01
5.37974596e-01 2.71933436e-01 -7.85351545e-02 6.32169485e-01
5.10897517e-01 -1.19147527e+00 7.14468837e-01 4.07815695e-01
1.68358624e-01 -4.13417667e-01 9.98786032e-01 7.65388072e-01
-6.90989137e-01 -1.58251315e-01 -3.58687162e-01 -4.10964966e-01
3.04283470e-01 2.36569420e-01 -5.60639560e-01 4.03970420e-01
1.04304326e+00 4.23294932e-01 -1.13047957e+00 8.16979289e-01
-9.62299585e-01 1.22968137e+00 -1.60444781e-01 -5.53290546e-01
3.05370927e-01 2.39292875e-01 3.58002603e-01 8.97483051e-01
-2.94921666e-01 1.77248955e-01 2.41079908e-02 8.42460334e-01
-3.61193478e-01 1.03761226e-01 -1.72866762e-01 -1.56167196e-02
2.94510782e-01 8.96429181e-01 -1.94365725e-01 -4.70840961e-01
-5.40392876e-01 8.32797289e-01 9.87581491e-01 1.23791527e-02
-5.62851727e-01 -4.55414325e-01 1.59020741e-02 -5.84488884e-02
1.57463565e-01 -1.24083884e-01 -1.27971664e-01 -1.45347989e+00
3.57132196e-01 -1.41680455e+00 8.02537084e-01 -1.26802862e+00
-1.91765845e+00 5.51802456e-01 -1.93786785e-01 -4.61832911e-01
-3.96514446e-01 -9.32751894e-01 -6.63027883e-01 7.56354272e-01
-1.63656831e+00 -7.77573824e-01 -3.82274896e-01 4.92537379e-01
8.91362607e-01 -1.04774207e-01 7.84977555e-01 6.07814267e-02
-1.93054080e-01 6.73212647e-01 -3.02124202e-01 3.08308721e-01
6.50958240e-01 -1.63091052e+00 5.21381915e-01 4.07892853e-01
3.93721044e-01 6.28064513e-01 5.87811768e-01 -3.63972187e-01
-1.19857609e+00 -3.34656000e-01 1.47743404e+00 -1.51971316e+00
1.09118807e+00 -1.30360052e-01 -1.29498339e+00 8.10133874e-01
1.06101489e+00 -5.47429919e-01 9.55311477e-01 5.07714152e-01
-8.10376644e-01 6.87217414e-02 -1.03493547e+00 4.50205952e-01
8.16558063e-01 -9.00287688e-01 -1.87718928e+00 5.60522676e-01
9.64203358e-01 -5.11974692e-01 -8.47697437e-01 2.76443273e-01
3.54372293e-01 -1.18915546e+00 6.79755151e-01 -1.06842804e+00
1.02805257e+00 -1.22077428e-01 -2.79300630e-01 -1.06728280e+00
-2.32282966e-01 -2.97650814e-01 -5.11242092e-01 1.06200993e+00
8.48661840e-01 -4.15238291e-01 6.33527458e-01 4.49854374e-01
-1.61670908e-01 -9.24905837e-01 -1.03644943e+00 -5.68218172e-01
7.95221567e-01 -3.55681628e-01 3.59431624e-01 6.57750070e-01
3.59738797e-01 1.09704280e+00 1.38397157e-01 -8.58580470e-02
3.34961742e-01 2.44214758e-01 6.91100061e-01 -9.63985622e-01
-4.78072554e-01 -4.75637823e-01 6.36247024e-02 -1.75488055e+00
4.93015379e-01 -1.08818710e+00 3.58982496e-02 -1.68855798e+00
2.07269236e-01 4.56216671e-02 6.47966713e-02 2.65364081e-01
-4.15023834e-01 -6.72837645e-02 -8.67754444e-02 -1.38533965e-01
-1.03003073e+00 6.23212636e-01 1.34423113e+00 -3.27070802e-01
4.88494813e-01 -1.43990830e-01 -8.17330658e-01 7.27730989e-01
9.40640032e-01 -1.05924472e-01 -4.54551816e-01 -1.01214755e+00
8.57292414e-01 1.53972968e-01 4.29276705e-01 -8.75033915e-01
4.97473568e-01 2.07753465e-01 2.10341558e-01 -8.23429883e-01
1.79445162e-01 -3.67499530e-01 -9.54697430e-01 3.21975410e-01
-8.24474812e-01 6.92063212e-01 5.62926531e-02 4.34123844e-01
-4.32003438e-01 -7.59739697e-01 4.73944098e-01 -3.75819713e-01
-4.10265177e-01 -4.87573028e-01 -4.45507556e-01 1.09710169e+00
4.35296476e-01 1.56174183e-01 -1.07281148e+00 -7.21972942e-01
-2.03035980e-01 8.39915872e-01 1.10332426e-02 5.09636223e-01
7.39251077e-01 -9.60905731e-01 -1.05246770e+00 -3.05558592e-01
3.71114552e-01 7.91747496e-02 1.03623569e-01 7.38891125e-01
-8.11966836e-01 6.45060122e-01 1.61966830e-01 -6.28323019e-01
-1.21157598e+00 3.56419861e-01 3.19502592e-01 -3.58376354e-01
-2.69290835e-01 1.32969773e+00 -3.70663166e-01 -9.80031848e-01
2.30716050e-01 -5.65458477e-01 -4.11636919e-01 1.27235383e-01
7.98495829e-01 5.07312715e-01 2.16186330e-01 -3.42870444e-01
-1.61920294e-01 5.92754066e-01 -2.03962967e-01 2.63794102e-02
9.81053114e-01 -1.17796347e-01 -4.20280427e-01 6.35803819e-01
1.02155089e+00 9.00373086e-02 -5.73394299e-01 -3.62889051e-01
4.22986895e-01 -1.47411630e-01 -3.26993376e-01 -1.32204735e+00
-4.06789094e-01 1.00067139e+00 -3.59165780e-02 3.72099638e-01
8.12618494e-01 6.32888079e-01 1.14240015e+00 1.00330508e+00
3.13854605e-01 -8.81538987e-01 7.35859632e-01 9.67734993e-01
1.30897355e+00 -1.33865154e+00 -2.42807597e-01 -1.71919733e-01
-3.34677696e-01 1.01670158e+00 9.78985250e-01 -1.08368151e-01
2.43339807e-01 -1.50046036e-01 2.87607044e-01 -5.58233619e-01
-1.21103001e+00 6.01649657e-02 5.21278858e-01 2.89422870e-01
5.40752232e-01 -1.45784870e-01 -1.00281276e-01 6.71748400e-01
-8.90033007e-01 -2.23754495e-01 5.40133715e-01 8.38185072e-01
-8.47153485e-01 -8.30970824e-01 -3.07754695e-01 6.17215753e-01
-2.32876629e-01 -2.90052891e-01 -7.27041721e-01 5.16353071e-01
-3.06058049e-01 1.57423115e+00 3.17273773e-02 -4.19792458e-02
4.12281483e-01 3.67487311e-01 7.55632281e-01 -5.35643339e-01
-1.08128917e+00 -8.83105159e-01 5.46129704e-01 -4.44290787e-01
-2.87431866e-01 -4.70932037e-01 -1.08596742e+00 -4.39535916e-01
-4.72375005e-01 5.57812154e-01 5.43499924e-02 1.04900062e+00
1.94459647e-01 3.81358951e-01 1.41067341e-01 1.53270200e-01
-1.05984020e+00 -1.23393726e+00 -5.87841496e-04 3.57548714e-01
6.01012468e-01 -3.15475762e-01 -6.52713418e-01 -2.31349185e-01] | [11.247316360473633, 8.047660827636719] |
85cf4885-f131-4bff-a594-b2135c767d9d | towards-explainable-dialogue-system | null | null | https://aclanthology.org/2021.icon-main.16 | https://aclanthology.org/2021.icon-main.16.pdf | Towards Explainable Dialogue System: Explaining Intent Classification using Saliency Techniques | Deep learning based methods have shown tremendous success in several Natural Language Processing (NLP) tasks. The recent trends in the usage of Deep Learning based models for natural language tasks have definitely produced incredible performance for several application areas. However, one major problem that most of these models face is the lack of transparency, i.e. the actual decision process of the underlying model is not explainable. In this paper, at first we solve a very fundamental problem of Natural Language Understanding (NLU), i.e. intent detection using a Bi-directional Long Short Term Memory (BiLSTM). In order to determine the defining features that lead to a specific intent class, we use the Layerwise Relevance Propagation (LRP) algorithm to find the defining feature(s). In the process, we conclude that saliency method of eLRP (epsilon Layerwise Relevance Propagation) is a prominent process for highlighting the important features of the input responsible for the current classification which results in significant insights to the inner workings, such as the reasons for misclassification by the black box model. | ['Asif Ekbal', 'Arindam Chatterjee', 'Ratnesh Joshi'] | null | null | null | null | icon-2021-12 | ['intent-detection', 'intent-classification'] | ['natural-language-processing', 'natural-language-processing'] | [ 3.88216585e-01 4.24937516e-01 -1.77099153e-01 -4.74698901e-01
-3.90549839e-01 -4.21733148e-02 6.59819186e-01 5.39762199e-01
-2.88889915e-01 6.08379602e-01 4.18030262e-01 -4.37653422e-01
-3.69579792e-01 -7.24280238e-01 -4.21771765e-01 -5.51546097e-01
2.47870982e-01 5.08331239e-01 5.59370816e-02 -3.19246531e-01
8.77990901e-01 -1.24119809e-02 -1.59803104e+00 6.03500962e-01
9.15802121e-01 1.03780925e+00 4.24811035e-01 4.99717057e-01
-6.19061410e-01 1.19097817e+00 -5.37399054e-01 -2.71536648e-01
-4.69854385e-01 -4.70062941e-01 -1.20133758e+00 -3.94520760e-01
7.55243599e-02 -1.30221918e-01 3.45185071e-01 1.10796297e+00
1.25089064e-01 9.40325186e-02 8.29457998e-01 -1.21841466e+00
-7.09010065e-01 6.64107084e-01 -3.01188797e-01 4.83540177e-01
2.46064410e-01 -7.52784014e-02 1.38653541e+00 -1.14551342e+00
4.91052449e-01 1.52772593e+00 4.04563963e-01 4.41611528e-01
-9.56671536e-01 -2.18907952e-01 3.60876709e-01 6.06818557e-01
-9.94301438e-01 -1.51997492e-01 8.05027246e-01 -5.22789836e-01
1.12054157e+00 3.74331847e-02 3.71897697e-01 9.75932956e-01
6.48677766e-01 9.95590091e-01 1.10916257e+00 -6.70894086e-01
1.96829259e-01 3.07824701e-01 8.05977821e-01 5.84537148e-01
1.63841456e-01 3.19630131e-02 -7.49547899e-01 -8.13768152e-03
4.28636730e-01 4.95800674e-02 3.21133994e-02 3.98412757e-02
-1.15874839e+00 1.06539166e+00 7.33077943e-01 5.52529931e-01
-5.81399858e-01 6.42987341e-02 3.07867199e-01 1.69327214e-01
5.10616779e-01 6.11174822e-01 -5.73599160e-01 4.55506407e-02
-8.11266124e-01 1.80508271e-01 4.99714077e-01 2.11839586e-01
9.08341050e-01 -9.00845677e-02 -3.72391969e-01 5.75935304e-01
6.86522543e-01 -6.49117678e-02 6.46230221e-01 -4.68213320e-01
4.11604166e-01 9.90569532e-01 -1.17883636e-02 -1.13840485e+00
-4.56568986e-01 -5.32303572e-01 -7.41705835e-01 2.47933224e-01
1.92229047e-01 -8.68306234e-02 -8.20207715e-01 1.55600119e+00
7.03995749e-02 -1.98431060e-01 2.52006680e-01 8.89565051e-01
7.25576520e-01 9.71225262e-01 4.87877548e-01 2.33567715e-01
1.54138625e+00 -9.11900461e-01 -9.22692418e-01 -6.61224842e-01
6.51546121e-01 -6.04595482e-01 1.16905308e+00 2.77761608e-01
-6.47682309e-01 -6.92189515e-01 -9.63962376e-01 -4.76151377e-01
-6.50914788e-01 1.35998920e-01 6.88285172e-01 1.63700983e-01
-9.71760213e-01 4.33122814e-01 -2.23096564e-01 -4.39925343e-01
2.37376958e-01 2.46799514e-01 4.56803441e-02 1.11531325e-01
-1.68467259e+00 1.28384697e+00 6.51327550e-01 2.60206223e-01
-7.58516490e-01 -6.13190591e-01 -7.77116299e-01 3.52976531e-01
2.56527305e-01 -7.32001960e-01 1.15490532e+00 -1.15610290e+00
-9.48518336e-01 8.61949861e-01 -4.61138129e-01 -8.84371638e-01
3.10841292e-01 -5.53352058e-01 -7.35827312e-02 -9.73271653e-02
2.60780185e-01 8.21277380e-01 9.11838174e-01 -1.29546893e+00
-9.41093147e-01 -4.19924647e-01 2.10404955e-02 2.93520600e-01
-1.10259369e-01 2.06524476e-01 1.88987046e-01 -7.10946381e-01
1.08500034e-01 -4.38397378e-01 1.12193655e-02 -1.29466474e-01
-4.24193501e-01 -6.81066155e-01 7.53227890e-01 -8.22330534e-01
1.35319138e+00 -2.01836276e+00 8.42383524e-05 -9.71839875e-02
3.11083555e-01 3.13414216e-01 2.14497134e-01 4.40266341e-01
-8.16396624e-02 2.19156221e-01 -3.51820081e-01 -2.94462651e-01
-3.58421169e-02 6.63569942e-02 -8.09883893e-01 -9.86725166e-02
4.92536217e-01 1.03361571e+00 -9.87415135e-01 -4.13012087e-01
3.09574485e-01 4.62742805e-01 -1.76146179e-01 2.31238201e-01
-4.09752905e-01 3.19998831e-01 -3.63034636e-01 1.82692692e-01
2.89860576e-01 -2.33640537e-01 -2.10132301e-01 -9.19848159e-02
-3.40791315e-01 9.08112347e-01 -9.31427538e-01 1.28243351e+00
-4.23170745e-01 8.34011734e-01 -2.02348188e-01 -8.68147433e-01
1.05229127e+00 2.10889786e-01 -1.02616325e-01 -6.62876129e-01
1.56091675e-02 2.90884972e-01 1.29504591e-01 -3.78676325e-01
5.57354212e-01 -3.29495192e-01 8.04361179e-02 7.72332847e-01
1.70771871e-02 4.36001718e-01 -8.86941552e-02 2.55556613e-01
5.72258472e-01 -1.96708024e-01 5.87838292e-01 -4.16529685e-01
6.12895429e-01 2.40478382e-01 4.27057326e-01 8.86198640e-01
-4.21480834e-01 5.73588848e-01 6.81319594e-01 -7.70021439e-01
-7.99186051e-01 -6.21721029e-01 7.46815279e-02 1.13934624e+00
-1.02772936e-01 -4.41487432e-02 -6.10352159e-01 -5.17063737e-01
-3.43046576e-01 1.31517708e+00 -8.28680873e-01 -4.65520173e-01
-3.95826548e-01 -7.30487823e-01 2.83566535e-01 3.44329447e-01
6.61358714e-01 -1.58778250e+00 -9.95430112e-01 3.05854887e-01
-2.94286311e-01 -7.82353759e-01 4.85124364e-02 3.82554829e-01
-9.79786992e-01 -9.00444746e-01 -5.03862500e-01 -9.44744051e-01
7.52421677e-01 1.98813111e-01 1.06003261e+00 1.35257632e-01
-1.17294274e-01 -1.73183039e-01 -3.21565181e-01 -5.95153034e-01
-4.68652517e-01 -3.06549482e-02 -2.00554103e-01 1.34759232e-01
7.58036494e-01 -3.00166346e-02 -5.52062213e-01 -1.65102094e-01
-7.50052631e-01 5.22920907e-01 6.91440642e-01 7.26700127e-01
3.25009972e-01 2.98932970e-01 6.60586238e-01 -9.20644224e-01
9.67091262e-01 -4.69051003e-01 -1.21824689e-01 3.86048377e-01
-8.75968814e-01 3.98239195e-01 5.81735611e-01 2.02476904e-01
-1.41020930e+00 -4.98111993e-01 -2.59502977e-01 1.32896021e-01
-4.14553344e-01 8.00253332e-01 1.44484907e-01 4.29200917e-01
4.81904447e-01 5.06306410e-01 -2.56647468e-01 -4.26165998e-01
1.76733986e-01 4.71307158e-01 1.32084951e-01 -1.01532355e-01
2.85245597e-01 4.15915221e-01 -2.80422091e-01 -7.10263491e-01
-1.30315936e+00 -2.97433048e-01 -6.06819749e-01 -1.52968928e-01
8.51257741e-01 -5.79998255e-01 -4.38224047e-01 3.95821691e-01
-1.62189770e+00 -8.60536173e-02 -1.53866425e-01 2.44028673e-01
-2.50510186e-01 2.93810163e-02 -4.83907580e-01 -1.01090753e+00
-6.39867425e-01 -8.73746991e-01 9.22367156e-01 3.20317566e-01
-5.41188955e-01 -1.25336039e+00 -8.53653327e-02 5.15757620e-01
4.34576929e-01 -1.97292238e-01 1.58221877e+00 -7.81213701e-01
-3.20280194e-01 3.94179160e-03 -6.12976491e-01 3.01633984e-01
-9.20411572e-02 -1.90528080e-01 -1.19892955e+00 2.49654576e-01
4.98123407e-01 -1.86074644e-01 1.12144101e+00 6.13665938e-01
9.28385973e-01 -4.20990467e-01 -2.16243982e-01 -2.27270722e-02
1.43052804e+00 1.41493261e-01 6.41008973e-01 4.36313331e-01
6.05771303e-01 1.11397767e+00 8.41835260e-01 1.73375160e-01
5.72525322e-01 3.78966331e-01 4.94228899e-01 -2.82059699e-01
-1.55482441e-01 -4.88638312e-01 5.00021219e-01 4.25170958e-01
2.96877921e-01 -5.16635962e-02 -1.14660001e+00 6.52904272e-01
-2.04591870e+00 -7.69596457e-01 -2.23456204e-01 1.98687184e+00
6.38731718e-01 3.92593622e-01 -4.39737976e-01 3.22953850e-01
8.11931908e-01 1.93567991e-01 -3.62920642e-01 -8.35452139e-01
-4.96208631e-02 -2.07256436e-01 -9.38151777e-02 7.37266183e-01
-9.63299930e-01 1.09307909e+00 5.91349268e+00 6.89651310e-01
-1.14933729e+00 5.68333268e-02 9.25593019e-01 5.24303675e-01
-3.58744740e-01 1.36592016e-01 -9.74498391e-01 2.26566985e-01
7.95380056e-01 -1.85163110e-01 4.83406708e-02 6.96923137e-01
5.93400657e-01 -5.45423388e-01 -1.02054572e+00 6.42905056e-01
1.16429508e-01 -1.24544835e+00 5.74615061e-01 -2.74799943e-01
4.64593738e-01 -7.50109032e-02 1.74317896e-01 2.67668903e-01
-8.94087330e-02 -9.65629399e-01 8.30603898e-01 4.99692023e-01
1.81639940e-01 -6.96835339e-01 8.86742175e-01 6.79987133e-01
-7.98009813e-01 -2.02072039e-01 -5.34541965e-01 -5.05317032e-01
4.06560302e-02 9.68544245e-01 -1.12610364e+00 2.02351227e-01
4.47787046e-01 6.35510206e-01 -4.97711122e-01 8.52926552e-01
-7.92747796e-01 7.38587558e-01 1.39997751e-01 -2.20401436e-01
5.05319059e-01 1.13996737e-01 6.70633852e-01 1.14895964e+00
5.59820198e-02 -1.98578015e-01 -3.16836029e-01 1.32649326e+00
7.84773082e-02 3.16978954e-02 -5.94219387e-01 -6.19163625e-02
3.87826450e-02 1.01011384e+00 -8.49237502e-01 -3.43925327e-01
-5.60847409e-02 9.69852388e-01 3.59867245e-01 2.58247554e-01
-5.84081709e-01 -2.93752819e-01 3.99242699e-01 -7.66494572e-02
7.70859197e-02 -1.42266288e-01 -8.35026205e-01 -9.45419252e-01
-1.02012806e-01 -5.47377408e-01 4.85569805e-01 -9.53928530e-01
-1.01749730e+00 5.08874893e-01 -2.68994808e-01 -6.97948635e-01
-2.50302762e-01 -5.36519408e-01 -9.68034863e-01 1.12860727e+00
-2.04863930e+00 -1.02129865e+00 6.70197681e-02 1.91739842e-01
1.04290879e+00 1.52482465e-01 9.48942721e-01 -6.45438060e-02
-4.99239415e-01 -3.92813347e-02 -1.35539681e-01 3.94955371e-03
4.11388725e-01 -1.18519199e+00 4.92956936e-01 9.91241395e-01
2.79569268e-01 8.04433942e-01 9.62616980e-01 -7.00312972e-01
-9.30038929e-01 -9.80902612e-01 1.89961314e+00 -3.62009585e-01
5.94956219e-01 -3.17752600e-01 -9.72166300e-01 4.49395150e-01
2.21678749e-01 -5.84670484e-01 6.51040733e-01 1.23203270e-01
-3.41931023e-02 3.94442528e-02 -9.87087548e-01 5.60984552e-01
3.54381621e-01 -4.92148429e-01 -1.18260968e+00 1.05380014e-01
7.14053094e-01 3.55457783e-01 -5.29214107e-02 2.70509958e-01
3.20779860e-01 -1.07890630e+00 7.05825746e-01 -6.83841765e-01
7.36108422e-01 -3.52749854e-01 2.71053165e-01 -1.23769855e+00
-4.31002170e-01 -2.35263392e-01 -1.48580492e-01 1.23926461e+00
6.94507003e-01 -6.04592621e-01 3.84461671e-01 6.64969444e-01
2.81176735e-02 -7.96344340e-01 -7.33996987e-01 -6.39836341e-02
-2.19815716e-01 -5.72917938e-01 1.31402791e-01 7.05234528e-01
-1.19508997e-01 7.61072934e-01 -3.24507415e-01 -1.44897569e-02
5.32762110e-01 3.26350361e-01 1.01894170e-01 -1.39872253e+00
1.63751528e-01 -4.46714789e-01 -8.06742162e-02 -8.89868617e-01
6.98407739e-02 -7.08755851e-01 1.20576680e-01 -2.08900213e+00
2.81841636e-01 -2.97233075e-01 -6.15066171e-01 4.61189866e-01
-5.26023626e-01 -2.65468776e-01 1.51877806e-01 3.76068950e-01
-5.51321268e-01 5.20900726e-01 8.73290956e-01 -1.58287153e-01
-1.10207625e-01 5.87272979e-02 -9.45306718e-01 9.12621379e-01
9.89648163e-01 -6.40212417e-01 -3.50331366e-01 -7.35299349e-01
6.36436641e-01 -3.52604151e-01 5.46621859e-01 -8.46422315e-01
2.81392366e-01 -4.96266186e-02 2.51741946e-01 -6.06283069e-01
4.79924791e-02 -6.13510132e-01 -4.74873424e-01 9.03511703e-01
-8.14018786e-01 -8.06466267e-02 6.69174343e-02 4.14656430e-01
-4.54301089e-01 -4.06746745e-01 6.15779519e-01 -2.94899851e-01
-1.22534120e+00 -1.45697027e-01 -6.00226104e-01 -9.86893624e-02
7.11748064e-01 -1.15145639e-01 -3.43417972e-01 -3.24814439e-01
-5.48809111e-01 1.98755994e-01 -9.09531713e-02 6.79192960e-01
8.24758530e-01 -8.76675844e-01 -7.43393779e-01 1.35177985e-01
7.46609420e-02 -1.04002886e-01 2.23895356e-01 8.17556500e-01
-3.43680084e-01 1.06340742e+00 -1.98188543e-01 -2.42942229e-01
-1.02411830e+00 2.78163016e-01 3.16155493e-01 -5.74450970e-01
-5.24017513e-01 1.02597082e+00 5.19724488e-01 -9.17011052e-02
2.10712180e-01 -2.79664487e-01 -8.66207838e-01 3.16482544e-01
7.83014417e-01 3.61261040e-01 1.19603939e-01 -7.48519540e-01
-5.16573727e-01 2.28434533e-01 -1.45446777e-01 -1.26056850e-01
1.28738022e+00 -3.81841183e-01 -3.59810561e-01 8.68035078e-01
9.12046552e-01 -6.66298866e-01 -9.87571597e-01 -1.52204230e-01
3.93030077e-01 -8.65013525e-02 2.59326041e-01 -1.03773105e+00
-5.84290147e-01 1.48136914e+00 5.29770434e-01 3.12839895e-01
7.44084120e-01 -4.59503531e-02 7.28717506e-01 3.06319505e-01
1.26651734e-01 -1.20461571e+00 -1.01571098e-01 8.17972600e-01
9.39888775e-01 -1.45573521e+00 -3.32829326e-01 -1.47409022e-01
-7.32664227e-01 1.33859539e+00 5.00779271e-01 -6.45578876e-02
4.22815114e-01 -2.63230056e-01 2.76707023e-01 -4.80265290e-01
-9.42232370e-01 -1.89009115e-01 2.92789370e-01 3.52820128e-01
6.86896265e-01 3.99026759e-02 -5.57864726e-01 5.51566482e-01
-7.03198016e-02 -1.15523309e-01 2.74906456e-01 7.73598254e-01
-8.70926857e-01 -7.41870582e-01 -4.62733626e-01 4.99087691e-01
-4.41064954e-01 -3.28669220e-01 -5.99735022e-01 1.67557508e-01
1.63747624e-01 1.10266721e+00 -9.04040784e-02 -8.51975679e-02
-1.82919085e-01 4.14968073e-01 -1.62501186e-01 -9.05450881e-01
-7.44419456e-01 -4.07924175e-01 -1.57949194e-01 -3.08215261e-01
-2.80894071e-01 -3.64652634e-01 -1.55116546e+00 1.45855755e-01
-2.18572453e-01 1.44152910e-01 7.69006431e-01 1.41567981e+00
4.02305573e-01 6.81608677e-01 2.65962481e-01 -4.11459535e-01
-3.73310357e-01 -9.29222822e-01 -1.23391017e-01 3.00928384e-01
4.18770581e-01 -5.34870327e-01 -4.41215813e-01 -7.73744285e-02] | [9.428832054138184, 6.496097087860107] |
301443a3-d866-4c4e-a6df-19c72b0bad79 | compressing-features-for-learning-with-noisy | 2206.13140 | null | https://arxiv.org/abs/2206.13140v1 | https://arxiv.org/pdf/2206.13140v1.pdf | Compressing Features for Learning with Noisy Labels | Supervised learning can be viewed as distilling relevant information from input data into feature representations. This process becomes difficult when supervision is noisy as the distilled information might not be relevant. In fact, recent research shows that networks can easily overfit all labels including those that are corrupted, and hence can hardly generalize to clean datasets. In this paper, we focus on the problem of learning with noisy labels and introduce compression inductive bias to network architectures to alleviate this over-fitting problem. More precisely, we revisit one classical regularization named Dropout and its variant Nested Dropout. Dropout can serve as a compression constraint for its feature dropping mechanism, while Nested Dropout further learns ordered feature representations w.r.t. feature importance. Moreover, the trained models with compression regularization are further combined with Co-teaching for performance boost. Theoretically, we conduct bias-variance decomposition of the objective function under compression regularization. We analyze it for both single model and Co-teaching. This decomposition provides three insights: (i) it shows that over-fitting is indeed an issue for learning with noisy labels; (ii) through an information bottleneck formulation, it explains why the proposed feature compression helps in combating label noise; (iii) it gives explanations on the performance boost brought by incorporating compression regularization into Co-teaching. Experiments show that our simple approach can have comparable or even better performance than the state-of-the-art methods on benchmarks with real-world label noise including Clothing1M and ANIMAL-10N. Our implementation is available at https://yingyichen-cyy.github.io/CompressFeatNoisyLabels/. | ['Johan A. K. Suykens', 'Chunrong Ai', 'Xi Shen', 'Shell Xu Hu', 'Yingyi Chen'] | 2022-06-27 | null | null | null | null | ['feature-compression', 'learning-with-noisy-labels', 'learning-with-noisy-labels'] | ['computer-vision', 'computer-vision', 'natural-language-processing'] | [ 4.66914296e-01 2.29597673e-01 -3.42377484e-01 -6.94023252e-01
-9.59398746e-01 -3.50489289e-01 1.39644101e-01 1.25817746e-01
-6.19011402e-01 7.63386071e-01 1.69203594e-01 -8.38467926e-02
-2.60049760e-01 -5.80466747e-01 -1.15231705e+00 -8.62951696e-01
2.20369935e-01 3.37335855e-01 -1.94101602e-01 1.85076967e-01
-1.39627457e-01 3.07340860e-01 -1.65152371e+00 2.32847363e-01
9.51452494e-01 9.74606156e-01 2.04155266e-01 2.50317872e-01
2.33832374e-02 9.77095842e-01 -4.44819123e-01 -4.75447357e-01
2.26379484e-01 -4.16595638e-01 -1.04642344e+00 2.04380631e-01
6.11459553e-01 -3.90879244e-01 -3.18635911e-01 1.13440645e+00
2.82728106e-01 2.07277685e-01 6.31533444e-01 -1.12014043e+00
-8.72160852e-01 9.72506702e-01 -3.67596716e-01 -1.81207165e-01
-2.28174835e-01 2.78471142e-01 1.13525152e+00 -8.28418493e-01
5.76669514e-01 1.23969698e+00 5.77374578e-01 8.79770935e-01
-1.59109855e+00 -5.92905760e-01 3.49367112e-01 7.75033161e-02
-1.08792269e+00 -2.42800519e-01 6.55574262e-01 -2.42049620e-01
6.03134394e-01 2.72533566e-01 3.48424762e-01 1.36015713e+00
-2.97620773e-01 1.20428932e+00 1.11735201e+00 -3.64355773e-01
4.11990248e-02 3.48620296e-01 6.67751074e-01 6.67215466e-01
3.33956361e-01 1.03920564e-01 -5.17185271e-01 1.66244686e-01
3.57493222e-01 4.04340893e-01 -4.38107163e-01 -1.88021764e-01
-8.56391132e-01 1.05664468e+00 6.99001133e-01 1.53700769e-01
-1.50478519e-02 4.29948807e-01 5.03374398e-01 7.14277923e-01
6.65403843e-01 5.66884696e-01 -6.38584971e-01 -6.74850345e-02
-8.64181936e-01 1.00228213e-01 6.49233580e-01 1.09512246e+00
9.53106463e-01 1.37169167e-01 -1.91689834e-01 9.76542294e-01
1.37396649e-01 2.08741620e-01 4.28929120e-01 -1.15471315e+00
4.26254749e-01 4.80805099e-01 -2.74326235e-01 -5.02324641e-01
-4.81870979e-01 -7.60561585e-01 -1.03977513e+00 8.31355676e-02
5.92239738e-01 -2.50305593e-01 -1.04577041e+00 2.09919381e+00
3.59311737e-02 6.96960986e-02 -3.92640717e-02 9.17988598e-01
9.72867548e-01 5.03836393e-01 6.10228889e-02 -2.38840774e-01
1.14296222e+00 -1.24003220e+00 -8.37987840e-01 -1.39981180e-01
9.10953343e-01 -4.60610896e-01 1.42489016e+00 5.12778282e-01
-9.81220305e-01 -5.96821249e-01 -9.44691777e-01 -3.69137228e-01
-2.34103933e-01 1.30501464e-01 7.29544342e-01 4.10726428e-01
-9.08842444e-01 1.13545036e+00 -8.36494029e-01 -1.74294144e-01
7.81735778e-01 5.21996677e-01 -2.61270493e-01 -5.08556545e-01
-1.06351459e+00 6.43531919e-01 1.76860109e-01 6.63157627e-02
-1.06900644e+00 -8.78642380e-01 -8.05897057e-01 3.28640819e-01
6.63428128e-01 -6.38420343e-01 1.31970417e+00 -1.08430123e+00
-1.35408902e+00 7.00866401e-01 -9.49230343e-02 -3.68860453e-01
5.20097554e-01 -5.18939495e-01 1.67111769e-01 5.00562489e-02
-1.64965183e-01 8.48719478e-01 7.82058120e-01 -1.22283173e+00
-3.53882581e-01 -3.09231818e-01 3.81909236e-02 1.43028423e-01
-6.00855470e-01 -1.84828386e-01 -2.68609285e-01 -5.96103311e-01
1.64699823e-01 -8.21279764e-01 -1.26675665e-01 1.85205117e-02
-6.64109647e-01 -4.43562478e-01 5.37340701e-01 -2.98235953e-01
1.02347362e+00 -2.21836162e+00 1.32334560e-01 4.22966257e-02
3.68609846e-01 2.41201609e-01 -2.75213003e-01 2.19334170e-01
-8.15394595e-02 4.52673823e-01 -3.52450699e-01 -7.67498732e-01
4.91363555e-02 4.11412060e-01 -2.62868255e-01 4.21283990e-01
5.04131377e-01 8.55165362e-01 -8.13860953e-01 -1.29589736e-01
-1.65816292e-01 7.34847903e-01 -6.61297500e-01 2.31360629e-01
-2.25256786e-01 4.55049932e-01 -3.57835770e-01 6.12300634e-01
8.21494818e-01 -4.50134099e-01 -2.21875380e-03 -1.49398729e-01
2.14124203e-01 5.28718829e-01 -1.01205862e+00 1.65788186e+00
-2.96838492e-01 4.38918650e-01 1.63509160e-01 -1.18797088e+00
7.96402514e-01 2.81855822e-01 1.71166450e-01 -3.74444962e-01
2.37141341e-01 2.80792981e-01 -1.54087514e-01 -5.57340384e-01
1.92859545e-01 -1.11105919e-01 2.17924222e-01 4.94259208e-01
4.76695985e-01 2.13777497e-02 2.97293007e-01 1.75328612e-01
1.12151206e+00 2.42547259e-01 -2.77919233e-01 -1.21603131e-01
1.28459364e-01 -4.29508895e-01 7.41201580e-01 8.66866171e-01
-8.08210969e-02 6.95089579e-01 7.30330646e-01 -3.26304287e-01
-8.51070046e-01 -6.93737626e-01 -2.56444544e-01 1.34479785e+00
-2.12028712e-01 -4.04437214e-01 -8.72657478e-01 -9.28226471e-01
8.87049269e-03 6.59677505e-01 -7.58386374e-01 -4.88480300e-01
-4.23319995e-01 -8.68839920e-01 3.51284504e-01 5.76648295e-01
3.13145041e-01 -1.12106717e+00 -3.32335651e-01 1.88967232e-02
-1.25877008e-01 -7.37153590e-01 -3.60087961e-01 9.78495061e-01
-1.17787445e+00 -9.46512222e-01 -5.18234968e-01 -7.92840004e-01
8.89676094e-01 2.35879615e-01 1.25203645e+00 5.46532810e-01
-9.90658775e-02 1.16915300e-01 -4.44589078e-01 -1.84847608e-01
-3.57212007e-01 4.51230139e-01 -3.29304338e-02 -1.58573702e-01
5.43374956e-01 -7.85896480e-01 -4.31530237e-01 1.53549537e-01
-1.15113890e+00 4.08074930e-02 5.97861588e-01 1.13082707e+00
6.71057165e-01 -6.39341176e-02 5.62167943e-01 -1.45819151e+00
3.81057739e-01 -5.08145809e-01 -4.34548736e-01 1.22769713e-01
-7.89462090e-01 3.76735330e-01 8.72296631e-01 -5.51575065e-01
-9.05077338e-01 2.29643911e-01 -2.27310732e-01 -4.93357003e-01
-2.98111826e-01 4.65416342e-01 -2.64194220e-01 1.35249540e-01
6.32484615e-01 2.82647684e-02 -3.57283875e-02 -9.58944738e-01
4.05609071e-01 5.15649319e-01 2.09974989e-01 -7.76188314e-01
7.36522019e-01 2.78172642e-01 -6.06566705e-02 -5.27055800e-01
-1.40954638e+00 -3.08459342e-01 -6.17870152e-01 1.46253452e-01
4.69240785e-01 -8.15538704e-01 -8.08328927e-01 3.42873603e-01
-9.50041890e-01 -5.13079047e-01 -7.05380678e-01 4.77538884e-01
-5.52396894e-01 8.88507590e-02 -1.09895909e+00 -5.82379460e-01
-2.18802959e-01 -1.33262885e+00 7.00844467e-01 3.58614504e-01
5.96385263e-02 -8.31059873e-01 -2.18706965e-01 3.86291057e-01
3.90292734e-01 4.38817171e-03 9.32855964e-01 -8.75437260e-01
-6.38982594e-01 -1.09499190e-02 -2.55327493e-01 7.86352634e-01
-1.45304382e-01 -1.28513724e-01 -1.37902021e+00 -4.64555413e-01
2.54601657e-01 -1.02185357e+00 1.42404783e+00 4.20933783e-01
1.43333292e+00 -3.75522822e-01 7.15970993e-02 9.74343240e-01
1.39197719e+00 -3.97021174e-01 3.67828786e-01 1.65637836e-01
9.34167862e-01 7.35092700e-01 4.18541223e-01 1.74541652e-01
1.32359922e-01 4.87834245e-01 5.83339512e-01 -2.21594125e-01
-2.97631919e-01 -3.25343966e-01 3.80802363e-01 1.07787812e+00
1.11369587e-01 -1.23316944e-01 -4.23106372e-01 2.72325933e-01
-1.78514993e+00 -5.52873135e-01 -2.46573120e-01 2.12528825e+00
1.26147652e+00 1.30962402e-01 -7.08791018e-02 1.51655912e-01
4.40208286e-01 -6.59238920e-02 -7.52707183e-01 -3.43121558e-01
-1.71212614e-01 1.10847980e-01 4.51457143e-01 4.69518930e-01
-1.20194340e+00 7.74405003e-01 5.08538628e+00 1.02051246e+00
-9.61729467e-01 3.88809949e-01 1.01339006e+00 -3.28781635e-01
-3.41677815e-01 -8.00834820e-02 -1.00274765e+00 3.96008849e-01
1.07438791e+00 2.76581198e-01 4.91606265e-01 8.57496083e-01
4.16354872e-02 1.10713139e-01 -1.44555092e+00 7.28898227e-01
-1.02258166e-02 -9.85608876e-01 -5.78320175e-02 -7.76428729e-02
7.54432559e-01 1.38373777e-01 2.48870134e-01 5.58138371e-01
3.96056771e-01 -1.13050187e+00 5.93051732e-01 4.43594605e-01
7.95884609e-01 -8.82436991e-01 8.83632660e-01 4.49838191e-01
-6.71254098e-01 -9.98053625e-02 -7.77596056e-01 -7.02435896e-02
-3.96322981e-02 8.76057804e-01 -4.63126153e-01 3.47141266e-01
6.18024766e-01 9.02007401e-01 -8.27950060e-01 1.00855029e+00
-6.47840202e-01 1.08022749e+00 -4.38389510e-01 1.43494904e-01
2.12055236e-01 -2.28393972e-01 2.00763777e-01 1.17488205e+00
1.20432623e-01 -2.19927892e-01 1.16793238e-01 1.14448678e+00
-6.30471766e-01 -4.36483808e-02 -5.04215240e-01 1.24954082e-01
3.44051749e-01 1.21789908e+00 -6.38955832e-01 -2.59464473e-01
-4.49954510e-01 7.91225255e-01 7.46533334e-01 4.77619618e-01
-6.35918617e-01 -2.78278023e-01 3.78927439e-01 -2.44939663e-02
3.62993896e-01 3.35631609e-01 -3.79522711e-01 -1.23160064e+00
2.98509113e-02 -8.76572251e-01 2.63933867e-01 -4.40418929e-01
-1.44294512e+00 5.61605036e-01 -1.76617503e-01 -1.03446949e+00
9.84537527e-02 -4.27628905e-01 -2.43853286e-01 6.21307492e-01
-1.85338581e+00 -9.39579070e-01 -1.24716774e-01 3.40723544e-01
6.58204675e-01 2.11370051e-01 8.75597477e-01 5.16492188e-01
-8.17409873e-01 9.91394579e-01 4.64409739e-01 1.24784540e-02
8.26554954e-01 -1.39470589e+00 1.91713963e-02 6.72259450e-01
2.88160086e-01 6.18579924e-01 5.10194719e-01 -3.97827834e-01
-1.23698676e+00 -1.35165679e+00 8.97202134e-01 -3.43236536e-01
3.62449259e-01 -5.88489711e-01 -1.34800768e+00 7.20090747e-01
9.40115303e-02 1.24976881e-01 7.18842745e-01 2.05529377e-01
-6.13081157e-01 -1.59619868e-01 -1.08149552e+00 3.33981931e-01
1.01948750e+00 -3.55008632e-01 -2.40649566e-01 6.92867279e-01
1.17873621e+00 -1.85471207e-01 -7.67915785e-01 2.49416322e-01
1.17422305e-01 -8.83555293e-01 6.58591509e-01 -7.12980628e-01
6.57302976e-01 1.53228670e-01 -1.04212172e-01 -1.29865074e+00
-2.26776704e-01 -6.50099099e-01 -1.05168492e-01 1.53010535e+00
6.80549979e-01 -3.01658988e-01 8.63991499e-01 7.09529996e-01
-2.15133414e-01 -1.06543112e+00 -7.03072369e-01 -8.59151006e-01
3.78667176e-01 -4.05533016e-01 3.37212026e-01 1.02761900e+00
-1.59397691e-01 4.75316554e-01 -5.97417355e-01 -2.85758853e-01
7.43149936e-01 -1.99179798e-01 4.43261057e-01 -1.30501115e+00
-3.72640997e-01 -2.85504788e-01 8.23175684e-02 -1.16685939e+00
2.53721833e-01 -1.25959623e+00 9.11459699e-02 -1.11090779e+00
4.89086390e-01 -5.08749545e-01 -4.43685710e-01 8.54338765e-01
-3.56335640e-01 2.59939641e-01 3.74231249e-01 3.00886333e-01
-6.88812912e-01 7.27596581e-01 1.31259322e+00 -8.57227966e-02
-3.81529666e-02 1.76384851e-01 -8.71818483e-01 7.41091549e-01
9.61690307e-01 -1.06750560e+00 -3.83242607e-01 -5.53253889e-01
2.50804573e-01 -3.81757617e-01 2.26280212e-01 -7.66149819e-01
1.02724299e-01 1.24750271e-01 1.47824779e-01 -2.58668005e-01
2.67997384e-01 -9.46306229e-01 -2.58873999e-01 4.84969497e-01
-9.21740592e-01 -2.80948728e-01 -7.36533776e-02 5.17587841e-01
-1.62216842e-01 -8.54645789e-01 8.51058066e-01 -1.69594198e-01
-6.77334294e-02 2.43163347e-01 -1.63341746e-01 3.41110289e-01
3.98532659e-01 1.42919406e-01 -5.07706463e-01 -3.13335478e-01
-8.85402322e-01 3.80765498e-01 4.22473252e-01 1.47195429e-01
5.81719935e-01 -1.27105319e+00 -7.69270718e-01 3.39435458e-01
-1.70454368e-01 3.20580840e-01 2.01960325e-01 1.02093649e+00
-1.87651619e-01 1.63157508e-01 2.56050408e-01 -6.11275256e-01
-1.13592148e+00 5.34820557e-01 1.42539516e-01 -4.60573882e-01
-6.89672232e-01 1.13263047e+00 1.35224387e-01 -6.71625137e-01
8.96401286e-01 -5.24255931e-01 -1.41187936e-01 1.10515527e-01
4.95031953e-01 2.60046035e-01 5.37829064e-02 -1.94732547e-01
3.45291086e-02 5.04054546e-01 -4.40213919e-01 3.91163290e-01
1.71897483e+00 -8.29949528e-02 -3.27407643e-02 6.73581421e-01
1.58762825e+00 -4.65323687e-01 -1.67947280e+00 -5.57477891e-01
1.15395911e-01 -2.25575343e-01 -5.71732037e-02 -8.26178432e-01
-1.46687317e+00 1.24768674e+00 5.34441710e-01 1.51649758e-01
1.13710821e+00 2.07797606e-02 6.29051626e-01 6.78070545e-01
8.31872597e-03 -1.06757486e+00 5.83348274e-02 6.04592621e-01
6.62327290e-01 -1.41981375e+00 2.02634688e-02 -4.28917289e-01
-6.09496236e-01 1.03818810e+00 5.58351398e-01 -1.43196642e-01
6.04305446e-01 1.45866692e-01 1.77313946e-02 -1.90935299e-01
-1.01692665e+00 -1.60967752e-01 8.88488367e-02 4.54236329e-01
5.79722583e-01 -7.50302225e-02 -1.39332578e-01 8.52059126e-01
-7.05639645e-02 2.22571008e-02 5.98070621e-01 7.13588893e-01
-4.74042058e-01 -1.12343061e+00 -1.61467686e-01 6.58374965e-01
-6.57090783e-01 -2.30327740e-01 -2.01093882e-01 6.92631125e-01
1.90840244e-01 8.83934677e-01 -2.38484323e-01 -3.04799557e-01
3.83065283e-01 3.57191384e-01 2.69327790e-01 -8.07279646e-01
-9.04488862e-01 2.66936690e-01 -8.69738236e-02 -6.48862362e-01
-4.33441997e-01 -4.85322773e-01 -1.13407898e+00 -2.58894831e-01
-5.28653383e-01 3.04323435e-01 5.14863133e-01 7.21938610e-01
2.22644195e-01 8.32358420e-01 5.71412981e-01 -6.98093593e-01
-1.09514332e+00 -1.17433143e+00 -6.70663416e-01 6.95892990e-01
5.10531187e-01 -6.19902849e-01 -9.73214269e-01 2.87396796e-02] | [9.273459434509277, 3.794708251953125] |
338e2e9e-4fa6-46b6-b498-d01b57a620a5 | shifting-more-attention-to-video-salient | null | null | http://openaccess.thecvf.com/content_CVPR_2019/html/Fan_Shifting_More_Attention_to_Video_Salient_Object_Detection_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Fan_Shifting_More_Attention_to_Video_Salient_Object_Detection_CVPR_2019_paper.pdf | Shifting More Attention to Video Salient Object Detection | The last decade has witnessed a growing interest in video salient object detection (VSOD). However, the research community long-term lacked a well-established VSOD dataset representative of real dynamic scenes with high-quality annotations. To address this issue, we elaborately collected a visual-attention-consistent Densely Annotated VSOD (DAVSOD) dataset, which contains 226 videos with 23,938 frames that cover diverse realistic-scenes, objects, instances and motions. With corresponding real human eye-fixation data, we obtain precise ground-truths. This is the first work that explicitly emphasizes the challenge of saliency shift, i.e., the video salient object(s) may dynamically change. To further contribute the community a complete benchmark, we systematically assess 17 representative VSOD algorithms over seven existing VSOD datasets and our DAVSOD with totally 84K frames (largest-scale). Utilizing three famous metrics, we then present a comprehensive and insightful performance analysis. Furthermore, we propose a baseline model. It is equipped with a saliency shift- aware convLSTM, which can efficiently capture video saliency dynamics through learning human attention-shift behavior. Extensive experiments open up promising future directions for model development and comparison.
| [' Jianbing Shen', ' Ming-Ming Cheng', ' Wenguan Wang', 'Deng-Ping Fan'] | 2019-06-01 | null | null | null | cvpr-2019-6 | ['video-salient-object-detection'] | ['computer-vision'] | [ 3.51323336e-01 -2.95059472e-01 -3.95316750e-01 -7.23529607e-02
-5.29356062e-01 -1.48039117e-01 3.48248303e-01 -1.88022122e-01
-2.70392865e-01 6.76607370e-01 4.49363500e-01 2.33897731e-01
3.14794749e-01 -6.52521104e-02 -8.74363363e-01 -6.34552598e-01
-4.12621386e-02 -3.01442266e-01 1.05538154e+00 -3.58195096e-01
3.96194518e-01 1.56261861e-01 -1.92674160e+00 3.01621974e-01
7.24098384e-01 1.08640456e+00 7.02340066e-01 4.04287130e-01
3.50477010e-01 9.70487118e-01 -3.65218818e-01 -1.63132459e-01
1.98453128e-01 -3.06264758e-01 -8.42824996e-01 1.46846399e-01
1.01712239e+00 -4.56100881e-01 -6.94139898e-01 1.11687171e+00
5.06697059e-01 2.78888971e-01 3.23813409e-02 -1.59785354e+00
-1.03578985e+00 6.76482975e-01 -7.64074981e-01 1.16697621e+00
2.43920907e-01 5.19576013e-01 1.15735483e+00 -1.26937425e+00
8.22959423e-01 1.27483928e+00 4.74468172e-01 6.51651323e-01
-9.01816964e-01 -4.52122718e-01 6.06171012e-01 8.16500306e-01
-1.12352741e+00 -4.73195732e-01 9.77578521e-01 -3.35847259e-01
7.99886703e-01 3.12587500e-01 9.46647048e-01 1.37688160e+00
1.40884340e-01 1.45929420e+00 7.07182229e-01 -6.38367236e-02
6.38301522e-02 -1.99312568e-01 2.09426790e-01 4.58719939e-01
4.41462427e-01 1.71262268e-02 -9.73043919e-01 3.00300568e-01
7.37027884e-01 2.82891005e-01 -5.38839638e-01 -9.02554512e-01
-1.71828592e+00 4.52470064e-01 6.55005991e-01 2.51747847e-01
-3.77278507e-01 1.83697164e-01 6.22332275e-01 -1.03027597e-02
2.70173311e-01 3.13817501e-01 -4.71757293e-01 -8.13361630e-02
-1.06490624e+00 4.28610593e-01 9.22967345e-02 1.29388249e+00
5.60187817e-01 4.70382154e-01 -6.42784476e-01 4.60331798e-01
7.75262490e-02 6.97260261e-01 5.11852741e-01 -1.03145003e+00
2.99566090e-01 5.54516613e-01 3.13914150e-01 -1.36021864e+00
-3.38804662e-01 -3.84893239e-01 -6.17169559e-01 -1.84877887e-01
1.81254491e-01 1.65801495e-01 -8.04478467e-01 1.76201916e+00
2.59437233e-01 6.18988276e-01 -2.04827532e-01 1.51419067e+00
1.25475371e+00 4.69429046e-01 1.42443851e-01 -2.64374077e-01
1.19762981e+00 -1.54338312e+00 -7.13552713e-01 -3.79775733e-01
5.02745844e-02 -5.25105715e-01 1.29371023e+00 1.11509077e-01
-1.29974866e+00 -7.45662630e-01 -8.92600894e-01 -3.62881333e-01
-8.65426362e-02 -3.14365402e-02 5.87515295e-01 -1.26439184e-01
-1.29957187e+00 1.46658897e-01 -7.51588762e-01 -5.48936725e-01
7.61745512e-01 9.40187126e-02 -1.21647734e-02 9.38005373e-02
-1.34001195e+00 7.88483679e-01 3.30903709e-01 2.70988382e-02
-1.47705901e+00 -8.61065626e-01 -9.93787706e-01 7.78394043e-02
7.10974157e-01 -7.24117577e-01 1.37763405e+00 -1.47963738e+00
-8.68541956e-01 8.53616774e-01 -3.56265515e-01 -6.10459328e-01
3.44405890e-01 -4.10436630e-01 -5.52814245e-01 3.81190777e-01
3.52134168e-01 9.55792010e-01 1.11273265e+00 -1.22330344e+00
-8.23890030e-01 -1.95409600e-02 1.53982326e-01 1.89599603e-01
-3.05938721e-01 3.82819027e-01 -4.75912601e-01 -9.76768374e-01
-2.18422547e-01 -8.41688216e-01 -1.89468451e-02 1.89448953e-01
-4.73519474e-01 -2.09674820e-01 1.06139481e+00 -5.04773200e-01
1.46047020e+00 -2.23524046e+00 3.83753240e-01 -5.87016165e-01
5.27345598e-01 5.77145517e-01 -2.50224948e-01 6.78294897e-02
-8.63423422e-02 -3.30902129e-01 -1.62099153e-01 -3.66142362e-01
-1.41086712e-01 -1.22964561e-01 -6.97967470e-01 4.17231351e-01
3.89514148e-01 1.28230381e+00 -1.31051958e+00 -6.26006484e-01
2.56079942e-01 3.47282499e-01 -4.89517778e-01 -1.49846764e-03
-2.93552250e-01 5.76016977e-02 -4.74363923e-01 9.51065063e-01
5.33830523e-01 -6.52415097e-01 -3.65306050e-01 -2.18467459e-01
-3.23008031e-01 -5.81422411e-02 -9.00781393e-01 1.92639637e+00
4.11374092e-01 1.14834929e+00 -2.50279009e-01 -7.48442471e-01
5.92802584e-01 -1.03403263e-01 5.22128105e-01 -6.53592348e-01
1.32774040e-01 1.96265608e-01 -1.93440229e-01 -5.41379690e-01
9.89292085e-01 3.91315103e-01 2.00359553e-01 5.93335256e-02
1.52040198e-01 3.85404259e-01 2.15803504e-01 4.27467555e-01
9.73897874e-01 6.20835051e-02 4.21882123e-01 -6.14740968e-01
5.76273263e-01 2.01251596e-01 6.67300224e-01 7.55774498e-01
-1.01277578e+00 8.32557321e-01 3.25425893e-01 -7.10016131e-01
-9.10114527e-01 -1.11682880e+00 3.21433358e-02 1.53849077e+00
9.77876365e-01 -1.48929924e-01 -6.99563563e-01 -5.98852456e-01
1.66725852e-02 3.88268858e-01 -9.86746013e-01 -2.45350391e-01
-5.88555932e-01 -2.95231611e-01 1.11104570e-01 5.74375451e-01
6.33417964e-01 -1.66857862e+00 -9.23005581e-01 2.17740741e-02
-3.83623481e-01 -1.07601607e+00 -9.87705052e-01 -2.88019508e-01
-5.95234871e-01 -1.16662192e+00 -1.01043475e+00 -1.20731246e+00
2.93265879e-01 1.38911128e+00 1.33143532e+00 1.49403721e-01
-2.37400815e-01 3.84266257e-01 -3.22724134e-01 -5.45531094e-01
2.46117637e-01 4.78304103e-02 2.44035631e-01 4.05465104e-02
6.93201780e-01 -4.25270349e-02 -9.37794626e-01 4.31552559e-01
-8.65744889e-01 3.07805330e-01 3.19476455e-01 6.42188311e-01
6.43686354e-01 -5.88114798e-01 7.46448636e-01 -2.22623020e-01
6.94605634e-02 -6.52742028e-01 -4.42859322e-01 2.93275416e-01
-3.14813733e-01 -3.09568614e-01 1.94816127e-01 -3.51756990e-01
-9.67506647e-01 -1.86652988e-01 4.35812205e-01 -8.69550407e-01
3.88862230e-02 1.28599539e-01 1.07413754e-01 -2.05803066e-02
5.92103541e-01 5.12273908e-01 -2.59175330e-01 -1.33825272e-01
1.48730353e-01 3.22912604e-01 8.31254721e-01 -1.27946138e-01
7.11227000e-01 7.37986386e-01 -3.96650672e-01 -8.24742436e-01
-1.45137429e+00 -5.91147661e-01 -5.59607625e-01 -3.45920682e-01
7.90850818e-01 -1.24735117e+00 -3.35595071e-01 7.16448903e-01
-1.13274384e+00 -4.47590530e-01 -2.84439683e-01 2.17893392e-01
-6.14138842e-01 5.34263074e-01 -4.44633961e-01 -3.44446301e-01
-3.76666933e-01 -1.14512777e+00 1.29660857e+00 3.94912541e-01
-6.84999451e-02 -7.62756765e-01 3.16482112e-02 1.77460715e-01
4.16187644e-01 1.48640677e-01 6.76855892e-02 -2.49359071e-01
-1.04929900e+00 4.57894534e-01 -5.53815126e-01 1.20985545e-01
1.64959371e-01 1.53916806e-01 -8.00585210e-01 -5.06945789e-01
-2.06303805e-01 -3.43446046e-01 1.24140489e+00 8.21662784e-01
1.41814113e+00 -6.48899470e-04 -4.12713915e-01 5.37030399e-01
1.08798456e+00 -5.03657945e-02 5.67737401e-01 6.05156302e-01
8.90047908e-01 2.95235872e-01 1.09552443e+00 3.88674945e-01
6.96927309e-01 8.07294250e-01 7.18943655e-01 -2.40151972e-01
-4.19440448e-01 -2.77987063e-01 5.72060943e-01 6.39771104e-01
-1.80380002e-01 -3.44523042e-01 -6.39325500e-01 8.66603971e-01
-2.10469604e+00 -1.38819027e+00 -1.96418166e-01 1.68887591e+00
6.91833615e-01 9.81441364e-02 4.64451015e-01 -1.84321404e-01
1.15735936e+00 6.57224000e-01 -9.66616631e-01 1.65159613e-01
-4.82374728e-01 -4.01932329e-01 2.26535246e-01 1.50701463e-01
-1.39391041e+00 1.19788599e+00 6.56205320e+00 6.69031620e-01
-1.17762125e+00 9.79879871e-02 4.17496413e-01 -4.86330926e-01
-2.92524815e-01 -3.01204801e-01 -7.79740036e-01 8.12504709e-01
4.17194575e-01 -4.17950422e-01 3.42309803e-01 1.09370899e+00
4.57697093e-01 -4.11940962e-02 -8.21380973e-01 1.32977319e+00
5.34632742e-01 -1.60244668e+00 2.56156027e-01 -3.72452348e-01
1.19374382e+00 4.38924551e-01 2.78913647e-01 2.15284586e-01
-5.14682441e-04 -5.99957764e-01 1.08823586e+00 5.95788598e-01
5.98228693e-01 -2.87217945e-01 5.55257082e-01 -1.72425687e-01
-1.26780772e+00 -2.46618316e-01 -4.95011210e-01 -8.03655162e-02
3.22613478e-01 2.29721308e-01 -3.23906094e-01 1.89840734e-01
1.23363149e+00 1.57538068e+00 -8.87946844e-01 1.24290562e+00
6.68182671e-02 4.87193942e-01 1.68105051e-01 -5.07026315e-02
4.53184426e-01 2.34377772e-01 9.63463604e-01 1.17702138e+00
2.51541436e-01 4.91865501e-02 5.12956642e-02 6.77260518e-01
-2.23197527e-02 -2.23393664e-01 -3.67079705e-01 7.47253895e-02
4.61972505e-01 1.09564376e+00 -6.91678166e-01 -4.91659790e-01
-6.78510010e-01 9.11022842e-01 2.43281081e-01 5.46721637e-01
-1.24392974e+00 1.51655199e-02 9.61574256e-01 1.47047982e-01
6.78973377e-01 -4.67390232e-02 -6.54871529e-03 -1.56112897e+00
5.40166982e-02 -8.99497032e-01 4.89127159e-01 -1.16691208e+00
-1.10775793e+00 5.13489723e-01 -1.28003821e-01 -1.28950453e+00
3.34065914e-01 -3.47485602e-01 -4.42283422e-01 3.33726466e-01
-1.89905298e+00 -1.06283832e+00 -7.94952750e-01 8.12775612e-01
1.14913404e+00 -1.95734695e-01 -5.44677973e-02 1.99368328e-01
-8.51407409e-01 3.69296193e-01 -1.59676135e-01 -8.36218819e-02
8.12855482e-01 -9.65887725e-01 4.94542867e-01 1.19623804e+00
9.20514017e-02 4.39889401e-01 8.52501452e-01 -6.01209521e-01
-1.33581722e+00 -1.25092483e+00 7.75430501e-01 -7.06473947e-01
9.12427485e-01 -1.94866315e-01 -1.11545634e+00 9.08651352e-01
3.17108124e-01 3.03304225e-01 4.20878679e-02 -3.17647815e-01
5.85575700e-02 -7.19810575e-02 -7.15486646e-01 7.38004446e-01
1.72382069e+00 -3.60235304e-01 -8.82167161e-01 3.26619437e-03
1.30536902e+00 -5.51423609e-01 -2.39908114e-01 5.04857421e-01
2.34530032e-01 -1.16834700e+00 1.13119972e+00 -7.85231113e-01
6.69612527e-01 -5.69565475e-01 -5.48039414e-02 -9.92207050e-01
-6.48845911e-01 -6.31211638e-01 -7.32260406e-01 9.13894951e-01
-1.48452252e-01 -9.04885456e-02 5.84978878e-01 3.73699218e-01
-6.07554674e-01 -7.82736659e-01 -8.15957963e-01 -5.72804093e-01
-5.08282304e-01 -2.15254441e-01 5.50134897e-01 9.20920730e-01
-1.85130239e-01 8.41680244e-02 -7.32486308e-01 1.27906844e-01
5.03933549e-01 3.65814149e-01 6.95834041e-01 -1.03564394e+00
2.40849257e-01 -5.11672378e-01 -6.69661283e-01 -1.29589081e+00
1.30985364e-01 -5.45121074e-01 8.44396129e-02 -1.30664062e+00
7.45446742e-01 6.45894483e-02 -7.80424774e-01 4.35832918e-01
-6.41766965e-01 3.64036560e-01 1.73462451e-01 1.83431730e-01
-1.63889372e+00 8.92339885e-01 1.63439965e+00 -1.91598237e-01
-1.69174388e-01 -3.92673552e-01 -9.78377283e-01 7.17987895e-01
6.22484744e-01 -1.55492902e-01 -5.25752306e-01 -6.37952924e-01
-3.22473384e-02 -3.37142944e-01 9.11042690e-01 -9.01650131e-01
1.62407935e-01 -6.47007644e-01 2.33285233e-01 -8.51217210e-01
1.47400439e-01 -5.17838001e-01 -2.72742122e-01 4.08006877e-01
-4.51617628e-01 1.56646848e-01 1.88417032e-01 7.78779864e-01
-3.00372303e-01 2.30127409e-01 8.65533173e-01 8.16835016e-02
-1.73320162e+00 7.61831284e-01 -1.27372295e-01 4.94851530e-01
1.37833655e+00 -3.13573301e-01 -5.72027087e-01 -1.90638602e-01
-4.08315629e-01 4.51434195e-01 6.95547760e-01 9.99260187e-01
8.35737824e-01 -1.47639203e+00 -6.63682342e-01 -3.66391726e-02
4.12903905e-01 1.38837460e-03 6.93429530e-01 9.26367044e-01
-1.75284117e-01 4.60245818e-01 -4.61037636e-01 -7.87502170e-01
-1.18715322e+00 1.17644465e+00 1.73640192e-01 4.84352648e-01
-7.75194526e-01 9.53816175e-01 5.49733520e-01 3.40729833e-01
2.30560690e-01 -5.04194081e-01 -3.52593303e-01 -1.06091157e-03
8.99531245e-01 4.17523414e-01 -3.55105966e-01 -9.15152669e-01
-5.68818152e-01 2.17249125e-01 3.55245061e-02 5.80006540e-01
1.30889082e+00 -5.59999049e-01 8.39502290e-02 4.92256850e-01
8.06821346e-01 -3.73696476e-01 -1.91788888e+00 -4.88166660e-01
-1.29039451e-01 -8.88470948e-01 -3.25198658e-02 -3.93388659e-01
-1.31710029e+00 6.13472700e-01 5.26350319e-01 1.72926441e-01
1.19412839e+00 2.69551963e-01 9.95717704e-01 2.90386826e-01
4.38446343e-01 -1.15220416e+00 5.49119711e-01 5.69335461e-01
1.16486371e+00 -1.66541898e+00 -9.85009149e-02 -2.78709292e-01
-1.04015219e+00 6.77509308e-01 1.08776581e+00 -1.62503302e-01
3.73362452e-01 -3.52039486e-01 -1.23537600e-01 -3.21049809e-01
-7.93829978e-01 -3.96540493e-01 5.51834762e-01 5.38424492e-01
5.10705672e-02 -1.95305347e-01 1.30169420e-02 5.28062224e-01
1.14099942e-02 9.40370113e-02 6.22303486e-01 6.17115557e-01
-8.28113019e-01 -2.57462412e-01 -7.01644346e-02 4.43948895e-01
-1.63096249e-01 -2.94654369e-01 -2.13707477e-01 6.85443461e-01
-6.94582090e-02 7.68108249e-01 1.69874102e-01 -1.25519007e-01
2.36112759e-01 -4.39730108e-01 1.90687492e-01 -5.01784861e-01
-1.75323859e-01 -3.51545066e-01 -3.16989809e-01 -8.53398263e-01
-8.59863043e-01 -1.08575654e+00 -1.12736142e+00 -3.44100952e-01
-9.34904367e-02 -2.35067129e-01 -6.95173889e-02 7.14367151e-01
5.48678160e-01 5.87541103e-01 5.27995944e-01 -1.26076603e+00
-2.62015551e-01 -6.36748016e-01 -6.09590352e-01 5.82924426e-01
8.16984475e-01 -1.18543267e+00 -3.90419871e-01 2.60675669e-01] | [9.719754219055176, -0.28002670407295227] |
49fe2602-f44e-43b4-8e43-c9e9709f40c3 | a-survey-for-efficient-open-domain-question | 2211.07886 | null | https://arxiv.org/abs/2211.07886v1 | https://arxiv.org/pdf/2211.07886v1.pdf | A Survey for Efficient Open Domain Question Answering | Open domain question answering (ODQA) is a longstanding task aimed at answering factual questions from a large knowledge corpus without any explicit evidence in natural language processing (NLP). Recent works have predominantly focused on improving the answering accuracy and achieved promising progress. However, higher accuracy often comes with more memory consumption and inference latency, which might not necessarily be efficient enough for direct deployment in the real world. Thus, a trade-off between accuracy, memory consumption and processing speed is pursued. In this paper, we provide a survey of recent advances in the efficiency of ODQA models. We walk through the ODQA models and conclude the core techniques on efficiency. Quantitative analysis on memory cost, processing speed, accuracy and overall comparison are given. We hope that this work would keep interested scholars informed of the advances and open challenges in ODQA efficiency research, and thus contribute to the further development of ODQA efficiency. | ['Meng Fang', 'Trevor Cohn', 'Xiaojun Chen', 'Qingqing Cao', 'Dongkuan Xu', 'Shangsi Chen', 'Qin Zhang'] | 2022-11-15 | null | null | null | null | ['open-domain-question-answering'] | ['natural-language-processing'] | [-1.33963048e-01 2.88039744e-01 -2.95823924e-02 -4.56368417e-01
-1.25726330e+00 -7.98373997e-01 4.46166366e-01 4.32007700e-01
-5.41879058e-01 1.07038605e+00 2.08372369e-01 -6.04302466e-01
-3.79591823e-01 -1.20723093e+00 -3.67435783e-01 -2.01347262e-01
2.17280611e-01 1.01315320e+00 6.02159917e-01 -5.61933994e-01
4.76685226e-01 3.49898160e-01 -1.58106828e+00 2.27305621e-01
1.25376928e+00 1.11721563e+00 3.75614874e-02 9.04853642e-01
-7.38794744e-01 1.38965142e+00 -8.10155272e-01 -1.03006899e+00
5.03333844e-02 -2.44670868e-01 -1.61009550e+00 -5.11566639e-01
4.07139480e-01 -3.84006083e-01 -2.03334317e-01 9.58810389e-01
6.64273977e-01 3.53359431e-01 3.80287588e-01 -1.18011653e+00
-7.43893266e-01 3.43028367e-01 -1.61808878e-02 6.33521914e-01
6.69200182e-01 5.86027689e-02 1.38441241e+00 -6.17751479e-01
4.45759982e-01 1.01902854e+00 3.12298596e-01 3.93781424e-01
-5.65144897e-01 -2.37609074e-01 3.98108829e-03 8.39851379e-01
-1.22982109e+00 -4.21210796e-01 3.85196358e-01 -5.13431132e-02
1.18248904e+00 4.68433619e-01 3.24339449e-01 4.60610509e-01
5.32906540e-02 7.84859359e-01 8.20673168e-01 -6.01902068e-01
4.54753309e-01 -1.45419398e-02 6.03974640e-01 7.29760706e-01
4.45596188e-01 -2.41536021e-01 -5.76833129e-01 -2.57675380e-01
4.62853134e-01 -4.20202971e-01 -2.74618920e-02 -2.15596572e-01
-7.20131159e-01 1.11071849e+00 2.45798394e-01 3.90669197e-01
-4.22204792e-01 -7.75942206e-02 3.79092127e-01 8.29129338e-01
4.08100665e-01 9.52638149e-01 -6.92019463e-01 -7.05485165e-01
-8.79641652e-01 7.11546719e-01 1.27332759e+00 6.92950726e-01
7.46409357e-01 -4.21483248e-01 -2.79530972e-01 8.70740891e-01
6.87299967e-02 5.24603188e-01 3.16472918e-01 -1.35123158e+00
4.03473347e-01 9.07405794e-01 3.05773318e-01 -9.04608190e-01
-2.67541796e-01 -1.92096885e-02 -2.03680143e-01 -2.48740286e-01
7.98663259e-01 -9.33335945e-02 -4.72194165e-01 1.34546781e+00
3.75197440e-01 -3.40038389e-01 1.80391401e-01 9.60746944e-01
1.01121914e+00 7.82057226e-01 2.01444253e-01 -1.91784188e-01
1.88091743e+00 -1.08956730e+00 -1.00842547e+00 -3.20986778e-01
8.29443038e-01 -9.52660084e-01 1.21156383e+00 2.87361473e-01
-1.46493065e+00 -1.60137400e-01 -7.68832922e-01 -7.76553094e-01
-4.97185290e-01 -1.78182393e-01 8.32054377e-01 8.21865857e-01
-7.30896473e-01 1.80200204e-01 -5.91390431e-01 -3.45539600e-01
4.05232877e-01 2.61176407e-01 1.77286521e-01 -4.70558047e-01
-1.46563876e+00 1.39342725e+00 2.37446979e-01 -2.55974770e-01
-5.29347420e-01 -7.43682086e-01 -5.12719870e-01 1.84134096e-01
8.70847285e-01 -9.93144214e-01 1.80950332e+00 -5.38069904e-01
-1.58297396e+00 8.31186950e-01 -3.62338841e-01 -7.55193949e-01
1.39288098e-01 -4.32614118e-01 -5.56754887e-01 2.75610685e-01
1.71547428e-01 4.69846636e-01 3.91192526e-01 -6.52236223e-01
-8.09378922e-01 -5.22171080e-01 7.35996246e-01 3.29912573e-01
-2.12311894e-01 2.26735219e-01 -2.25114360e-01 -2.18670517e-01
-9.02766734e-02 -6.16049409e-01 -1.64357752e-01 3.07445265e-02
3.66026461e-01 -9.27521825e-01 4.71245795e-01 -7.03714967e-01
1.62037361e+00 -1.68977261e+00 -2.78480828e-01 -1.20103732e-01
2.75324196e-01 6.18827224e-01 -5.80199920e-02 6.76623046e-01
5.56513011e-01 -1.22718327e-01 -1.95104599e-01 3.49633038e-01
2.31208175e-01 4.98825639e-01 -7.15238214e-01 8.37351661e-03
2.98577368e-01 1.28722000e+00 -1.03463435e+00 -7.49589264e-01
8.86705890e-03 -7.90014341e-02 -5.96088588e-01 4.40509379e-01
-7.83683419e-01 -4.34841402e-02 -6.46092713e-01 6.30279720e-01
5.40907741e-01 -4.49676543e-01 -4.17005792e-02 2.81384438e-01
-2.97028404e-02 9.53580618e-01 -1.03999007e+00 1.62386107e+00
-6.06039286e-01 4.93342638e-01 -1.13303684e-01 -9.78048503e-01
8.93659532e-01 2.83953130e-01 2.29273457e-02 -1.16638005e+00
1.88529551e-01 4.82692599e-01 1.16234638e-01 -8.15684915e-01
8.05497527e-01 -1.93945512e-01 -3.54301035e-02 6.83734238e-01
1.18958823e-01 -4.96520996e-01 5.38801908e-01 3.00278276e-01
1.06796312e+00 -2.58925676e-01 4.92022485e-01 -4.16859180e-01
8.40652287e-01 5.40126443e-01 5.34568243e-02 8.07129085e-01
-5.18962681e-01 3.65599953e-02 5.09488225e-01 -6.54737949e-01
-8.61549377e-01 -1.07893205e+00 -7.14209303e-02 1.39401579e+00
1.16918698e-01 -4.00557637e-01 -7.37958133e-01 -7.00455785e-01
-3.23092014e-01 7.74668217e-01 -2.14450821e-01 6.88962340e-02
-7.38590598e-01 -5.00258803e-01 8.53242278e-01 3.95788282e-01
7.12029636e-01 -1.02257013e+00 -8.45902920e-01 2.47977212e-01
-8.12556326e-01 -1.12072384e+00 5.02889268e-02 -2.22562149e-01
-1.12007177e+00 -1.25963879e+00 -5.89494109e-01 -7.90804267e-01
1.93210334e-01 3.05225134e-01 1.67182004e+00 2.69371480e-01
8.32350343e-04 4.45967674e-01 -5.18704057e-01 -8.23647439e-01
-1.50880039e-01 3.74718457e-01 -3.14860672e-01 -6.70652866e-01
1.03293347e+00 -2.41847664e-01 -7.66303360e-01 3.11580122e-01
-1.00862074e+00 -4.22986656e-01 4.23608959e-01 6.35586441e-01
1.98167637e-01 -1.07954368e-01 9.54224944e-01 -8.54099452e-01
1.06087613e+00 -5.28137565e-01 -5.14990270e-01 5.56087017e-01
-7.13108897e-01 3.67650598e-01 4.68739003e-01 5.65557703e-02
-1.38352573e+00 -6.52354419e-01 -6.63475215e-01 3.56786519e-01
-6.42614737e-02 5.28098464e-01 1.66924074e-02 -2.32623927e-02
1.01267946e+00 1.86055377e-02 -6.74038827e-02 -2.96244472e-01
5.31893611e-01 7.06499517e-01 3.92441362e-01 -8.50473464e-01
4.43852395e-01 4.49951947e-01 -3.54021043e-01 -6.55071318e-01
-1.32572174e+00 -7.28037953e-01 -1.58471137e-01 1.20092183e-01
5.33691823e-01 -6.19358778e-01 -6.12613261e-01 4.21954654e-02
-1.20284021e+00 -3.13203819e-02 -5.24096966e-01 1.29781052e-01
-5.53451359e-01 5.55323422e-01 -6.37423694e-01 -9.15453613e-01
-5.42706549e-01 -6.86559498e-01 9.33764398e-01 3.13596845e-01
-4.54370409e-01 -1.10937929e+00 2.78301865e-01 1.07917249e+00
5.60545146e-01 -4.48124439e-01 9.65246975e-01 -7.84601450e-01
-6.95995569e-01 -1.30960375e-01 -3.88873547e-01 9.05488804e-02
-3.52178901e-01 -4.55396235e-01 -8.49654317e-01 2.63265781e-02
1.55295372e-01 -6.58885062e-01 4.18651789e-01 1.08731806e-01
1.01759207e+00 -4.13436174e-01 1.72443852e-01 -2.14575723e-01
1.34056413e+00 1.95278049e-01 8.39865744e-01 4.16528463e-01
-2.70932205e-02 8.64220977e-01 9.72930253e-01 2.57972062e-01
6.42367184e-01 3.07859004e-01 1.84051394e-01 3.90326321e-01
-6.79042041e-02 -2.39169493e-01 3.54156271e-02 9.04150903e-01
-2.59501159e-01 -3.63515943e-01 -1.09498048e+00 8.83364379e-01
-1.73357189e+00 -9.32989180e-01 -1.85070738e-01 1.95499492e+00
8.05665553e-01 -8.21165815e-02 1.14258891e-02 2.83654541e-01
3.20766866e-01 8.38676542e-02 -4.48095977e-01 -8.15341353e-01
-9.21128616e-02 6.82801187e-01 1.06948748e-01 5.96037149e-01
-5.99529922e-01 8.95057023e-01 7.14552116e+00 9.02415872e-01
-5.48972189e-01 2.09490240e-01 3.92153203e-01 1.18076116e-01
-4.36973065e-01 6.40977174e-02 -8.10582876e-01 4.12043482e-02
1.29250026e+00 -4.52949554e-01 4.36908394e-01 7.93327987e-01
-2.24974692e-01 -4.17725235e-01 -9.59523559e-01 8.86919200e-01
-9.68460832e-03 -1.41452146e+00 1.89105794e-01 -1.73355922e-01
6.40816391e-01 -3.80238593e-02 -3.43054146e-01 7.52889276e-01
5.67397654e-01 -1.01668584e+00 1.66928738e-01 3.11493963e-01
2.22035095e-01 -7.71734834e-01 1.11355329e+00 7.30677426e-01
-8.32914293e-01 -3.84924591e-01 -6.22068107e-01 -7.25631475e-01
2.60907501e-01 4.32742417e-01 -6.43861830e-01 7.06479371e-01
7.94672847e-01 -5.32079265e-02 -3.98938626e-01 7.33165264e-01
-2.52126575e-01 7.13850379e-01 -3.49941432e-01 -4.81983900e-01
4.12328899e-01 -1.07981123e-01 2.07301423e-01 9.10887361e-01
-3.71047482e-02 7.36524999e-01 -2.82596350e-01 5.26924074e-01
-1.11111186e-01 3.84498179e-01 -3.04154575e-01 -1.22391351e-01
4.34659541e-01 8.20421219e-01 -3.61964524e-01 -5.48443794e-01
-6.93403244e-01 6.36750579e-01 6.77711666e-01 -7.54810125e-02
-8.33976090e-01 -3.77590746e-01 6.72326803e-01 9.51940715e-02
1.17458507e-01 -1.96412042e-01 -5.12984812e-01 -1.15765989e+00
1.75550088e-01 -1.09988904e+00 1.08047485e+00 -8.14990520e-01
-1.51603520e+00 3.61962616e-01 -4.63623777e-02 -6.36553228e-01
-1.80993363e-01 -6.19232416e-01 -2.32610330e-01 6.51565790e-01
-1.89793444e+00 -5.68589270e-01 -1.89760566e-01 5.71162999e-01
6.65246189e-01 2.52952784e-01 1.05224800e+00 4.57518131e-01
-5.91996238e-02 5.06584525e-01 -2.09128670e-02 -1.20481718e-02
6.09133661e-01 -9.19113159e-01 2.49440655e-01 6.29972279e-01
3.20682943e-01 6.89315200e-01 6.36064529e-01 -2.33504832e-01
-1.39009488e+00 -6.30108237e-01 1.53474033e+00 -9.60946798e-01
8.17666948e-01 7.56278932e-02 -1.04184318e+00 4.19796556e-01
2.98361212e-01 -2.27171004e-01 8.59063685e-01 3.92512232e-01
-2.49348074e-01 -2.50637922e-02 -1.21041274e+00 5.60426235e-01
9.48992550e-01 -6.17098212e-01 -1.38630164e+00 2.37041488e-01
7.73048878e-01 -3.64453644e-01 -1.01800764e+00 3.01444471e-01
4.66201216e-01 -8.96768749e-01 9.88499701e-01 -8.39168191e-01
4.72257525e-01 -1.02786124e-01 -1.47693336e-01 -7.57251740e-01
1.42432079e-02 -3.14965546e-01 -4.66156989e-01 8.28052521e-01
3.60596418e-01 -6.89984739e-01 8.71528268e-01 1.07302606e+00
2.06743721e-02 -8.80857050e-01 -1.11862028e+00 -5.85839987e-01
4.83095884e-01 -6.35838985e-01 8.08172643e-01 7.49761999e-01
2.27405757e-01 6.71730995e-01 5.30365147e-02 8.46403390e-02
1.72031939e-01 3.79281878e-01 6.48628175e-01 -1.27120137e+00
3.21323052e-02 -2.82142848e-01 -1.61704049e-01 -1.60155487e+00
-2.16110125e-01 -4.40327853e-01 -2.19857916e-01 -1.83156896e+00
-2.79208571e-02 -3.57691765e-01 -1.55296465e-02 1.90285906e-01
-2.54233122e-01 1.77677467e-01 5.63236438e-02 1.29936025e-01
-1.11302412e+00 5.11930883e-01 1.48501813e+00 -2.02920125e-03
1.47235757e-02 -1.43572897e-01 -9.57837999e-01 6.57057643e-01
8.86246741e-01 -3.22441936e-01 -5.88185251e-01 -6.62365615e-01
5.60384750e-01 3.87322724e-01 8.89146626e-02 -8.07561278e-01
4.61691737e-01 -2.10223615e-01 -2.75182217e-01 -6.12862229e-01
2.35839218e-01 -8.25624406e-01 -5.58520794e-01 4.30703163e-01
-2.61413187e-01 1.80300042e-01 -7.25820020e-04 3.62672299e-01
-5.92062294e-01 -5.59857130e-01 5.68670750e-01 -3.71903121e-01
-9.37911153e-01 1.85295269e-01 -4.30562973e-01 6.66112006e-01
9.91422832e-01 -4.11679074e-02 -5.98057568e-01 -5.37335932e-01
-4.66482282e-01 5.95476329e-01 4.33474667e-02 4.45439249e-01
3.83231372e-01 -8.76822114e-01 -6.91190124e-01 -2.13073477e-01
2.31523141e-01 1.03780501e-01 3.39811981e-01 4.91788119e-01
-9.34147477e-01 1.00564122e+00 -5.40582649e-02 -1.91539481e-01
-1.02394366e+00 6.20616019e-01 2.64634013e-01 -8.19523036e-01
-1.29210785e-01 7.66134918e-01 -1.60685569e-01 -6.34551823e-01
1.27828807e-01 -7.38327671e-03 -3.69339645e-01 -5.25505655e-02
8.06887746e-01 8.08472335e-01 3.65466624e-01 -1.21875092e-01
-2.60628521e-01 2.55573183e-01 4.93866578e-02 -2.33700480e-02
9.57558155e-01 -2.78653860e-01 -3.20351809e-01 2.43767992e-01
7.92258620e-01 -8.18240866e-02 -5.48343956e-01 -4.28790808e-01
3.08222681e-01 -5.49881458e-01 -1.47501975e-01 -8.96725535e-01
-5.13810694e-01 9.14605021e-01 3.46307337e-01 5.14483750e-01
1.12476397e+00 1.51631743e-01 1.26011074e+00 9.38410819e-01
6.24099791e-01 -1.27617383e+00 1.08292334e-01 7.62233794e-01
6.57534301e-01 -1.34449029e+00 8.52891505e-02 -4.90874499e-01
-4.60377723e-01 1.07096434e+00 7.20976949e-01 -9.09591317e-02
5.15712440e-01 4.27711522e-03 2.94059664e-01 -5.20984411e-01
-8.74190867e-01 -3.88469130e-01 7.69750997e-02 4.40295547e-01
5.60356617e-01 -9.93990973e-02 -8.51924539e-01 6.00981534e-01
-5.56995451e-01 1.55864045e-01 2.70092338e-01 9.70406651e-01
-7.78424799e-01 -1.23221874e+00 -3.96231264e-01 5.06869435e-01
-7.13479578e-01 -1.82367906e-01 -3.88379216e-01 6.33274436e-01
-9.97099429e-02 1.47356045e+00 1.97441265e-01 3.08800310e-01
4.87150759e-01 3.87975127e-01 5.78829408e-01 -3.92379999e-01
-6.09936833e-01 -9.16770279e-01 5.09298146e-01 -4.36765432e-01
-5.17607212e-01 -2.98294723e-01 -1.19786274e+00 -8.37458670e-01
-5.02817631e-01 8.05334508e-01 4.53661710e-01 9.39478338e-01
4.80339974e-01 1.15592234e-01 3.78315188e-02 3.37611318e-01
-7.42529571e-01 -7.90506601e-01 -2.37825811e-01 9.55281556e-02
1.67293578e-01 -4.25467402e-01 -1.40497267e-01 -1.50279313e-01] | [11.235349655151367, 7.9423346519470215] |
eb710abb-d344-4a0f-954e-f14de8cbbde5 | a-skip-connection-architecture-for | null | null | http://openaccess.thecvf.com/content_CVPRW_2019/html/Media_Forensics/Mazaheri_A_Skip_Connection_Architecture_for_Localization_of_Image_Manipulations_CVPRW_2019_paper.html | http://openaccess.thecvf.com/content_CVPRW_2019/papers/Media%20Forensics/Mazaheri_A_Skip_Connection_Architecture_for_Localization_of_Image_Manipulations_CVPRW_2019_paper.pdf | A Skip Connection Architecture for Localization of Image Manipulations | Detection and localization of image manipulations are becoming of increasing interest to researchers in recent years due to the significant rise of malicious content-changing image tampering on the web. One of the major challenges for an image manipulation detection method is to discriminate between the tampered regions and other regions in an image. We observe that most of the manipulated images leave some traces near boundaries of manipulated regions including blurred edges. In order to exploit these traces in localizing the tampered regions, we propose an encoder-decoder based network where we fuse representations from early layers in the encoder (which are richer in low-level spatial cues, like edges) by skip pooling with representations of the last layer of the decoder and use for manipulation detection. In addition, we utilize resampling features extracted from patches of images by feeding them to LSTM cells to capture the transition between manipulated and non manipulated blocks in the frequency domain and combine the output of the LSTM with our encoder. The overall framework is capable of detecting different types of image manipulations simultaneously including copy-move, removal, and splicing. Experimental results on two standard benchmark datasets (CASIA 1.0 and NIST’16) demonstrate that the proposed method can achieve a significantly better performance than the state-of-the-art methods and baselines. | ['Amit K. Roy-Chowdhury', 'Niluthpol Chowdhury Mithun', 'Ghazal Mazaheri', 'Jawadul H. Bappy'] | 2019-06-01 | null | null | null | ieee-conference-on-computer-vision-and-2 | ['image-manipulation-detection'] | ['computer-vision'] | [ 8.32271934e-01 -5.60487866e-01 -1.82478294e-01 9.53945145e-02
-6.68386877e-01 -5.59423268e-01 5.24607420e-01 -3.45409811e-02
-3.05680782e-01 1.37200311e-01 2.32296571e-01 9.27439183e-02
3.77627581e-01 -6.47306085e-01 -1.13453591e+00 -7.71132886e-01
-1.61720797e-01 -7.82595277e-01 5.05804300e-01 2.27003787e-02
6.86681390e-01 2.12069944e-01 -1.37633479e+00 1.00541627e+00
5.90292037e-01 1.28926730e+00 3.80608439e-01 5.78178465e-01
1.34304315e-01 9.84125912e-01 -5.64786434e-01 -2.97974885e-01
3.24966937e-01 -2.46096224e-01 -4.76909280e-01 5.51104784e-01
5.03208518e-01 -7.68626094e-01 -9.95522022e-01 1.62558568e+00
2.51232296e-01 -8.16907734e-02 3.41938615e-01 -9.34282720e-01
-1.18025303e+00 4.73583788e-01 -1.17196107e+00 6.90889001e-01
3.20934743e-01 2.47214764e-01 3.83602232e-01 -9.86895382e-01
7.08183885e-01 1.33684826e+00 3.52307856e-01 8.67700577e-02
-7.83314764e-01 -6.05399489e-01 8.35826620e-02 8.35347116e-01
-1.32017696e+00 -5.68702757e-01 8.86780918e-01 -3.29214841e-01
7.99040079e-01 1.22146182e-01 -3.69326174e-02 1.27597296e+00
5.42113364e-01 9.98286903e-01 1.04919481e+00 -4.35138315e-01
-2.01487094e-01 1.25745967e-01 8.94702077e-02 9.24406707e-01
3.50056626e-02 -1.51458485e-02 -6.32996321e-01 -1.38623249e-02
7.27947056e-01 2.81184196e-01 -7.46799052e-01 -9.56146047e-02
-1.33102179e+00 4.54722047e-01 5.70799947e-01 3.25969219e-01
-4.27624136e-01 2.65684634e-01 6.88333392e-01 2.19988108e-01
1.69737622e-01 -4.24451753e-02 -1.37890354e-01 4.48384881e-02
-8.28427851e-01 -1.34222165e-01 1.51729062e-01 9.25411105e-01
5.84597588e-01 -2.19651476e-01 -3.62241715e-01 7.78027773e-01
2.00802740e-02 2.50805289e-01 7.25593567e-01 -6.63096189e-01
8.16331208e-01 5.39658129e-01 2.10154541e-02 -1.32660222e+00
4.15652156e-01 -2.85213739e-01 -9.29514050e-01 -7.67245069e-02
8.45060721e-02 2.11034909e-01 -1.08777869e+00 1.25449538e+00
8.29646811e-02 4.19852227e-01 -2.67671317e-01 8.44002247e-01
3.98222655e-01 7.56231487e-01 -4.19073291e-02 -7.36215338e-02
1.65523911e+00 -1.13321984e+00 -8.26166928e-01 -3.85691166e-01
4.08940196e-01 -8.39293420e-01 7.93263316e-01 3.32085133e-01
-9.06174123e-01 -8.14141691e-01 -1.23056674e+00 -2.03752711e-01
-7.21650004e-01 2.39324421e-01 9.51025933e-02 5.68559349e-01
-7.35040128e-01 5.69634140e-01 -4.91822422e-01 -9.34699923e-02
6.83993161e-01 2.57990528e-02 -4.53862846e-01 -2.76623160e-01
-1.34717596e+00 6.90295339e-01 5.48101723e-01 3.62227798e-01
-1.07018006e+00 -2.76606500e-01 -9.87275243e-01 2.93080568e-01
4.72307563e-01 -1.44557180e-02 8.44210327e-01 -1.21169591e+00
-9.24705267e-01 9.14506376e-01 -2.56349146e-01 -3.64166439e-01
3.97702903e-01 -1.24678522e-01 -6.94825768e-01 3.82547677e-01
-2.01407392e-02 5.40368378e-01 1.43027127e+00 -1.42471182e+00
-7.37209022e-01 -4.97974247e-01 -9.06976089e-02 -1.16942257e-01
-4.44152117e-01 3.17291021e-01 -7.50627339e-01 -9.34709430e-01
-1.93122476e-02 -6.12030208e-01 2.60818541e-01 1.31746739e-01
-5.11264622e-01 1.60528421e-01 1.54207706e+00 -1.46217334e+00
1.49404287e+00 -2.38932562e+00 -1.46289423e-01 -4.75901365e-02
5.16590662e-02 6.90364182e-01 -3.96600306e-01 4.71906841e-01
-1.74608171e-01 2.37505987e-01 -3.30736578e-01 -9.02347714e-02
-2.44302750e-01 -1.75438777e-01 -6.47259116e-01 9.44171190e-01
3.30989361e-01 9.21293557e-01 -7.27573752e-01 -4.17798430e-01
4.76256251e-01 5.38560331e-01 -2.03875542e-01 2.37666327e-03
-1.45340383e-01 2.44197682e-01 -1.88439384e-01 8.74937296e-01
1.09942770e+00 -6.54023364e-02 -3.99535075e-02 -4.47778821e-01
1.72096281e-03 2.17869610e-01 -8.89094532e-01 1.60493457e+00
-2.96012461e-01 8.90491366e-01 2.17832670e-01 -8.87723029e-01
3.06472421e-01 1.81379959e-01 -5.64223193e-02 -8.79078865e-01
4.56739545e-01 2.22923830e-02 -3.04179396e-02 -1.04288530e+00
4.54772532e-01 6.61172926e-01 -2.11068485e-02 3.49787444e-01
8.21656287e-02 6.17757857e-01 1.10039264e-01 4.99137975e-02
1.05209219e+00 5.97948581e-02 6.29788265e-02 1.81497291e-01
8.86623740e-01 -5.30361474e-01 3.42549860e-01 7.26606011e-01
-5.99685371e-01 4.92390752e-01 3.77473176e-01 -2.05171138e-01
-8.43691766e-01 -8.50771725e-01 2.91136149e-02 1.06300282e+00
4.44520116e-01 -1.26139939e-01 -9.22272861e-01 -8.44310582e-01
-4.54549156e-02 5.39216638e-01 -7.88393259e-01 -4.46421176e-01
-8.70832384e-01 -3.30628157e-01 4.53766882e-01 2.53849089e-01
1.21760607e+00 -1.19893527e+00 -5.37974179e-01 4.74131713e-03
-5.08591115e-01 -1.51749098e+00 -8.48343253e-01 -9.37991962e-02
-4.93234009e-01 -1.07749021e+00 -6.73504055e-01 -1.26806486e+00
6.64863050e-01 8.13902497e-01 4.31068897e-01 4.95473556e-02
-5.52730262e-01 -1.77405495e-02 -3.57693017e-01 2.67245471e-02
-4.09382761e-01 -2.92665303e-01 -4.77725655e-01 4.69203681e-01
3.69459003e-01 -2.38372356e-01 -7.67135978e-01 2.41093397e-01
-1.32761335e+00 -1.40743226e-01 7.02168941e-01 5.55402279e-01
2.23990127e-01 5.64583063e-01 1.89842895e-01 -6.77093625e-01
4.12968814e-01 -5.40630460e-01 -2.90832400e-01 3.95800024e-01
1.51011586e-01 2.14306321e-02 6.77586615e-01 -5.72303832e-01
-1.10526001e+00 -1.75999850e-01 2.49801785e-01 -6.10797107e-01
-1.52805969e-01 1.25750139e-01 -3.07710201e-01 -3.49852055e-01
3.62970501e-01 8.09897065e-01 -2.01922238e-01 -4.31137741e-01
3.07721734e-01 1.14490497e+00 6.95183575e-01 3.48753259e-02
8.45920980e-01 6.37742817e-01 -4.50214237e-01 -8.02030742e-01
-3.53242397e-01 -4.87500101e-01 -4.68019634e-01 -3.75959247e-01
7.40096450e-01 -8.73132527e-01 -5.67361414e-01 1.01536274e+00
-1.28420031e+00 1.46590739e-01 2.80601293e-01 1.19566657e-02
-1.65425628e-01 1.19353843e+00 -9.68880951e-01 -6.05819464e-01
-3.74719739e-01 -1.45861733e+00 1.28131235e+00 2.05270678e-01
3.48599136e-01 -5.90537071e-01 -6.05362177e-01 3.98983955e-01
3.91144276e-01 3.43006134e-01 1.09514415e+00 -2.27133855e-01
-7.57872224e-01 -4.08602148e-01 -4.66888577e-01 5.09933770e-01
3.30126286e-01 -1.66877016e-01 -1.10604191e+00 -3.17651331e-01
4.22288567e-01 -1.93897694e-01 1.34187627e+00 9.01218429e-02
1.68927598e+00 -4.85615015e-01 -4.46452379e-01 4.53650415e-01
1.56478572e+00 3.07144284e-01 1.02456582e+00 2.91315705e-01
6.30129755e-01 3.68865877e-01 3.80958349e-01 1.98172584e-01
1.43656917e-02 5.26678085e-01 6.03709102e-01 9.85087976e-02
-3.27617586e-01 -2.39929184e-01 8.54126334e-01 6.32241189e-01
4.92293239e-01 -3.33137453e-01 -5.78921974e-01 6.35783195e-01
-1.66739726e+00 -1.10229123e+00 -3.49691287e-02 2.29599643e+00
6.87539995e-01 1.92952186e-01 -4.03132558e-01 2.94042706e-01
1.36868834e+00 6.93897367e-01 -5.38021147e-01 -4.68817651e-01
-7.69914910e-02 9.72692445e-02 9.13211524e-01 1.29673868e-01
-1.60292888e+00 9.14961159e-01 5.53133106e+00 1.02046621e+00
-1.19297373e+00 1.09205872e-01 7.38911092e-01 2.60101855e-01
3.66040111e-01 -3.21578443e-01 -4.92455006e-01 9.71831858e-01
7.66083360e-01 3.78520578e-01 7.89151192e-01 5.94963968e-01
3.61160338e-01 -3.34145516e-01 -8.25135946e-01 1.00394535e+00
5.38304269e-01 -1.07942200e+00 1.30032927e-01 -8.87211263e-02
7.67976761e-01 -1.29544407e-01 3.95783126e-01 2.07437694e-01
-1.91670746e-01 -8.86508942e-01 7.75044560e-01 2.84998417e-01
7.82127380e-01 -6.28214896e-01 6.01478875e-01 4.28172290e-01
-1.20363140e+00 -3.79665226e-01 -3.99741024e-01 2.39663884e-01
-2.56618652e-02 4.88521785e-01 -6.15620971e-01 3.26740682e-01
7.22841859e-01 6.98558867e-01 -5.79074800e-01 9.52728748e-01
-2.43276820e-01 4.61096406e-01 1.81308821e-01 3.77342492e-01
3.08718532e-01 1.05804093e-02 4.07018781e-01 1.30971408e+00
2.07091779e-01 -3.40665191e-01 -1.10192187e-01 8.92388225e-01
-6.23828888e-01 -2.87250578e-01 -6.14929914e-01 -1.77989110e-01
4.22128379e-01 1.12649500e+00 -5.95481634e-01 -3.82320851e-01
-5.96361816e-01 1.56967568e+00 1.76284522e-01 4.95052844e-01
-1.24439442e+00 -6.84143662e-01 4.81128305e-01 5.62496036e-02
7.52738178e-01 -2.33174697e-01 -4.58409786e-02 -1.26777458e+00
3.41538817e-01 -1.14154577e+00 1.63603976e-01 -7.92509496e-01
-1.04021394e+00 1.26408577e-01 -4.14242893e-01 -1.16588461e+00
2.18090981e-01 -5.99071026e-01 -4.11550462e-01 8.40783358e-01
-1.69869840e+00 -9.99396384e-01 -4.36740726e-01 5.58238268e-01
9.42201138e-01 2.11544901e-01 4.36603516e-01 4.47207719e-01
-7.10845351e-01 5.66443145e-01 1.09482132e-01 7.08097696e-01
7.37691998e-01 -6.96342766e-01 6.71376586e-01 1.40420234e+00
-3.43937762e-02 6.25665069e-01 1.98277012e-01 -7.37903953e-01
-1.63188410e+00 -1.35050905e+00 6.04627848e-01 6.77151456e-02
5.14414310e-01 -6.09871566e-01 -9.31956649e-01 6.24673367e-01
6.99869156e-01 4.71166484e-02 1.34755701e-01 -7.88289189e-01
-6.98865592e-01 1.66971043e-01 -1.18519640e+00 4.51972872e-01
7.26828814e-01 -9.63752449e-01 -5.43406248e-01 3.33626151e-01
3.80964577e-01 -2.67339259e-01 -4.06881869e-01 3.74896109e-01
4.11599040e-01 -9.68656063e-01 9.14757669e-01 -1.63506255e-01
5.92664301e-01 -3.66275668e-01 -1.79672152e-01 -9.54758644e-01
-3.99235010e-01 -5.68891108e-01 -2.70565480e-01 1.12926567e+00
-1.14745311e-01 -3.80473822e-01 4.53216612e-01 5.07941879e-02
5.31247258e-02 -5.56476235e-01 -7.33307779e-01 -4.50977117e-01
-3.74687225e-01 -8.39150921e-02 1.40684858e-01 9.08219278e-01
-8.15202072e-02 -1.81453258e-01 -5.31525075e-01 5.03800213e-01
5.76644421e-01 -8.66537914e-02 3.83738697e-01 -2.97145307e-01
-2.06631511e-01 -3.45157892e-01 -5.91687620e-01 -1.39020264e+00
3.12369466e-02 -7.22195029e-01 1.05935201e-01 -1.27690101e+00
5.08595765e-01 2.91082025e-01 -3.04767549e-01 3.90289336e-01
-3.87408048e-01 3.74534428e-01 1.26476973e-01 1.99753195e-01
-4.56732780e-01 3.01649839e-01 1.11860085e+00 -4.86123979e-01
2.17658579e-01 -5.70211709e-01 -3.56102556e-01 6.83624387e-01
6.10404134e-01 -3.21379155e-01 -8.84631798e-02 -6.96671307e-01
-3.27976465e-01 -4.32364419e-02 5.16994655e-01 -1.20052111e+00
7.94340596e-02 2.12203100e-01 7.64567673e-01 -6.14490390e-01
5.97695485e-02 -7.81167388e-01 -2.77566016e-01 6.56107128e-01
-4.89438534e-01 -1.57455981e-01 1.41334862e-01 8.10929120e-01
-3.40363145e-01 -3.70430976e-01 1.05944932e+00 -2.48710424e-01
-1.02882111e+00 8.39108974e-02 -5.67274272e-01 -2.85655171e-01
1.24203575e+00 -2.53622979e-01 -6.63648784e-01 -2.33997956e-01
-3.39712143e-01 -8.34977403e-02 3.78181577e-01 7.85925269e-01
8.05394948e-01 -1.04771531e+00 -4.00063485e-01 4.72010136e-01
6.96817227e-03 -5.44903636e-01 5.57909250e-01 7.01866746e-01
-6.56246245e-01 4.34745282e-01 -2.37698585e-01 -4.39942390e-01
-1.34235156e+00 1.04538286e+00 3.00712377e-01 -1.42165497e-01
-4.25970197e-01 7.16461897e-01 2.34010369e-01 3.74373138e-01
3.05765986e-01 -4.59531754e-01 6.35348679e-03 -1.79278344e-01
8.64478052e-01 3.10363084e-01 8.81145969e-02 -8.90983045e-01
-2.95597166e-01 3.40150923e-01 -4.15428311e-01 3.50690871e-01
9.59982932e-01 -3.35336268e-01 -1.87285691e-01 -1.03383558e-02
1.80433095e+00 -1.16685174e-01 -1.21261728e+00 -4.75428730e-01
-4.10126783e-02 -8.37169170e-01 2.59638429e-01 -6.91002905e-01
-1.28631401e+00 1.02605593e+00 8.55048537e-01 1.34317085e-01
1.32494533e+00 -3.24925870e-01 1.23869550e+00 -1.75166670e-02
3.23298454e-01 -1.04778743e+00 2.12693110e-01 2.33933493e-01
8.85884166e-01 -1.13995576e+00 -1.65159747e-01 -5.75128376e-01
-4.10389781e-01 1.09362233e+00 4.10860807e-01 -4.20413315e-01
3.72993857e-01 2.29236767e-01 -2.05826148e-01 1.48167759e-01
-4.29941237e-01 -3.62935103e-02 9.08125937e-02 4.40559924e-01
9.32246000e-02 -2.67132312e-01 -1.21466972e-01 9.98895615e-02
4.91360694e-01 -1.46931529e-01 5.31581581e-01 1.14120471e+00
-6.30184531e-01 -7.68842518e-01 -6.82270885e-01 5.41810751e-01
-7.39221215e-01 -1.95807979e-01 -4.37159538e-01 5.45700192e-01
4.64811116e-01 1.01519489e+00 4.90121692e-02 -4.42976445e-01
1.40481163e-02 4.05683406e-02 4.46417898e-01 -2.58804947e-01
-6.74208820e-01 2.00518370e-01 -1.97155863e-01 -7.29236722e-01
-2.41576985e-01 -5.19309402e-01 -8.23496699e-01 -9.12770182e-02
-4.44890380e-01 -4.51223373e-01 6.16969764e-01 6.30229712e-01
5.88589966e-01 5.83480060e-01 8.46222460e-01 -8.86074543e-01
-7.39254057e-01 -1.07268095e+00 -5.50223827e-01 8.26226413e-01
5.58239579e-01 -5.08899450e-01 -4.12460148e-01 4.68643844e-01] | [12.30740737915039, 0.9365178346633911] |
cecafb23-0e4a-4cd2-9d96-39cf97ce407e | imageeye-batch-image-processing-using-program | 2304.03253 | null | https://arxiv.org/abs/2304.03253v3 | https://arxiv.org/pdf/2304.03253v3.pdf | ImageEye: Batch Image Processing Using Program Synthesis | This paper presents a new synthesis-based approach for batch image processing. Unlike existing tools that can only apply global edits to the entire image, our method can apply fine-grained edits to individual objects within the image. For example, our method can selectively blur or crop specific objects that have a certain property. To facilitate such fine-grained image editing tasks, we propose a neuro-symbolic domain-specific language (DSL) that combines pre-trained neural networks for image classification with other language constructs that enable symbolic reasoning. Our method can automatically learn programs in this DSL from user demonstrations by utilizing a novel synthesis algorithm. We have implemented the proposed technique in a tool called ImageEye and evaluated it on 50 image editing tasks. Our evaluation shows that ImageEye is able to automate 96% of these tasks. | ['Isil Dillig', 'Roopsha Samanta', 'Qiaochu Chen', 'Celeste Barnaby'] | 2023-04-06 | null | null | null | null | ['program-synthesis'] | ['computer-code'] | [ 6.38307810e-01 3.16911727e-01 1.22740448e-01 -6.05091870e-01
-1.92697823e-01 -6.53800547e-01 6.84648037e-01 -4.69201058e-02
-4.59987074e-01 5.51217556e-01 -4.46798742e-01 -4.69650894e-01
2.67131180e-02 -7.26426482e-01 -1.15179646e+00 -7.30240643e-02
2.67277539e-01 1.73942685e-01 5.27781367e-01 -6.02849126e-02
5.44814408e-01 8.96049917e-01 -1.75742495e+00 7.02789903e-01
1.18363154e+00 7.19115198e-01 6.37493849e-01 9.23184931e-01
-2.99721062e-01 1.15976417e+00 -9.93205070e-01 -9.21672359e-02
1.25562072e-01 -3.04668993e-01 -8.95081043e-01 -3.65125798e-02
7.24769890e-01 -5.18663466e-01 1.48138344e-01 1.35247660e+00
1.57848019e-02 1.69465505e-02 6.33715212e-01 -1.46123731e+00
-8.34314108e-01 6.96442306e-01 -2.10838616e-01 -2.77750522e-01
4.55154419e-01 3.60085964e-01 2.45299950e-01 -6.97807491e-01
9.32770967e-01 1.32738662e+00 5.71066022e-01 6.34278715e-01
-1.30574465e+00 -5.20054638e-01 -3.85741442e-02 5.19315042e-02
-1.14839375e+00 -1.82866037e-01 5.35191000e-01 -6.30441248e-01
1.44466102e+00 3.02989155e-01 6.13995731e-01 6.34393394e-01
2.29614884e-01 5.42668641e-01 1.37741399e+00 -8.30722809e-01
2.17603713e-01 2.70108432e-01 1.01023138e-01 1.21172297e+00
-2.54521463e-02 1.63396761e-01 -3.72848421e-01 2.23440856e-01
1.28041363e+00 -1.30214110e-01 1.00653246e-01 -3.83372575e-01
-1.39621210e+00 3.65305364e-01 2.74985641e-01 2.02019975e-01
-1.85121462e-01 6.23451650e-01 5.45416832e-01 5.93384266e-01
-8.33634939e-03 8.69988084e-01 -3.16385835e-01 -1.14012793e-01
-9.62206244e-01 2.39559650e-01 9.34695542e-01 1.30062068e+00
9.37033296e-01 5.25557399e-02 -4.56378996e-01 5.80864727e-01
-1.56816244e-01 1.24780416e-01 4.18225944e-01 -1.57533395e+00
5.29329814e-02 6.46219730e-01 2.50424534e-01 -7.58001387e-01
-1.40891612e-01 6.37296289e-02 -3.57584894e-01 1.18059027e+00
3.40519071e-01 -1.73626319e-02 -1.04839420e+00 1.57476556e+00
4.79760244e-02 -8.30933377e-02 -8.03661495e-02 6.17285728e-01
7.07428098e-01 5.02507448e-01 3.25747281e-01 2.23670870e-01
1.12107944e+00 -1.25970387e+00 -5.87328553e-01 2.32329462e-02
6.93360209e-01 -3.72647971e-01 1.59351218e+00 5.95825732e-01
-1.30063617e+00 -8.29213500e-01 -1.11464667e+00 -2.84161717e-01
-8.10084581e-01 5.14305770e-01 6.96125209e-01 6.14499152e-01
-1.32244170e+00 7.75020242e-01 -6.73476458e-01 -4.16581541e-01
3.55550468e-01 4.57241386e-01 -3.28702390e-01 3.31830978e-01
-8.77598345e-01 1.15540171e+00 7.45739937e-01 -1.50540397e-01
-1.08800066e+00 -7.95582235e-01 -9.28173065e-01 2.15138227e-01
3.79936069e-01 -7.38120079e-01 1.63421810e+00 -1.44647753e+00
-1.87227798e+00 1.13051939e+00 7.96706378e-02 -6.75643086e-01
5.55285692e-01 -3.61015946e-01 -1.36761278e-01 3.98207992e-01
-1.18957542e-01 1.12535262e+00 1.21011865e+00 -1.23244405e+00
-6.43813789e-01 2.10980684e-01 6.94851935e-01 -1.26298368e-01
-1.51852325e-01 3.53636980e-01 -2.34007612e-01 -7.43924677e-01
-6.13467455e-01 -7.20191002e-01 5.73871993e-02 5.32811463e-01
-2.36627564e-01 -5.76479360e-02 9.89747524e-01 -5.64352751e-01
9.99681830e-01 -2.17585230e+00 3.56533468e-01 1.37202501e-01
1.03242189e-01 4.25203085e-01 -1.03253774e-01 2.08394513e-01
-1.80498153e-01 2.85763424e-02 -3.52463156e-01 -1.67750970e-01
2.35634252e-01 2.34716088e-01 -3.90928626e-01 -1.56151384e-01
3.63835543e-01 9.06655967e-01 -8.90948594e-01 -6.71920955e-01
5.43365955e-01 2.02512637e-01 -6.46386743e-01 1.22065619e-01
-7.20290422e-01 -3.48833539e-02 -6.75237924e-02 5.65202236e-01
4.68152285e-01 -9.25856922e-03 1.05473489e-01 -3.24002415e-01
-1.35603547e-01 -2.95360893e-01 -9.94329691e-01 2.06657481e+00
-7.75737762e-01 7.17171729e-01 1.67107861e-02 -8.35759342e-01
7.52826333e-01 7.70101622e-02 -2.54064620e-01 -3.92013997e-01
-1.04935236e-01 -1.95800010e-02 -2.58568048e-01 -6.88966930e-01
5.95803440e-01 2.99617440e-01 -1.00238532e-01 5.40153444e-01
2.67634839e-01 -7.81040907e-01 4.69903946e-01 2.10744649e-01
1.11630380e+00 9.10129547e-01 4.76887733e-01 -1.57939881e-01
6.30233467e-01 3.14185470e-01 -6.95976168e-02 1.22898400e+00
-1.05746612e-01 2.93454111e-01 5.92236638e-01 -4.57650989e-01
-1.14497149e+00 -9.95902538e-01 4.96110350e-01 1.34578550e+00
-2.61324644e-02 -2.28666365e-01 -1.41656411e+00 -7.06943750e-01
-2.21562654e-01 1.06819844e+00 -5.35365880e-01 -2.39876986e-01
-4.59996611e-01 2.16276571e-01 7.88204193e-01 5.34545064e-01
7.47389853e-01 -1.63781524e+00 -1.23072207e+00 -6.70551695e-03
5.43130934e-01 -1.05446327e+00 -4.86381531e-01 1.23927742e-01
-8.02579105e-01 -9.29885983e-01 -4.65871185e-01 -1.15418756e+00
1.09926498e+00 -7.81060383e-02 1.01675797e+00 2.26064786e-01
-3.95877093e-01 5.00818968e-01 -1.80248559e-01 -3.43979001e-01
-8.44394147e-01 4.99469712e-02 -3.46529305e-01 -4.07769233e-01
1.04581818e-01 -5.29593587e-01 2.72863097e-02 2.73665898e-02
-1.09159279e+00 6.72280550e-01 5.98959684e-01 6.97691143e-01
3.01407009e-01 1.24293648e-01 4.50518876e-01 -1.15937471e+00
1.04818773e+00 3.10568392e-01 -1.13000131e+00 5.77005625e-01
-2.50031114e-01 3.89779121e-01 6.67162478e-01 -6.26293302e-01
-1.34564531e+00 5.87117732e-01 2.41028711e-01 -4.78240699e-01
-5.03289342e-01 5.30252516e-01 8.83995295e-02 -3.09021324e-01
8.42685878e-01 1.00658454e-01 7.73554668e-02 -1.70392618e-01
6.07430696e-01 5.00614166e-01 9.91125703e-01 -8.72394085e-01
7.80816793e-01 1.03340603e-01 4.38162545e-03 -4.84221250e-01
-2.60356665e-01 3.15255553e-01 -7.16155112e-01 -2.92740226e-01
8.61177146e-01 -6.24013245e-01 -8.20689380e-01 7.12481081e-01
-1.46651244e+00 -9.31524992e-01 -3.79160315e-01 2.70354990e-02
-9.88026023e-01 5.49714752e-02 -5.39207518e-01 -5.06519854e-01
-1.52079254e-01 -1.43637216e+00 1.00772142e+00 1.95028633e-01
-7.01675475e-01 -7.37035513e-01 -4.35210496e-01 -2.90967822e-01
4.75048006e-01 4.07996893e-01 1.09465170e+00 -3.05625051e-01
-7.77582943e-01 -5.28651588e-02 -3.58221710e-01 4.71434236e-01
3.15866992e-02 4.26042646e-01 -8.02793801e-01 2.76141524e-01
-3.32067668e-01 -4.23836619e-01 4.51581329e-01 2.52977032e-02
1.69451761e+00 -2.56629497e-01 -3.39464098e-01 5.92357457e-01
1.27154171e+00 4.34316605e-01 7.64513731e-01 5.59377551e-01
6.57464385e-01 4.81349438e-01 7.63198197e-01 2.45601572e-02
-4.27070484e-02 6.98269010e-01 2.99661040e-01 2.52600163e-02
-2.47420952e-01 -1.93107739e-01 5.09049296e-01 4.04681377e-02
-1.59504429e-01 2.93682426e-01 -6.72930717e-01 3.77826899e-01
-1.74580181e+00 -9.94684517e-01 1.99169815e-01 1.72479820e+00
1.35449052e+00 2.29655027e-01 -1.09146118e-01 -9.18026194e-02
7.76008308e-01 -2.48193592e-01 -4.98335540e-01 -1.08570540e+00
3.50162327e-01 6.22054517e-01 3.38269651e-01 7.55992711e-01
-1.00912273e+00 1.40746498e+00 6.89491272e+00 8.44575167e-01
-1.25088823e+00 -1.18752949e-01 -1.15887541e-02 3.89441773e-02
-1.67121273e-02 -1.68332368e-01 -4.48910862e-01 1.97470665e-01
7.51848638e-01 -1.90270379e-01 8.50828826e-01 1.18883300e+00
1.86855406e-01 -3.26355398e-01 -1.48317730e+00 6.68061495e-01
3.11041363e-02 -1.66164494e+00 1.91136822e-01 -6.38407946e-01
7.54052520e-01 -5.43606639e-01 -6.52087554e-02 2.73304790e-01
5.26164651e-01 -1.18866718e+00 1.00139272e+00 8.91793072e-01
1.21492898e+00 -5.13816595e-01 7.89873302e-03 4.23855931e-01
-5.51092744e-01 1.13304332e-01 -4.14912365e-02 -1.91413626e-01
-3.06763589e-01 2.22257078e-01 -1.07994878e+00 9.94304419e-02
6.92413211e-01 5.09492934e-01 -8.06125164e-01 8.10802758e-01
-6.35914743e-01 3.68492082e-02 -1.20964751e-01 -8.35075527e-02
-8.42334479e-02 2.98102517e-02 2.82206029e-01 1.45835292e+00
3.92267406e-01 -2.67916381e-01 -1.82105571e-01 1.34916520e+00
-2.47179456e-02 -3.91371608e-01 -7.94609308e-01 -2.73872703e-01
4.47514266e-01 1.18299711e+00 -7.79495895e-01 -8.37971151e-01
-3.14946026e-01 1.53624117e+00 3.04406703e-01 4.75591272e-01
-1.00437915e+00 -1.18125057e+00 4.00995821e-01 -1.91693962e-01
4.33201730e-01 -4.38893795e-01 -4.20890570e-01 -9.08241570e-01
-2.24603012e-01 -1.19422400e+00 -1.59215987e-01 -1.74342811e+00
-4.54930037e-01 5.76391637e-01 3.78723502e-01 -8.44391048e-01
-5.66466808e-01 -1.00767159e+00 -5.30070245e-01 8.94020021e-01
-1.03981590e+00 -1.21160686e+00 -5.70030689e-01 5.73764622e-01
5.50581694e-01 -2.45057911e-01 1.03611922e+00 8.72014090e-02
-7.17017129e-02 4.35437232e-01 -3.85412902e-01 -1.28030721e-02
6.12147033e-01 -1.44045281e+00 3.74739051e-01 8.46669555e-01
-1.64004281e-01 1.05670595e+00 8.80345464e-01 -6.78885937e-01
-1.11783075e+00 -1.10081947e+00 6.10648692e-01 -3.41563225e-01
5.16692996e-01 -4.22544181e-01 -7.49230385e-01 1.12172246e+00
5.78173041e-01 -1.34227321e-01 -4.39579338e-02 -5.50568700e-01
-4.92738128e-01 -3.75522152e-02 -1.38841116e+00 8.21931720e-01
1.15320206e+00 -9.06325996e-01 -8.95410955e-01 3.28217000e-01
9.46631253e-01 -8.66460383e-01 -6.78850472e-01 1.40357977e-02
6.50656521e-01 -9.73132968e-01 9.29665089e-01 -5.35185039e-01
8.26958895e-01 -7.57694423e-01 1.61789149e-01 -1.53922760e+00
1.22671410e-01 -5.93219697e-01 -8.22200775e-02 1.00656962e+00
1.74182326e-01 -6.72144115e-01 3.29874516e-01 7.29490221e-01
-2.97990441e-01 -2.28950411e-01 -3.06293607e-01 -8.02360952e-01
-2.15180710e-01 -4.62989956e-01 8.22584331e-01 6.30629480e-01
4.55012247e-02 -1.88993916e-01 -7.88149759e-02 1.70342147e-01
3.12802404e-01 4.00776006e-02 9.30698037e-01 -9.06710863e-01
-4.88200635e-01 -5.11534631e-01 -2.74449229e-01 -6.70386374e-01
5.58664560e-01 -8.53275657e-01 1.20610006e-01 -1.39322662e+00
-9.99868512e-02 -1.08953238e-01 8.83654580e-02 9.26504016e-01
2.74475902e-01 1.78147495e-01 2.68730283e-01 -9.48506519e-02
-4.17877614e-01 -1.32694915e-01 1.22113264e+00 -2.30634615e-01
-2.25261152e-01 -4.59861934e-01 -4.61946130e-01 9.62006330e-01
7.21403122e-01 -4.00714815e-01 -4.52242881e-01 -5.03846407e-01
2.30480626e-01 -1.34355858e-01 8.56173337e-01 -1.35031664e+00
3.76175135e-01 -4.44415212e-01 3.73970240e-01 -3.55832398e-01
7.75964186e-02 -1.05363703e+00 4.21842754e-01 6.27582014e-01
-6.97516024e-01 -1.10431738e-01 4.66499090e-01 -6.60272837e-02
-2.37610906e-01 -6.06586218e-01 6.88615382e-01 -4.17501062e-01
-1.32204044e+00 -4.20613319e-01 -5.39605200e-01 -4.45663124e-01
1.34059930e+00 -3.60722721e-01 -5.01683950e-01 -2.05521911e-01
-8.98665726e-01 -1.32281259e-01 8.33502889e-01 2.15799585e-01
5.49122572e-01 -1.12169075e+00 4.76252325e-02 3.70887607e-01
1.35092720e-01 -1.58322290e-01 -2.23491594e-01 3.73569459e-01
-1.15310991e+00 2.98303604e-01 -8.64806056e-01 -3.98952246e-01
-1.45654023e+00 7.85831809e-01 5.87366521e-01 6.51009902e-02
-4.90869105e-01 6.78160012e-01 2.24246368e-01 -5.90302289e-01
1.71388954e-01 -1.10840261e+00 1.30364656e-01 -3.96120727e-01
5.44135034e-01 5.81718311e-02 -1.25952020e-01 -2.05491595e-02
-9.14927572e-02 5.64605534e-01 6.83039054e-03 -2.53288031e-01
1.16336381e+00 1.55875087e-01 -4.31177229e-01 3.55741024e-01
8.28409314e-01 -1.46397114e-01 -1.41785514e+00 8.45604166e-02
-6.21474497e-02 -2.32499883e-01 -1.93800122e-01 -1.18890584e+00
-6.25193238e-01 8.35402906e-01 4.31528032e-01 1.94746077e-01
1.34093297e+00 -3.56354564e-01 1.35322079e-01 8.79505038e-01
4.46683764e-01 -1.23004472e+00 7.66099989e-02 6.45431697e-01
1.31598258e+00 -7.80565977e-01 -9.37663093e-02 -6.00832880e-01
-6.01072133e-01 1.45942855e+00 7.38967776e-01 -3.64543498e-01
1.89037755e-01 5.53337932e-01 -1.67698994e-01 -1.44808963e-01
-5.14208376e-01 1.45953774e-01 -3.93771380e-02 1.01240301e+00
2.22291172e-01 -8.77804756e-02 -6.00618571e-02 2.56502241e-01
-1.18978091e-01 9.61717606e-01 6.96588159e-01 1.27749801e+00
-4.09343064e-01 -1.04813635e+00 -3.98112774e-01 2.80008674e-01
-9.56125557e-02 -2.10040703e-01 -4.01089519e-01 1.00092483e+00
2.35783547e-01 3.11671108e-01 6.35063425e-02 -1.53661281e-01
2.72999257e-01 3.16898823e-01 1.00173175e+00 -8.27006519e-01
-8.19340289e-01 -5.53823829e-01 1.99115634e-01 -7.73962080e-01
-5.77551186e-01 -5.28399587e-01 -1.48564827e+00 -2.20754165e-02
3.57475072e-01 -4.22838420e-01 8.10786486e-01 8.67340982e-01
4.23105806e-01 8.46943855e-01 -2.01733172e-01 -1.04008126e+00
-3.61614376e-01 -7.70497143e-01 -3.59815747e-01 3.12651247e-01
3.70303303e-01 -4.99453485e-01 7.75454119e-02 7.96520710e-01] | [11.368924140930176, -0.25948551297187805] |
738aade1-cdbe-4e5e-8d8c-913f5a727f5a | quartile-based-seasonality-decomposition-for | 2306.05989 | null | https://arxiv.org/abs/2306.05989v1 | https://arxiv.org/pdf/2306.05989v1.pdf | Quartile-Based Seasonality Decomposition for Time Series Forecasting and Anomaly Detection | The timely detection of anomalies is essential in the telecom domain as it facilitates the identification and characterization of irregular patterns, abnormal behaviors, and network anomalies, contributing to enhanced service quality and operational efficiency. Precisely forecasting and eliminating predictable time series patterns constitutes a vital component of time series anomaly detection. While the state-of-the-art methods aim to maximize forecasting accuracy, the computational performance takes a hit. In a system composed of a large number of time series variables, e.g., cell Key Performance Indicators (KPIs), the time and space complexity of the forecasting employed is of crucial importance. Quartile-Based Seasonality Decomposition (QBSD) is a live forecasting method proposed in this paper to make an optimal trade-off between computational complexity and forecasting accuracy. This paper compares the performance of QBSD to the state-of-the-art forecasting methods and their applicability to practical anomaly detection. To demonstrate the efficacy of the proposed solution, experimental evaluation was conducted using publicly available datasets as well as a telecom KPI dataset. | ['Bulbul Singh', 'Ebenezer RHP Isaac'] | 2023-06-09 | null | null | null | null | ['time-series-anomaly-detection'] | ['time-series'] | [ 4.33795266e-02 -5.62267125e-01 2.81340271e-01 -2.19542757e-01
-2.99424171e-01 -4.23043698e-01 4.36759830e-01 6.22130692e-01
-9.13491622e-02 4.93563145e-01 -2.30274081e-01 -6.82386577e-01
-7.68223882e-01 -7.25524902e-01 -1.61360539e-02 -8.40071678e-01
-3.96841019e-01 4.21750993e-01 -6.38701841e-02 -2.64812022e-01
5.54157436e-01 8.39159191e-01 -1.55244541e+00 -1.14703901e-01
9.89775658e-01 1.65895140e+00 -4.45436060e-01 4.52369839e-01
-4.03583497e-01 4.11498457e-01 -8.18293035e-01 -3.14003080e-02
3.78371894e-01 -2.03853533e-01 -2.03454152e-01 1.74107507e-01
-3.64256442e-01 9.70165953e-02 -1.75341256e-02 8.07654083e-01
3.25068861e-01 2.74379790e-01 4.16599572e-01 -1.51102901e+00
1.14707023e-01 3.65090854e-02 -6.44045234e-01 9.59449589e-01
6.38398156e-02 -2.07264304e-01 8.00131500e-01 -6.28846765e-01
2.39447355e-02 4.50840414e-01 6.79821312e-01 -2.06591666e-01
-1.16674018e+00 -4.99181747e-01 2.48787552e-01 4.19627398e-01
-1.29775620e+00 -2.47650430e-01 9.84281421e-01 -5.51422656e-01
9.21887636e-01 4.16581929e-01 5.91285706e-01 2.68289506e-01
5.21087110e-01 3.66940379e-01 6.56449020e-01 -3.15804899e-01
3.35400730e-01 -2.18967050e-01 1.56636238e-01 -4.85914275e-02
2.35847637e-01 -8.26997682e-02 -2.67054290e-01 -4.86325353e-01
4.17019695e-01 5.98552406e-01 -1.62135124e-01 1.46977708e-01
-8.63536119e-01 4.71307606e-01 -1.93018600e-01 6.90420151e-01
-9.84631121e-01 -3.17786276e-01 6.06024683e-01 7.29165316e-01
7.13945270e-01 3.77840668e-01 -5.46684504e-01 -5.91707408e-01
-1.16402137e+00 1.85854301e-01 6.85329914e-01 4.81315047e-01
2.47138903e-01 5.20099699e-01 -1.11517392e-01 4.36707258e-01
-1.57383084e-01 4.02405113e-01 6.12887263e-01 -4.97336656e-01
4.17981774e-01 7.45461166e-01 2.28640243e-01 -1.32749748e+00
-5.71460366e-01 -9.61990714e-01 -1.07126188e+00 -4.67032306e-02
5.95988035e-01 -6.22179806e-02 -4.63305652e-01 1.21322703e+00
2.03416497e-01 2.30787501e-01 -6.30747601e-02 5.08868217e-01
-9.33755487e-02 8.50367308e-01 -3.41994643e-01 -8.30220997e-01
9.49873388e-01 -2.75026560e-01 -7.84411848e-01 2.45339006e-01
5.68984449e-01 -1.00687659e+00 5.27948141e-01 5.87282956e-01
-6.32848918e-01 -2.69427299e-01 -6.16197169e-01 8.61233294e-01
-2.83214241e-01 -1.00198291e-01 4.04364496e-01 4.48390901e-01
-5.88035405e-01 4.86745059e-01 -9.06850100e-01 -2.44047314e-01
-1.01679638e-01 1.46466091e-01 1.99130997e-02 2.67991543e-01
-9.76430118e-01 5.58418810e-01 2.44929209e-01 4.12279725e-01
-2.04473987e-01 -7.56229758e-01 -2.37292290e-01 4.20086533e-01
3.63451451e-01 -4.25113626e-02 1.10328722e+00 -9.38421667e-01
-1.01416302e+00 2.00348005e-01 -2.57541388e-01 -6.80130363e-01
5.06523550e-01 6.48862273e-02 -1.16660869e+00 -1.32569298e-03
-2.17134506e-01 -8.40397239e-01 9.66623247e-01 -5.48232377e-01
-8.68840039e-01 -4.90042150e-01 -4.49543208e-01 -2.73116082e-01
-3.98717970e-01 9.02288705e-02 1.69252872e-01 -9.22544241e-01
6.11740172e-01 -7.37318814e-01 -2.29218781e-01 -7.34426200e-01
-5.85511513e-02 -1.34438068e-01 1.02253544e+00 -9.01940346e-01
1.71654963e+00 -2.17698622e+00 -3.57940197e-01 9.44660008e-01
-1.03797214e-02 1.41086474e-01 3.77707124e-01 8.72126877e-01
-2.30921537e-01 -3.68082851e-01 -1.13972463e-01 1.64344721e-02
-1.95412427e-01 2.56255746e-01 -6.55614078e-01 6.79763854e-01
9.35794860e-02 3.86862457e-01 -4.06072468e-01 6.62734583e-02
3.25997591e-01 1.92978233e-01 -2.44066671e-01 1.80923179e-01
2.14864109e-02 6.70458317e-01 -4.61688757e-01 8.14346671e-01
3.79988134e-01 -1.58049136e-01 -1.09385036e-01 1.25957638e-01
-4.48321879e-01 -3.29673104e-02 -1.25131559e+00 7.67758787e-01
-5.93614638e-01 6.61692739e-01 -1.75554156e-01 -1.44433427e+00
1.31407750e+00 5.34124732e-01 1.11029017e+00 -9.42046762e-01
3.71478796e-02 4.76086289e-01 1.04050472e-01 -2.27514520e-01
4.90192741e-01 1.72253102e-01 3.91332209e-02 5.36695659e-01
-5.29943109e-01 4.11628664e-01 4.06080246e-01 -1.56025127e-01
1.12373054e+00 -4.92274195e-01 5.26993454e-01 -3.49724144e-01
8.26168001e-01 -1.50684893e-01 8.24224472e-01 2.74266422e-01
-6.82910234e-02 2.14519277e-02 7.87827492e-01 -7.55601704e-01
-9.34543788e-01 -7.03448057e-01 -1.15794063e-01 6.99524581e-01
-3.67633581e-01 -2.02545673e-01 -1.57473966e-01 -1.18240170e-01
1.55722648e-01 7.96914816e-01 -3.73409569e-01 1.07211933e-01
-5.63958406e-01 -9.01777983e-01 -6.32002428e-02 1.91859961e-01
1.14754446e-01 -7.63725162e-01 -7.68195808e-01 6.70589149e-01
2.22003646e-02 -9.83799338e-01 -1.26182288e-01 4.56406511e-02
-1.08922219e+00 -9.25380349e-01 -5.65888107e-01 -1.33365661e-01
8.30523193e-01 1.24416672e-01 1.05952859e+00 5.42423222e-03
-1.69880569e-01 2.81675458e-01 -4.88780767e-01 -6.25544786e-01
-3.08999598e-01 4.26632129e-02 2.06091657e-01 6.03722930e-01
3.47488791e-01 -9.95933115e-01 -7.20027626e-01 6.15765452e-01
-1.02919221e+00 -5.34483373e-01 3.72931331e-01 6.70881510e-01
7.32181728e-01 6.84654295e-01 8.69091868e-01 -6.83655381e-01
8.77109885e-01 -5.76815665e-01 -1.21192479e+00 1.74371243e-01
-1.18285942e+00 -2.18540981e-01 1.03983712e+00 -1.36749208e-01
-7.54805148e-01 -2.77404517e-01 1.09118395e-01 -2.35404700e-01
-3.07293120e-03 9.38079059e-01 3.22694927e-01 6.00311570e-02
1.89544603e-01 5.43215394e-01 -1.18702222e-02 -6.22910500e-01
-3.58610481e-01 6.34633839e-01 4.55055118e-01 -2.57722676e-01
5.27135491e-01 4.14419770e-01 3.23013246e-01 -1.03640580e+00
-4.71280605e-01 -1.07936883e+00 -3.57870281e-01 -3.77741367e-01
-3.25991102e-02 -4.20538843e-01 -8.28276038e-01 5.81063688e-01
-8.10225964e-01 3.94541293e-01 -3.04622382e-01 3.59571964e-01
-1.73555657e-01 2.46991590e-01 -2.69524306e-01 -1.15584683e+00
-3.78441036e-01 -7.19290018e-01 3.91332775e-01 4.97804284e-02
-3.42953950e-01 -1.00008214e+00 1.17619686e-01 -1.66762576e-01
8.39060903e-01 6.40147328e-01 9.50524151e-01 -1.06452525e+00
-2.33688235e-01 -8.73093605e-01 -8.05598721e-02 2.67389923e-01
2.56279767e-01 3.43190372e-01 -5.81749439e-01 -4.18359756e-01
3.10086399e-01 8.15466106e-01 3.49606544e-01 5.10124683e-01
1.27484107e+00 -4.68087286e-01 -4.22987826e-02 4.73599166e-01
1.33124971e+00 6.51453435e-01 5.36450505e-01 5.14665723e-01
2.17563331e-01 6.55561268e-01 7.73471832e-01 1.10495901e+00
1.04745351e-01 4.80267406e-01 5.62724769e-01 1.18201911e-01
8.74765337e-01 3.02266717e-01 -3.83788794e-02 1.15235591e+00
-2.42311358e-01 -1.10010549e-01 -1.23544741e+00 6.92136645e-01
-1.87177491e+00 -1.02897799e+00 -1.50799572e-01 2.34709811e+00
2.53992565e-02 4.31949228e-01 3.03010225e-01 1.03860652e+00
5.24515808e-01 8.19766000e-02 -5.07447779e-01 -5.18087447e-01
-4.80061844e-02 -6.04222268e-02 4.14636105e-01 2.95156650e-02
-8.29061389e-01 1.34863025e-02 5.08603859e+00 6.37040496e-01
-1.41303957e+00 -2.24040970e-01 6.06908500e-01 -6.10141903e-02
-1.16975173e-01 -1.81682780e-01 -4.70316142e-01 7.83747792e-01
1.30813932e+00 -4.08727199e-01 3.62143993e-01 8.99238288e-01
5.88143706e-01 -1.95626672e-02 -6.50995255e-01 1.15917695e+00
-2.63972551e-01 -1.08573759e+00 -3.88979405e-01 2.08476469e-01
7.28818238e-01 -1.08343825e-01 -5.81971779e-02 -3.39989588e-02
-6.09805107e-01 -7.74995565e-01 3.27213347e-01 7.22503245e-01
2.31033295e-01 -1.13162255e+00 1.16768992e+00 1.98338896e-01
-1.21104872e+00 -4.95354742e-01 1.07709542e-01 -2.53853053e-01
4.18867290e-01 1.27663362e+00 -4.31521416e-01 6.03084445e-01
7.18113601e-01 5.67136228e-01 -8.46664160e-02 1.19496477e+00
2.87718982e-01 8.10040414e-01 -5.83471239e-01 1.34938791e-01
2.10278958e-01 -6.04887784e-01 7.03338027e-01 8.76065969e-01
7.49488652e-01 8.33408907e-02 5.39196394e-02 1.40154824e-01
3.73346716e-01 3.89211804e-01 -2.21341133e-01 -3.26847672e-01
4.23604608e-01 1.05808794e+00 -8.11342597e-01 -7.20060617e-02
-5.17465532e-01 4.53058004e-01 -4.39590991e-01 4.61886406e-01
-4.66115534e-01 -5.37709832e-01 8.28815460e-01 4.12506282e-01
2.37212092e-01 -5.01962841e-01 -5.74589491e-01 -9.48280573e-01
3.49835247e-01 -8.08852077e-01 6.59424067e-01 1.40252076e-02
-1.27387774e+00 7.69351661e-01 -3.49040508e-01 -1.87371612e+00
-6.95238411e-01 -2.97360212e-01 -7.77370036e-01 9.32097077e-01
-1.46515381e+00 -4.85101759e-01 -3.66011083e-01 6.14613056e-01
3.90585601e-01 -4.49607700e-01 7.24485219e-01 5.60099900e-01
-7.23026454e-01 2.74396956e-01 8.05177689e-01 -1.99267000e-01
1.79673105e-01 -9.00038183e-01 4.75285053e-01 1.19933653e+00
-3.28697190e-02 1.64200872e-01 1.03495979e+00 -3.26451212e-01
-1.11239493e+00 -7.10078418e-01 9.93520260e-01 7.94898942e-02
9.10673738e-01 1.92144111e-01 -1.09334397e+00 2.65144020e-01
-3.43908995e-01 5.97697683e-02 7.06034124e-01 6.74048066e-02
1.78373888e-01 -7.01229215e-01 -9.92115319e-01 3.15719664e-01
3.12310576e-01 -3.43983084e-01 -2.40060478e-01 7.75181577e-02
-1.48595162e-02 -2.04556331e-01 -9.59219515e-01 3.81177664e-01
5.13274968e-01 -1.02840543e+00 5.78022659e-01 -3.47616792e-01
-1.16516471e-01 -3.38340938e-01 -1.81719780e-01 -1.44110763e+00
-4.92382497e-02 -9.33072686e-01 -4.38014746e-01 1.11240661e+00
2.15539604e-01 -1.12886810e+00 6.12010121e-01 5.58766842e-01
1.51634499e-01 -8.25526893e-01 -1.20574510e+00 -9.50436175e-01
-6.78511441e-01 -7.00907469e-01 1.00838661e+00 9.04410481e-01
-1.39311865e-01 -2.62017876e-01 -2.56953806e-01 2.78032631e-01
5.89537024e-01 5.63714206e-01 6.34986818e-01 -1.64472580e+00
-1.14557363e-01 -4.38063473e-01 -7.81329930e-01 -3.59395057e-01
-2.88291693e-01 -2.61450410e-01 -3.62249345e-01 -1.00241959e+00
-7.29179978e-01 -4.21011299e-01 -8.91012430e-01 7.87758604e-02
2.19085947e-01 -9.56992358e-02 -2.71308441e-02 5.39163351e-01
-1.30021885e-01 4.62076455e-01 4.83534336e-01 2.52287328e-01
-5.20949483e-01 7.15568006e-01 -2.59837240e-01 5.36047697e-01
1.10090184e+00 -4.13400114e-01 -4.19508815e-01 1.68629438e-02
4.43397999e-01 3.55766714e-01 6.51798621e-02 -9.65480924e-01
3.04513484e-01 -2.09504396e-01 1.03528172e-01 -8.91378641e-01
1.80631727e-02 -1.13883030e+00 3.22168618e-01 4.99160737e-01
1.35619879e-01 8.49124968e-01 8.34207311e-02 6.38664722e-01
-8.18868637e-01 2.44844601e-01 3.81151944e-01 3.86216283e-01
-6.98908567e-01 3.07077765e-01 -3.69621307e-01 -1.10533752e-01
1.09412515e+00 -3.05121839e-01 -7.67960325e-02 -5.78152597e-01
-5.65906882e-01 2.57407725e-01 5.58260977e-02 3.31934035e-01
4.24234331e-01 -9.90610182e-01 -6.08653843e-01 4.95394349e-01
1.67595372e-01 -2.99181551e-01 3.05569828e-01 1.25970232e+00
-5.80383599e-01 6.96027040e-01 -1.41346216e-01 -6.79534316e-01
-1.05764246e+00 2.91872352e-01 1.87842861e-01 -3.98541123e-01
-5.00931144e-01 1.97865650e-01 -3.98649991e-01 1.36416525e-01
1.86834469e-01 -3.98566604e-01 -1.70687750e-01 3.89485896e-01
5.86553514e-01 9.10615861e-01 7.47459531e-01 -3.67790341e-01
-4.58477616e-01 3.19646329e-01 -5.44558167e-02 2.10436121e-01
1.40775061e+00 -2.90245086e-01 -3.33358616e-01 4.98571336e-01
8.67934763e-01 -5.56774959e-02 -9.40416932e-01 -3.60367537e-01
5.73086083e-01 -5.64610243e-01 1.77779406e-01 -3.27392638e-01
-1.07004833e+00 5.85083365e-01 5.49066842e-01 9.98823225e-01
1.56449091e+00 -5.41149139e-01 9.89720881e-01 1.45497113e-01
2.22871765e-01 -1.19042051e+00 -4.03416961e-01 4.68755573e-01
6.17348254e-01 -1.07311141e+00 -2.03379676e-01 -1.57184690e-01
-4.42369163e-01 1.14273095e+00 8.95202681e-02 -8.63279626e-02
1.03595197e+00 2.30700389e-01 2.36093298e-01 -4.48145904e-03
-7.37257004e-01 -5.59045747e-02 3.15540940e-01 5.75422011e-02
2.81567544e-01 1.26012966e-01 -5.92775643e-01 5.34964383e-01
-2.77989537e-01 -1.73133820e-01 2.64026433e-01 8.96751881e-01
-4.62401472e-02 -7.54237473e-01 -4.94892091e-01 8.68751049e-01
-8.57558966e-01 1.84246480e-01 1.93648070e-01 5.44163465e-01
-3.47528011e-01 1.09932995e+00 4.16468263e-01 -1.39618516e-01
5.88535428e-01 1.49190485e-01 -3.50114733e-01 1.05547085e-01
-5.28048813e-01 2.36549061e-02 -2.35758469e-01 -5.86702347e-01
-1.92530677e-01 -7.61934817e-01 -1.01107466e+00 -6.30787790e-01
-1.59754366e-01 4.65908557e-01 9.49403584e-01 1.05752802e+00
4.89539027e-01 6.74160480e-01 1.17519987e+00 -3.07842851e-01
-6.59223616e-01 -7.35190392e-01 -9.05841351e-01 4.18102205e-01
4.63392496e-01 -3.96986276e-01 -5.33154488e-01 -2.67864913e-01] | [7.208133220672607, 2.7911462783813477] |
da293f86-95d5-477e-ad3d-20a7b1625c27 | unsupervised-classification-in-hyperspectral | 1604.08182 | null | http://arxiv.org/abs/1604.08182v2 | http://arxiv.org/pdf/1604.08182v2.pdf | Unsupervised Classification in Hyperspectral Imagery with Nonlocal Total Variation and Primal-Dual Hybrid Gradient Algorithm | In this paper, a graph-based nonlocal total variation method (NLTV) is
proposed for unsupervised classification of hyperspectral images (HSI). The
variational problem is solved by the primal-dual hybrid gradient (PDHG)
algorithm. By squaring the labeling function and using a stable simplex
clustering routine, an unsupervised clustering method with random
initialization can be implemented. The effectiveness of this proposed algorithm
is illustrated on both synthetic and real-world HSI, and numerical results show
that the proposed algorithm outperforms other standard unsupervised clustering
methods such as spherical K-means, nonnegative matrix factorization (NMF), and
the graph-based Merriman-Bence-Osher (MBO) scheme. | ['Andrea L. Bertozzi', 'Wei Zhu', 'Da Kuang', 'Dominique Zosso', 'Devin Dahlberg', 'Alexandre Tiard', 'Victoria Chayes', 'Stephanie Sanchez', 'Stanley Osher'] | 2016-04-27 | null | null | null | null | ['classification-of-hyperspectral-images'] | ['computer-vision'] | [ 3.22710574e-01 -3.58453244e-01 -1.09621152e-01 -7.55110160e-02
-6.72871530e-01 -6.14059925e-01 3.46323609e-01 -2.78915673e-01
-2.32482314e-01 9.05699670e-01 -2.78378785e-01 -2.67388463e-01
-5.91993690e-01 -6.29329085e-01 1.11371102e-02 -1.45922315e+00
-5.76509628e-04 4.60807323e-01 -1.92563936e-01 -1.18858136e-01
3.30369502e-01 6.70337021e-01 -1.26094759e+00 -4.26871806e-01
1.26293743e+00 8.43246222e-01 1.70472443e-01 6.45912468e-01
1.11241393e-01 4.81766224e-01 -1.34014636e-01 -2.44735479e-02
4.75173295e-01 -6.21641874e-01 -5.99211991e-01 7.42494047e-01
3.21964435e-02 4.48942222e-02 -3.01833861e-02 1.37059045e+00
4.90635157e-01 8.38313818e-01 1.19309509e+00 -1.41726005e+00
-5.90680420e-01 3.19496542e-01 -1.26956761e+00 -2.86853969e-01
-1.09950572e-01 -1.46689638e-01 9.00219262e-01 -1.15982509e+00
4.86270875e-01 1.07298613e+00 6.18983626e-01 -6.53284565e-02
-1.29740739e+00 -2.79082775e-01 -2.43930086e-01 2.50317454e-01
-1.80171347e+00 -6.93942606e-02 1.02069604e+00 -5.29225111e-01
6.39320970e-01 5.50032794e-01 5.34005880e-01 1.04775220e-01
-1.95979118e-01 6.82764888e-01 1.32153988e+00 -7.14058220e-01
5.39103627e-01 -3.25159878e-01 3.38949561e-01 8.35781455e-01
3.40656430e-01 -1.00525796e-01 -8.08203667e-02 -5.65807819e-01
3.78996223e-01 -3.27066556e-02 -2.98931867e-01 -6.91793084e-01
-1.09920669e+00 1.21627033e+00 4.32621121e-01 1.64800256e-01
-4.99908119e-01 -1.79758564e-01 1.11009113e-01 -1.19708680e-01
6.58490896e-01 5.37805036e-02 8.53468031e-02 5.46412230e-01
-1.35536838e+00 -2.50893712e-01 5.72643816e-01 7.08541036e-01
1.24473441e+00 3.94789189e-01 1.25406712e-01 9.06579196e-01
7.66465425e-01 1.01607525e+00 3.64995360e-01 -1.08298790e+00
-1.05665244e-01 5.86392641e-01 1.05205387e-01 -1.57134831e+00
-5.07205367e-01 -5.07171787e-02 -1.26115453e+00 1.26563251e-01
-3.56989242e-02 -2.61610478e-01 -9.94376540e-01 1.06423819e+00
5.47480702e-01 3.76108140e-01 1.39505208e-01 1.08791077e+00
8.11314821e-01 8.51703227e-01 -2.11027756e-01 -9.55443919e-01
5.78250408e-01 -9.68966842e-01 -9.82711494e-01 1.94756448e-01
3.10450494e-01 -6.48653090e-01 5.87967455e-01 5.30647755e-01
-6.92577541e-01 -3.49920809e-01 -9.68042493e-01 1.80503964e-01
-3.23855788e-01 3.54051918e-01 7.15322733e-01 9.94543612e-01
-1.15616143e+00 2.03890905e-01 -9.84850943e-01 -4.51561242e-01
-3.11083421e-02 2.78703988e-01 -2.23283052e-01 -1.79486409e-01
-6.63816214e-01 2.03883156e-01 6.24480665e-01 5.40957212e-01
-6.87279046e-01 -3.89504373e-01 -7.88560927e-01 -3.63294721e-01
4.13925052e-01 -1.86449558e-01 2.84234494e-01 -9.11702991e-01
-1.70086718e+00 8.02711487e-01 -6.78199649e-01 8.76072943e-02
1.05344735e-01 5.38446426e-01 -2.42116913e-01 5.87858140e-01
1.44757673e-01 2.13950142e-01 1.04639578e+00 -1.75316536e+00
-1.79004803e-01 -5.29289067e-01 -4.26129609e-01 4.19202983e-01
-2.42862701e-01 -2.69466579e-01 -2.42457241e-01 -5.01049638e-01
9.63821411e-01 -1.08934200e+00 -5.85802972e-01 -3.85875642e-01
-7.77912736e-01 -4.39478606e-02 1.10993028e+00 -8.62514734e-01
1.16064477e+00 -2.15218520e+00 5.94229519e-01 9.63823140e-01
1.40953645e-01 8.62808302e-02 1.36510804e-01 4.66078192e-01
-1.75756231e-01 3.34218815e-02 -9.70819592e-01 -1.88528046e-01
-5.95918559e-02 2.76093543e-01 2.58585900e-01 9.27147269e-01
-5.24818122e-01 6.18477702e-01 -1.14421725e+00 -6.14394844e-01
5.46468079e-01 3.72143090e-01 -1.07455462e-01 -7.24735111e-02
1.45857677e-01 3.50004643e-01 -4.88911942e-02 8.20464790e-01
1.22613764e+00 -2.80593872e-01 5.23015261e-01 -5.13920426e-01
-4.07067537e-01 -8.94799769e-01 -1.67837155e+00 1.31970096e+00
2.52540588e-01 3.69071811e-01 5.05528986e-01 -1.35297585e+00
7.82347322e-01 1.18464574e-01 9.46611464e-01 1.11032188e-01
7.38081634e-02 9.94749442e-02 -6.02383375e-01 -3.37435871e-01
4.81433302e-01 -1.51004717e-01 5.29137373e-01 3.81525099e-01
-5.46955876e-02 -3.44244301e-01 5.78205645e-01 5.25217235e-01
4.04789418e-01 -1.83146805e-01 3.81731451e-01 -7.60547876e-01
7.96990693e-01 4.13437694e-01 3.04785281e-01 5.14403403e-01
-2.99661309e-01 4.93529826e-01 2.33452648e-01 1.86112270e-01
-6.13973796e-01 -9.60565746e-01 -3.13224345e-01 7.45800734e-01
1.54144064e-01 -1.52887791e-01 -7.67634094e-01 -2.09240422e-01
-1.60945281e-01 5.71449637e-01 -2.47936919e-01 2.90228665e-01
2.54942417e-01 -1.64508677e+00 1.54058501e-01 -1.93232268e-01
7.34546304e-01 -4.36074287e-01 2.83630520e-01 7.90135935e-02
-4.92264777e-01 -8.74136865e-01 -1.23584501e-01 3.33721712e-02
-9.10900593e-01 -1.12621272e+00 -6.63929403e-01 -8.38006198e-01
1.06315470e+00 7.52863646e-01 5.70405304e-01 -1.27803594e-01
-3.32791805e-01 7.78784156e-01 -4.39428300e-01 2.01625764e-01
-1.40428722e-01 -3.35608304e-01 2.24812388e-01 5.66354871e-01
3.04401010e-01 -3.39686722e-01 -5.79078078e-01 3.79334450e-01
-1.09709084e+00 -2.50769585e-01 1.25979170e-01 1.09324348e+00
1.14534938e+00 8.17081451e-01 -1.79422554e-02 -8.09747696e-01
6.19537354e-01 -4.00755525e-01 -9.47835624e-01 4.22513753e-01
-1.02780151e+00 -5.32166541e-01 3.18876386e-01 -8.87416825e-02
-1.15337610e+00 5.42955756e-01 5.36688030e-01 -6.88585937e-01
1.06748447e-01 9.76167083e-01 -6.91009983e-02 -7.51728237e-01
7.82696128e-01 4.48133767e-01 1.69490188e-01 2.11405344e-02
6.71229362e-01 8.08624566e-01 4.84050065e-01 -3.89703959e-01
1.09148347e+00 9.17355716e-01 4.13050830e-01 -1.49736547e+00
-6.00767195e-01 -1.00379539e+00 -9.19788122e-01 -5.58125973e-01
1.15568566e+00 -8.48148942e-01 -7.33431458e-01 6.21062875e-01
-4.94100571e-01 -3.29902142e-01 3.29865143e-02 6.59770906e-01
-4.70239639e-01 1.01140678e+00 -3.91271472e-01 -1.22258961e+00
-3.33946347e-02 -1.24755538e+00 7.56947517e-01 1.18143991e-01
5.28839409e-01 -1.34198511e+00 1.09170720e-01 5.15306890e-01
9.79727060e-02 6.50705040e-01 6.04994237e-01 4.32900824e-02
-3.86017859e-01 7.26320818e-02 -4.08230960e-01 5.52577376e-01
3.41751844e-01 5.34255028e-01 -6.65138781e-01 -4.72450376e-01
2.02674955e-01 -8.11429322e-02 8.20163488e-01 9.77454960e-01
7.99575508e-01 -1.59217760e-01 -1.94581687e-01 1.06605315e+00
2.21524167e+00 8.72194245e-02 5.36897898e-01 6.32523522e-02
1.23141444e+00 3.52154791e-01 5.12923300e-01 6.59085691e-01
2.78326452e-01 -4.20011543e-02 4.37828809e-01 -3.69920075e-01
4.42060590e-01 3.46522540e-01 4.81815711e-02 1.05591249e+00
-4.75862533e-01 -1.19744971e-01 -1.01169384e+00 4.40930247e-01
-2.07882833e+00 -1.08384240e+00 -8.42089891e-01 1.92566025e+00
3.79344106e-01 -6.45688772e-01 -1.63653065e-02 3.99453133e-01
1.08519423e+00 2.91125089e-01 -5.09875059e-01 -1.39686823e-01
-4.51621801e-01 -9.76543501e-02 9.88690078e-01 9.08887506e-01
-1.29265797e+00 9.47821856e-01 6.82598162e+00 1.00601220e+00
-9.15888906e-01 1.09943308e-01 5.34696102e-01 4.08493161e-01
-1.13502368e-01 9.52233151e-02 -2.80238956e-01 1.47016466e-01
3.43526155e-01 -2.02565908e-01 1.11913085e+00 5.95302522e-01
5.44235766e-01 -4.07305598e-01 -3.09469011e-02 1.26383424e+00
4.00669008e-01 -9.31577742e-01 -3.92010659e-02 -1.95227340e-02
1.37950647e+00 -4.30343486e-03 -6.73662696e-04 -2.20638290e-01
3.72497231e-01 -6.02259040e-01 -5.35055250e-03 4.94119257e-01
7.11588860e-01 -9.75902259e-01 4.11220700e-01 6.48741648e-02
-1.45072734e+00 -2.69229896e-02 -6.88298523e-01 2.70932347e-01
4.15737703e-02 1.16046464e+00 -3.43159258e-01 9.36690211e-01
5.11520386e-01 8.24625313e-01 -4.36803311e-01 8.39112818e-01
-1.81567490e-01 7.76812494e-01 -3.94316167e-01 1.34795785e-01
5.79148531e-01 -1.46866858e+00 8.33648562e-01 9.51000333e-01
2.52386242e-01 5.95492721e-01 4.45355356e-01 6.43271565e-01
2.85333335e-01 4.33745325e-01 -4.76487279e-01 -4.26018238e-01
3.42181846e-02 1.62453461e+00 -1.28532207e+00 -2.42891833e-01
-2.64745116e-01 9.98844445e-01 -2.37857983e-01 9.26952839e-01
-5.46424806e-01 -3.30924183e-01 1.06317289e-01 -4.25926954e-01
2.85584211e-01 -5.45852423e-01 -2.83614516e-01 -1.23154616e+00
-4.28384602e-01 -6.41384780e-01 5.91887236e-01 -9.75717664e-01
-1.24756145e+00 2.38849372e-01 6.80204704e-02 -1.10089540e+00
1.07093200e-01 -7.01061368e-01 -4.45066452e-01 7.09931314e-01
-1.47424746e+00 -1.11823487e+00 -5.06251872e-01 1.06035125e+00
1.93170328e-02 -2.60462612e-01 6.28359139e-01 6.38153553e-02
-7.86352277e-01 9.73661989e-02 9.64186311e-01 -8.59906804e-03
3.60890001e-01 -1.45107317e+00 -7.28970885e-01 1.22285366e+00
-2.45633826e-01 3.67276341e-01 6.13703132e-01 -7.44489431e-01
-1.64927447e+00 -1.17731345e+00 2.18742162e-01 1.55225381e-01
8.52034211e-01 2.16599569e-01 -5.58838427e-01 6.24620378e-01
2.58763164e-01 -7.14384615e-02 1.03799844e+00 -2.25973338e-01
3.43417563e-02 -9.39336494e-02 -1.40868282e+00 4.43688780e-01
5.99386632e-01 -4.35601205e-01 6.25132397e-02 1.17066717e+00
2.39108577e-01 -5.16998433e-02 -9.54825342e-01 3.91260594e-01
-1.61075965e-02 -9.47877705e-01 8.49859834e-01 -9.08205062e-02
-1.60307344e-02 -8.03069949e-01 -5.06876111e-01 -1.37002885e+00
-6.44454062e-01 -8.72355819e-01 2.01475725e-01 9.76920068e-01
2.99257696e-01 -6.25049293e-01 6.99525595e-01 3.59676182e-01
4.03513992e-03 -2.49105975e-01 -6.72092617e-01 -6.94557905e-01
-3.60527545e-01 -2.42965803e-01 1.41204104e-01 1.48085320e+00
1.29597262e-01 1.25468329e-01 -3.39804798e-01 6.15789711e-01
1.46860647e+00 1.35288775e-01 4.25326258e-01 -1.23504066e+00
-3.50440480e-02 -2.76583999e-01 -4.29816753e-01 -4.42007840e-01
4.19288367e-01 -1.09589708e+00 -7.07748858e-03 -1.84202158e+00
2.09910139e-01 -1.01553738e-01 -2.40851402e-01 2.64996469e-01
-1.83166608e-01 5.27572751e-01 -6.59083501e-02 3.84798914e-01
-4.98385936e-01 3.92153740e-01 1.11582279e+00 -3.79350573e-01
-4.99754995e-01 -2.67031789e-01 -2.85097539e-01 5.75545728e-01
7.49744713e-01 -2.78712392e-01 -5.58206677e-01 4.84815389e-02
8.30714852e-02 1.58066884e-01 3.50341834e-02 -8.04156065e-01
3.29824179e-01 -4.39516604e-01 2.88307607e-01 -8.92648220e-01
1.40806079e-01 -7.60903537e-01 2.97953725e-01 2.17744946e-01
6.30475953e-02 -2.58087367e-01 -2.35558152e-01 7.13786542e-01
-2.64187008e-01 -3.59959871e-01 1.06302059e+00 -1.46151245e-01
-6.63760543e-01 4.87544596e-01 -5.16181886e-01 -3.59739631e-01
1.18194151e+00 -2.42117941e-01 -8.29021409e-02 -3.43613088e-01
-9.38888669e-01 2.85541683e-01 3.70677233e-01 -5.86317837e-01
7.31304765e-01 -1.49866354e+00 -8.21042955e-01 2.37951428e-01
-4.10741456e-02 -7.29520619e-02 3.33859593e-01 1.18888855e+00
-9.04636681e-01 5.67394011e-02 1.67434946e-01 -9.23862219e-01
-9.94463563e-01 5.91814756e-01 3.58766764e-01 8.04334581e-02
-1.23958357e-01 7.81897545e-01 -1.32404581e-01 -7.40872979e-01
-1.33917145e-02 2.84598380e-01 -2.97709763e-01 3.82946193e-01
2.51276106e-01 1.06159902e+00 -1.19942360e-01 -1.18531919e+00
-1.92534700e-01 6.65082097e-01 6.54993892e-01 -3.75956625e-01
1.22220612e+00 -3.88148844e-01 -1.10641325e+00 4.25962150e-01
1.35704148e+00 -1.37891695e-01 -7.88440704e-01 -2.43873522e-02
-3.48697901e-01 -4.04562116e-01 6.52842164e-01 -2.33386964e-01
-1.31884336e+00 4.99510109e-01 6.93767250e-01 5.76835990e-01
1.39999211e+00 -6.91053152e-01 2.19008371e-01 6.34299219e-01
2.76497692e-01 -1.37513959e+00 -3.27476621e-01 3.11983198e-01
5.72646677e-01 -1.34927392e+00 4.85097289e-01 -6.12695932e-01
-7.19279647e-01 1.12908101e+00 1.91803217e-01 -1.26293451e-01
1.24665153e+00 -2.03711241e-01 2.28866801e-01 -2.27473482e-01
2.67506093e-02 -5.95449865e-01 2.88841993e-01 7.37458050e-01
9.33304131e-02 4.06583637e-01 -5.39877892e-01 -2.30659559e-01
1.96048439e-01 -3.39706182e-01 6.20977819e-01 7.67155886e-01
-5.06883144e-01 -6.04443192e-01 -9.10427570e-01 5.98145723e-01
-3.05963550e-02 -6.80926591e-02 -4.99860495e-01 5.17574847e-01
-2.70447910e-01 1.42688656e+00 -4.03016925e-01 -1.85013324e-01
-2.99479395e-01 -5.28780892e-02 2.69538194e-01 -3.31066370e-01
8.75023380e-02 6.36301577e-01 -4.48672086e-01 -4.11887348e-01
-1.23623466e+00 -6.54607594e-01 -1.25834727e+00 -2.53004491e-01
-8.19172919e-01 5.39137363e-01 8.06460440e-01 8.32744539e-01
-1.54304817e-01 3.06937303e-02 1.10509217e+00 -1.19207180e+00
-1.36430457e-01 -7.14297295e-01 -1.33558774e+00 9.26883891e-02
2.70388350e-02 -6.15651190e-01 -1.20009077e+00 3.42895716e-01] | [10.074843406677246, -2.009645462036133] |
488ec3e8-80f5-4648-be1e-a483a7e2e042 | salesforce-causalai-library-a-fast-and | 2301.10859 | null | https://arxiv.org/abs/2301.10859v1 | https://arxiv.org/pdf/2301.10859v1.pdf | Salesforce CausalAI Library: A Fast and Scalable Framework for Causal Analysis of Time Series and Tabular Data | We introduce the Salesforce CausalAI Library, an open-source library for causal analysis using observational data. It supports causal discovery and causal inference for tabular and time series data, of both discrete and continuous types. This library includes algorithms that handle linear and non-linear causal relationships between variables, and uses multi-processing for speed-up. We also include a data generator capable of generating synthetic data with specified structural equation model for both the aforementioned data formats and types, that helps users control the ground-truth causal process while investigating various algorithms. Finally, we provide a user interface (UI) that allows users to perform causal analysis on data without coding. The goal of this library is to provide a fast and flexible solution for a variety of problems in the domain of causality. This technical report describes the Salesforce CausalAI API along with its capabilities, the implementations of the supported algorithms, and experiments demonstrating their performance and speed. Our library is available at \url{https://github.com/salesforce/causalai}. | ['Juan Carlos Niebles', 'Kun Zhang', 'Caiming Xiong', 'Stephen Hoi', 'Huan Wang', 'Eric Hu', 'Shelby Heinecke', 'Paul Josel', 'Wenzhuo Yang', 'Weiran Yao', 'Chenghao Liu', 'Matthew Fernandez', 'Devansh Arpit'] | 2023-01-25 | null | null | null | null | ['causal-discovery'] | ['knowledge-base'] | [-1.60243988e-01 1.17337964e-01 -6.66506767e-01 -4.63862151e-01
-4.46018934e-01 -6.23521566e-01 8.64802539e-01 2.86510020e-01
2.21971497e-01 1.05987895e+00 6.49402678e-01 -8.10484886e-01
-6.23795867e-01 -1.17487419e+00 -6.27889514e-01 -5.61248124e-01
-8.84266376e-01 5.95009804e-01 -8.68568122e-02 2.16110945e-02
2.71081775e-01 3.08096111e-01 -1.65254128e+00 3.58574629e-01
9.38381135e-01 3.63235861e-01 -1.30289957e-01 7.40477920e-01
2.45563105e-01 8.13939929e-01 -3.70279998e-01 3.14577706e-02
-1.09293088e-01 -5.02285480e-01 -6.71590030e-01 -8.35791290e-01
3.75922807e-02 -1.68822810e-01 -2.95133054e-01 4.47899014e-01
4.25574780e-01 -3.73305440e-01 6.58157051e-01 -2.00892758e+00
-5.23039281e-01 9.92153287e-01 -3.25296432e-01 4.29779947e-01
9.55201209e-01 1.87312707e-01 9.42548096e-01 -4.27217841e-01
8.05832922e-01 1.70729208e+00 4.08320457e-01 8.16878378e-02
-1.47641802e+00 -1.19691515e+00 -5.67045622e-03 2.95510173e-01
-1.16093874e+00 -2.74028093e-01 2.93722063e-01 -6.63219750e-01
8.10275674e-01 7.50598073e-01 6.05524719e-01 1.43822324e+00
5.01549602e-01 4.01863486e-01 1.28737772e+00 -2.64140010e-01
3.56687367e-01 -3.81398439e-01 3.76701593e-01 5.27286768e-01
3.16092014e-01 8.70370269e-01 -8.93657327e-01 -8.99826646e-01
7.60043502e-01 -1.45000055e-01 -1.14154458e-01 -6.82374090e-02
-1.46840096e+00 9.72344041e-01 1.94252223e-01 -1.33808181e-01
-4.36413199e-01 4.99447167e-01 3.91178131e-01 2.92347312e-01
3.31803292e-01 1.06844507e-01 -6.33808017e-01 -1.26159966e-01
-4.43833411e-01 8.78803551e-01 8.79880369e-01 1.02098012e+00
-2.55556917e-03 -3.05518895e-01 -2.93693423e-01 3.13730001e-01
4.77600247e-01 8.58703673e-01 -2.02234730e-01 -9.09388840e-01
-5.97480051e-02 6.73529983e-01 3.63530010e-01 -9.00846422e-01
-6.90908730e-01 1.39446273e-01 -7.64925063e-01 1.26604602e-01
3.49230438e-01 -3.45369667e-01 -7.43560493e-01 1.70716047e+00
7.38213420e-01 3.18879336e-01 -2.23388910e-01 8.18454504e-01
9.90095317e-01 5.56948304e-01 4.94472265e-01 -7.88387239e-01
1.42287695e+00 -7.49662071e-02 -1.16940928e+00 3.99373531e-01
3.58945310e-01 -7.28392959e-01 1.13303506e+00 4.85443771e-01
-8.93195391e-01 3.84217203e-02 -7.22297311e-01 7.90833011e-02
-4.38460499e-01 -3.54533285e-01 1.15149295e+00 6.06314480e-01
-6.92430735e-01 4.91424352e-01 -9.59075451e-01 -3.76736343e-01
-2.41060858e-03 2.36845344e-01 -9.94082391e-02 1.43301502e-01
-1.72732913e+00 5.46606004e-01 1.92618653e-01 -4.31591541e-01
-1.02895141e+00 -1.39240313e+00 -6.25187397e-01 4.04375000e-03
4.94566172e-01 -8.99707496e-01 1.17462385e+00 -1.44688085e-01
-8.59812200e-01 2.74455696e-01 -2.89021343e-01 -1.82696149e-01
5.15564263e-01 -9.83418003e-02 -7.49088287e-01 -4.36973006e-01
4.03872639e-01 7.32566044e-02 2.46465877e-01 -1.14240706e+00
-7.10157931e-01 -5.50684869e-01 -2.53962874e-01 -2.79992938e-01
1.92336142e-01 4.69108641e-01 2.89110281e-02 -6.34822905e-01
-3.43948573e-01 -7.53937006e-01 -2.43268177e-01 -4.21467692e-01
-6.69593215e-01 -4.35168058e-01 7.00144768e-01 -4.23797280e-01
1.74041951e+00 -1.72491479e+00 -7.30160102e-02 3.05543780e-01
-4.71439101e-02 -6.00432694e-01 1.36539131e-01 1.02178872e+00
-6.91936731e-01 4.69893843e-01 -2.70239294e-01 2.58464575e-01
-1.20651841e-01 3.01709980e-01 -3.76395822e-01 5.35618126e-01
1.28257692e-01 7.20257461e-01 -1.13846314e+00 -5.27818561e-01
3.89063627e-01 -2.91615929e-02 -6.30492032e-01 4.24080193e-01
-4.56737727e-01 5.07217824e-01 -5.17464161e-01 5.35070539e-01
4.97546226e-01 -1.64699759e-02 6.88535810e-01 1.90100044e-01
-7.91660070e-01 6.12174571e-01 -1.59807515e+00 1.32312930e+00
-2.73521990e-01 1.46817401e-01 -9.12919194e-02 -5.78857243e-01
6.12653315e-01 7.19224691e-01 6.89236045e-01 -4.02572095e-01
6.31814674e-02 -5.65588176e-02 -4.19304445e-02 -8.29703689e-01
5.58270589e-02 1.56286731e-01 -3.07541251e-01 7.89433479e-01
-3.60315263e-01 4.39910404e-03 5.56606531e-01 2.92229444e-01
1.53961682e+00 1.07296459e-01 6.36934221e-01 -6.44239306e-01
-4.47813869e-02 5.36333144e-01 6.37331426e-01 6.39983118e-01
3.76008570e-01 6.39895350e-02 1.01289618e+00 -5.61392546e-01
-9.21143889e-01 -1.31428647e+00 -5.11160731e-01 7.83609986e-01
-2.67889559e-01 -7.20219493e-01 -1.62725389e-01 -2.40568444e-01
3.80980611e-01 1.29702926e+00 -9.73812461e-01 1.81790292e-01
-3.41041267e-01 -1.22095144e+00 4.34910476e-01 1.77331924e-01
-4.01588567e-02 -9.40430403e-01 -6.02767348e-01 5.93608357e-02
-3.40456069e-01 -4.61724997e-01 1.39807044e-02 -1.98815335e-02
-8.05286050e-01 -1.62670147e+00 3.82187128e-01 6.73644915e-02
2.92124003e-01 -1.13957047e-01 1.10897803e+00 -2.66943604e-01
-5.17698884e-01 1.53372705e-01 -2.49452308e-01 -6.44816697e-01
-6.22088432e-01 -3.85270417e-01 -5.57712987e-02 -4.71858919e-01
1.74889356e-01 -7.90413558e-01 -6.14517689e-01 3.97784978e-01
-7.77420878e-01 1.01502217e-01 -1.99627697e-01 7.11157441e-01
2.11792246e-01 2.08521143e-01 6.66877747e-01 -8.42270195e-01
1.00168240e+00 -1.21744406e+00 -9.73101318e-01 1.23323046e-01
-9.33147490e-01 1.35967717e-01 2.67388135e-01 -1.43467888e-01
-1.13461280e+00 -1.07108496e-01 1.03335552e-01 1.45774963e-03
-2.89795220e-01 9.49898839e-01 5.92487156e-02 8.86539698e-01
9.79176164e-01 -6.21081591e-01 3.82637158e-02 -3.63883406e-01
6.03683472e-01 4.85763043e-01 4.21064615e-01 -6.50845945e-01
2.46809721e-01 4.81557786e-01 1.25137791e-01 -3.83684672e-02
-3.22717816e-01 -1.49643108e-01 -3.80263358e-01 -3.46251577e-01
6.82325602e-01 -7.30835974e-01 -1.35250795e+00 9.09158438e-02
-1.11361039e+00 -5.95413804e-01 1.27015099e-01 4.76670057e-01
-4.51955676e-01 -6.08959138e-01 -3.38593006e-01 -9.89933431e-01
-1.52960747e-01 -7.81099260e-01 8.85382175e-01 -7.99574554e-02
-7.60210812e-01 -1.06600761e+00 4.47940022e-01 -1.05057068e-01
-1.78232044e-01 8.97569060e-01 1.16740096e+00 -2.32180044e-01
-4.24855173e-01 -1.08321175e-01 -1.82516813e-01 -7.38497734e-01
1.31338626e-01 8.39080870e-01 -6.42503560e-01 1.24587737e-01
-5.42559206e-01 2.19348520e-02 3.78054410e-01 8.83578777e-01
1.12667859e+00 -5.03685653e-01 -9.34172273e-01 1.54910296e-01
1.15036178e+00 4.07199174e-01 6.20457053e-01 3.77582274e-02
3.09078872e-01 8.59312534e-01 9.92140293e-01 9.00240064e-01
6.80293798e-01 6.61206067e-01 5.95184326e-01 -2.37551346e-01
1.23205543e-01 -3.01184028e-01 6.68413937e-02 5.44654876e-02
-3.03051203e-01 -3.15440983e-01 -1.22200274e+00 4.94465888e-01
-2.24975896e+00 -1.31248760e+00 -1.43335533e+00 2.33371663e+00
1.04297447e+00 -2.70381361e-01 4.83006805e-01 2.62913853e-02
4.61911649e-01 -3.21584851e-01 -2.64805257e-01 -6.55175567e-01
2.39975572e-01 1.67263538e-01 7.65670538e-01 5.88197589e-01
-9.32760715e-01 5.78727186e-01 7.61813068e+00 3.70489419e-01
-7.52863109e-01 3.69691014e-01 3.82651299e-01 -2.58326679e-01
-5.55512547e-01 2.07032546e-01 -5.23046494e-01 3.19083780e-01
1.52625787e+00 -7.91544914e-01 3.06966215e-01 3.24028194e-01
1.38315582e+00 -4.67322141e-01 -1.07681739e+00 1.84074819e-01
-8.28642964e-01 -1.65128720e+00 -3.57909650e-01 9.40682963e-02
5.97635090e-01 -1.68376211e-02 -4.26527500e-01 -1.17683679e-01
1.14834726e+00 -1.08170497e+00 7.02872396e-01 4.77678925e-01
9.41756845e-01 -5.75742066e-01 3.19012344e-01 1.20203812e-02
-8.06020617e-01 -3.19661856e-01 2.50615776e-01 -5.24628937e-01
4.04933095e-01 1.02512908e+00 -8.55687797e-01 8.43845427e-01
1.11453795e+00 6.82781875e-01 -4.29757573e-02 8.13534796e-01
-4.72903460e-01 9.98163462e-01 -3.22146893e-01 4.77924794e-02
-5.03173709e-01 2.09394246e-02 5.46763241e-01 1.18448174e+00
2.99755484e-01 4.45884198e-01 -4.76370119e-02 1.07990134e+00
5.57179332e-01 -6.40452430e-02 -7.87572026e-01 7.63129890e-02
8.65249991e-01 7.79024661e-01 -4.13465679e-01 -2.67430365e-01
-8.95221978e-02 6.94527626e-02 -1.08445272e-01 2.52569526e-01
-1.14364469e+00 1.48854598e-01 8.68846416e-01 3.88511091e-01
-5.91441453e-01 -3.83921772e-01 -8.06906343e-01 -6.78726614e-01
-6.76755309e-01 -9.22141373e-01 1.28044248e+00 -1.06886840e+00
-9.64706838e-01 -1.25514984e-01 9.32832479e-01 -6.54695332e-01
-4.56974953e-01 -2.93384075e-01 -5.68877518e-01 9.34699178e-01
-6.09718263e-01 -9.04098749e-01 -2.05941021e-01 7.58388460e-01
1.58497289e-01 2.85395443e-01 1.08820748e+00 4.65430319e-01
-7.79019773e-01 -9.40307379e-02 -2.62524456e-01 -4.29279119e-01
8.50995541e-01 -1.46098804e+00 3.99748147e-01 6.58334434e-01
-6.64621294e-01 9.60863411e-01 1.23209882e+00 -1.36037529e+00
-1.50485897e+00 -9.20230210e-01 1.04339266e+00 -4.85882431e-01
1.17861891e+00 -5.20426750e-01 -4.91926581e-01 7.74925828e-01
3.26126784e-01 -5.48194110e-01 6.58311188e-01 8.92435193e-01
-2.47860640e-01 7.15594813e-02 -9.38273132e-01 8.83689523e-01
1.18881166e+00 1.35732964e-01 -3.77466053e-01 5.96758008e-01
7.11404264e-01 -2.80992657e-01 -1.27114642e+00 4.01243091e-01
5.13111532e-01 -8.93958926e-01 8.85824442e-01 -6.49421811e-01
8.90667617e-01 -5.06882548e-01 3.36288512e-01 -1.34908891e+00
-3.50623667e-01 -7.47813702e-01 9.38579366e-02 1.16729414e+00
6.08319998e-01 -7.64227152e-01 1.36026621e-01 7.55653799e-01
2.95253340e-02 -3.76671180e-02 -8.04661751e-01 -4.55231458e-01
-1.50398552e-01 -9.68364060e-01 8.66386354e-01 1.41274977e+00
3.67227674e-01 3.01287234e-01 -3.83182019e-01 5.68667293e-01
7.14206040e-01 3.20262372e-01 8.71896267e-01 -1.14627087e+00
4.99031059e-02 -1.58392832e-01 7.77746812e-02 6.37141913e-02
-2.12673157e-01 -7.06449986e-01 -4.53436047e-01 -1.52595055e+00
1.56218261e-01 -8.37505937e-01 2.03303650e-01 9.91003811e-01
-1.98774874e-01 -2.41515100e-01 -4.77653205e-01 1.18222743e-01
4.35949653e-01 3.15554202e-01 1.07666171e+00 1.27998471e-01
-4.73045588e-01 -3.65306661e-02 -3.96549106e-01 2.12121710e-01
1.02275360e+00 -9.56943154e-01 -6.13186657e-01 1.67105317e-01
4.53264058e-01 6.05042279e-01 1.02508640e+00 -2.05781505e-01
-5.95815405e-02 -9.09458935e-01 3.88046913e-02 -7.13859260e-01
-3.62799227e-01 -5.03049374e-01 1.10631049e+00 6.98086858e-01
-4.33158129e-01 4.13774967e-01 4.08092648e-01 3.16637278e-01
6.75841654e-03 1.76120386e-01 2.19162583e-01 5.66875264e-02
-2.73400962e-01 4.96467613e-02 -4.69045281e-01 -2.58891076e-01
1.12618303e+00 5.58167040e-01 -7.81789541e-01 -3.56644064e-01
-7.24638045e-01 7.24234223e-01 2.17183474e-02 8.18662524e-01
1.69008046e-01 -1.33762956e+00 -9.46079016e-01 -7.16848001e-02
2.01092184e-01 -3.88721853e-01 1.63869649e-01 1.14304054e+00
-3.22411865e-01 4.75962698e-01 -7.17820302e-02 -4.27379251e-01
-1.22107983e+00 9.11843657e-01 1.67859316e-01 -1.58799626e-02
-4.03314114e-01 2.86638230e-01 8.89690146e-02 -5.47167420e-01
-1.97818428e-01 -3.21173608e-01 7.54410699e-02 1.95580557e-01
6.29387498e-01 7.34705627e-01 -1.30216792e-01 2.99917221e-01
-4.85208750e-01 -2.83982337e-01 6.20411396e-01 -2.80749112e-01
1.35339439e+00 -6.75256401e-02 -6.02571189e-01 8.10049236e-01
5.47462344e-01 -4.44087666e-03 -7.41819859e-01 4.46239710e-01
1.40758276e-01 -3.76694828e-01 5.26546612e-02 -1.33464825e+00
-5.04583061e-01 1.18744887e-01 4.28046077e-01 6.86072588e-01
1.03472281e+00 5.95572926e-02 -2.22355217e-01 -5.32275021e-01
2.81097770e-01 -5.67414999e-01 -7.35737145e-01 1.00438967e-01
1.56792152e+00 -8.59508395e-01 3.00006211e-01 -8.01918149e-01
-1.23416796e-01 8.66054416e-01 1.51799977e-01 -1.31093552e-02
8.35572898e-01 8.23157668e-01 1.29204872e-03 -5.66557229e-01
-1.43118680e+00 -4.96501045e-04 2.87832133e-02 6.48071289e-01
9.01436567e-01 6.07851863e-01 -1.15770936e+00 3.23851258e-01
-3.97139758e-01 3.73829186e-01 6.38106704e-01 6.38607264e-01
3.01487505e-01 -1.39046049e+00 -1.01231742e+00 5.27654350e-01
-3.21335316e-01 -2.19721213e-01 -2.82303780e-01 1.20587969e+00
8.06206614e-02 1.44888568e+00 2.39354059e-01 -1.00863136e-01
4.58506733e-01 -1.46644488e-01 1.48142606e-01 -3.12400579e-01
-5.70324183e-01 1.94685012e-01 7.92698979e-01 -9.44118738e-01
-3.36828411e-01 -1.27982235e+00 -1.41705346e+00 -8.40649009e-01
1.29937366e-01 9.44477692e-02 6.86485767e-01 4.55613524e-01
2.98495680e-01 8.37866127e-01 5.30018866e-01 -3.63565087e-01
8.20906535e-02 -9.32998002e-01 -3.72652948e-01 1.64154857e-01
-1.93887074e-02 -1.10900819e+00 -2.02748820e-01 2.60512441e-01] | [7.828372955322266, 5.3903584480285645] |
06d2761d-5f0b-4399-909a-2d6ec510132d | the-algonauts-project-2021-challenge-how-the | 2104.13714 | null | https://arxiv.org/abs/2104.13714v1 | https://arxiv.org/pdf/2104.13714v1.pdf | The Algonauts Project 2021 Challenge: How the Human Brain Makes Sense of a World in Motion | The sciences of natural and artificial intelligence are fundamentally connected. Brain-inspired human-engineered AI are now the standard for predicting human brain responses during vision, and conversely, the brain continues to inspire invention in AI. To promote even deeper connections between these fields, we here release the 2021 edition of the Algonauts Project Challenge: How the Human Brain Makes Sense of a World in Motion (http://algonauts.csail.mit.edu/). We provide whole-brain fMRI responses recorded while 10 human participants viewed a rich set of over 1,000 short video clips depicting everyday events. The goal of the challenge is to accurately predict brain responses to these video clips. The format of our challenge ensures rapid development, makes results directly comparable and transparent, and is open to all. In this way it facilitates interdisciplinary collaboration towards a common goal of understanding visual intelligence. The 2021 Algonauts Project is conducted in collaboration with the Cognitive Computational Neuroscience (CCN) conference. | ['A. Oliva', 'G. Roig', 'K. Kay', 'N. A. R. Murty', 'A. Andonian', 'M. Graumann', 'P. Iamshchinina', 'A. Lascelles', 'B. Lahner', 'K. Dwivedi', 'R. M. Cichy'] | 2021-04-28 | null | null | null | null | ['human-fmri-response-prediction'] | ['computer-vision'] | [ 1.89284936e-01 -1.09701961e-01 4.15064186e-01 -3.24848086e-01
9.49377716e-02 -5.31880200e-01 7.81690538e-01 -4.32189792e-01
-5.83433509e-01 6.75073802e-01 1.54880509e-01 -2.42290758e-02
3.65693606e-02 -3.26059222e-01 -5.64399242e-01 -4.23924655e-01
-2.00152308e-01 1.88071802e-02 1.60698533e-01 -2.55357265e-01
5.02172410e-01 6.28658116e-01 -1.93910050e+00 5.42587221e-01
4.90980327e-01 8.86818171e-01 6.26832485e-01 9.09471869e-01
5.47210336e-01 1.09206486e+00 -1.78242773e-01 -1.41293317e-01
3.48806053e-01 -7.73251414e-01 -9.51780438e-01 -2.79762000e-01
5.37946939e-01 -1.55001357e-02 -5.35393834e-01 9.75385010e-01
4.53001291e-01 3.56630296e-01 3.86079609e-01 -1.29384220e+00
-6.76644146e-01 -1.20416783e-01 -2.38896683e-01 9.38291430e-01
4.28196639e-01 5.27208805e-01 8.52115154e-01 -9.00670052e-01
1.01890898e+00 1.08584142e+00 3.12749982e-01 6.35119140e-01
-1.03284872e+00 -2.56548196e-01 9.21729803e-02 6.40680134e-01
-9.47640717e-01 -8.18285048e-01 4.92529094e-01 -7.53866136e-01
1.24950433e+00 5.19203484e-01 1.16735268e+00 1.30174708e+00
6.16108179e-01 5.74163854e-01 1.35562587e+00 -3.66778314e-01
2.18321279e-01 -3.38813104e-02 -3.42960432e-02 5.58244824e-01
-1.03174532e-02 5.83241701e-01 -9.01039124e-01 1.27787039e-01
9.70855415e-01 -3.18436861e-01 -5.83566785e-01 -1.68052152e-01
-1.80270612e+00 4.18109000e-01 4.40906078e-01 6.00474000e-01
-4.83447939e-01 -5.81521401e-03 2.06567913e-01 3.22861552e-01
1.97276995e-01 7.82994390e-01 -1.08039074e-01 -1.36962473e-01
-7.08915889e-01 4.79617625e-01 5.67241490e-01 6.99909568e-01
3.00661981e-01 3.09346855e-01 2.12723047e-01 7.77754962e-01
1.40432373e-01 3.98245186e-01 5.79555929e-01 -1.46561968e+00
-2.39498004e-01 1.75360020e-03 1.98355973e-01 -1.18645692e+00
-7.86605299e-01 -2.00174630e-01 -3.98161799e-01 6.72236204e-01
5.74918509e-01 -1.71308801e-01 -6.90878153e-01 1.67666566e+00
1.49708074e-02 -1.12427890e-01 -2.51439571e-01 1.57488704e+00
6.85660660e-01 4.55099016e-01 8.23841840e-02 -2.36900970e-01
1.43624532e+00 -7.21155822e-01 -5.35918415e-01 -6.68914855e-01
1.52501896e-01 -4.50611115e-01 8.65471542e-01 6.40275538e-01
-1.41800177e+00 -8.71545732e-01 -1.03261602e+00 -1.78863943e-01
-3.41288090e-01 -3.26648474e-01 9.37201560e-01 5.43711483e-01
-1.27984488e+00 2.86870331e-01 -8.14456701e-01 -6.89366519e-01
3.53727043e-01 1.60603613e-01 -6.22727275e-01 1.79242402e-01
-1.11940992e+00 1.42801142e+00 2.87419856e-01 3.49633731e-02
-8.50528240e-01 -7.45559931e-01 -5.25303841e-01 -4.24926907e-01
1.30558275e-02 -7.32648432e-01 1.27300870e+00 -1.32376313e+00
-1.15991437e+00 1.50985682e+00 -1.80761531e-01 -5.72121620e-01
2.41687581e-01 -1.13757268e-01 -6.69226408e-01 4.57110614e-01
-6.35620281e-02 1.18613029e+00 5.94793260e-01 -9.25237000e-01
-5.75173914e-01 -4.98472035e-01 -2.77577825e-02 2.81703115e-01
6.56787157e-02 3.24784607e-01 -2.16992060e-03 -7.40370214e-01
4.20455076e-02 -1.02928591e+00 1.90537628e-02 1.61827356e-01
1.66592956e-01 1.13566322e-02 3.11899900e-01 -5.22951961e-01
4.14419740e-01 -2.03898311e+00 2.98625261e-01 -3.63992929e-01
6.24167562e-01 -3.09714321e-02 -1.61388397e-01 2.99875677e-01
-2.94222027e-01 -1.07234992e-01 -2.46182516e-01 2.27131620e-01
-1.85577888e-02 -3.08265209e-01 -3.32170039e-01 6.34494245e-01
1.14946194e-01 1.06359911e+00 -9.66254711e-01 -1.56060532e-01
2.47251958e-01 4.12721068e-01 -4.17192906e-01 -7.67600313e-02
-1.33978024e-01 8.52202892e-01 -2.00283110e-01 3.99481595e-01
4.79041308e-01 -1.21104601e-03 8.56189895e-03 -2.74030894e-01
-6.21580601e-01 -6.61683530e-02 -3.08566958e-01 1.83060050e+00
3.81198078e-02 1.34098041e+00 1.79564267e-01 -1.00380015e+00
6.81652129e-01 3.46626759e-01 3.78178626e-01 -1.14946151e+00
2.30698287e-01 -5.61259724e-02 5.68300545e-01 -8.40918303e-01
8.98787752e-02 -5.06610930e-01 2.58633941e-01 3.51176172e-01
1.48117930e-01 -5.12044072e-01 2.15883180e-01 4.62191515e-02
1.16985500e+00 2.75392920e-01 3.41036171e-01 -6.32436097e-01
3.33977461e-01 1.94640830e-01 4.63201821e-01 6.77725196e-01
-6.68509066e-01 5.71430564e-01 2.34248757e-01 -9.65313494e-01
-1.08076179e+00 -1.24793100e+00 -8.04636925e-02 9.31336820e-01
-2.93047875e-02 1.02617204e-01 -7.71396041e-01 -1.87822653e-03
-3.56220156e-01 7.70011127e-01 -6.99317336e-01 -1.05111383e-01
-3.68121535e-01 -3.80300194e-01 3.34209740e-01 3.57458681e-01
6.06832027e-01 -1.73225713e+00 -1.32910347e+00 1.54875489e-02
-2.43498862e-01 -1.48263705e+00 -7.85983056e-02 -3.51453602e-01
-6.30045652e-01 -1.03612757e+00 -7.71140695e-01 -8.25360775e-01
3.11234295e-01 2.77408987e-01 1.12614596e+00 2.73794141e-02
-9.64581311e-01 7.09080994e-01 -1.35495782e-01 -6.08369350e-01
-1.22984767e-01 -5.80560684e-01 4.27817613e-01 -1.08022474e-01
4.34763193e-01 -5.90632498e-01 -6.80920303e-01 2.24823669e-01
-6.55808747e-01 3.65386575e-01 2.71077842e-01 4.98463809e-01
2.37373993e-01 -1.98551625e-01 6.50753260e-01 -4.88538712e-01
4.62968588e-01 -4.21565861e-01 -6.58143461e-01 4.84051332e-02
-1.67272717e-01 -4.95124578e-01 5.58791459e-01 -2.09769189e-01
-1.24570835e+00 -3.66656333e-02 1.62407234e-01 -1.44083098e-01
-5.10019839e-01 3.78141135e-01 2.44372874e-01 -3.24215382e-01
9.49747741e-01 1.48532435e-01 -3.94007005e-02 6.75203800e-02
2.27378324e-01 3.37382257e-01 1.06399977e+00 -2.59906083e-01
4.94986296e-01 6.67744160e-01 -1.48187101e-01 -1.34589458e+00
-3.66714388e-01 1.97974995e-01 -6.64042771e-01 -8.09373915e-01
1.15232754e+00 -6.84576690e-01 -9.71094370e-01 5.30750215e-01
-1.20258725e+00 -4.30848062e-01 -1.01775602e-01 8.13719034e-01
-8.38535905e-01 7.89997280e-02 -3.77783537e-01 -8.67804110e-01
9.42720696e-02 -9.88242626e-01 4.97382730e-01 5.69848239e-01
-5.73825479e-01 -9.26511288e-01 1.84389487e-01 5.63853085e-01
3.88981253e-01 4.24736857e-01 6.68835580e-01 -2.10834637e-01
-4.94948179e-01 -1.84840839e-02 -2.57280171e-01 1.10193081e-01
-2.15786323e-01 -9.73036215e-02 -1.34896374e+00 -7.36044422e-02
5.72556913e-01 -4.71375972e-01 6.67829335e-01 6.62409067e-01
9.70193923e-01 3.05871636e-01 -3.59433666e-02 3.52599651e-01
1.08041835e+00 5.75106800e-01 9.45930183e-01 2.24148914e-01
2.61573344e-01 1.25290978e+00 2.81481475e-01 1.49512634e-01
3.16214472e-01 4.47925270e-01 1.67438433e-01 2.03981936e-01
-1.21079892e-01 2.85111398e-01 3.69687408e-01 5.02299488e-01
-5.81813991e-01 4.46755141e-02 -1.22827291e+00 4.21540678e-01
-1.45596433e+00 -1.31351221e+00 -2.71701187e-01 2.24884796e+00
2.97789186e-01 1.00094460e-01 8.08369517e-02 -1.67210132e-01
4.84963149e-01 9.13686529e-02 -6.61734998e-01 -2.46160746e-01
-2.54554212e-01 1.43137977e-01 -2.04279665e-02 4.29794759e-01
-8.21252763e-01 9.44805205e-01 7.26934052e+00 8.58735815e-02
-1.12093210e+00 1.72885731e-01 4.80945617e-01 -4.67047960e-01
-8.18964094e-02 -1.97685584e-01 -7.20799640e-02 2.01531410e-01
1.29062462e+00 -5.88676870e-01 1.00455952e+00 1.59392193e-01
5.38953245e-01 -4.50683594e-01 -9.55886185e-01 1.02309239e+00
2.88690597e-01 -1.48702061e+00 -5.26449978e-01 4.23037671e-02
4.22676384e-01 4.25860822e-01 1.21013194e-01 -2.79071052e-02
-2.62023866e-01 -1.26153922e+00 9.46578205e-01 1.00440526e+00
6.27853572e-01 -4.10458326e-01 6.82175234e-02 3.12250853e-01
-8.18707645e-01 -1.45135567e-01 -4.49298173e-01 -5.56364417e-01
1.13879487e-01 2.89272338e-01 -2.16894746e-01 -2.87450738e-02
9.36846614e-01 5.59486926e-01 -4.35704201e-01 1.17366755e+00
1.19273119e-01 2.67850399e-01 1.42427832e-01 -1.21990204e-01
-7.04446286e-02 -6.05127662e-02 9.15799320e-01 8.13067198e-01
-1.98771283e-02 5.88461041e-01 -2.44545147e-01 1.13113260e+00
2.11758196e-01 -1.61198899e-01 -9.31034744e-01 -4.37880009e-01
2.45777354e-01 1.19918942e+00 -9.25858438e-01 -8.77055749e-02
-5.28588891e-01 9.17355418e-01 2.41815850e-01 5.77867568e-01
-7.49344528e-01 -2.07368404e-01 7.07190573e-01 -8.68527815e-02
-2.38721251e-01 -5.21566629e-01 -2.75423974e-01 -1.00965285e+00
-1.13453679e-01 -9.09383237e-01 -1.45884240e-02 -1.63573289e+00
-1.06351292e+00 7.78670788e-01 4.86787148e-02 -8.66746962e-01
-2.21563876e-01 -8.98422062e-01 -4.82632905e-01 8.43143523e-01
-8.64139795e-01 -6.74238682e-01 -6.10652149e-01 3.53155911e-01
6.49739563e-01 -2.75715470e-01 9.80726779e-01 -7.67168179e-02
-3.17840397e-01 -1.83403835e-01 -6.54514253e-01 -2.37993583e-01
5.71986675e-01 -8.82019341e-01 7.50147343e-01 6.98352754e-01
3.01826984e-01 5.27192831e-01 8.69072139e-01 -5.81728160e-01
-1.50283146e+00 -3.85144353e-01 6.94911003e-01 -6.66942835e-01
7.83051312e-01 -4.28445727e-01 -7.04360485e-01 7.29876101e-01
5.99058628e-01 1.48006584e-02 6.50988579e-01 -3.55791241e-01
-8.61041695e-02 1.71187252e-01 -1.02849102e+00 7.45599031e-01
1.28133690e+00 -7.23952353e-01 -7.91267693e-01 6.72351658e-01
4.01982337e-01 -1.41265213e-01 -6.63883567e-01 1.80683628e-01
8.64562392e-01 -1.32254303e+00 9.41710293e-01 -8.21188986e-01
3.89210790e-01 -2.65517652e-01 -1.34670466e-01 -1.30628216e+00
-6.17063046e-01 -7.87355483e-01 2.32243910e-01 3.82668555e-01
8.00887942e-02 -7.85031021e-01 6.11789465e-01 8.27625334e-01
-1.48851648e-01 -2.76668578e-01 -1.00615180e+00 -5.92924953e-01
2.41674557e-02 -6.86039448e-01 2.62993369e-02 8.75517130e-01
3.43908727e-01 4.03667465e-02 -2.30686173e-01 3.52513120e-02
6.69093192e-01 -2.89264619e-01 3.93713713e-01 -1.37766683e+00
-5.44670671e-02 -5.16673982e-01 -8.04892123e-01 -3.75427395e-01
1.60271436e-01 -8.32247794e-01 -3.81618850e-02 -1.25059903e+00
9.30386037e-03 2.85712242e-01 -2.51362711e-01 2.91776538e-01
2.44627565e-01 6.53633416e-01 3.85523409e-01 2.11326838e-01
-1.27801046e-01 2.16364861e-01 1.42022479e+00 1.61997437e-01
2.16463879e-01 -3.97530645e-01 -7.54676640e-01 7.74739861e-01
9.79355335e-01 -4.03920710e-02 -5.55485904e-01 -5.44228077e-01
1.08454332e-01 1.07543051e-01 9.22913134e-01 -1.54817975e+00
3.81179571e-01 -2.58978605e-01 7.31110513e-01 -1.21304810e-01
5.66775441e-01 -5.43578148e-01 2.73265958e-01 5.50753713e-01
-2.36029565e-01 2.33161584e-01 5.83983302e-01 1.89462021e-01
1.45917200e-03 6.64402395e-02 1.09213793e+00 -4.99159276e-01
-1.16893089e+00 -4.58073914e-02 -8.27763557e-01 2.82122791e-01
1.14952278e+00 -3.55232239e-01 -6.41109228e-01 -2.91671395e-01
-1.00016367e+00 8.91629755e-02 3.89568597e-01 5.32547176e-01
8.44199836e-01 -1.09071946e+00 -7.24331856e-01 4.74470466e-01
-6.53960705e-02 -7.70670593e-01 6.12926543e-01 1.19642448e+00
-7.01241612e-01 6.46870017e-01 -1.05296934e+00 -3.60433608e-01
-9.79449391e-01 2.65209287e-01 6.14701986e-01 6.04848444e-01
-8.14758897e-01 1.17936563e+00 5.79596519e-01 -2.01719552e-02
-2.22651124e-01 1.38841838e-01 -2.23719731e-01 -7.13547841e-02
1.13857865e+00 2.99695194e-01 -1.75031513e-01 -6.99497521e-01
-4.84420210e-01 1.85216829e-01 3.57485376e-02 -6.36771858e-01
1.39541233e+00 -9.14229229e-02 -2.73293108e-01 7.81740844e-01
7.66504347e-01 -2.23946035e-01 -1.25228250e+00 3.40266466e-01
-6.94373846e-02 -3.89684349e-01 2.76235104e-01 -1.00095618e+00
-9.76302147e-01 9.25942957e-01 7.17041194e-01 2.41785988e-01
1.20860517e+00 -2.63397127e-01 4.80762482e-01 3.45751375e-01
5.91085613e-01 -1.06650996e+00 -1.07877478e-01 4.11919415e-01
1.32585537e+00 -1.05818045e+00 -7.17793964e-03 -5.41649014e-02
-9.83170033e-01 1.04522109e+00 8.01970065e-01 -2.77273893e-01
3.47145051e-01 2.05157071e-01 1.27666026e-01 -3.92865658e-01
-9.00803983e-01 -1.08032443e-01 4.31981921e-01 1.02546120e+00
6.72874928e-01 -4.65230606e-02 -2.25962058e-01 3.45132977e-01
-4.28370982e-01 2.85355538e-01 5.06493211e-01 8.46850455e-01
-6.12455904e-01 -5.58063567e-01 -5.09456396e-01 2.93566227e-01
-3.51668219e-03 -8.78090560e-02 -6.56414390e-01 7.85874367e-01
5.50825410e-02 9.34887826e-01 2.43347082e-02 -3.05644900e-01
3.01393420e-01 3.11791509e-01 9.93742108e-01 -4.27674681e-01
-4.06444639e-01 -2.59265929e-01 1.81177706e-01 -8.56725872e-01
-5.23656011e-01 -8.96741688e-01 -1.35771012e+00 -2.49142215e-01
4.28581506e-01 -1.57960551e-03 6.92787766e-01 7.90561020e-01
3.63957047e-01 4.08923119e-01 1.45749971e-01 -1.42787170e+00
1.99797094e-01 -8.38233769e-01 -6.06468260e-01 5.00792563e-02
1.22892998e-01 -8.10924232e-01 -3.25327992e-01 4.14198786e-01] | [10.490528106689453, 2.4560608863830566] |
2d8f09c4-840d-40f6-9183-341583fe2328 | stylization-based-architecture-for-fast-deep | null | null | http://openaccess.thecvf.com/content_CVPR_2020/html/Xu_Stylization-Based_Architecture_for_Fast_Deep_Exemplar_Colorization_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Xu_Stylization-Based_Architecture_for_Fast_Deep_Exemplar_Colorization_CVPR_2020_paper.pdf | Stylization-Based Architecture for Fast Deep Exemplar Colorization | Exemplar-based colorization aims to add colors to a grayscale image guided by a content related reference im- age. Existing methods are either sensitive to the selection of reference images (content, position) or extremely time and resource consuming, which limits their practical applica- tion. To tackle these problems, we propose a deep exemplar colorization architecture inspired by the characteristics of stylization in feature extracting and blending. Our coarse- to-fine architecture consists of two parts: a fast transfer sub-net and a robust colorization sub-net. The transfer sub- net obtains a coarse chrominance map via matching basic feature statistics of the input pairs in a progressive way. The colorization sub-net refines the map to generate the final re- sults. The proposed end-to-end network can jointly learn faithful colorization with a related reference and plausible color prediction with unrelated reference. Extensive exper- imental validation demonstrates that our approach outper- forms the state-of-the-art methods in less time whether in exemplar-based colorization or image stylization tasks.
| [' Guixu Zhang', ' Yun Sheng', ' Faming Fang', ' Tingting Wang', 'Zhongyou Xu'] | 2020-06-01 | null | null | null | cvpr-2020-6 | ['image-stylization'] | ['computer-vision'] | [ 2.14500099e-01 -3.55933756e-01 6.24747090e-02 -3.13057125e-01
-7.31098711e-01 -4.05097246e-01 5.91467440e-01 -1.43092990e-01
-4.39282149e-01 6.72941148e-01 -3.94231752e-02 1.01429775e-01
1.02630399e-01 -7.92410553e-01 -9.54208374e-01 -5.90096772e-01
4.38247740e-01 3.04455608e-01 1.28164232e-01 -2.37560540e-01
5.31781852e-01 5.57789207e-01 -1.46800137e+00 3.10835749e-01
1.16779912e+00 1.00888944e+00 1.10110395e-01 5.01483917e-01
-3.05618763e-01 5.58135390e-01 -3.71241391e-01 -3.91208857e-01
4.47041243e-01 -7.99522817e-01 -6.00747108e-01 3.33460391e-01
7.91564822e-01 -1.87306687e-01 -1.37075737e-01 1.28764904e+00
3.58716279e-01 3.87438804e-01 6.44481242e-01 -1.33956254e+00
-1.34884357e+00 4.30542022e-01 -1.07306552e+00 -1.47182032e-01
1.70635253e-01 1.10096879e-01 8.37247729e-01 -9.85313892e-01
6.46531105e-01 1.48968041e+00 5.29111087e-01 6.67635143e-01
-1.59552765e+00 -7.75197387e-01 3.74565542e-01 1.61778823e-01
-1.29357994e+00 -1.09482765e-01 1.07605731e+00 -1.52076572e-01
3.04360062e-01 1.57946751e-01 8.96336436e-01 7.68380642e-01
1.37242168e-01 8.02796006e-01 1.58108115e+00 -4.70222950e-01
2.84154683e-01 3.96033116e-02 -2.81138986e-01 8.48291278e-01
2.05050841e-01 1.85458675e-01 -6.35893524e-01 2.25323588e-01
1.10168457e+00 2.52800703e-01 -2.98866987e-01 -5.65771639e-01
-1.31311357e+00 5.05853772e-01 1.04357219e+00 2.23159894e-01
-4.56555098e-01 4.99010563e-01 -1.17781125e-01 2.33240888e-01
3.47506046e-01 6.40176773e-01 -2.28332072e-01 2.32164264e-01
-1.16950011e+00 1.44950837e-01 3.95777702e-01 9.00951087e-01
1.19605613e+00 3.02375108e-01 -2.91100025e-01 8.39763343e-01
1.11925997e-01 3.40106487e-01 4.18364942e-01 -8.87781262e-01
2.43875951e-01 4.11269993e-01 2.26590976e-01 -6.99110627e-01
-2.80673593e-01 -3.30584466e-01 -9.23583388e-01 8.10802102e-01
5.32494903e-01 -1.43789515e-01 -1.19685531e+00 1.65032279e+00
2.60567605e-01 4.02178198e-01 -1.01104870e-01 1.11385953e+00
5.62428117e-01 6.08704507e-01 1.67642802e-01 1.47304758e-01
1.23635924e+00 -1.23209870e+00 -4.56610173e-01 -8.63618925e-02
-1.16965376e-01 -1.07913411e+00 1.42225397e+00 4.36420023e-01
-1.24242127e+00 -9.79708970e-01 -1.10607457e+00 -1.91554204e-01
-5.27741134e-01 2.14189187e-01 5.98405540e-01 5.46598077e-01
-1.31448066e+00 7.84787238e-01 -4.66468334e-01 -4.15925235e-01
4.99786943e-01 1.99723884e-01 -3.03930581e-01 -2.95788497e-02
-7.98829675e-01 6.84021294e-01 4.58115518e-01 2.19429731e-01
-7.34169066e-01 -8.61090004e-01 -6.17009163e-01 -5.21379635e-02
9.72583070e-02 -7.89377868e-01 7.69951224e-01 -1.64898872e+00
-1.83189774e+00 7.71192193e-01 4.90619652e-02 -4.36490439e-02
6.75767243e-01 -3.84075850e-01 -4.38769609e-01 2.19965074e-02
6.58110380e-02 9.96026993e-01 1.21981132e+00 -1.74423146e+00
-9.48175728e-01 -9.64047983e-02 -3.00995111e-02 3.41744155e-01
-1.11732282e-01 -2.11905316e-01 -8.81865978e-01 -1.09483731e+00
1.39546499e-01 -7.62988865e-01 -8.79878178e-02 3.52125496e-01
-5.89997947e-01 3.19514066e-01 7.71340311e-01 -4.76429313e-01
8.34219038e-01 -2.00678754e+00 2.14732021e-01 2.81953096e-01
1.71750769e-01 1.83620125e-01 -5.44486523e-01 3.69812071e-01
-4.34814453e-01 -8.59216079e-02 -1.59061506e-01 -2.08681509e-01
4.23115306e-02 -1.94419041e-01 -2.22324148e-01 3.86562377e-01
4.14786607e-01 9.09415126e-01 -1.05971718e+00 -4.52934951e-01
4.54002559e-01 5.68204343e-01 -5.15995443e-01 4.05165404e-01
-3.27471495e-01 4.27014053e-01 -2.63623416e-01 6.93796813e-01
9.11543727e-01 -4.29230928e-02 -1.80480644e-01 -7.22084165e-01
-1.56304821e-01 -4.28018302e-01 -1.21778929e+00 2.02044559e+00
-5.88270187e-01 5.91527820e-01 -1.74110383e-01 -5.70724785e-01
1.03519917e+00 -2.22303733e-01 2.66001523e-01 -8.12025070e-01
9.89368036e-02 1.14521019e-01 -6.26552105e-02 -1.39835626e-01
6.92367613e-01 -1.95600286e-01 1.47386506e-01 5.29653609e-01
-1.77134387e-02 -4.09467548e-01 1.40490010e-01 1.55922621e-01
3.99620384e-01 7.24783063e-01 1.46926761e-01 -3.12816620e-01
5.52019417e-01 -2.73941815e-01 4.04107898e-01 5.52578568e-01
-4.53145690e-02 8.75868618e-01 4.36293989e-01 -4.99907255e-01
-1.07613564e+00 -1.26756251e+00 2.21067041e-01 1.28572428e+00
5.83497167e-01 -6.57502636e-02 -9.80782211e-01 -6.94527090e-01
1.25202551e-01 7.66110539e-01 -1.16565871e+00 -2.01113731e-01
-5.14918804e-01 -4.93203372e-01 1.08488321e-01 5.95967591e-01
8.19927514e-01 -1.22601330e+00 -4.10193294e-01 1.36794224e-01
9.78924260e-02 -7.06518292e-01 -8.78725708e-01 2.11917907e-01
-6.68087006e-01 -9.31676924e-01 -1.00451469e+00 -9.10290778e-01
9.64864731e-01 3.62647951e-01 1.08409631e+00 1.65500212e-02
-2.12136507e-01 2.96576709e-01 -3.61782104e-01 -1.38999254e-01
-2.55160391e-01 -1.65281326e-01 -2.75454938e-01 5.07577479e-01
1.88389033e-01 -4.45422143e-01 -1.11252630e+00 1.29078954e-01
-1.01683378e+00 5.03133357e-01 9.20880198e-01 9.34624553e-01
9.44230139e-01 -1.84888437e-01 2.70071357e-01 -9.62268531e-01
6.92096591e-01 6.01177476e-02 -5.55029452e-01 5.98383546e-01
-6.40751898e-01 4.40574080e-01 8.82896423e-01 -4.11567867e-01
-1.20015967e+00 3.24399531e-01 9.07415301e-02 -5.13738573e-01
-1.26370564e-01 6.98467940e-02 -1.24686517e-01 -2.77728766e-01
5.48249602e-01 2.28630647e-01 -2.20535979e-01 -4.09505635e-01
9.89675343e-01 8.31265822e-02 8.43824029e-01 -7.27976263e-01
1.28585351e+00 5.22378027e-01 -1.63635500e-02 -4.45612162e-01
-7.05961883e-01 7.48794619e-03 -9.00101125e-01 -3.49360406e-01
1.00858104e+00 -7.60166287e-01 -5.08423150e-01 5.12703240e-01
-8.96598935e-01 -5.11932492e-01 -4.38243359e-01 2.27939203e-01
-6.04117274e-01 1.86391279e-01 -4.34792221e-01 -4.33961242e-01
-3.90758097e-01 -1.03831971e+00 1.14534009e+00 6.24964118e-01
1.27477616e-01 -7.34377563e-01 5.19392863e-02 -2.01253340e-01
4.98769462e-01 5.41884601e-01 9.42406952e-01 1.53190345e-01
-6.46668375e-01 -1.52937174e-01 -7.88456500e-01 1.34295925e-01
3.43302190e-01 3.07660252e-01 -9.40289080e-01 -2.97222286e-01
-5.62980115e-01 -3.20948780e-01 1.14357209e+00 3.01000029e-01
1.36721444e+00 -4.97850068e-02 7.73078650e-02 9.86064970e-01
1.82942951e+00 1.38069034e-01 6.57878041e-01 4.21311915e-01
9.39389169e-01 4.40603107e-01 5.87924421e-01 3.78534049e-01
7.45270699e-02 5.67270935e-01 4.63914603e-01 -7.90982723e-01
-6.12798393e-01 -4.00479704e-01 8.10881779e-02 2.15084121e-01
-2.96939194e-01 1.19689599e-01 -4.10689682e-01 2.98015565e-01
-1.71916449e+00 -9.49740112e-01 2.97343433e-01 2.17146325e+00
1.02924025e+00 -4.88406932e-03 2.66300917e-01 -1.98620245e-01
7.88854122e-01 1.70688629e-01 -7.85718322e-01 -3.33054096e-01
-1.76345080e-01 4.69815284e-01 3.62793595e-01 3.44405591e-01
-9.86308396e-01 1.33400309e+00 5.02039671e+00 8.44318032e-01
-1.39695632e+00 -7.58643970e-02 8.76701236e-01 6.84734136e-02
-5.58686554e-01 -1.79880381e-01 -2.05087259e-01 4.26262170e-01
1.73399016e-01 -3.99840176e-02 9.10032392e-01 6.22947812e-01
4.61344309e-02 -1.85559243e-01 -1.09936929e+00 1.15752041e+00
1.44296601e-01 -1.31573975e+00 5.10260522e-01 -5.43971717e-01
1.05821884e+00 -3.17622930e-01 5.86199105e-01 3.13644335e-02
5.44442594e-01 -8.65355074e-01 1.17643940e+00 8.92187774e-01
1.21740365e+00 -9.08264279e-01 2.64993608e-01 -4.50378150e-01
-1.37998223e+00 1.25724720e-02 -4.22760487e-01 4.07778233e-01
-2.17860695e-02 1.41339913e-01 -2.55566061e-01 5.15383542e-01
7.02216208e-01 6.63655698e-01 -9.04861629e-01 1.11273789e+00
-3.70196193e-01 1.93195730e-01 1.65467769e-01 -3.92780639e-02
4.47299898e-01 -5.77345729e-01 -6.08015731e-02 1.31292617e+00
2.70362526e-01 -7.13806376e-02 -3.97197902e-02 1.23217642e+00
-2.57184774e-01 -1.71991549e-02 -4.56496738e-02 1.36183485e-01
4.86772150e-01 1.54289496e+00 -1.01918018e+00 -3.46404105e-01
-4.32517588e-01 1.59129488e+00 5.49767256e-01 8.60717177e-01
-8.35869253e-01 -7.46468961e-01 5.66146493e-01 -1.89176761e-02
4.45737213e-01 4.91692796e-02 -3.35991800e-01 -1.07120442e+00
-3.77256125e-01 -8.31800044e-01 8.02870020e-02 -1.14561903e+00
-1.44290841e+00 8.72913897e-01 -2.52017081e-01 -1.40490997e+00
-1.17847351e-02 -6.03374958e-01 -8.94921005e-01 1.09118247e+00
-1.66359985e+00 -1.59763324e+00 -7.77245104e-01 8.31853330e-01
3.68916333e-01 -1.39824405e-01 6.98248863e-01 4.37171344e-04
-5.73571980e-01 7.93900371e-01 3.01567703e-01 1.77103400e-01
1.08159769e+00 -1.69146180e+00 5.77406406e-01 1.10729313e+00
2.61006117e-01 5.85558712e-01 5.57103038e-01 -4.82324898e-01
-1.51829314e+00 -1.35732913e+00 1.61112562e-01 -2.61195991e-02
4.99432147e-01 -2.80892313e-01 -7.73399770e-01 1.52621746e-01
4.32834923e-01 1.19298540e-01 3.11473638e-01 -8.29772502e-02
-6.94266379e-01 -6.15348160e-01 -1.01355064e+00 9.04935777e-01
7.37013519e-01 -5.90216517e-01 -3.72157753e-01 -2.14490537e-02
4.37057704e-01 -2.85439909e-01 -5.35830617e-01 -2.04771131e-01
7.95923293e-01 -1.15180600e+00 1.01336181e+00 -5.57466686e-01
6.12512112e-01 -6.33812606e-01 1.01837598e-01 -1.48685944e+00
-6.78580284e-01 -7.21350670e-01 4.00712222e-01 1.20387077e+00
2.13415042e-01 -3.42374563e-01 7.67185092e-01 6.39865816e-01
-9.14030895e-02 -6.06247723e-01 -4.67073232e-01 -5.85613191e-01
2.08421066e-01 -4.49752696e-02 8.14376831e-01 8.55474353e-01
-3.26804101e-01 6.34189919e-02 -4.46111113e-01 -1.19311504e-01
8.96391869e-01 6.26328468e-01 8.91324103e-01 -9.76588428e-01
-2.84583002e-01 -7.54134178e-01 -6.08183555e-02 -7.10373223e-01
4.89174686e-02 -6.32511437e-01 1.94959804e-01 -1.55600953e+00
2.93260843e-01 -4.46177423e-01 -6.29435539e-01 4.41250741e-01
-6.19416595e-01 7.10874379e-01 4.43529069e-01 5.86339906e-02
-6.67746425e-01 5.95629036e-01 1.60135353e+00 -2.44968265e-01
-3.99440259e-01 -3.97808015e-01 -8.94575477e-01 5.79508245e-01
8.05445015e-01 -9.29116532e-02 -3.55708808e-01 -3.92729700e-01
-2.69867387e-02 -1.32483959e-01 3.46966803e-01 -1.11994064e+00
-1.00564891e-02 -3.50256592e-01 1.01598954e+00 -5.12825072e-01
2.42588654e-01 -8.91884804e-01 1.34723291e-01 4.02901679e-01
-4.83776510e-01 1.56570509e-01 1.77023306e-01 6.03691876e-01
-2.24671841e-01 -3.14988382e-02 1.20279419e+00 -1.43732548e-01
-1.01836216e+00 4.95042533e-01 8.73681083e-02 1.04361914e-01
1.05527163e+00 -5.05231798e-01 -2.49709442e-01 -2.91656911e-01
-4.25377995e-01 -2.48744696e-01 7.99601972e-01 3.69369179e-01
7.12192118e-01 -1.55373979e+00 -7.60579765e-01 3.97628784e-01
1.71057835e-01 -4.54268873e-01 2.87136644e-01 4.84334290e-01
-7.82340705e-01 -1.61120668e-02 -8.82924914e-01 -2.28557318e-01
-9.21418488e-01 8.68239224e-01 4.25173759e-01 1.74530104e-01
-3.96125525e-01 1.02104604e+00 5.96517444e-01 6.15774877e-02
1.50888339e-01 -4.14255470e-01 2.92334966e-02 -1.18621185e-01
3.62021923e-01 2.84277380e-01 -2.48471603e-01 -6.73194289e-01
-2.10405160e-02 1.03028405e+00 -1.87829092e-01 -2.86243737e-01
1.18605435e+00 -1.59870148e-01 -1.49374321e-01 2.13832408e-01
1.20974839e+00 -1.92668959e-01 -1.73895586e+00 -3.17533165e-01
-2.95863181e-01 -7.06588566e-01 -3.31463888e-02 -1.06783330e+00
-1.49361777e+00 8.73044431e-01 7.64679551e-01 -1.29820421e-01
1.57848382e+00 -3.58490855e-01 6.19561732e-01 1.45833388e-01
2.55509615e-01 -1.23671365e+00 3.79385382e-01 1.27453551e-01
1.20445561e+00 -1.18544424e+00 3.92254665e-02 -1.54769748e-01
-7.89820254e-01 1.26601160e+00 8.94934356e-01 -6.12897217e-01
5.71355283e-01 -6.84147254e-02 3.62348884e-01 -1.27271878e-05
-2.56329000e-01 -4.02988881e-01 7.31080353e-01 7.85056233e-01
5.03855646e-01 -4.15243860e-03 -9.26476866e-02 2.72948921e-01
-5.57792373e-02 -2.31335387e-01 2.72384703e-01 4.19764042e-01
-2.93080568e-01 -1.02340364e+00 -5.23976922e-01 2.98658580e-01
-1.13229729e-01 -2.92716771e-01 -4.70401436e-01 9.49162662e-01
3.09641212e-01 6.43009484e-01 1.28330216e-01 -4.34042603e-01
4.03894961e-01 -2.74206728e-01 7.03553200e-01 -2.30497047e-01
-6.85941875e-01 3.68837625e-01 -3.76082689e-01 -9.59578216e-01
-4.54704911e-01 -4.87719446e-01 -1.18159211e+00 -4.01636213e-01
-7.96060413e-02 -2.02020612e-02 3.13648671e-01 4.11145955e-01
1.62901863e-01 6.88542008e-01 7.05369413e-01 -1.29058862e+00
-1.45482942e-01 -7.98741221e-01 -9.17453349e-01 8.80499363e-01
3.45197201e-01 -4.36585009e-01 1.53349534e-01 4.02452260e-01] | [11.468721389770508, -0.9932047724723816] |
0aa81ba5-cc22-48ed-9d93-143a755fcd4d | a-multi-size-neural-network-with-attention | 2105.03278 | null | https://arxiv.org/abs/2105.03278v1 | https://arxiv.org/pdf/2105.03278v1.pdf | A Multi-Size Neural Network with Attention Mechanism for Answer Selection | Semantic matching is of central significance to the answer selection task which aims to select correct answers for a given question from a candidate answer pool. A useful method is to employ neural networks with attention to generate sentences representations in a way that information from pair sentences can mutually influence the computation of representations. In this work, an effective architecture,multi-size neural network with attention mechanism (AM-MSNN),is introduced into the answer selection task. This architecture captures more levels of language granularities in parallel, because of the various sizes of filters comparing with single-layer CNN and multi-layer CNNs. Meanwhile it extends the sentence representations by attention mechanism, thus containing more information for different types of questions. The empirical study on three various benchmark tasks of answer selection demonstrates the efficacy of the proposed model in all the benchmarks and its superiority over competitors. The experimental results show that (1) multi-size neural network (MSNN) is a more useful method to capture abstract features on different levels of granularities than single/multi-layer CNNs; (2) the attention mechanism (AM) is a better strategy to derive more informative representations; (3) AM-MSNN is a better architecture for the answer selection task for the moment. | ['Jie Huang'] | 2021-04-24 | null | null | null | null | ['answer-selection'] | ['natural-language-processing'] | [ 6.80386182e-03 -5.29336371e-02 1.57413229e-01 -4.39818650e-01
-5.07499218e-01 8.74326304e-02 4.66024250e-01 1.77999258e-01
-4.90759164e-01 6.46746933e-01 6.63876057e-01 -5.41937053e-02
-3.17035586e-01 -1.31719792e+00 -3.90896171e-01 -2.05781162e-01
4.48758423e-01 3.23594719e-01 6.62213564e-01 -8.26969624e-01
6.75733685e-01 1.82350457e-01 -1.81357825e+00 8.45519602e-01
9.29324031e-01 1.16705942e+00 5.25445342e-01 4.24275339e-01
-9.93617058e-01 9.99393106e-01 -7.28509307e-01 -5.23197472e-01
-7.40185380e-02 -8.38068008e-01 -1.24794567e+00 -4.72160727e-01
2.00155765e-01 -8.61633122e-02 -2.99624652e-01 9.97831881e-01
6.19265556e-01 3.38637084e-01 5.26128054e-01 -6.81761563e-01
-1.10704756e+00 8.52987289e-01 -2.46192321e-01 7.08360493e-01
6.02465212e-01 1.71471402e-01 1.12042046e+00 -8.87858033e-01
2.08222091e-01 1.83158541e+00 4.35985565e-01 4.72007096e-01
-6.35626435e-01 -7.05344021e-01 3.52752179e-01 5.95062017e-01
-1.07462752e+00 -2.23011635e-02 6.55695438e-01 -1.53757900e-01
9.71186280e-01 4.65717912e-01 3.99457604e-01 8.60089839e-01
2.13052094e-01 9.46818173e-01 8.72279286e-01 -3.44867915e-01
-9.20341462e-02 1.69163480e-01 7.34315634e-01 4.77044255e-01
1.70826539e-01 -2.92927593e-01 -3.72120798e-01 -1.27929166e-01
5.10499477e-01 1.92355365e-01 -3.81619811e-01 4.47509706e-01
-1.16241133e+00 1.15866089e+00 9.84352648e-01 6.82170153e-01
-4.19801563e-01 3.96641009e-02 5.14963150e-01 6.43263817e-01
2.37766549e-01 7.99165964e-01 -3.44069690e-01 4.09139246e-01
-3.21912050e-01 3.86799574e-01 5.04966915e-01 7.64504671e-01
8.18322778e-01 1.39494672e-01 -8.95765603e-01 7.49263525e-01
3.24953258e-01 6.16189353e-02 1.18697417e+00 -4.75427568e-01
8.81937027e-01 1.30846357e+00 -1.88995391e-01 -1.28738701e+00
-4.79060352e-01 -6.26655996e-01 -9.93083596e-01 -1.32557467e-01
-1.40987774e-02 -1.55683443e-01 -5.43855667e-01 1.63004196e+00
1.46536559e-01 -1.49691895e-01 2.41820529e-01 9.95559037e-01
1.92132688e+00 9.66032803e-01 1.51040554e-01 -3.42398062e-02
1.72029924e+00 -1.06369388e+00 -8.76236558e-01 -4.38661218e-01
2.10599110e-01 -7.70887315e-01 1.18734312e+00 -1.28629550e-01
-1.07801473e+00 -1.32117260e+00 -1.08705807e+00 -2.39040032e-01
-6.20265365e-01 2.13064492e-01 5.15352666e-01 2.77252495e-01
-8.37862313e-01 5.19574404e-01 2.14213565e-01 -4.34372753e-01
2.32971773e-01 4.88888443e-01 -5.73779307e-02 1.41587093e-01
-1.94476926e+00 1.04337692e+00 7.83325791e-01 1.91615164e-01
-7.54086971e-02 -2.45682478e-01 -6.71996117e-01 5.11018932e-01
1.29538104e-01 -1.01188815e+00 9.70501959e-01 -1.13641524e+00
-1.16065824e+00 6.60605967e-01 -2.03210264e-01 -4.61510926e-01
-5.57913668e-02 -8.30724929e-03 -3.95035148e-01 2.87275016e-01
4.56733964e-02 8.58453393e-01 8.03119302e-01 -7.52880335e-01
-6.71869636e-01 -3.09086621e-01 3.81061047e-01 3.76351237e-01
-5.60875475e-01 2.40344599e-01 -1.58730134e-01 -7.08821595e-01
1.99555978e-01 -3.57626736e-01 -1.61239415e-01 -5.51940918e-01
-1.38451278e-01 -9.15387034e-01 5.94553232e-01 -5.83763182e-01
1.62724257e+00 -1.56907368e+00 1.47257820e-01 -9.08467621e-02
1.70210883e-01 5.99093497e-01 -2.29635790e-01 3.67316216e-01
-4.91681434e-02 7.40271136e-02 -3.29707153e-02 1.64567560e-01
-7.80814290e-02 -1.50125310e-01 -3.57777476e-01 -2.94454068e-01
3.89457524e-01 1.17927814e+00 -6.97133303e-01 -4.89941180e-01
-6.22528605e-02 -1.49359524e-01 -4.36338246e-01 4.29987013e-01
-4.36505675e-01 2.28585213e-01 -8.05833697e-01 2.46564627e-01
6.18377030e-01 -2.32239038e-01 -3.65416199e-01 -2.76407182e-01
2.29210317e-01 2.58818418e-01 -1.16490018e+00 1.38970232e+00
-5.25141716e-01 3.46119285e-01 -8.24184418e-02 -8.46610308e-01
1.60832489e+00 3.44561458e-01 -2.65739173e-01 -9.22918081e-01
3.67372364e-01 3.36362064e-01 2.48384774e-01 -1.08864570e+00
5.38808823e-01 -3.08701750e-02 -1.34528741e-01 4.84619141e-01
2.37147644e-01 1.03373691e-01 2.53835231e-01 7.38264918e-02
9.17226732e-01 -1.76454708e-01 2.28221446e-01 -4.48127687e-01
1.31931722e+00 -4.20404792e-01 5.28097153e-01 6.69986963e-01
5.04129790e-02 4.97999668e-01 2.63408542e-01 -6.93817973e-01
-5.69220841e-01 -5.17117321e-01 9.14485529e-02 1.04039335e+00
1.81118309e-01 -2.02254727e-01 -6.67169213e-01 -7.17862904e-01
-4.96077500e-02 5.30239642e-01 -6.21056378e-01 -4.70731765e-01
-6.96437955e-01 -5.96519709e-01 2.51911014e-01 6.21508539e-01
1.10189700e+00 -1.81417704e+00 -3.94277632e-01 1.54972717e-01
-4.61528540e-01 -4.71240282e-01 -4.45486784e-01 -2.19656661e-01
-8.87629747e-01 -1.04975104e+00 -7.16901958e-01 -1.10182571e+00
5.50045431e-01 2.44061396e-01 1.32185280e+00 5.44027627e-01
2.23948315e-01 3.67189646e-02 -5.57752252e-01 -4.76812154e-01
-1.85715154e-01 4.50032592e-01 -4.56550032e-01 4.92414653e-01
8.03104341e-01 -3.62096936e-01 -8.49788308e-01 1.73047870e-01
-9.79602933e-01 -1.63107321e-01 6.46896720e-01 8.95092547e-01
2.33740196e-01 -1.74681544e-01 1.36221814e+00 -9.95129883e-01
1.45188069e+00 -8.81116211e-01 -1.30611658e-01 2.68632203e-01
-3.54705453e-01 2.86174923e-01 9.69518900e-01 -1.65113732e-01
-1.42716348e+00 -4.42770541e-01 -5.37224472e-01 4.60788384e-02
-1.01334013e-01 5.20007253e-01 -1.65246516e-01 8.40602368e-02
6.66149020e-01 2.63391793e-01 -1.47274872e-02 -6.05952740e-01
1.43361285e-01 8.26512098e-01 9.82510522e-02 -4.85544324e-01
3.59669209e-01 -9.18036699e-02 -2.38994047e-01 -4.04829890e-01
-8.26994658e-01 -3.91234845e-01 -3.71870846e-01 -8.00334737e-02
1.03735852e+00 -5.71321070e-01 -8.05207729e-01 2.76679158e-01
-1.62491894e+00 4.09024477e-01 2.58079674e-02 4.59720731e-01
-5.62397093e-02 1.55339226e-01 -7.35662460e-01 -6.85147941e-01
-9.11307275e-01 -1.06734967e+00 6.35356963e-01 8.70879650e-01
-2.50550389e-01 -9.30807531e-01 -4.29090530e-01 6.13455176e-01
8.43338549e-01 -1.13365740e-01 1.20524609e+00 -1.04951859e+00
-6.34414077e-01 -6.28142059e-02 -3.92083317e-01 3.45086426e-01
-8.48109201e-02 -4.77056623e-01 -9.05067980e-01 6.24866597e-02
4.87392098e-01 -3.74649554e-01 1.34148335e+00 7.52014965e-02
1.42121255e+00 -4.03639466e-01 4.78765881e-03 1.81007162e-02
1.34033835e+00 3.72191459e-01 7.89442122e-01 3.98429364e-01
4.90155429e-01 1.01739955e+00 5.08333981e-01 2.25583818e-02
4.40772355e-01 3.81846249e-01 5.05334318e-01 -1.22650012e-01
-3.81062090e-01 -1.87031016e-01 4.41734307e-02 1.16305637e+00
3.27990174e-01 -1.39386982e-01 -5.39963484e-01 5.26413083e-01
-2.01274729e+00 -1.05042708e+00 -5.99157929e-01 2.07444787e+00
7.31229246e-01 2.61033565e-01 -4.88993302e-02 1.45216629e-01
9.62430239e-01 2.20816687e-01 -4.69849825e-01 -6.65435791e-01
-1.56995758e-01 5.81667483e-01 -1.97769165e-01 3.83086950e-01
-8.77282977e-01 5.83756804e-01 5.24159145e+00 1.23531222e+00
-7.09580839e-01 8.55751783e-02 7.78669655e-01 2.79001415e-01
-5.46397150e-01 -1.02822371e-01 -1.04998362e+00 6.88533068e-01
6.87442005e-01 -1.81772441e-01 -7.60801584e-02 5.65605998e-01
-9.55623537e-02 7.03898817e-02 -7.60970592e-01 7.24614143e-01
2.57584840e-01 -1.54003561e+00 5.95517695e-01 -6.99766695e-01
6.52171373e-01 -3.79544079e-01 -1.68131664e-01 7.36706913e-01
-2.45753434e-02 -1.07907033e+00 4.62358356e-01 9.66330469e-01
1.95304379e-01 -9.92533803e-01 1.29808128e+00 4.65013981e-01
-1.40826786e+00 -4.67428654e-01 -9.31502104e-01 -4.35331047e-01
-4.76331115e-02 2.56089151e-01 -1.78311914e-01 7.68599629e-01
7.16758013e-01 4.14228588e-01 -1.01855505e+00 1.02431035e+00
-3.95898759e-01 3.86170208e-01 8.31859857e-02 -7.70943105e-01
4.41127211e-01 -8.57084766e-02 2.42068902e-01 1.22858405e+00
2.56267399e-01 3.15231502e-01 -8.53667483e-02 1.12179518e+00
-2.59312689e-01 5.01556337e-01 -5.13782322e-01 4.74352658e-01
5.22458196e-01 1.23296332e+00 -5.32303035e-01 -3.65795702e-01
-4.22659785e-01 5.78605473e-01 3.47482979e-01 2.69171029e-01
-4.74905521e-01 -1.00969577e+00 2.03264564e-01 7.76511570e-03
2.43492976e-01 4.20798898e-01 -3.51484239e-01 -8.61578763e-01
3.48202884e-02 -9.03994381e-01 8.56719315e-01 -9.22661960e-01
-1.51529634e+00 9.43373561e-01 -2.99330592e-01 -1.09037006e+00
-6.07699016e-03 -5.53275764e-01 -1.10761905e+00 1.38057470e+00
-1.65621531e+00 -9.92511690e-01 -4.06303287e-01 6.25437021e-01
1.02765751e+00 -3.92856032e-01 8.91755998e-01 2.89842099e-01
-4.44384038e-01 5.29480696e-01 -2.79833287e-01 2.02126160e-01
3.25387269e-01 -1.04537666e+00 2.11810231e-01 6.39337361e-01
-3.72519977e-02 9.05694067e-01 2.03654230e-01 -3.79636258e-01
-1.08526146e+00 -9.86409903e-01 1.11673701e+00 -1.47510275e-01
2.59506464e-01 1.69082999e-01 -1.13560688e+00 3.59403253e-01
7.14486539e-01 -4.95059639e-01 6.22760832e-01 1.11048453e-01
-4.91880029e-02 -4.59591925e-01 -1.06797528e+00 6.02791131e-01
4.94288683e-01 -3.20057303e-01 -1.13467169e+00 1.95900202e-01
1.18537676e+00 -2.12974802e-01 -6.36721671e-01 5.12180686e-01
1.11845002e-01 -1.18441248e+00 9.27572608e-01 -7.13105261e-01
6.55586243e-01 -4.74407971e-01 1.34064347e-01 -1.03494895e+00
-6.14812076e-01 -1.75666943e-01 -1.08626708e-01 1.52923632e+00
4.62283522e-01 -7.72976518e-01 5.25200486e-01 3.42867494e-01
-8.62946436e-02 -1.26242948e+00 -7.23736465e-01 -2.90298104e-01
2.16547906e-01 2.09902078e-01 1.08490109e+00 6.05769575e-01
-2.79753447e-01 9.51162398e-01 -3.73963974e-02 -3.55372459e-01
-5.80030680e-02 4.23520446e-01 3.75345409e-01 -1.30537152e+00
-1.68533266e-01 -8.35561275e-01 -1.82649136e-01 -1.37878573e+00
2.78790087e-01 -1.03340638e+00 -2.87455618e-01 -1.82672954e+00
1.39342815e-01 -1.97814375e-01 -5.31293452e-01 -3.86801735e-02
-9.00766015e-01 -1.74862728e-01 2.00312868e-01 2.49220096e-02
-6.80575788e-01 7.78432131e-01 1.49684572e+00 -3.38740706e-01
1.27146915e-01 4.82860923e-01 -1.15692508e+00 6.35848701e-01
9.47474599e-01 -2.99517214e-01 -6.28998280e-01 -6.34347737e-01
4.13279980e-01 1.51055688e-02 6.26190081e-02 -9.60414767e-01
3.95097643e-01 5.73654696e-02 5.93692303e-01 -6.59882963e-01
4.94910143e-02 -5.06841481e-01 -3.91598523e-01 6.38125241e-01
-7.18090475e-01 3.47700566e-01 2.46487305e-01 4.11427736e-01
-6.02821290e-01 -9.07242417e-01 5.31582534e-01 -6.47021890e-01
-8.94017637e-01 2.67209500e-01 -1.74010709e-01 2.37028420e-01
5.71329236e-01 -2.21369714e-01 -4.41554099e-01 -4.27507311e-01
-2.96515614e-01 6.05822086e-01 -3.93516153e-01 8.47606957e-01
1.00210202e+00 -1.73398733e+00 -9.80290890e-01 1.23734772e-01
1.03401095e-01 -4.57455888e-02 5.04420280e-01 4.25272107e-01
-6.09720707e-01 5.07606506e-01 -1.44496739e-01 -1.52920842e-01
-1.08699858e+00 4.86466616e-01 3.90972883e-01 -5.67550004e-01
-2.75490701e-01 1.26335967e+00 2.52037823e-01 -6.28137052e-01
1.46225810e-01 -3.01796257e-01 -1.29217470e+00 2.22941130e-01
6.84480429e-01 3.59053999e-01 9.55745578e-02 -4.95285541e-01
5.89328865e-03 5.58177590e-01 -1.51713833e-01 3.42334211e-01
1.00519025e+00 -6.99470192e-02 -4.62918967e-01 1.66826844e-01
1.05293858e+00 -3.04863006e-01 -5.56711674e-01 -4.51914191e-01
1.06114827e-01 -2.71482110e-01 -3.90178889e-01 -6.16646111e-01
-1.06717968e+00 1.15276384e+00 2.99433559e-01 4.63598073e-01
1.13177931e+00 -1.50165841e-01 1.30274260e+00 4.39998746e-01
3.90976630e-02 -1.09904420e+00 5.70600569e-01 7.79617906e-01
1.33632207e+00 -1.02524829e+00 -3.09466422e-01 -1.79178253e-01
-5.16686082e-01 1.43630660e+00 1.25771821e+00 -3.14673632e-01
5.17530680e-01 -3.61494809e-01 -1.84688464e-01 -4.04104590e-01
-7.68786788e-01 -4.32835549e-01 3.79889905e-01 1.77437797e-01
6.43139064e-01 -2.78131068e-01 -8.50331426e-01 1.19551682e+00
-1.88302740e-01 -3.29989821e-01 1.95305616e-01 5.64927638e-01
-9.61590946e-01 -1.03243554e+00 -3.97099316e-01 8.19430470e-01
-3.06405753e-01 -1.77828863e-01 -4.50459629e-01 2.71704227e-01
5.61230540e-01 1.41527569e+00 -6.96975440e-02 -5.79863310e-01
4.47622776e-01 3.96219380e-02 1.44322798e-01 -8.29373658e-01
-1.34457111e+00 -7.13469267e-01 -1.88868144e-03 -3.08478892e-01
-3.23476344e-01 -8.09402764e-02 -1.28382075e+00 -2.02134728e-01
-4.82748479e-01 4.74070340e-01 9.64730084e-02 1.11249566e+00
5.08795977e-01 1.00249445e+00 5.36422789e-01 -2.07602203e-01
-7.46838987e-01 -1.31006205e+00 -1.57657042e-01 6.23557806e-01
-4.32983972e-02 -3.95673573e-01 -1.47738725e-01 -5.03428936e-01] | [11.003519058227539, 8.15639591217041] |
c5623d5e-21fa-4258-b6a4-855461f5b510 | contrastive-cross-domain-sequential | 2304.03891 | null | https://arxiv.org/abs/2304.03891v1 | https://arxiv.org/pdf/2304.03891v1.pdf | Contrastive Cross-Domain Sequential Recommendation | Cross-Domain Sequential Recommendation (CDSR) aims to predict future interactions based on user's historical sequential interactions from multiple domains. Generally, a key challenge of CDSR is how to mine precise cross-domain user preference based on the intra-sequence and inter-sequence item interactions. Existing works first learn single-domain user preference only with intra-sequence item interactions, and then build a transferring module to obtain cross-domain user preference. However, such a pipeline and implicit solution can be severely limited by the bottleneck of the designed transferring module, and ignores to consider inter-sequence item relationships. In this paper, we propose C^2DSR to tackle the above problems to capture precise user preferences. The main idea is to simultaneously leverage the intra- and inter- sequence item relationships, and jointly learn the single- and cross- domain user preferences. Specifically, we first utilize a graph neural network to mine inter-sequence item collaborative relationship, and then exploit sequential attentive encoder to capture intra-sequence item sequential relationship. Based on them, we devise two different sequential training objectives to obtain user single-domain and cross-domain representations. Furthermore, we present a novel contrastive cross-domain infomax objective to enhance the correlation between single- and cross- domain user representations by maximizing their mutual information. To validate the effectiveness of C^2DSR, we first re-split four e-comerce datasets, and then conduct extensive experiments to demonstrate the effectiveness of our approach C^2DSR. | ['Bin Wang', 'Tingwen Liu', 'Jiawei Sheng', 'Xin Cong', 'Jiangxia Cao'] | 2023-04-08 | null | null | null | null | ['sequential-recommendation'] | ['miscellaneous'] | [ 3.15544009e-01 -5.37469089e-01 -4.84792352e-01 -5.80697417e-01
-5.16270280e-01 -6.77154064e-01 2.85348445e-01 -2.86129452e-02
-3.30458045e-01 6.41537309e-01 4.90128607e-01 -1.48558363e-01
-5.28260767e-01 -6.29266679e-01 -5.49677968e-01 -1.31662458e-01
-1.24344990e-01 3.83633643e-01 8.09363090e-03 -3.97751272e-01
3.69856209e-02 4.97080944e-02 -1.23968363e+00 5.65488935e-01
1.04890823e+00 9.94390607e-01 5.29672623e-01 5.32077670e-01
-1.39666766e-01 5.64054966e-01 -2.67403871e-01 -3.76397640e-01
8.30603391e-02 -5.92942238e-01 -7.67999411e-01 1.82440937e-01
-2.89526999e-01 -4.13934618e-01 -4.95282263e-01 7.42337167e-01
6.23348951e-01 7.23551869e-01 6.19194150e-01 -9.28292453e-01
-1.04030621e+00 7.64901400e-01 -6.71330631e-01 1.22095212e-01
6.64187610e-01 -6.97798505e-02 1.40291095e+00 -7.43886650e-01
3.51754755e-01 1.18364942e+00 4.24635410e-01 5.04112244e-01
-1.10920978e+00 -6.78471625e-01 7.80548036e-01 -4.71129231e-02
-1.13728154e+00 -6.46751747e-02 9.54694629e-01 -2.05983907e-01
8.27994108e-01 1.38145402e-01 3.12584162e-01 1.48874593e+00
-4.11323518e-01 1.23427534e+00 7.29291379e-01 -4.29454632e-02
8.08956288e-03 9.10134986e-02 4.78507906e-01 3.91667813e-01
-8.60820785e-02 -8.67374390e-02 -4.80124414e-01 -2.08338395e-01
8.91060114e-01 5.91538727e-01 -4.05885041e-01 -2.02351600e-01
-1.05842137e+00 8.00001383e-01 3.32109869e-01 4.75204319e-01
-2.56895214e-01 -4.41935897e-01 3.62780303e-01 5.77677548e-01
5.28886914e-01 4.64425206e-01 -9.10066366e-01 -1.43126577e-01
-7.14071572e-01 1.02751859e-01 7.43311405e-01 1.06122804e+00
6.90088809e-01 -3.12051862e-01 -2.66589820e-01 1.36220551e+00
4.36147779e-01 2.54317939e-01 9.93490696e-01 -4.52456594e-01
6.89998746e-01 6.17760181e-01 2.35290751e-01 -1.19709814e+00
-2.18250111e-01 -5.18592298e-01 -8.73429775e-01 -6.52984619e-01
1.79746479e-01 -4.81359512e-01 -6.90981746e-01 1.79160988e+00
2.67231688e-02 2.32515410e-01 1.50614753e-01 1.20737326e+00
7.65514731e-01 8.09025884e-01 6.62153810e-02 -3.58801961e-01
9.86096084e-01 -1.11829770e+00 -4.40148085e-01 -2.49650538e-01
7.00622141e-01 -4.28555578e-01 1.04887497e+00 2.39421979e-01
-8.68686557e-01 -8.16289306e-01 -9.26548183e-01 -5.14177531e-02
-6.49039894e-02 3.26154053e-01 5.32242358e-01 1.73766583e-01
-4.26600963e-01 6.90727890e-01 -4.57874119e-01 -2.73249179e-01
1.34504065e-01 6.93113327e-01 -2.62378216e-01 -2.09385425e-01
-1.63614190e+00 4.62124318e-01 5.28837204e-01 -1.40570253e-01
-2.94994235e-01 -8.18552732e-01 -7.22852647e-01 2.24576518e-01
4.00557816e-01 -7.35944450e-01 1.08422136e+00 -1.33986962e+00
-1.52976608e+00 3.56149435e-01 -2.30282173e-01 -1.42975226e-01
2.28092402e-01 -3.50388169e-01 -7.97138155e-01 -3.97463113e-01
-1.60407320e-01 1.35440335e-01 4.62263227e-01 -1.14100671e+00
-7.89258599e-01 -4.67165858e-01 2.54917264e-01 7.17855394e-01
-6.58700824e-01 -2.14835897e-01 -9.46725726e-01 -9.37134981e-01
-2.22478867e-01 -7.35135376e-01 -2.87484020e-01 -6.21741176e-01
-8.96558017e-02 -2.95351833e-01 4.43331689e-01 -6.90265954e-01
1.62203312e+00 -2.32112122e+00 2.39044040e-01 2.78458387e-01
2.85288729e-02 4.00430620e-01 -4.80782241e-01 4.25790101e-01
-8.14844593e-02 -1.20092884e-01 -8.04979205e-02 -3.46318752e-01
8.42250213e-02 3.21399957e-01 -2.83016205e-01 3.50902230e-02
4.86967489e-02 7.86699414e-01 -1.07682669e+00 -1.22513168e-01
1.04204696e-02 5.05442441e-01 -7.43877470e-01 6.89625025e-01
-3.19340527e-01 4.10953492e-01 -8.62147510e-01 1.83257714e-01
6.50118232e-01 -5.98177731e-01 7.27464974e-01 -2.76224196e-01
2.25408554e-01 4.33039963e-01 -1.25859737e+00 1.98452830e+00
-6.80825770e-01 3.37470695e-02 -2.33218610e-01 -1.14398742e+00
1.06011665e+00 1.42185047e-01 6.89129472e-01 -9.23224330e-01
2.47680023e-02 1.99481681e-01 -2.32639223e-01 -4.64889526e-01
4.18884665e-01 2.12071706e-02 -1.44309968e-01 6.02794707e-01
2.02775583e-01 6.86737657e-01 -3.63904163e-02 2.88134485e-01
9.21915948e-01 1.92296386e-01 4.90504056e-01 8.94049034e-02
6.39986396e-01 -5.15676558e-01 5.33431828e-01 4.93941814e-01
1.06561542e-01 7.69636214e-01 2.38311991e-01 -2.87248641e-01
-4.53178316e-01 -8.27658355e-01 5.18260062e-01 1.51390696e+00
5.87953269e-01 -3.80354404e-01 -1.30855560e-01 -1.27162838e+00
1.57593340e-01 7.01335907e-01 -6.54077768e-01 -3.67282242e-01
-4.75608289e-01 -7.18485773e-01 9.28895548e-02 6.09145761e-01
1.73685223e-01 -9.01841342e-01 1.00638449e-01 3.60890836e-01
-3.57757717e-01 -7.56079435e-01 -1.28416920e+00 1.36429057e-01
-6.39302731e-01 -8.63728762e-01 -9.51211214e-01 -7.43244767e-01
4.08782363e-01 5.50892949e-01 1.17724991e+00 -1.27582969e-02
2.04265058e-01 4.24971431e-01 -7.80818224e-01 2.74666220e-01
2.84159064e-01 3.09338987e-01 1.82311505e-01 3.39841932e-01
5.86131215e-01 -8.60779881e-01 -8.03563297e-01 5.78327954e-01
-7.78671265e-01 -9.63560119e-02 7.52698004e-01 9.47004259e-01
4.62731481e-01 2.20706710e-03 7.12313354e-01 -1.22810781e+00
1.01807976e+00 -1.07715547e+00 -1.49956897e-01 5.12938261e-01
-7.18923032e-01 1.93969026e-01 8.75244737e-01 -6.17248774e-01
-1.19040513e+00 2.66607609e-02 -2.35801473e-01 -6.00813746e-01
-1.80380002e-01 7.58719683e-01 -4.94799584e-01 5.88634431e-01
2.53825337e-01 3.66798639e-01 -3.75174195e-01 -7.22104549e-01
4.07463849e-01 9.50024724e-01 3.22736740e-01 -7.50291049e-01
3.04648280e-01 -3.04997060e-02 -8.31207991e-01 -4.27894622e-01
-1.13772511e+00 -9.92099047e-01 -4.53189760e-01 2.57838994e-01
5.56422353e-01 -9.01614487e-01 -7.43487060e-01 1.00324571e-01
-9.39327061e-01 -3.15399766e-01 -7.51555562e-02 6.20355487e-01
-3.42554361e-01 5.70991814e-01 -4.40940827e-01 -7.84148932e-01
-3.68283808e-01 -8.61070275e-01 9.07267928e-01 2.86974967e-01
-2.47381002e-01 -1.34509563e+00 1.29624218e-01 2.89083660e-01
1.49283767e-01 -3.85680735e-01 5.97219467e-01 -1.20809424e+00
8.72591212e-02 -4.21471409e-02 -4.93104428e-01 2.74216831e-01
5.32671928e-01 -5.41423678e-01 -5.26050508e-01 -4.05921221e-01
-2.85516113e-01 -3.27707410e-01 7.04261601e-01 1.76488519e-01
1.39662886e+00 -3.29337507e-01 -4.34529454e-01 4.34577703e-01
1.21016288e+00 3.33566487e-01 1.41819134e-01 -5.19057177e-02
7.48555720e-01 5.16907632e-01 7.87789822e-01 8.21048319e-01
6.55633867e-01 8.58489871e-01 -1.06407873e-01 1.36989787e-01
1.17261559e-01 -6.21633828e-01 4.50442225e-01 9.66001451e-01
-4.42810543e-02 -5.74045062e-01 -3.65325868e-01 4.57232296e-01
-2.13969207e+00 -8.45371306e-01 3.07492793e-01 2.17158246e+00
8.71241450e-01 2.12655887e-02 3.84770334e-01 -2.96683341e-01
7.38287210e-01 1.15985945e-01 -7.47543335e-01 -1.81171775e-01
1.33737639e-01 9.07044187e-02 3.72138292e-01 3.36975962e-01
-9.62598085e-01 7.17144668e-01 5.04202557e+00 9.77432311e-01
-1.07973659e+00 -1.09429248e-01 5.81398368e-01 -1.83284700e-01
-6.59695208e-01 -3.97380978e-01 -7.65357673e-01 8.47071052e-01
7.97816217e-01 -1.56606421e-01 7.49637067e-01 7.76503265e-01
5.61821535e-02 5.56309581e-01 -1.25882697e+00 1.20471883e+00
1.72437280e-01 -9.11426246e-01 1.07047126e-01 -8.98420885e-02
6.68193102e-01 -1.20751098e-01 1.76250756e-01 6.66559637e-01
7.04963624e-01 -8.11244965e-01 1.77280545e-01 3.42004776e-01
8.07732344e-01 -8.69629622e-01 6.21500015e-01 4.60160106e-01
-1.43619883e+00 -1.64639518e-01 -2.45241314e-01 -6.26097526e-03
2.42740974e-01 5.15791714e-01 -4.90843534e-01 1.08469975e+00
5.42241991e-01 1.32315636e+00 -1.66972429e-01 6.34544075e-01
1.38892189e-01 5.52895010e-01 -1.92635745e-01 -2.86632162e-02
2.47155160e-01 -4.30685818e-01 2.10295409e-01 1.37295043e+00
5.81548810e-01 4.93937284e-01 4.69321340e-01 3.88884455e-01
-3.45696360e-01 1.27646536e-01 -2.90163308e-01 -2.44643912e-01
5.19539893e-01 1.05279016e+00 -1.73057735e-01 -2.01835066e-01
-8.06538939e-01 1.57080805e+00 5.84694326e-01 4.77845132e-01
-9.06560361e-01 -3.79744262e-01 9.89471138e-01 -1.09433189e-01
5.67386270e-01 -1.01690166e-01 -2.09666193e-01 -1.49872649e+00
-1.60183266e-01 -9.28003550e-01 8.62102687e-01 -2.93658733e-01
-1.87890112e+00 4.63298082e-01 -3.15595776e-01 -1.32124114e+00
-1.20381311e-01 -4.66375232e-01 -5.25145292e-01 8.75663340e-01
-1.59507310e+00 -1.14546025e+00 5.13685644e-02 9.03628945e-01
8.35545421e-01 -1.98613673e-01 6.96867704e-01 6.17742598e-01
-6.80328369e-01 1.03743041e+00 3.17695349e-01 1.17604628e-01
1.00682580e+00 -1.11022019e+00 2.88179070e-01 3.92883301e-01
1.09107986e-01 9.11706805e-01 3.41066241e-01 -6.42267406e-01
-1.34829545e+00 -1.35605550e+00 8.28394949e-01 -3.24869990e-01
4.44185644e-01 -1.83410004e-01 -8.58653307e-01 7.72758663e-01
-2.34653265e-03 1.40677122e-02 1.13535750e+00 6.04102910e-01
-4.12572891e-01 -6.06299154e-02 -7.99696147e-01 5.50987363e-01
1.36293054e+00 -4.82760251e-01 -7.51956105e-01 8.84564519e-02
9.34157789e-01 -1.32826641e-01 -9.59515095e-01 3.67757976e-01
7.14991212e-01 -6.69956088e-01 1.11615396e+00 -8.64859819e-01
5.03092527e-01 -1.14034921e-01 -3.31884027e-01 -1.60312581e+00
-5.98602414e-01 -5.29081881e-01 -2.72628695e-01 1.32772195e+00
6.04887426e-01 -5.29554248e-01 7.00661659e-01 6.51194096e-01
-1.43988162e-01 -7.77819395e-01 -1.01556294e-01 -5.81699193e-01
-2.71558344e-01 -2.98100382e-01 7.29203045e-01 1.22077155e+00
3.24292392e-01 6.63026869e-01 -9.33216810e-01 2.11416245e-01
1.58561230e-01 6.00683928e-01 4.85783875e-01 -1.07930708e+00
-9.79294956e-01 -1.73481539e-01 3.10473502e-01 -1.85941994e+00
1.72108635e-01 -7.80419767e-01 -1.87947210e-02 -1.37025559e+00
1.71313137e-01 -5.28555214e-01 -1.02843797e+00 2.18593284e-01
-5.00376165e-01 -6.99545024e-03 -1.52004194e-02 2.32841834e-01
-1.09353459e+00 6.90720558e-01 1.34741437e+00 8.15934390e-02
-6.54080749e-01 2.87723303e-01 -1.09887064e+00 3.85912359e-01
6.11996651e-01 -1.32677898e-01 -9.27785099e-01 -6.59102976e-01
1.66653439e-01 2.22110942e-01 -1.97502300e-01 -4.81433123e-01
2.30734035e-01 -4.34045553e-01 4.75720227e-01 -2.87949950e-01
2.98885316e-01 -9.16232407e-01 -9.15916264e-02 -4.64235768e-02
-6.91792727e-01 -1.98287740e-01 -1.14324249e-01 1.00076306e+00
-2.73293406e-01 1.54856101e-01 3.99151206e-01 -2.15854183e-01
-8.51749361e-01 7.23951399e-01 -5.19907810e-02 3.31716388e-02
8.39129150e-01 -1.65632907e-02 1.43322140e-01 -3.43445659e-01
-8.87501657e-01 5.54709554e-01 2.10521787e-01 7.92883754e-01
5.76495945e-01 -1.53950441e+00 -4.69617903e-01 2.79948592e-01
2.04819813e-01 -2.97019720e-01 5.89851618e-01 4.45244402e-01
2.84755290e-01 2.78709650e-01 -1.12683401e-01 -1.22604422e-01
-1.18586755e+00 7.82466531e-01 3.35089192e-02 -6.51407719e-01
-2.46949762e-01 1.04631257e+00 3.25334847e-01 -7.96742857e-01
1.59963697e-01 8.29917863e-02 -6.35878623e-01 2.46899739e-01
4.76174921e-01 1.74417645e-01 -1.63091734e-01 -4.46715415e-01
-3.00677001e-01 3.58996421e-01 -5.40365040e-01 1.13845341e-01
1.40309906e+00 -5.49459696e-01 2.56777376e-01 4.28268135e-01
1.49792469e+00 -4.03043926e-02 -1.26601017e+00 -6.83890998e-01
-1.19140871e-01 -6.35785758e-01 -1.94955453e-01 -8.92043412e-01
-1.00838172e+00 6.65976286e-01 2.49218971e-01 -1.77954175e-02
1.33354759e+00 -1.75706849e-01 1.27554345e+00 1.67208835e-01
1.50569186e-01 -1.03362167e+00 2.02735350e-01 6.22820854e-01
5.84013641e-01 -1.30374360e+00 -3.54784995e-01 -2.76357502e-01
-1.12539673e+00 8.99480939e-01 8.76302421e-01 -6.85598329e-02
8.98068726e-01 -2.08248898e-01 -2.05936700e-01 1.63608968e-01
-7.81935096e-01 -2.39491716e-01 7.57332861e-01 2.41150245e-01
8.71224463e-01 1.86658129e-01 -3.72095376e-01 1.56381154e+00
2.80944645e-01 2.90295631e-01 -1.25726163e-01 6.36735320e-01
-1.20558612e-01 -1.57909918e+00 3.91626477e-01 6.02966309e-01
-2.85414457e-01 -1.39647394e-01 -3.62002730e-01 1.71074331e-01
1.18803456e-01 9.43331242e-01 -1.38407871e-01 -9.19452310e-01
5.71917295e-01 -5.59539534e-02 1.71218768e-01 -6.39080048e-01
-5.71129978e-01 1.01655766e-01 2.08831020e-03 -4.98604834e-01
-2.66590178e-01 -5.16351998e-01 -1.09335780e+00 -2.68421173e-01
-2.23624364e-01 3.29367608e-01 4.57897149e-02 9.79872882e-01
7.84451365e-01 5.32007694e-01 9.83900249e-01 -5.81085384e-01
-6.50554836e-01 -9.19340730e-01 -7.42837429e-01 8.93529058e-01
3.05253059e-01 -4.97451872e-01 -1.37902243e-04 -8.11858624e-02] | [10.168294906616211, 5.593576908111572] |
4a8622f2-8af6-4034-bbcd-34069ad3ef86 | efficient-and-robust-question-answering-from | 1805.08092 | null | http://arxiv.org/abs/1805.08092v1 | http://arxiv.org/pdf/1805.08092v1.pdf | Efficient and Robust Question Answering from Minimal Context over Documents | Neural models for question answering (QA) over documents have achieved
significant performance improvements. Although effective, these models do not
scale to large corpora due to their complex modeling of interactions between
the document and the question. Moreover, recent work has shown that such models
are sensitive to adversarial inputs. In this paper, we study the minimal
context required to answer the question, and find that most questions in
existing datasets can be answered with a small set of sentences. Inspired by
this observation, we propose a simple sentence selector to select the minimal
set of sentences to feed into the QA model. Our overall system achieves
significant reductions in training (up to 15 times) and inference times (up to
13 times), with accuracy comparable to or better than the state-of-the-art on
SQuAD, NewsQA, TriviaQA and SQuAD-Open. Furthermore, our experimental results
and analyses show that our approach is more robust to adversarial inputs. | ['Sewon Min', 'Richard Socher', 'Victor Zhong', 'Caiming Xiong'] | 2018-05-21 | efficient-and-robust-question-answering-from-1 | https://aclanthology.org/P18-1160 | https://aclanthology.org/P18-1160.pdf | acl-2018-7 | ['triviaqa'] | ['miscellaneous'] | [ 2.78316230e-01 2.88763791e-01 4.34165955e-01 -6.74890578e-01
-1.40640974e+00 -1.07941711e+00 6.15264833e-01 1.75988674e-01
-5.54881096e-01 8.28444302e-01 2.93014377e-01 -6.51885569e-01
-4.30135019e-02 -9.30583417e-01 -9.49085236e-01 -1.39054403e-01
2.30759919e-01 7.64019668e-01 6.69149458e-01 -7.49871314e-01
1.85136080e-01 1.45126000e-01 -1.24357271e+00 5.75531006e-01
9.73885536e-01 9.16979790e-01 -1.12280525e-01 1.15118420e+00
-1.63611397e-01 1.17703497e+00 -1.06211352e+00 -9.16568458e-01
-2.58377101e-02 -4.43819284e-01 -1.66179371e+00 -4.73732710e-01
9.39886332e-01 -5.51145971e-01 -3.37833643e-01 8.21075678e-01
6.58053458e-01 2.38031104e-01 4.11256462e-01 -8.03864896e-01
-7.59965479e-01 6.59448564e-01 2.52651721e-01 5.40523708e-01
6.87483490e-01 5.20814694e-02 1.31779134e+00 -6.05089664e-01
5.85651100e-01 1.20749247e+00 4.89179820e-01 9.34788406e-01
-1.12371838e+00 -2.45912671e-01 4.25670594e-02 3.84671360e-01
-7.81714916e-01 -5.96832335e-01 5.08876145e-01 1.70031507e-02
1.15846908e+00 7.42354214e-01 -1.00557663e-01 1.17855179e+00
1.98278978e-01 7.61311829e-01 8.39262009e-01 -4.63338763e-01
3.29966635e-01 9.00688544e-02 3.72989774e-01 5.58887541e-01
-2.68300027e-01 -4.03632551e-01 -2.90890455e-01 -3.90198201e-01
-4.99873944e-02 -6.27528191e-01 -2.22820356e-01 1.64800420e-01
-8.40126038e-01 1.11453426e+00 3.48364353e-01 3.28819633e-01
-1.72938347e-01 5.77787347e-02 5.85048318e-01 8.04934084e-01
3.53188306e-01 9.68643129e-01 -7.34783947e-01 -1.91478893e-01
-7.85807312e-01 7.00626791e-01 1.34737861e+00 7.39757121e-01
3.25059623e-01 -2.39992440e-01 -4.56264168e-01 7.77977705e-01
6.23326423e-03 5.45782745e-01 3.08388710e-01 -1.27024543e+00
9.20579910e-01 3.52085441e-01 1.51192307e-01 -9.97498035e-01
-3.73100042e-01 -2.84341693e-01 -4.50319439e-01 -3.10117990e-01
7.12268054e-01 -2.73260713e-01 -6.18682861e-01 1.74401605e+00
1.55626133e-01 -2.66911536e-01 3.53467584e-01 6.45016909e-01
9.84581470e-01 7.76308954e-01 -2.42448244e-02 1.18845671e-01
1.54570735e+00 -1.03775597e+00 -6.37832761e-01 -5.88811994e-01
6.66038930e-01 -6.69473469e-01 1.25443602e+00 2.78791577e-01
-1.40787971e+00 -3.88857007e-01 -8.28452110e-01 -4.03864294e-01
-3.42306823e-01 -3.91840309e-01 3.95386547e-01 7.16216028e-01
-1.10162330e+00 3.87581795e-01 -5.06959319e-01 -2.72832304e-01
2.84948796e-01 3.87475401e-01 -1.07619278e-01 -2.52082705e-01
-1.59556866e+00 1.19690454e+00 1.66331772e-02 2.15332899e-02
-9.52380955e-01 -5.66193879e-01 -8.02787125e-01 1.57631934e-01
5.26165187e-01 -8.91086757e-01 1.78577292e+00 -6.54899359e-01
-1.50496042e+00 5.74155748e-01 -4.03349161e-01 -9.17536438e-01
3.05656940e-01 -4.41033989e-01 -3.81517649e-01 6.54261112e-01
-2.57213041e-03 5.34583151e-01 5.69426239e-01 -1.02084863e+00
-3.46398115e-01 -2.98013806e-01 9.08375919e-01 -4.45978576e-03
-3.27188522e-01 3.99654835e-01 -2.96180159e-01 -2.97634333e-01
-2.16811702e-01 -7.47779548e-01 -3.01979333e-01 -2.93121129e-01
-4.02190804e-01 -3.40358824e-01 5.29505670e-01 -9.61456120e-01
1.22564518e+00 -1.63637292e+00 7.93929920e-02 -4.95473929e-02
-8.10940489e-02 5.27944863e-01 -3.09682280e-01 6.03475273e-01
3.11676949e-01 2.19804302e-01 -3.79880905e-01 -3.71561378e-01
1.85842559e-01 4.60682303e-01 -7.74824321e-01 -7.73879066e-02
4.49011654e-01 1.05050886e+00 -8.29761684e-01 -5.28112113e-01
-1.87724173e-01 1.51733890e-01 -7.68583715e-01 4.03303534e-01
-7.19365180e-01 2.36714676e-01 -6.74705207e-01 2.97836274e-01
4.37392592e-01 -2.09869578e-01 -3.97125892e-02 1.83700487e-01
5.30052781e-01 9.14863050e-01 -6.69088066e-01 1.52939737e+00
-6.17623150e-01 5.66280723e-01 1.91453278e-01 -8.47224057e-01
8.26341450e-01 4.89134461e-01 -2.63251364e-01 -6.79388762e-01
1.64964035e-01 1.84968904e-01 -1.13405855e-02 -8.30716431e-01
6.44597471e-01 -6.03453927e-02 -1.33285657e-01 4.90736157e-01
2.33764067e-01 -3.51822019e-01 5.05559206e-01 4.12732750e-01
1.33588505e+00 -3.12324673e-01 -2.34922886e-01 -3.34255733e-02
9.52514470e-01 4.25053649e-02 9.12561193e-02 1.09034097e+00
-2.45550796e-01 6.59589708e-01 5.98245323e-01 -3.52761269e-01
-7.56217480e-01 -9.85587180e-01 -1.55862877e-02 1.24262953e+00
-1.83972836e-01 -2.15641618e-01 -1.20703828e+00 -9.77456570e-01
-3.06436062e-01 1.08404434e+00 -5.41318834e-01 -1.17192112e-01
-9.29010808e-01 -3.86170238e-01 1.03115416e+00 5.44180393e-01
4.66225654e-01 -1.29481518e+00 -5.41609228e-01 2.27007583e-01
-4.74963903e-01 -1.21982896e+00 -4.53550011e-01 -1.31063461e-01
-8.99286926e-01 -1.02821910e+00 -4.54174370e-01 -8.66014004e-01
4.55728531e-01 -2.48267725e-01 1.52390134e+00 3.18386927e-02
1.96468711e-01 4.27492470e-01 -4.35671955e-01 -3.28914165e-01
-8.81783485e-01 4.19950694e-01 -3.46692801e-01 -2.24964306e-01
4.07268137e-01 -2.68640727e-01 -5.46635747e-01 2.41480991e-01
-1.19849837e+00 -4.77877766e-01 2.61190087e-01 8.43905389e-01
1.82890758e-01 -2.28524402e-01 1.06514990e+00 -1.02450550e+00
1.05566299e+00 -3.93222243e-01 -3.77978951e-01 4.94791120e-01
-3.52152616e-01 1.51473105e-01 1.00195348e+00 -1.75182343e-01
-1.10468566e+00 -3.81395608e-01 -7.25812018e-01 1.86539337e-01
-3.79922390e-01 4.85756516e-01 -1.45940915e-01 6.26469553e-02
9.67031479e-01 4.57154810e-02 -3.39755230e-03 -4.03047174e-01
5.19488156e-01 6.68295681e-01 5.16001165e-01 -5.41577518e-01
7.90793121e-01 3.55671525e-01 -2.63928980e-01 -4.62423533e-01
-1.22281539e+00 -2.48135686e-01 -3.78150612e-01 8.08255076e-02
8.78822744e-01 -4.76746321e-01 -7.28657305e-01 1.11319602e-01
-1.39249265e+00 -1.25156343e-02 -2.27648020e-01 5.46427704e-02
-5.10738850e-01 5.44983983e-01 -9.79067624e-01 -7.30605602e-01
-9.10863638e-01 -9.97426212e-01 8.07792246e-01 1.44065306e-01
-4.23077077e-01 -1.05399477e+00 6.28134459e-02 1.08698595e+00
7.04674959e-01 3.85362618e-02 1.06924927e+00 -1.07470238e+00
-4.21410412e-01 -4.36862797e-01 1.29982993e-01 7.93119848e-01
-3.04488212e-01 -3.36496323e-01 -1.05910885e+00 -2.70896763e-01
3.46461296e-01 -7.96747208e-01 6.84641778e-01 -7.18560442e-02
1.14416933e+00 -5.29946208e-01 2.61265308e-01 5.26541024e-02
1.05252039e+00 -1.73848253e-02 7.63859987e-01 4.14234877e-01
1.70620322e-01 8.04844618e-01 4.99117136e-01 -2.09422782e-01
4.76732045e-01 4.41099435e-01 5.16363680e-01 2.96775281e-01
5.79382330e-02 -1.92064241e-01 4.46625799e-01 8.64180744e-01
3.98540169e-01 -4.90421385e-01 -8.12288940e-01 8.02231848e-01
-1.52986562e+00 -9.89079297e-01 -2.34279677e-01 1.84099281e+00
1.05108893e+00 3.24074060e-01 -1.87498093e-01 1.69955388e-01
2.79284120e-01 3.00708920e-01 -4.28405046e-01 -9.37155902e-01
-1.61060676e-01 5.86920917e-01 4.23059650e-02 7.88550377e-01
-1.04188788e+00 8.82841289e-01 7.51961946e+00 6.43543243e-01
-6.04813337e-01 1.21900879e-01 5.01473248e-01 -1.16818689e-01
-5.35229027e-01 -1.10297620e-01 -5.72494447e-01 2.45391786e-01
1.43088102e+00 -4.38437127e-02 4.35900301e-01 5.83551824e-01
-1.95703268e-01 6.13541193e-02 -1.15716183e+00 1.38936162e-01
3.93713146e-01 -1.24829710e+00 1.84078857e-01 -5.04366398e-01
6.19736612e-01 7.80702457e-02 -4.14191782e-02 4.91690636e-01
3.55788618e-01 -1.16751051e+00 5.12186468e-01 3.63999397e-01
1.96149752e-01 -7.90777147e-01 1.03633416e+00 5.77464163e-01
-3.80754262e-01 -2.72412211e-01 -5.23771286e-01 -8.98412094e-02
3.12387824e-01 2.41148785e-01 -9.53190386e-01 6.09448552e-01
7.10248291e-01 -2.37693593e-01 -8.68879616e-01 5.98513782e-01
-3.78885359e-01 1.08152664e+00 -2.22379550e-01 -3.59059423e-01
3.98384064e-01 2.83463389e-01 4.21817064e-01 9.53875303e-01
-5.39284945e-02 5.30069284e-02 -3.02776724e-01 7.42514610e-01
-3.01660150e-01 1.42803743e-01 -4.38836694e-01 4.13938165e-02
2.50923723e-01 9.50484216e-01 4.16146731e-03 -2.85497487e-01
-4.84440327e-01 9.91068840e-01 5.87918520e-01 3.31471175e-01
-6.26957417e-01 -7.19404697e-01 3.14459234e-01 -6.94639310e-02
4.58683401e-01 -8.03201795e-02 -1.98997870e-01 -9.71038163e-01
5.00684977e-01 -1.29399014e+00 6.13100231e-01 -7.91003048e-01
-1.39992809e+00 9.67766464e-01 -6.88415170e-02 -6.05251014e-01
-5.56166589e-01 -6.22499049e-01 -6.87653840e-01 9.30500805e-01
-1.48128545e+00 -9.31474447e-01 1.87019899e-01 4.94997591e-01
4.07383353e-01 -5.20692952e-02 1.23592126e+00 2.94284105e-01
-2.54192144e-01 8.09516966e-01 1.46649703e-01 2.96540529e-01
5.72027266e-01 -1.34053564e+00 8.16836059e-01 8.36955488e-01
2.66100645e-01 7.43175685e-01 9.53809083e-01 -1.55512616e-01
-1.46857035e+00 -8.44450474e-01 1.44515121e+00 -1.13728607e+00
7.84624159e-01 -4.73554373e-01 -1.14508426e+00 6.52057230e-01
5.76711357e-01 -3.13672394e-01 6.91983521e-01 2.36636162e-01
-5.89656055e-01 -1.75968513e-01 -1.32098973e+00 6.34709597e-01
5.72514176e-01 -7.99449444e-01 -1.18113935e+00 4.84363437e-01
1.16871107e+00 -5.23643553e-01 -8.79396915e-01 3.70884389e-01
2.29919001e-01 -8.50227356e-01 9.34767723e-01 -1.10467684e+00
6.93618774e-01 -3.48605923e-02 -2.05097124e-01 -1.10671890e+00
4.44982760e-02 -5.58821142e-01 -2.97785670e-01 1.22145045e+00
8.32303107e-01 -5.95528662e-01 7.27908611e-01 7.83629894e-01
-8.29835907e-02 -9.06942666e-01 -1.22046852e+00 -5.42648017e-01
6.14772499e-01 -4.05979961e-01 6.66366637e-01 5.62841117e-01
-7.55295828e-02 6.40990615e-01 -2.03155935e-01 3.11398625e-01
9.67739820e-02 1.59423754e-01 6.19879663e-01 -8.27622175e-01
-6.76264107e-01 -9.26376283e-02 -5.21683209e-02 -1.27382302e+00
3.15803528e-01 -7.84701347e-01 1.09652959e-01 -1.81108224e+00
-1.54922396e-01 -5.36118150e-02 -1.13943040e-01 2.20820740e-01
-4.40779567e-01 1.59813896e-01 1.56752869e-01 -3.06496263e-01
-7.36056149e-01 4.22695518e-01 1.28236818e+00 -2.60457307e-01
2.80260295e-01 1.59415439e-01 -9.18828428e-01 5.15111983e-01
9.32182312e-01 -4.26251620e-01 -3.98470491e-01 -1.10097933e+00
5.26124656e-01 1.09762281e-01 2.86688715e-01 -7.40143955e-01
2.77585179e-01 2.68004924e-01 -4.83839698e-02 -5.61402977e-01
4.98188019e-01 -5.71004212e-01 -6.68368042e-01 2.86687851e-01
-8.90425861e-01 1.06906474e-01 3.57484341e-01 4.32938933e-01
-4.08715010e-01 -8.14658344e-01 4.16097641e-01 -1.53996095e-01
-1.11728877e-01 -6.79266378e-02 -4.72347736e-01 6.55846179e-01
5.93780100e-01 4.12443489e-01 -6.64753914e-01 -9.04134035e-01
-5.53917408e-01 6.35541022e-01 -1.43824488e-01 4.91338700e-01
4.62924331e-01 -8.56711864e-01 -8.54470432e-01 -1.74494147e-01
-7.50517547e-02 1.98348835e-01 3.35981339e-01 3.38658422e-01
-6.77590072e-01 7.67097771e-01 2.86134243e-01 -1.61264360e-01
-1.31770802e+00 4.67508703e-01 4.74730164e-01 -5.37270308e-01
-1.22103870e-01 1.10850751e+00 -2.61835307e-01 -8.70613575e-01
2.42459059e-01 -2.78942704e-01 -2.39053085e-01 -2.22566366e-01
7.29873717e-01 3.11661839e-01 3.02746773e-01 -4.34066564e-01
-2.82621115e-01 2.07881644e-01 -3.62837315e-01 -1.58006027e-01
1.10253572e+00 3.39017212e-02 -2.21447304e-01 1.28691450e-01
1.26092732e+00 1.03879251e-01 -7.53689706e-01 -3.66313308e-01
1.16576232e-01 -1.48273125e-01 -3.13034564e-01 -1.02167571e+00
-6.14140749e-01 1.14180982e+00 1.79552034e-01 6.49675131e-01
1.03942335e+00 2.43359998e-01 1.23827732e+00 9.52778041e-01
1.69745758e-01 -9.58279669e-01 5.13729639e-02 9.18022573e-01
1.16283381e+00 -1.08464479e+00 -2.92871088e-01 -2.62895197e-01
-5.23598790e-01 8.61424923e-01 5.92583358e-01 -1.26744166e-01
3.19676787e-01 -6.25492306e-03 4.34334844e-01 -1.73496619e-01
-1.02212346e+00 1.98008895e-01 2.81060457e-01 4.21396255e-01
2.66164750e-01 -2.10132524e-01 -1.80036888e-01 6.43471956e-01
-6.73186600e-01 -3.20407182e-01 5.60429394e-01 8.88578713e-01
-5.30447602e-01 -1.25603271e+00 -3.11604142e-01 4.58192796e-01
-1.03492069e+00 -2.55196869e-01 -6.47028744e-01 4.56467956e-01
-3.99296671e-01 1.59599769e+00 -1.05390906e-01 -2.01989129e-01
6.53354347e-01 4.74602312e-01 3.90804499e-01 -5.43091357e-01
-1.17443955e+00 -7.51320064e-01 6.72585309e-01 -4.82651085e-01
-1.66243225e-01 -4.30688828e-01 -1.11043453e+00 -1.92813084e-01
-3.99290472e-01 4.55973774e-01 4.14976984e-01 1.07932556e+00
4.25458997e-01 4.03655171e-01 5.03439546e-01 -5.41041559e-03
-1.25770235e+00 -1.10983706e+00 -3.03255487e-02 4.04738098e-01
4.12669778e-01 1.28431156e-01 -4.94400740e-01 -1.15778707e-01] | [11.242973327636719, 8.072137832641602] |
54219d37-7441-4af8-b514-6e8ada97c905 | beyond-toxic-toxicity-detection-datasets-are | 2303.15110 | null | https://arxiv.org/abs/2303.15110v1 | https://arxiv.org/pdf/2303.15110v1.pdf | Beyond Toxic: Toxicity Detection Datasets are Not Enough for Brand Safety | The rapid growth in user generated content on social media has resulted in a significant rise in demand for automated content moderation. Various methods and frameworks have been proposed for the tasks of hate speech detection and toxic comment classification. In this work, we combine common datasets to extend these tasks to brand safety. Brand safety aims to protect commercial branding by identifying contexts where advertisements should not appear and covers not only toxicity, but also other potentially harmful content. As these datasets contain different label sets, we approach the overall problem as a binary classification task. We demonstrate the need for building brand safety specific datasets via the application of common toxicity detection datasets to a subset of brand safety and empirically analyze the effects of weighted sampling strategies in text classification. | ['Isaac Kwan Yin Chung', 'Elizaveta Korotkova'] | 2023-03-27 | null | null | null | null | ['hate-speech-detection', 'toxic-comment-classification'] | ['natural-language-processing', 'natural-language-processing'] | [ 4.38981563e-01 -2.23025531e-02 -2.47265846e-01 -2.29522854e-01
-8.75379682e-01 -7.98824608e-01 7.51329541e-01 9.11055982e-01
-2.62957096e-01 4.50013727e-01 5.77089727e-01 -2.32619986e-01
2.10976340e-02 -6.58262789e-01 -2.80729085e-01 -4.42388922e-01
1.25353307e-01 -9.01144668e-02 3.32495630e-01 -2.71970123e-01
4.13922548e-01 1.72057450e-01 -1.35543287e+00 7.52590477e-01
7.05642819e-01 8.30267549e-01 -2.86104351e-01 3.58009249e-01
1.38491228e-01 1.00754547e+00 -8.93837214e-01 -8.40670645e-01
1.05268188e-01 -2.65355259e-01 -5.27172625e-01 3.15399945e-01
7.53715813e-01 -1.01242132e-01 -1.96055695e-01 1.11966383e+00
7.25142121e-01 -5.25444895e-02 7.02718556e-01 -1.35395288e+00
-8.17202628e-01 1.01138079e+00 -4.78009284e-01 3.87797445e-01
4.34679121e-01 3.59845487e-03 1.11986136e+00 -2.74877042e-01
5.83722651e-01 1.33400655e+00 6.41551852e-01 3.94963533e-01
-1.35866952e+00 -8.35497856e-01 1.00806169e-01 1.30887404e-02
-1.14942384e+00 -3.04659188e-01 7.52465606e-01 -1.00990236e+00
4.55954939e-01 4.59491700e-01 3.19990724e-01 1.59213758e+00
2.18331575e-01 4.85344559e-01 1.28823030e+00 -5.15596457e-02
3.93327922e-01 5.83849907e-01 3.47125143e-01 3.07751417e-01
3.01628202e-01 -1.96982458e-01 -5.09403467e-01 -6.05351388e-01
-4.53279555e-01 -8.53736624e-02 -4.30845395e-02 2.32011244e-01
-3.94563437e-01 1.46215737e+00 6.90346584e-02 1.49291426e-01
-8.64514485e-02 -5.30596897e-02 7.29624331e-01 1.56689920e-02
1.14939773e+00 5.95166624e-01 -2.82450289e-01 2.09739268e-01
-7.76289463e-01 6.86147869e-01 8.54205906e-01 6.42445862e-01
3.30181032e-01 -3.42349350e-01 -3.34231228e-01 1.01512265e+00
4.19721544e-01 3.47703069e-01 2.51050323e-01 -5.12400329e-01
4.47837681e-01 4.64073867e-01 1.74414745e-04 -1.43330407e+00
-4.72451568e-01 -5.00711024e-01 -1.63495794e-01 -6.14959039e-02
2.18038648e-01 -2.79415041e-01 -6.84857309e-01 1.57599020e+00
3.00297499e-01 -2.07477361e-01 -4.34396625e-01 3.87663871e-01
6.06374025e-01 6.06845617e-01 6.94354534e-01 -2.57657170e-01
1.46743834e+00 -4.14049238e-01 -7.98878133e-01 -9.22167897e-02
9.35593903e-01 -9.17832553e-01 8.24764132e-01 6.65468991e-01
-5.92486858e-01 -6.80461410e-04 -1.19346988e+00 8.45789909e-02
-8.75348985e-01 -6.61827445e-01 2.38024235e-01 1.17847633e+00
-4.95262921e-01 5.19069552e-01 6.52828738e-02 -3.20191920e-01
7.16355264e-01 4.95450047e-04 7.62306005e-02 -1.21180452e-01
-1.47098708e+00 1.04572487e+00 5.41992262e-02 -3.77842337e-01
-7.49221444e-01 -7.90663302e-01 -7.06475317e-01 -3.81698489e-01
3.91408652e-01 6.58651069e-02 1.22264743e+00 -8.48868549e-01
-6.46921098e-01 7.46564984e-01 4.54879135e-01 -3.85188699e-01
2.91450292e-01 -6.41440451e-02 -8.86728287e-01 -1.10286519e-01
4.48768467e-01 2.65524894e-01 6.73693597e-01 -1.13756323e+00
-5.48606694e-01 -5.88815570e-01 4.83092628e-02 -3.00606042e-01
-7.52287328e-01 6.03496194e-01 2.82290667e-01 -9.03495729e-01
-5.80630004e-01 -1.00608540e+00 -5.52788116e-02 -3.70175302e-01
-6.33207738e-01 -3.52069765e-01 1.05713546e+00 -6.98386908e-01
1.74833393e+00 -2.27824807e+00 -2.50869811e-01 2.02568695e-01
3.83694410e-01 1.41515762e-01 -1.00669444e-01 5.34380674e-01
-2.29126457e-02 7.71503985e-01 -1.59708843e-01 -2.71960557e-01
1.30106553e-01 3.37841585e-02 -4.06928897e-01 6.00818574e-01
1.45972997e-01 3.19102138e-01 -8.70698631e-01 -4.47960407e-01
-1.67422801e-01 4.98795182e-01 -5.29873073e-01 -2.60686159e-01
-5.40950119e-01 5.76360263e-02 -4.65249032e-01 5.31225562e-01
6.85775578e-01 5.41363582e-02 -1.11848935e-01 8.97833407e-02
-2.34005019e-01 5.05275428e-01 -7.77600646e-01 9.21388268e-01
-8.07512403e-02 6.21252179e-01 -1.24639407e-01 -4.56519604e-01
6.88486516e-01 2.29544550e-01 5.57979047e-01 -6.63219452e-01
3.75421524e-01 -8.42258185e-02 9.56764594e-02 -7.79539347e-01
4.13032830e-01 -4.44830149e-01 -5.10919869e-01 5.23870647e-01
-3.18452537e-01 2.20590428e-01 2.85011798e-01 3.22403580e-01
1.18572545e+00 -3.47150356e-01 -1.13665588e-01 -2.99494684e-01
3.55460823e-01 1.93002686e-01 2.91196644e-01 4.68526632e-01
-3.86067480e-01 5.72486222e-01 6.88536823e-01 -1.95415884e-01
-7.32954979e-01 -5.50919056e-01 -4.71532077e-01 1.51883638e+00
-1.57946706e-01 -9.28349376e-01 -1.04493320e+00 -9.76120293e-01
2.00842902e-01 1.01560283e+00 -8.71325016e-01 -2.05366194e-01
-1.51222602e-01 -1.03181815e+00 6.74729943e-01 -5.29609099e-02
6.53789937e-02 -9.52992558e-01 -2.28615269e-01 2.67442256e-01
-2.90468279e-02 -1.07345617e+00 -6.77532792e-01 3.05281252e-01
-1.96814597e-01 -1.19530392e+00 -2.72274703e-01 -3.81128013e-01
6.97994903e-02 2.15005681e-01 9.05142426e-01 -1.42415404e-01
-2.90751070e-01 2.70288587e-01 -5.59791505e-01 -8.00988019e-01
-7.37740636e-01 2.87934452e-01 -2.43914023e-01 3.61538261e-01
3.90274763e-01 -2.28455052e-01 -2.88600743e-01 3.10745806e-01
-1.40739822e+00 -3.12909901e-01 -1.09678671e-01 1.88108981e-01
-8.52308720e-02 1.98114410e-01 6.69073641e-01 -1.42354071e+00
1.09440601e+00 -1.04798067e+00 -1.88786820e-01 -8.12847912e-02
-6.72882676e-01 -1.88890368e-01 3.38714093e-01 -6.15640998e-01
-8.16679835e-01 -1.23108841e-01 -1.37489676e-01 1.64403200e-01
-6.08195364e-02 6.82956398e-01 -1.32679433e-01 1.33885562e-01
1.03678727e+00 -3.47099185e-01 -2.80509815e-02 -4.96955514e-01
2.03665912e-01 1.05905998e+00 -2.33507112e-01 -2.66577452e-01
6.67105019e-01 1.78441569e-01 -3.14851761e-01 -9.54018295e-01
-1.31225288e+00 -4.69157100e-01 -9.91144180e-02 -3.49096566e-01
9.86153126e-01 -6.88061476e-01 -5.34482300e-01 2.05655366e-01
-1.11370838e+00 2.22355872e-02 1.04039736e-01 -4.37017046e-02
1.42184660e-01 4.56383258e-01 -5.83716571e-01 -9.69085276e-01
-2.14434326e-01 -1.00471950e+00 7.61784852e-01 -3.54276180e-01
-7.39229202e-01 -8.39671075e-01 3.11328232e-01 6.14337742e-01
2.89665163e-01 6.50870502e-01 1.01207864e+00 -1.00399494e+00
9.03163850e-02 -3.68167132e-01 1.18777320e-01 3.94234657e-01
1.47679765e-02 -3.35474610e-02 -1.06361604e+00 2.12484803e-02
1.24391064e-01 -4.63084638e-01 8.13562930e-01 1.31994620e-01
1.00909388e+00 -6.01589680e-01 -1.88290760e-01 -9.34557095e-02
1.25422168e+00 2.05338851e-01 6.47083402e-01 4.00340110e-01
6.98949993e-01 1.17429030e+00 4.38549370e-01 7.28331685e-01
1.36183515e-01 6.64629579e-01 5.25131524e-01 1.95642039e-01
1.60768673e-01 -1.51722655e-01 5.17436326e-01 2.98151433e-01
5.47911942e-01 -6.30199313e-01 -8.23568165e-01 2.75427103e-01
-1.69756436e+00 -1.10170317e+00 -4.54107940e-01 1.87158990e+00
9.22186911e-01 1.56255931e-01 6.91367328e-01 3.64601195e-01
9.60261464e-01 2.86920279e-01 -1.53265655e-01 -7.04259872e-01
-8.67819116e-02 -8.86103585e-02 7.07493246e-01 3.92370015e-01
-1.34203434e+00 5.43243527e-01 6.76836109e+00 8.18633139e-01
-9.96708572e-01 3.46375942e-01 9.94942188e-01 -2.46114671e-01
-4.64141220e-01 -3.63856554e-01 -8.55072379e-01 9.38731492e-01
1.16536736e+00 2.00607404e-02 1.45982027e-01 6.48201942e-01
3.41321796e-01 -2.20311973e-02 -8.25419128e-01 6.77075863e-01
3.54165316e-01 -1.01832867e+00 -1.11925051e-01 2.68721700e-01
5.58637559e-01 -9.60609913e-02 3.79798412e-01 2.11582467e-01
3.20307612e-01 -9.38692689e-01 9.31896806e-01 -1.26665682e-01
3.31804812e-01 -7.64789164e-01 3.96705538e-01 1.99837551e-01
-4.24023360e-01 -3.62705767e-01 9.20225400e-03 -8.26509222e-02
5.55511005e-02 5.85335732e-01 -6.53518379e-01 -2.10035518e-01
5.61498582e-01 7.88928151e-01 -8.98405194e-01 9.18411374e-01
1.56343475e-01 6.63947701e-01 -1.16624013e-01 -3.50572556e-01
3.27146411e-01 9.68041047e-02 3.08124691e-01 1.52559590e+00
-1.47478795e-02 -2.25758865e-01 9.30904821e-02 5.57777941e-01
-4.39185761e-02 3.43279004e-01 -6.29661500e-01 -4.74569887e-01
1.45987064e-01 1.21345198e+00 -6.25755370e-01 -6.02189004e-02
-6.54739201e-01 6.23045087e-01 -6.46678789e-04 2.99688000e-02
-1.05554461e+00 -1.14397824e-01 6.71435058e-01 7.06997335e-01
2.57017892e-02 1.18333086e-01 -4.63069737e-01 -6.19033456e-01
-3.58482927e-01 -1.17920041e+00 6.26147807e-01 -4.71351475e-01
-1.60736477e+00 3.10978323e-01 1.84654206e-01 -9.98866081e-01
3.28565329e-01 -4.83792871e-01 -2.81423926e-01 5.85731983e-01
-1.08655882e+00 -7.73107529e-01 -1.25413407e-02 3.78849685e-01
5.13777256e-01 1.08922869e-01 5.03689826e-01 5.33686996e-01
-6.98612750e-01 4.54143703e-01 -1.57486722e-01 4.45949212e-02
8.97598982e-01 -9.23193693e-01 8.14616308e-02 3.86483997e-01
-2.06892610e-01 3.16857159e-01 9.89075959e-01 -1.03632998e+00
-9.46533322e-01 -1.30494821e+00 9.82946694e-01 -7.45484233e-01
1.07825971e+00 -6.68119371e-01 -6.18328631e-01 2.25226209e-01
3.87692034e-01 -5.80479026e-01 1.22246790e+00 2.68536266e-02
-7.46879101e-01 1.46394983e-01 -1.38827395e+00 6.32066905e-01
8.40532064e-01 -5.69987476e-01 -1.80325672e-01 6.54558480e-01
7.24029124e-01 2.77455091e-01 -7.97917545e-01 -1.06658861e-01
3.00635189e-01 -5.95580935e-01 5.79622567e-01 -6.97521925e-01
7.34516799e-01 -5.52365184e-02 -2.26614714e-01 -1.17040086e+00
-5.60682416e-01 -6.40138388e-01 2.01574057e-01 1.49636805e+00
6.61143482e-01 -2.99939841e-01 5.61489642e-01 9.82664764e-01
8.43040049e-02 -3.55981588e-01 -6.30772173e-01 -2.88090914e-01
5.90757243e-02 -4.91828024e-01 1.97738335e-01 1.12436891e+00
1.63323298e-01 8.03667784e-01 -5.01130939e-01 -5.41881751e-03
5.38714111e-01 -3.92281562e-01 1.89072326e-01 -1.49480414e+00
-1.27030164e-01 -3.25672299e-01 -3.80240262e-01 -2.50190198e-01
4.62278229e-04 -9.20807719e-01 -1.44485816e-01 -1.03532267e+00
4.90658760e-01 -2.41335601e-01 -2.15067610e-01 3.27115655e-01
5.93251921e-02 6.08643591e-01 2.17792585e-01 4.09739232e-03
-6.65589869e-01 1.73145905e-02 8.80844414e-01 -5.29073060e-01
1.01054604e-04 -1.27519757e-01 -1.33189154e+00 5.17263412e-01
9.08592463e-01 -7.68737257e-01 -3.41313928e-01 1.02792054e-01
7.61019766e-01 -5.17629027e-01 9.46211070e-02 -7.39151299e-01
-2.45657414e-01 -1.37641251e-01 8.39086398e-02 -3.88732374e-01
7.92785659e-02 -7.68169463e-01 1.49205402e-01 3.95567179e-01
-9.45421517e-01 -5.35686277e-02 1.35536894e-01 7.89754212e-01
8.07798356e-02 -3.18294466e-01 7.63968825e-01 -5.82285263e-02
-2.39669070e-01 1.97993085e-01 -9.48271573e-01 2.72384107e-01
1.01024294e+00 -1.23758018e-01 -6.15373135e-01 -3.52985382e-01
-8.40997517e-01 -3.99579629e-02 2.39303246e-01 6.85447931e-01
2.15013012e-01 -1.06326973e+00 -8.25121522e-01 -3.81793261e-01
4.09302086e-01 -7.58137405e-01 -1.56842768e-02 6.32609904e-01
-1.70495406e-01 1.51496157e-01 1.50437251e-01 -1.25138760e-01
-1.51181650e+00 9.91495788e-01 1.42702997e-01 1.02050804e-01
-8.62807184e-02 5.65399766e-01 2.37235263e-01 6.67246580e-02
3.70719731e-01 6.12584576e-02 -5.26016653e-01 6.27696574e-01
7.63098776e-01 5.51690459e-01 1.25041470e-01 -8.27159524e-01
-2.12484404e-01 -1.26232291e-02 -3.69796962e-01 -7.97606818e-03
1.25521469e+00 6.55574277e-02 -1.27340227e-01 4.23849910e-01
1.61510193e+00 2.60424912e-01 -8.28772783e-01 6.82973191e-02
4.17337328e-01 -3.54787558e-01 3.73328000e-01 -1.07267797e+00
-8.52937043e-01 4.95270461e-01 7.29327500e-01 9.76728380e-01
7.33365536e-01 1.51299611e-01 7.28598773e-01 -2.45397404e-01
-1.55252129e-01 -1.25597620e+00 2.29181334e-01 2.57362068e-01
8.93159807e-01 -1.05716085e+00 1.55305535e-01 -8.12600136e-01
-6.98466301e-01 6.79831266e-01 2.85341799e-01 1.64161026e-01
6.81657851e-01 2.35702068e-01 1.58771425e-02 -4.72540110e-01
-5.40868700e-01 -2.01759264e-01 2.26921752e-01 5.29080212e-01
6.51483238e-01 3.39137278e-02 -7.87465870e-01 7.25159287e-01
4.25432250e-02 -3.63486260e-01 7.03303277e-01 8.84569943e-01
-4.35769826e-01 -1.21088731e+00 -3.65596890e-01 6.24524176e-01
-1.04215622e+00 -2.48355865e-01 -1.15008402e+00 4.90699261e-01
5.50688446e-01 1.43412066e+00 -2.41363987e-01 -6.13313913e-01
2.64479727e-01 4.31080312e-02 1.15519956e-01 -8.93153369e-01
-1.22589076e+00 2.81942457e-01 6.74863517e-01 -1.10995859e-01
-3.12052459e-01 -9.10899282e-01 -5.14343441e-01 -4.75288302e-01
-3.85233313e-01 1.00672632e-01 7.46942759e-01 6.94128811e-01
4.09317851e-01 3.51533949e-01 6.10528529e-01 -3.37228775e-01
-5.84866047e-01 -9.47536528e-01 -7.08694696e-01 9.17301118e-01
1.97868437e-01 -5.91660917e-01 -6.20818794e-01 5.57625294e-02] | [8.849251747131348, 10.562220573425293] |
fd0b2efa-7e03-47b4-a77b-594b9ca13ecd | refining-a-k-nearest-neighbor-graph-for-a | 2302.11296 | null | https://arxiv.org/abs/2302.11296v1 | https://arxiv.org/pdf/2302.11296v1.pdf | Refining a $k$-nearest neighbor graph for a computationally efficient spectral clustering | Spectral clustering became a popular choice for data clustering for its ability of uncovering clusters of different shapes. However, it is not always preferable over other clustering methods due to its computational demands. One of the effective ways to bypass these computational demands is to perform spectral clustering on a subset of points (data representatives) then generalize the clustering outcome, this is known as approximate spectral clustering (ASC). ASC uses sampling or quantization to select data representatives. This makes it vulnerable to 1) performance inconsistency (since these methods have a random step either in initialization or training), 2) local statistics loss (because the pairwise similarities are extracted from data representatives instead of data points). We proposed a refined version of $k$-nearest neighbor graph, in which we keep data points and aggressively reduce number of edges for computational efficiency. Local statistics were exploited to keep the edges that do not violate the intra-cluster distances and nullify all other edges in the $k$-nearest neighbor graph. We also introduced an optional step to automatically select the number of clusters $C$. The proposed method was tested on synthetic and real datasets. Compared to ASC methods, the proposed method delivered a consistent performance despite significant reduction of edges. | ['Masahiro Takatsuka', 'John Stavrakakis', 'Mashaan Alshammari'] | 2023-02-22 | null | null | null | null | ['graph-clustering', 'spectral-graph-clustering', 'graph-partitioning'] | ['graphs', 'graphs', 'graphs'] | [-4.66063386e-03 -2.27790400e-01 6.09203577e-02 -3.17115515e-01
-6.09688640e-01 -6.93729520e-01 2.99399853e-01 4.61816072e-01
-4.98244673e-01 5.74487329e-01 -9.56628695e-02 3.10610458e-02
-6.57782972e-01 -8.58681560e-01 -2.90377140e-01 -1.13566518e+00
-3.90428752e-01 5.34157038e-01 4.98927921e-01 3.09729308e-01
6.78990841e-01 6.38660669e-01 -1.72654641e+00 3.43437679e-02
1.27483642e+00 7.11590588e-01 3.09035182e-01 2.14910179e-01
-4.00026858e-01 9.49476734e-02 -3.35408121e-01 1.36802280e-02
5.85542619e-01 -3.92329425e-01 -6.12256587e-01 3.17680776e-01
-7.82205686e-02 3.40877235e-01 2.12031975e-01 1.20086575e+00
4.83265400e-01 4.32865351e-01 7.34084189e-01 -1.44812584e+00
-2.34467193e-01 5.73564947e-01 -1.09021842e+00 1.73218660e-02
1.56360388e-01 -7.58302435e-02 8.28351498e-01 -8.91052783e-01
5.53746700e-01 9.66271877e-01 7.95444965e-01 2.75893379e-02
-1.39450955e+00 -7.99399018e-01 -6.33128062e-02 1.16664618e-01
-2.27862310e+00 -3.84262890e-01 1.01112723e+00 -2.78366268e-01
5.66138744e-01 3.02008450e-01 5.27475953e-01 2.30166212e-01
-2.36496791e-01 1.40482903e-01 1.22638214e+00 -3.88524711e-01
5.02379060e-01 3.10378790e-01 2.98106313e-01 5.79551578e-01
2.80897439e-01 -2.71899015e-01 -1.45640254e-01 -5.17346263e-01
2.37621412e-01 5.97631978e-03 -2.91623175e-01 -6.76656187e-01
-9.25989628e-01 8.72826099e-01 4.36874777e-01 2.80643195e-01
-3.60440761e-01 -2.82070935e-01 2.04021737e-01 -4.19650711e-02
-5.25788888e-02 3.55212986e-02 -2.06211075e-01 5.74077554e-02
-1.10818911e+00 -7.30680674e-02 3.05957526e-01 8.83766353e-01
1.20275700e+00 -1.15022130e-01 3.03017437e-01 9.03823793e-01
2.59948760e-01 1.59385055e-01 6.15133584e-01 -7.45281458e-01
3.50858152e-01 1.14008963e+00 -7.84777012e-03 -1.52116013e+00
-4.22121376e-01 -4.29009587e-01 -1.04488349e+00 1.51583210e-01
2.56539285e-01 2.89523136e-02 -9.10543919e-01 1.52446067e+00
4.46639717e-01 3.41081262e-01 -9.09877494e-02 8.22450399e-01
4.04349566e-01 5.82909048e-01 7.87426382e-02 -6.09766364e-01
1.01629567e+00 -2.38976255e-01 -4.64157432e-01 3.74713272e-01
5.60703993e-01 -9.44711685e-01 9.28354084e-01 3.14966649e-01
-8.92633855e-01 -5.94908357e-01 -1.06903696e+00 6.15224361e-01
-5.91883004e-01 1.55600786e-01 3.52495909e-01 8.35117042e-01
-1.08532369e+00 6.42638505e-01 -7.48332500e-01 -4.59409177e-01
2.37524912e-01 7.81948090e-01 -3.29623252e-01 2.50533462e-01
-7.45747507e-01 2.23014638e-01 7.95477808e-01 -3.54145579e-02
-1.95735902e-01 -2.04743460e-01 -5.72848439e-01 -8.71186778e-02
4.18857455e-01 -1.01608418e-01 1.69780731e-01 -9.06376123e-01
-1.06055999e+00 5.94106376e-01 -2.86545247e-01 -2.99204350e-01
1.94796711e-01 4.27207112e-01 -4.04272705e-01 1.85737014e-01
1.12368383e-01 5.76016307e-01 6.81106389e-01 -1.54876125e+00
-6.50040627e-01 -4.96478736e-01 -6.56956315e-01 2.53486663e-01
-4.13344949e-01 -3.07712972e-01 -5.97028732e-01 -4.63614076e-01
8.54945719e-01 -1.15135741e+00 -2.50943273e-01 -6.15978241e-01
-5.87893844e-01 -3.96274686e-01 1.12872910e+00 -2.38070890e-01
1.53162289e+00 -2.20411253e+00 -1.90236062e-01 1.07417190e+00
1.03223912e-01 1.61739275e-01 2.43925586e-01 6.35297656e-01
-3.22504133e-01 3.75377685e-01 -5.77794194e-01 -1.21254213e-01
-2.53051221e-01 -8.06605257e-03 1.55188143e-01 7.20782638e-01
-1.94821462e-01 1.04690664e-01 -6.67420983e-01 -9.13056076e-01
3.95353228e-01 2.31206477e-01 -5.34619868e-01 -1.37249693e-01
1.93422988e-01 2.32004628e-01 -4.36682522e-01 5.15041053e-01
1.16463697e+00 -1.98556796e-01 2.00788453e-01 -5.47521114e-01
-3.67460966e-01 -2.65207708e-01 -2.20872331e+00 1.30020082e+00
9.97842178e-02 -5.64916758e-03 1.53700069e-01 -1.30741584e+00
1.20992601e+00 2.43970245e-01 9.03116167e-01 -1.97957560e-01
1.08546942e-01 2.00354204e-01 1.99385330e-01 -1.94318980e-01
3.53484303e-01 1.55053958e-01 4.26947996e-02 3.80053520e-01
-3.13880712e-01 2.51404703e-01 2.70574838e-01 1.95569083e-01
7.87523031e-01 -1.87472641e-01 3.34940374e-01 -7.43204236e-01
8.41137171e-01 2.26508215e-01 8.67925048e-01 5.28446019e-01
-1.92843884e-01 6.74681127e-01 -3.75466235e-02 -1.75391212e-01
-9.17225838e-01 -1.23440945e+00 -2.18426675e-01 6.38981402e-01
4.14329201e-01 -5.42110682e-01 -7.52814889e-01 -5.45639217e-01
-1.24300085e-01 4.80783194e-01 -3.75584871e-01 -7.84684271e-02
-5.08938789e-01 -1.04999423e+00 3.10754716e-01 2.37511527e-02
8.13721955e-01 -8.52001309e-01 -3.77541929e-01 1.07007362e-01
5.98568656e-03 -5.86872220e-01 -5.41358411e-01 3.26583743e-01
-9.00443316e-01 -1.27643132e+00 -1.73453197e-01 -8.45091403e-01
1.02547944e+00 5.93943417e-01 5.01243711e-01 2.50767976e-01
-3.04239839e-01 -2.88341623e-02 -6.80343449e-01 1.03691570e-01
-1.12061888e-01 9.66871977e-02 2.11748928e-01 3.17467719e-01
6.11967146e-01 -7.62721896e-01 -7.94069648e-01 4.78146046e-01
-8.35681498e-01 -3.88245881e-01 4.45455819e-01 6.57708764e-01
8.73333991e-01 9.27088261e-01 6.23464286e-01 -1.01251483e+00
6.25731409e-01 -5.51071584e-01 -6.15834594e-01 2.90016197e-02
-8.28218102e-01 1.25898942e-01 1.13476515e+00 -7.52957836e-02
-8.31080794e-01 4.99285430e-01 1.19810745e-01 -3.49569440e-01
-2.73575902e-01 3.56719911e-01 -1.72841251e-01 2.92965639e-02
4.68144685e-01 3.74583930e-01 -1.46056116e-01 -4.60302979e-01
1.30039006e-01 9.12653744e-01 3.05840194e-01 -5.16764522e-01
9.04046774e-01 5.66170573e-01 1.77080140e-01 -1.02311921e+00
-2.42529940e-02 -8.97242308e-01 -9.25737917e-01 4.52186912e-02
9.36445653e-01 -5.59427679e-01 -9.38796461e-01 4.12820093e-02
-5.87081730e-01 4.10027385e-01 1.04788788e-01 4.96295720e-01
-2.28691220e-01 7.04401672e-01 -1.13589168e-01 -1.10196543e+00
-2.14272439e-01 -1.18546319e+00 6.06264591e-01 2.67382652e-01
-2.29313314e-01 -6.79964423e-01 -1.74621463e-01 6.72614388e-03
-4.81044315e-03 4.50137556e-01 1.08913553e+00 -7.09581912e-01
-3.14457506e-01 -1.61295265e-01 -1.94990084e-01 -2.70477962e-02
5.44318199e-01 3.65711808e-01 -6.60863876e-01 -5.87436199e-01
-1.94696382e-01 1.54091790e-01 5.65502048e-01 4.16892290e-01
1.32296395e+00 -2.25035131e-01 -7.51060367e-01 5.17360091e-01
1.69712770e+00 7.37875104e-01 5.00163913e-01 1.40856221e-01
5.05452991e-01 5.75966120e-01 4.86634135e-01 5.20485342e-01
8.49828273e-02 5.80814242e-01 2.19682753e-01 -1.32260352e-01
1.13491796e-01 -1.70573175e-01 -5.71895950e-02 9.92488861e-01
9.46021825e-03 -1.84531540e-01 -9.57535982e-01 7.22277284e-01
-1.71229005e+00 -1.06856072e+00 -5.09463787e-01 2.70823359e+00
5.88317096e-01 2.02018246e-01 5.14077306e-01 6.56287313e-01
1.18443477e+00 -1.88011289e-01 -3.25236291e-01 -2.89046437e-01
-1.05702862e-01 3.40696067e-01 6.79000199e-01 5.18039465e-01
-1.07738221e+00 7.43464351e-01 5.26879454e+00 9.94775295e-01
-8.06635737e-01 -1.59010679e-01 5.98099291e-01 2.55769908e-01
-1.59390699e-02 2.28689939e-01 -7.83667088e-01 8.57546329e-01
5.95642984e-01 1.40045628e-01 2.90546179e-01 5.92708886e-01
4.86626834e-01 -3.81611645e-01 -6.06770694e-01 9.91140306e-01
-2.90167660e-01 -1.01247120e+00 1.32231981e-01 6.50569350e-02
7.14340866e-01 -3.60877842e-01 -1.40549004e-01 -1.89296663e-01
3.58571380e-01 -8.98087978e-01 1.72707155e-01 2.46201530e-01
4.85680848e-01 -1.30381668e+00 6.73802853e-01 5.37733197e-01
-1.62907434e+00 -9.99190509e-02 -3.94407719e-01 2.89339900e-01
-1.49516389e-01 5.52726150e-01 -9.78570342e-01 7.29712784e-01
9.48121488e-01 3.62344444e-01 -5.81292272e-01 1.06046319e+00
5.92476368e-01 5.12996614e-01 -6.30118370e-01 3.40799540e-02
3.95809382e-01 -8.32585037e-01 4.49637026e-01 1.15029287e+00
6.06475294e-01 1.95829242e-01 4.49881256e-01 8.26281846e-01
2.93983400e-01 4.13700581e-01 -6.29875481e-01 2.60283768e-01
1.20501363e+00 1.19903040e+00 -1.26561451e+00 -1.47312775e-01
-2.69244581e-01 7.90931404e-01 1.63187549e-01 2.02557519e-01
-5.82755804e-01 -6.75915778e-01 2.07346290e-01 5.13742745e-01
2.23694935e-01 -3.90276790e-01 -2.54263699e-01 -3.55431527e-01
-2.11542398e-01 -6.54873729e-01 7.41052330e-01 -3.31572682e-01
-1.21021819e+00 4.88514781e-01 1.40996426e-01 -1.45263827e+00
9.50192451e-04 -4.11720425e-02 -5.22309422e-01 7.88069665e-01
-7.51858056e-01 -7.28408277e-01 -4.22523826e-01 9.16042328e-01
3.09172899e-01 5.14250696e-02 6.00496590e-01 3.86871099e-01
-5.52102983e-01 5.61294377e-01 4.20777172e-01 5.21591976e-02
7.21410632e-01 -1.26489973e+00 -3.62739474e-01 8.08204770e-01
-2.92342715e-02 1.02363300e+00 7.14579284e-01 -7.49474108e-01
-1.14581370e+00 -8.85511935e-01 4.69375283e-01 1.46191001e-01
2.33751938e-01 -3.09132457e-01 -8.74549448e-01 4.01067734e-01
1.03514813e-01 -1.32326514e-01 7.93729722e-01 -2.02582747e-01
2.19214559e-01 -3.65697116e-01 -1.40937984e+00 6.87667251e-01
8.71564865e-01 -9.52321291e-03 -3.53708863e-01 -1.17072254e-01
2.09545240e-01 2.14005753e-01 -8.68168771e-01 4.20351774e-01
1.56383321e-01 -1.30491149e+00 9.70622540e-01 1.07241333e-01
-3.01922917e-01 -1.07306778e+00 -1.93595603e-01 -1.22716939e+00
-5.36844850e-01 -6.17833853e-01 5.98815501e-01 1.58550942e+00
3.62491786e-01 -5.82241535e-01 1.11016536e+00 3.86810809e-01
-8.86527225e-02 -4.58441973e-01 -9.49004292e-01 -6.78032696e-01
-2.29248077e-01 6.84389984e-03 5.52460790e-01 1.18753564e+00
1.05394244e-01 5.56543246e-02 -1.05900474e-01 3.68814737e-01
9.33565736e-01 2.93163568e-01 6.81782067e-01 -1.41240120e+00
7.30467308e-03 -4.81187940e-01 -3.96308303e-01 -6.81206882e-01
-1.28584415e-01 -9.97068524e-01 -5.05376495e-02 -1.18843198e+00
1.30078673e-01 -1.05499470e+00 -2.89226234e-01 1.77336544e-01
-1.96180627e-01 3.55304897e-01 -8.30805600e-02 4.37209517e-01
-4.32278812e-01 1.94800302e-01 8.15616190e-01 1.91661879e-01
-8.00705373e-01 7.53466925e-03 -3.78600627e-01 6.49252355e-01
9.14093137e-01 -5.47975540e-01 -6.42325282e-01 2.13437021e-01
-1.14102669e-01 3.13567556e-02 -9.54623595e-02 -1.15452003e+00
5.57935715e-01 -1.53379038e-01 4.43776459e-01 -1.07426620e+00
1.51412860e-01 -1.28280425e+00 6.57235384e-01 3.46393138e-01
1.28666991e-02 2.81435162e-01 2.53035203e-02 6.29611075e-01
-2.44402364e-01 -2.92847663e-01 1.06329823e+00 -1.70827657e-01
-6.18364990e-01 1.56235158e-01 -2.77242929e-01 -2.63318390e-01
1.45530403e+00 -8.69522095e-01 7.08532855e-02 -9.99270827e-02
-7.66205966e-01 2.34675273e-01 6.73625290e-01 -8.58649090e-02
4.34336007e-01 -1.30273867e+00 -4.06634182e-01 3.11248779e-01
1.24041595e-01 9.26673934e-02 1.14094131e-01 7.88354874e-01
-5.79915822e-01 2.07859799e-01 -9.49079990e-02 -7.99762070e-01
-1.41879714e+00 7.53568232e-01 8.81748796e-02 8.26124325e-02
-5.65441072e-01 5.44899702e-01 -1.42832786e-01 -2.82985091e-01
9.96948928e-02 5.57463467e-02 -2.55242676e-01 9.97047052e-02
8.82291608e-03 7.11867332e-01 5.01547270e-02 -8.36300492e-01
-6.45003140e-01 1.03025746e+00 1.45963281e-02 1.66464169e-02
1.10228515e+00 -4.33071375e-01 -2.18114555e-01 2.85945356e-01
1.34429812e+00 3.72008890e-01 -9.29109514e-01 3.26091386e-02
2.72318095e-01 -5.74247897e-01 -9.50870067e-02 -3.29668522e-01
-1.03158343e+00 5.28623760e-01 7.66642570e-01 6.05882227e-01
1.35313427e+00 -2.72330433e-01 5.61758518e-01 3.35393846e-02
3.72215003e-01 -1.39715970e+00 -2.56700546e-01 -4.92834859e-02
1.35505930e-01 -1.08451402e+00 1.30711675e-01 -6.59619451e-01
-5.46395183e-01 1.01911521e+00 5.50311387e-01 -4.10339624e-01
8.97911668e-01 1.30751565e-01 -1.77381918e-01 -3.05992424e-01
-2.44463772e-01 -1.09679021e-01 1.25681106e-02 6.05546474e-01
3.74237925e-01 9.11766663e-03 -6.82312965e-01 3.96284819e-01
-2.58961201e-01 -3.66004467e-01 3.22726697e-01 8.19475830e-01
-6.49237633e-01 -1.09122109e+00 -6.71824276e-01 5.66484749e-01
-2.81284869e-01 1.18590578e-01 -4.23365563e-01 1.04328144e+00
5.89111567e-01 1.35766900e+00 1.94907919e-01 -5.32203794e-01
1.20973110e-01 2.34452307e-01 -8.09928551e-02 -2.74216264e-01
-5.65035760e-01 3.17968756e-01 -2.74651021e-01 -4.02408272e-01
-5.31295419e-01 -7.05766022e-01 -1.66433978e+00 -4.37985092e-01
-4.90994662e-01 8.23010266e-01 4.78253871e-01 5.17139852e-01
4.82795715e-01 5.72822094e-02 1.08769703e+00 -3.37326974e-01
-2.30108544e-01 -6.85264349e-01 -8.67754996e-01 6.81482434e-01
-1.41045406e-01 -7.32897997e-01 -6.00114167e-01 2.22272456e-01] | [7.551070690155029, 4.632050514221191] |
5227ebf0-6821-4fa3-b0d7-08cc1be6557a | one-model-is-not-enough-ensembles-for | null | null | https://www.mdpi.com/1424-8220/22/13/5043 | https://www.mdpi.com/1424-8220/22/13/5043 | One Model is Not Enough: Ensembles for Isolated Sign Language Recognition | In this paper, we dive into sign language recognition, focusing on the recognition of isolated signs. The task is defined as a classification problem, where a sequence of frames (i.e., images) is recognized as one of the given sign language glosses. We analyze two appearance-based approaches, I3D and TimeSformer, and one pose-based approach, SPOTER. The appearance-based approaches are trained on a few different data modalities, whereas the performance of SPOTER is evaluated on different types of preprocessing. All the methods are tested on two publicly available datasets: AUTSL and WLASL300. We experiment with ensemble techniques to achieve new state-of-the-art results of 73.84% accuracy on the WLASL300 dataset by using the CMA-ES optimization method to find the best ensemble weight parameters. Furthermore, we present an ensembling technique based on the Transformer model, which we call Neural Ensembler. | ['Zdeněk Krňoul', 'Miroslav Hlaváč', 'Matyáš Boháček', 'Jakub Kanis', 'Ivan Gruber', 'Marek Hrúz'] | 2022-07-04 | null | null | null | sensors-2022-7 | ['sign-language-recognition'] | ['computer-vision'] | [ 3.01615179e-01 -6.05713308e-01 1.71168093e-02 -4.18223500e-01
-7.23316014e-01 -3.88730466e-01 6.61521018e-01 -9.50046003e-01
-5.35009682e-01 3.01843196e-01 2.07572833e-01 5.20112626e-02
-1.06162004e-01 9.57385600e-02 -3.51757169e-01 -1.24381065e+00
9.62627232e-02 2.93738425e-01 3.25425565e-01 -1.26387551e-01
3.30954641e-01 2.92054623e-01 -1.89110291e+00 4.16816473e-01
9.31556225e-01 1.16578269e+00 -1.17616504e-01 7.27149487e-01
-5.60559556e-02 9.25411284e-01 -5.15002847e-01 -3.92725527e-01
2.25989446e-01 -5.56159794e-01 -4.67064232e-01 3.55037898e-02
1.06364644e+00 -3.34180027e-01 -1.50166109e-01 8.93856585e-01
8.46871734e-01 8.29393044e-02 8.93893600e-01 -1.20497191e+00
-3.32390815e-01 2.12524638e-01 -4.59791839e-01 3.43264975e-02
2.91747302e-01 3.89563084e-01 8.78596365e-01 -1.14788389e+00
7.88046002e-01 1.10748696e+00 3.66753280e-01 8.31273437e-01
-9.37032402e-01 -5.64820766e-01 3.65421683e-01 6.70336902e-01
-1.14977074e+00 -5.71490943e-01 7.74143517e-01 -6.12559140e-01
6.91897810e-01 2.70584613e-01 6.41807973e-01 1.21961474e+00
-1.56886399e-01 1.46968317e+00 1.71732557e+00 -7.29083955e-01
1.15691073e-01 -2.61752814e-01 5.12757301e-01 7.76555181e-01
1.56647578e-01 1.63906679e-01 -7.93024123e-01 1.52177259e-01
4.56195176e-01 -3.78374785e-01 -5.17460465e-01 -4.04136300e-01
-1.36113930e+00 1.88845053e-01 -7.12416833e-03 4.79454368e-01
-3.42602283e-01 8.58084708e-02 1.85909376e-01 3.24137568e-01
2.99247950e-01 -9.96259898e-02 -4.64385122e-01 -1.57323405e-01
-8.35111380e-01 5.38984984e-02 6.82883024e-01 6.81153595e-01
2.12995950e-02 1.79466411e-01 -3.14757138e-01 1.01293719e+00
6.44963324e-01 9.40407515e-01 7.20832586e-01 -4.32669967e-01
4.91358399e-01 6.25590742e-01 -1.11342900e-01 -3.08123201e-01
-2.18701228e-01 -1.25873938e-01 -6.19108856e-01 9.52240765e-01
7.70098805e-01 -2.07713768e-01 -1.74362421e+00 1.59338033e+00
6.18357621e-02 3.90965730e-01 6.04350641e-02 1.25007546e+00
1.13942659e+00 3.36887956e-01 1.28591657e-01 1.31654546e-01
1.08890891e+00 -1.12830043e+00 -5.41555643e-01 1.48835406e-01
2.42114916e-01 -6.21489942e-01 8.36450815e-01 8.89161944e-01
-1.01532638e+00 -4.95001674e-01 -8.52425396e-01 2.51258820e-01
-4.29052770e-01 1.00385451e+00 3.05931628e-01 6.48617387e-01
-9.95300114e-01 3.53577673e-01 -8.64171028e-01 -4.68740731e-01
7.96414614e-02 2.55202442e-01 -4.91746336e-01 1.45804724e-02
-4.65212762e-01 1.09722006e+00 2.01175377e-01 4.32942957e-01
-5.82951367e-01 -2.44801924e-01 -4.82030809e-01 -6.55769765e-01
1.43776715e-01 -5.46627223e-01 1.04449499e+00 -1.36747611e+00
-1.87073684e+00 1.30492234e+00 -2.95112580e-01 -5.69992661e-02
9.04424429e-01 -3.10767680e-01 -7.41818249e-01 -9.52961445e-02
-6.15982413e-01 3.74283850e-01 1.25112975e+00 -1.41861713e+00
-7.23112047e-01 -3.71851414e-01 -2.68281400e-01 2.18823820e-01
1.19071767e-01 5.75866044e-01 -5.49068868e-01 -7.17412710e-01
3.73280019e-01 -9.14560437e-01 1.83875293e-01 1.73275262e-01
-1.97455540e-01 -3.54896009e-01 8.60847712e-01 -1.16450965e+00
9.89029408e-01 -2.02024412e+00 6.86096847e-01 3.98532599e-01
1.13880917e-01 7.51790643e-01 -3.85642767e-01 1.13660693e-01
-1.59381509e-01 -2.18362555e-01 -1.51335672e-01 -5.22636831e-01
6.69166073e-02 1.65767968e-01 -1.45717427e-01 5.09651363e-01
3.26396502e-03 8.24398875e-01 -6.18516862e-01 -5.90167999e-01
4.84363228e-01 5.12576878e-01 -2.13167116e-01 3.34979296e-01
-8.42096955e-02 6.82352126e-01 -2.19778866e-01 1.06548440e+00
5.85993886e-01 -5.97934723e-02 4.40403000e-02 -5.86630583e-01
-2.74872482e-01 -1.24648124e-01 -1.21088934e+00 1.77050745e+00
-3.57978344e-01 5.86621404e-01 -1.17636442e-01 -8.55374217e-01
7.77972102e-01 3.41582298e-01 5.11707008e-01 -5.07915616e-01
3.27329099e-01 7.26330757e-01 2.86663443e-01 -8.04890275e-01
-5.53264134e-02 2.89590359e-01 3.86041373e-01 3.87271762e-01
3.88023973e-01 8.18525851e-02 5.13224006e-01 -3.23787302e-01
1.00220764e+00 6.21995211e-01 8.91566128e-02 -2.62617134e-02
7.32206523e-01 -2.34032124e-01 2.72779882e-01 6.51800036e-01
-2.75735080e-01 6.95933223e-01 -3.77627946e-02 -3.24326217e-01
-7.17829764e-01 -1.18912768e+00 -7.67734498e-02 9.54735696e-01
1.32946134e-01 -1.01357251e-02 -4.94083911e-01 -8.73127401e-01
4.45827879e-02 6.82037175e-01 -5.87129712e-01 2.22107530e-01
-8.14348996e-01 -8.17593455e-01 4.94892120e-01 6.64835334e-01
7.37860560e-01 -1.23968983e+00 -8.40114772e-01 -1.20457217e-01
-8.06398764e-02 -9.51675415e-01 -5.25966108e-01 -2.20191032e-01
-5.03432572e-01 -1.17682135e+00 -1.34320617e+00 -8.81938577e-01
7.19229460e-01 -2.92461932e-01 8.34773421e-01 9.40815452e-03
-3.12247902e-01 7.29996502e-01 -7.66705871e-01 -5.36142468e-01
-2.46148303e-01 -4.28630829e-01 6.23786934e-02 7.69637585e-01
2.93516874e-01 -4.99316037e-01 -5.54443240e-01 4.90127295e-01
-5.24137259e-01 2.50706077e-01 8.63185585e-01 1.00322986e+00
5.28491378e-01 -6.90916538e-01 4.24061865e-02 -7.27807060e-02
4.46843296e-01 2.16438040e-01 -5.37262022e-01 9.42996562e-01
-3.85486484e-01 3.79441917e-01 1.85570240e-01 -7.86998212e-01
-1.18084061e+00 2.84342170e-01 -3.07505056e-02 -6.18757069e-01
-3.49552602e-01 2.57196218e-01 -1.45828724e-01 -4.96479005e-01
3.71658921e-01 7.01458633e-01 2.25116741e-02 -7.97622502e-01
3.47695887e-01 8.44540179e-01 6.55933499e-01 -4.39487815e-01
5.74508667e-01 3.89961630e-01 -1.87368169e-01 -8.00123692e-01
-5.34884453e-01 -6.82467759e-01 -8.42209816e-01 -9.50271189e-01
8.93602371e-01 -3.80724102e-01 -5.57393074e-01 1.27334213e+00
-9.94458199e-01 -4.09749776e-01 -2.46658161e-01 7.29978442e-01
-5.13066471e-01 4.43333894e-01 -4.35437225e-02 -1.01109064e+00
-4.31033343e-01 -1.18979788e+00 1.29190218e+00 1.61279470e-01
1.94760971e-02 -5.85706890e-01 5.24737120e-01 3.35576534e-01
3.54038507e-01 3.10032338e-01 4.44425136e-01 -7.00362861e-01
-5.05031586e-01 -3.03736567e-01 -2.46175781e-01 7.49062896e-01
1.59402058e-01 2.81907320e-01 -1.00039518e+00 -2.66397119e-01
-3.98663223e-01 -4.66059178e-01 1.07463336e+00 3.89958322e-01
9.35082972e-01 -5.08680753e-02 -1.26835138e-01 6.55235827e-01
1.22651541e+00 5.01482129e-01 6.25916541e-01 2.07307655e-02
5.08583844e-01 2.59212047e-01 4.73812550e-01 1.89565092e-01
1.14785969e-01 9.82723594e-01 9.75854397e-02 -6.71119466e-02
-6.36097133e-01 -1.64145511e-02 6.27829731e-01 9.96880472e-01
-1.09341025e+00 -3.34000230e-01 -8.82990956e-01 2.85968661e-01
-1.99492335e+00 -1.09001458e+00 -5.08027487e-02 2.06598496e+00
4.64079827e-01 -3.65054846e-01 2.96919465e-01 2.63590038e-01
4.66006726e-01 2.31021136e-01 -6.74516201e-01 4.74029966e-02
-4.57968444e-01 3.33452433e-01 3.91147852e-01 1.43330872e-01
-1.27568233e+00 7.99959779e-01 6.29286432e+00 5.78085423e-01
-1.43837714e+00 1.89543255e-02 1.43065006e-01 -1.46057494e-02
4.35434759e-01 -3.37600857e-01 -7.60426521e-01 5.59767485e-01
3.82403582e-01 9.82545763e-02 5.70617914e-01 6.72207952e-01
1.00134514e-01 -2.02585399e-01 -9.47806776e-01 1.29888582e+00
7.73403227e-01 -7.96243191e-01 1.35431709e-02 -2.44158238e-01
7.08745360e-01 3.81864727e-01 -4.09564935e-02 4.61111031e-02
3.00509870e-01 -8.24223459e-01 9.65401232e-01 1.11133492e+00
1.10295975e+00 1.07061602e-01 7.44825244e-01 2.23003402e-01
-1.30934131e+00 5.06137386e-02 4.14262444e-01 5.34846842e-01
4.88744289e-01 2.53542070e-03 -2.54731119e-01 6.13543451e-01
7.07822204e-01 8.97614837e-01 -6.70930982e-01 1.63086033e+00
-5.54330111e-01 7.84862578e-01 -5.82541466e-01 -3.87580097e-01
-1.17393315e-01 -2.99395055e-01 8.14142227e-01 1.31465650e+00
3.35611790e-01 9.45508629e-02 -6.04792964e-04 6.03088737e-01
3.78077239e-01 9.52374265e-02 -2.34531641e-01 1.55036598e-01
-2.73252636e-01 1.00846434e+00 -2.96213686e-01 -4.69730288e-01
-2.56489664e-01 1.19304621e+00 -1.96175277e-01 5.22915959e-01
-7.71862209e-01 -1.72178775e-01 4.39015567e-01 -5.08265555e-01
2.87116468e-01 -2.42748424e-01 5.34663983e-02 -1.58676183e+00
4.34511840e-01 -1.18687522e+00 4.10677880e-01 -1.15695500e+00
-1.29932034e+00 5.69008052e-01 1.57113001e-01 -1.63197935e+00
-3.62804532e-01 -1.29344034e+00 -7.79827654e-01 7.92922318e-01
-1.44142830e+00 -1.47483230e+00 -6.61921561e-01 6.87556624e-01
5.65249085e-01 -4.54336315e-01 6.50650024e-01 4.19366270e-01
-5.57562470e-01 7.12144852e-01 1.68111578e-01 2.64313757e-01
6.76796377e-01 -1.29897809e+00 4.71411236e-02 1.04394960e+00
5.54031074e-01 6.81382120e-02 3.72392863e-01 -4.09765273e-01
-1.27409244e+00 -7.14221895e-01 9.36568677e-01 -4.61594373e-01
4.61667567e-01 3.81933972e-02 -5.42033970e-01 5.24888813e-01
2.93186656e-03 9.13827419e-02 2.37384811e-01 -2.98657417e-01
-3.70772988e-01 -1.32960854e-02 -1.05970371e+00 5.55297852e-01
1.51238334e+00 -2.13345438e-01 -8.64605784e-01 2.21592441e-01
-3.15702051e-01 -6.49087369e-01 -6.30444884e-01 5.91140032e-01
1.19557750e+00 -8.06771874e-01 9.21979308e-01 -8.86787653e-01
-1.89024583e-02 -5.73556304e-01 -2.10866690e-01 -1.47493339e+00
5.44828437e-02 -2.22050279e-01 -1.59131348e-01 9.60929692e-01
3.60341638e-01 -8.14165711e-01 5.75006902e-01 4.28745925e-01
2.08915193e-02 -5.64401329e-01 -1.13629234e+00 -1.13859200e+00
-3.54937017e-01 -4.18239802e-01 3.53667766e-01 4.03829008e-01
-2.65749007e-01 4.91190739e-02 -3.97364259e-01 -1.14302799e-01
9.44152057e-01 1.86237022e-01 9.45457995e-01 -1.05429149e+00
-2.52440691e-01 -7.41242886e-01 -7.43874490e-01 -1.03706694e+00
1.20996729e-01 -7.67150164e-01 1.83027506e-01 -1.46044874e+00
2.33825035e-02 -1.82307139e-01 -5.57586491e-01 6.08207047e-01
1.63281918e-01 6.59557730e-02 4.45437938e-01 2.25862429e-01
-5.86673081e-01 3.88619184e-01 1.16801417e+00 -4.26957905e-01
-1.01078674e-01 2.51647949e-01 2.71304429e-01 9.46379662e-01
6.04097426e-01 1.70576394e-01 -7.85753578e-02 -5.08806586e-01
-3.75186592e-01 -2.83143938e-01 6.74533546e-01 -1.21680927e+00
1.57972321e-01 -5.39771933e-03 5.29477932e-02 -9.69782114e-01
6.03367448e-01 -7.68660009e-01 1.68461353e-01 4.33081001e-01
-2.24066809e-01 -1.40173197e-01 -1.14671767e-01 2.43802339e-01
-3.15621436e-01 -1.26039669e-01 8.96750867e-01 2.08818376e-01
-1.28651428e+00 1.51643157e-01 -2.08900645e-01 -1.75786689e-02
8.09692562e-01 -4.09444988e-01 -1.73911601e-01 -2.41023943e-01
-9.32400644e-01 1.51693784e-02 8.40892121e-02 4.45938498e-01
7.65015066e-01 -1.27311003e+00 -1.03476655e+00 3.04416746e-01
4.30872113e-01 -4.75488216e-01 8.75444338e-02 1.14455175e+00
-4.62314159e-01 3.15952629e-01 -2.65447557e-01 -7.58159697e-01
-1.97252584e+00 -1.57543525e-01 7.50644803e-01 -3.23006570e-01
-7.23404229e-01 9.79130089e-01 -2.61069417e-01 -3.54288697e-01
5.69532037e-01 -6.67185187e-01 -3.81803662e-01 -1.44508719e-01
5.91243863e-01 4.51721430e-01 7.04000816e-02 -1.03358269e+00
-6.10852361e-01 1.18054461e+00 2.97461241e-01 -3.73833448e-01
1.22975361e+00 4.85325992e-01 1.27105474e-01 5.07056892e-01
8.80784154e-01 -2.48762369e-01 -1.09646082e+00 -2.57175475e-01
-3.87545954e-03 -3.58700514e-01 1.89436655e-02 -1.60804176e+00
-1.24407065e+00 7.77182221e-01 1.27857018e+00 -5.21158576e-01
1.43025732e+00 -8.06740820e-02 3.88369381e-01 3.76366109e-01
5.19865990e-01 -1.20482862e+00 -1.28517061e-01 4.40794259e-01
1.44348896e+00 -1.32266986e+00 -2.13518634e-01 -6.12665713e-03
-8.02532971e-01 1.15619075e+00 4.92569566e-01 -1.25768974e-01
7.20304489e-01 1.18299074e-01 4.05012459e-01 -1.32116184e-01
-4.93601799e-01 -6.91484809e-01 1.01779091e+00 5.08598506e-01
2.49411598e-01 1.62865654e-01 -5.60410023e-01 3.70866925e-01
1.11583434e-01 3.49556208e-01 -2.23893076e-01 8.18819582e-01
-2.36374773e-02 -1.11878145e+00 -4.20406461e-01 5.07106900e-01
7.58013800e-02 1.38016865e-01 -7.64241874e-01 8.66040707e-01
2.98458695e-01 7.12181866e-01 -1.84343994e-01 -5.32317579e-01
6.79137111e-01 5.35369098e-01 1.04801512e+00 -2.97454242e-02
-4.50117558e-01 -1.13976412e-01 3.18970352e-01 -5.16182899e-01
-9.90007222e-01 -1.17601538e+00 -8.13216686e-01 3.18944842e-01
-4.85015422e-01 -4.88284469e-01 6.90540433e-01 1.09719789e+00
-1.15559526e-01 2.45670155e-01 2.73967087e-01 -1.10313427e+00
-7.47919559e-01 -1.15098131e+00 -4.43599701e-01 6.67554617e-01
2.76233703e-01 -9.76604879e-01 -4.17068958e-01 2.37776548e-01] | [9.160294532775879, -6.47140645980835] |
f8d57e7b-25fa-406a-b702-1f6ecea273b2 | information-structure-syntax-and-pragmatics | null | null | https://aclanthology.org/W16-3801 | https://aclanthology.org/W16-3801.pdf | Information structure, syntax, and pragmatics and other factors in resolving scope ambiguity | The paper is a corpus study of the factors involved in disambiguating potential scope ambiguity in sentences with negation and universal quantifier, such as {``}I don{'}t want talk to all these people{''}, which can alternatively mean {`}I don{'}t want to talk to any of these people{'} and {`}I don{'}t want to talk to some of these people{'}. The relevant factors are demonstrated to be largely different from those involved in disambiguating lexical polysemy. They include the syntactic function of the constituent containing {``}all{''} quantifier (subject, direct complement, adjunct), as well as the deepness of its embedding; the status of the main predicate and {``}all{''} constituent with respect to the information structure of the 6utterance (topic vs. focus, given vs. new information); pragmatic implicatures pertaining to the situations described in the utterances. | ['Valentina Apresjan'] | 2016-12-01 | null | null | null | ws-2016-12 | ['implicatures'] | ['natural-language-processing'] | [ 2.18423575e-01 5.56705475e-01 1.28240675e-01 -5.41596055e-01
-7.14227974e-01 -6.49107158e-01 4.65456277e-01 6.63795352e-01
-7.72500992e-01 8.09633374e-01 8.53201687e-01 -5.16354620e-01
-6.30489290e-01 -5.57424247e-01 -2.19716996e-01 -5.43572843e-01
-5.56357913e-02 8.31319690e-01 6.71891212e-01 -7.14855492e-01
2.60577410e-01 3.66930425e-01 -1.44293344e+00 2.87152022e-01
1.27358884e-01 7.31573164e-01 6.06418371e-01 1.66441843e-01
-3.28205287e-01 6.18967295e-01 -1.13436663e+00 -3.66364479e-01
-1.53092131e-01 -2.72149473e-01 -1.19606030e+00 1.46570310e-01
2.40585178e-01 -3.08473110e-01 8.01678672e-02 1.26369596e+00
1.88450620e-01 2.21857324e-01 6.02321684e-01 -7.95992494e-01
-5.25836229e-01 8.71610224e-01 -4.72992025e-02 8.50099266e-01
5.04336298e-01 1.26078919e-01 1.55563438e+00 -2.31108785e-01
7.78227091e-01 1.56856787e+00 1.65067643e-01 6.98365152e-01
-1.14223158e+00 -2.75517553e-01 3.79799753e-01 -3.51600260e-01
-1.12587178e+00 -5.73314786e-01 6.72073305e-01 -5.39588034e-01
1.03060758e+00 4.92188215e-01 3.77297699e-01 9.45179522e-01
6.03844337e-02 6.04943812e-01 9.93084967e-01 -4.59315747e-01
2.15354726e-01 3.66220862e-01 6.70336485e-01 -8.63775834e-02
2.69054592e-01 -1.47173434e-01 -2.78810501e-01 -4.59562153e-01
1.69811055e-01 -4.36603904e-01 -2.97675133e-01 2.64435410e-01
-1.13545871e+00 8.17070067e-01 1.38963684e-01 8.26564908e-01
-5.57178438e-01 1.64602585e-02 5.24438143e-01 2.78510004e-02
-5.61395008e-03 4.66866642e-01 -5.77022254e-01 9.02226269e-02
-2.26791240e-02 7.94016302e-01 6.73158467e-01 8.22191179e-01
5.94913840e-01 -1.21947609e-01 -1.67805865e-01 4.57684338e-01
5.59548974e-01 2.39295378e-01 1.14965484e-01 -1.41079211e+00
7.62123883e-01 4.80476648e-01 5.46222150e-01 -7.51286685e-01
-3.43288511e-01 5.30848652e-03 2.41094306e-02 -1.69542789e-01
5.39395809e-01 -1.93721130e-01 -3.98264080e-02 2.14387012e+00
3.55777949e-01 -7.49388099e-01 5.27277887e-01 8.42790842e-01
9.43573058e-01 7.36615181e-01 7.35180795e-01 -7.73774922e-01
2.12553287e+00 -3.23648825e-02 -9.70525563e-01 -5.45100212e-01
4.86439496e-01 -8.37894857e-01 1.46629107e+00 -2.59530455e-01
-1.47291291e+00 -2.04416215e-02 -8.40540230e-01 -3.00792754e-02
5.04691564e-02 -4.74383622e-01 5.44709206e-01 4.02784407e-01
-8.08963776e-01 3.57133865e-01 -4.30699170e-01 -4.32820857e-01
-2.60319680e-01 4.53155369e-01 -3.54641229e-01 2.20136791e-01
-1.39605284e+00 8.32079887e-01 6.50404930e-01 2.61989683e-01
-2.28915274e-01 -1.56218275e-01 -1.00226390e+00 2.73959279e-01
2.35683277e-01 -4.06330973e-01 1.35412753e+00 -1.29733384e+00
-9.18323040e-01 1.21678996e+00 -2.25354314e-01 -1.32559896e-01
4.01679315e-02 -3.61876749e-02 -5.33640742e-01 3.50234628e-01
8.24798703e-01 3.85133505e-01 9.40152779e-02 -1.03919578e+00
-8.15872788e-01 -7.98266828e-01 6.89231336e-01 4.39397871e-01
6.65282756e-02 1.14134634e+00 4.00186926e-01 -3.22146982e-01
6.23447239e-01 -6.19184017e-01 3.69013250e-01 -3.81431967e-01
-3.94321471e-01 -7.54964709e-01 7.18138218e-01 -4.70709741e-01
1.23067462e+00 -2.36187291e+00 -4.44788225e-02 -2.25499272e-01
9.99729559e-02 3.59072894e-01 3.40629280e-01 7.37404704e-01
-9.06558484e-02 9.72948000e-02 -1.42368451e-01 2.19643950e-01
4.35917258e-01 6.92019403e-01 2.87576821e-02 3.58331919e-01
1.29147023e-01 4.07758445e-01 -6.44445300e-01 -1.66950494e-01
-5.44848060e-03 4.89504755e-01 -2.40513101e-01 -1.58554703e-01
-4.94495422e-01 2.34317720e-01 -7.22195148e-01 2.21526235e-01
2.97050864e-01 2.76758391e-02 3.74049902e-01 3.04535311e-02
-3.71731102e-01 1.19420111e+00 -1.41596413e+00 7.29714870e-01
3.68830055e-01 3.05942714e-01 6.74626648e-01 -6.19748831e-01
5.11242807e-01 6.60055995e-01 -2.94818729e-01 -5.19922256e-01
3.33794981e-01 4.80650008e-01 5.96501708e-01 -3.95294189e-01
1.25036016e-01 -8.53702426e-01 -4.65168178e-01 3.85844409e-01
-6.10816061e-01 3.63278314e-02 1.70830816e-01 2.52794921e-01
8.86859953e-01 1.17813912e-03 3.38216156e-01 -8.56596947e-01
9.80435371e-01 -3.03218663e-01 8.20001781e-01 3.29201519e-01
-4.93573993e-01 9.16052014e-02 7.73217320e-01 -3.32107484e-01
-6.29070401e-01 -1.17485666e+00 -3.92187446e-01 1.28858662e+00
3.59557122e-01 -2.98965663e-01 -7.15558112e-01 -1.28677964e-01
-3.74474257e-01 1.44412768e+00 -4.09936726e-01 3.99977863e-01
-8.91389847e-01 -4.89194363e-01 2.21074432e-01 4.85913724e-01
3.51378322e-01 -1.47584975e+00 -1.14818108e+00 4.07087117e-01
-6.25065267e-01 -9.86016333e-01 -3.14514935e-01 3.18071365e-01
-4.52324748e-01 -7.63787031e-01 2.44290590e-01 -5.36075771e-01
3.61246675e-01 -3.47282946e-01 1.13764155e+00 4.59914386e-01
1.30203934e-02 4.88656700e-01 -2.16333449e-01 -6.11639082e-01
-5.17279744e-01 -6.79471552e-01 -1.86242685e-01 -9.24090147e-02
5.56784213e-01 -5.62111020e-01 -5.49699426e-01 -2.55690217e-02
-1.02728748e+00 -6.25906169e-01 3.54145728e-02 2.47721091e-01
3.79364491e-01 1.28031284e-01 1.21362753e-01 -9.96689260e-01
8.47564638e-01 -1.57638222e-01 -2.49020547e-01 -1.43503189e-01
1.89383581e-01 -5.40558398e-02 -1.03627138e-01 -1.97652832e-01
-1.07245922e+00 -8.47716272e-01 -3.37219685e-01 4.75973755e-01
-5.74313164e-01 1.84471115e-01 -8.17524254e-01 7.89774418e-01
4.30765778e-01 -2.43148983e-01 -2.05159426e-01 -3.17027777e-01
-2.07011834e-01 4.54582930e-01 5.78885555e-01 -8.86560440e-01
2.90652722e-01 3.77846777e-01 9.68171633e-04 -8.05324018e-01
-1.01180565e+00 -3.62398714e-01 -3.53276491e-01 1.81852132e-01
1.27722406e+00 -8.56569886e-01 -1.01659644e+00 -8.41232762e-02
-1.33238173e+00 1.61147311e-01 -3.53717744e-01 5.66099524e-01
-4.84902114e-01 3.37829381e-01 -8.63352656e-01 -9.36053932e-01
-2.41439328e-01 -1.25568819e+00 1.10031652e+00 -7.80075118e-02
-9.07131076e-01 -1.02161729e+00 -6.91889584e-01 3.98831159e-01
2.64896989e-01 3.02868605e-01 1.43999064e+00 -1.38309813e+00
-4.83580291e-01 -1.37621745e-01 -1.53437510e-01 2.94008434e-01
3.48405987e-01 -1.38101757e-01 -6.75078332e-01 -1.27702758e-01
5.74557781e-01 -1.41178099e-02 -8.44474882e-02 1.94731832e-01
-8.62017348e-02 -6.99749529e-01 -1.22919679e-01 -3.81024152e-01
1.45414841e+00 3.72118801e-01 5.89826465e-01 3.83158356e-01
2.02513728e-02 1.01997685e+00 5.26709616e-01 8.67664963e-02
7.10982323e-01 5.89099646e-01 2.81093836e-01 8.19682300e-01
-1.45732284e-01 2.74940908e-01 3.90678287e-01 4.26215440e-01
1.23136878e-01 -1.70529902e-01 -7.63329625e-01 7.37584293e-01
-1.45922649e+00 -9.18753028e-01 -3.86666745e-01 1.90115047e+00
7.69988894e-01 3.58138442e-01 -1.75235912e-01 1.45521507e-01
8.77686083e-01 3.41459572e-01 2.37246335e-01 -6.57840490e-01
-1.77923486e-01 -3.02121826e-02 -2.78226227e-01 9.58079875e-01
-6.55735314e-01 9.42096055e-01 4.52355337e+00 4.30004567e-01
-6.21527851e-01 7.34085813e-02 2.06477329e-01 2.57736653e-01
-7.76299894e-01 6.58255994e-01 -1.17824423e+00 6.34014249e-01
7.00250506e-01 8.08527395e-02 -1.00842446e-01 3.32777798e-01
8.64834860e-02 -5.28438091e-01 -1.42466772e+00 2.24302083e-01
-1.40222400e-01 -1.01145065e+00 -5.87647073e-02 -4.76361904e-03
-3.65283736e-03 -4.92267162e-01 -2.89232910e-01 2.82682572e-02
1.10646479e-01 -2.98506200e-01 1.13746488e+00 -8.20694640e-02
2.35266641e-01 -4.77994084e-01 8.03383350e-01 6.51453495e-01
-8.33689213e-01 -7.05167800e-02 -1.53437406e-01 -5.60440123e-01
3.47127229e-01 1.90977812e-01 -3.20286304e-01 2.05547065e-01
7.28753626e-01 -7.47787237e-01 1.50884107e-01 1.41841918e-01
-7.14656174e-01 3.90579820e-01 -4.51687694e-01 -4.54115570e-01
3.22934896e-01 -1.74952492e-01 1.31458569e+00 8.98534477e-01
-2.25845844e-01 8.78285527e-01 2.00199187e-01 8.18309009e-01
5.55891573e-01 2.91395009e-01 -2.85004854e-01 2.73554921e-01
5.36428630e-01 6.97144508e-01 -9.56777692e-01 -1.48313463e-01
-2.56408334e-01 4.64172423e-01 5.41574135e-02 3.58143926e-01
-3.09370309e-01 -2.25118324e-01 6.58585012e-01 8.90455008e-01
8.44818801e-02 2.39222258e-01 4.07495707e-01 -4.76671636e-01
3.22977394e-01 -4.03810859e-01 5.47590196e-01 -7.63873935e-01
-1.18565190e+00 3.59263003e-01 4.06518966e-01 -1.71121985e-01
-1.18689530e-01 -5.60781837e-01 -4.52809066e-01 1.14679170e+00
-1.18489361e+00 -9.33101952e-01 7.31334984e-01 3.70730430e-01
2.50526309e-01 1.23764090e-01 1.01792037e+00 -9.73586589e-02
8.58638808e-02 -1.24756783e-01 -5.14048040e-01 1.89545006e-01
3.21012065e-02 -9.82899368e-01 -1.03593618e-01 6.98716640e-01
-4.78223443e-01 1.04889071e+00 1.13421679e+00 -5.35208285e-01
-9.13574576e-01 -3.98508459e-01 2.02236438e+00 -6.03614330e-01
8.56893361e-01 -1.42604440e-01 -9.80804682e-01 1.21488690e+00
5.12648761e-01 -3.11789662e-01 7.60646820e-01 7.67120495e-02
-7.67458081e-02 -1.05909780e-01 -1.50716496e+00 6.44772530e-01
6.73492312e-01 -3.88646394e-01 -1.54029107e+00 3.65212619e-01
1.13955033e+00 -3.96976978e-01 -4.06086087e-01 6.12299323e-01
-1.54597135e-02 -1.33798814e+00 7.99090505e-01 -6.47727728e-01
-1.50987998e-01 -2.32500821e-01 -7.16660082e-01 -3.01581025e-01
-3.71479958e-01 -5.80394804e-01 6.99973345e-01 1.63674819e+00
3.51485759e-01 -1.01931059e+00 1.78203776e-01 1.39616346e+00
-3.27150255e-01 -1.28398925e-01 -1.58107114e+00 -4.52702343e-02
3.53611350e-01 -3.75393480e-01 5.78588665e-01 6.15285039e-01
2.76873767e-01 6.49170637e-01 4.44243610e-01 4.58312891e-02
4.66309756e-01 1.26916617e-01 -8.72818828e-02 -1.36595333e+00
-3.84189993e-01 -2.93115586e-01 -2.45674461e-01 -7.26313472e-01
2.89630264e-01 -6.02214456e-01 -8.93105287e-03 -1.57818604e+00
1.40187934e-01 -6.51204586e-01 3.14410799e-03 2.14239806e-01
-1.98301792e-01 -4.68826920e-01 3.17042291e-01 2.49724343e-01
-3.64024103e-01 1.98502541e-01 1.28441072e+00 2.21399516e-01
-1.00995295e-01 1.93791553e-01 -1.21568513e+00 1.18231380e+00
5.57550430e-01 -4.04858977e-01 -4.46270883e-01 -4.47485954e-01
5.77099800e-01 8.94627988e-01 4.82616067e-01 -4.07963336e-01
6.14322871e-02 -2.41193920e-01 5.52520901e-03 -9.11059007e-02
4.53021586e-01 -8.90518129e-01 1.33388981e-01 2.71021873e-01
-4.15301085e-01 2.88457006e-01 3.74784470e-01 1.03912875e-01
-8.03626552e-02 -6.80992603e-01 7.06513762e-01 -6.39420152e-01
-4.46777135e-01 -2.88983822e-01 -6.48381114e-01 3.51963431e-01
9.72201169e-01 -1.59476008e-02 -3.24614674e-01 -1.21606499e-01
-1.23063052e+00 3.58494446e-02 1.06067099e-01 1.53651774e-01
2.88569152e-01 -1.01094091e+00 -4.23573762e-01 -2.77364671e-01
-2.64583528e-01 1.86363772e-01 6.32789955e-02 6.05940580e-01
-4.23834801e-01 4.26848650e-01 2.88734213e-02 -2.59955674e-01
-1.26358771e+00 6.49783075e-01 1.61748990e-01 1.14488319e-01
-5.38636506e-01 1.12786233e+00 5.81012368e-01 -3.05431992e-01
1.15413994e-01 -1.59315690e-01 -1.64386511e-01 2.03867629e-01
3.47244889e-01 1.81271300e-01 -2.68330395e-01 -1.32871079e+00
-5.97870946e-01 1.47518903e-01 2.02288315e-01 -2.71093488e-01
8.69238019e-01 -6.48835778e-01 -9.27665770e-01 8.40440214e-01
9.91064250e-01 2.40404591e-01 -4.87049222e-01 -2.84226626e-01
1.12895839e-01 -4.20775771e-01 -7.61417449e-01 -7.21203983e-01
-1.86203152e-01 7.39146829e-01 -9.29138660e-02 2.84386158e-01
6.43525779e-01 7.98250437e-01 4.13926691e-01 8.16943347e-02
5.85216224e-01 -1.00970745e+00 -1.36396289e-01 2.92192012e-01
1.03217506e+00 -4.83123630e-01 1.29393339e-01 -8.59825075e-01
-6.41608119e-01 8.31189156e-01 3.38325828e-01 1.98762104e-01
2.98636585e-01 3.78895700e-02 1.84724450e-01 -6.01979613e-01
-7.11312056e-01 -2.22225681e-01 -9.51456353e-02 2.25763530e-01
5.59048235e-01 2.90891141e-01 -9.91103470e-01 5.67797661e-01
-4.04294401e-01 -8.66920292e-01 3.65619868e-01 8.30582678e-01
-7.61146069e-01 -1.13663816e+00 -5.47478974e-01 2.33551711e-02
-9.13046002e-01 -1.25663742e-01 -2.67778456e-01 1.12056351e+00
5.87343633e-01 1.08392990e+00 2.55609632e-01 6.93669677e-01
4.53318924e-01 2.04955950e-01 4.89443332e-01 -9.73559320e-01
-8.75386417e-01 6.84184074e-01 6.24580264e-01 -8.52818862e-02
-7.13016212e-01 -1.43426859e+00 -1.87909663e+00 7.74276257e-02
-2.30117947e-01 5.15296459e-01 1.77579448e-01 1.32683599e+00
-3.26557577e-01 2.79942597e-03 -5.65280244e-02 -2.91071177e-01
-7.13872313e-01 -8.11471879e-01 -7.39255607e-01 7.16536701e-01
2.40446299e-01 -4.87000167e-01 -8.83656085e-01 -2.98362046e-01] | [10.161881446838379, 9.285039901733398] |
3c8e92d3-caeb-4721-824d-100c8604153b | local-energy-distribution-based | 2304.11839 | null | https://arxiv.org/abs/2304.11839v3 | https://arxiv.org/pdf/2304.11839v3.pdf | Local Energy Distribution Based Hyperparameter Determination for Stochastic Simulated Annealing | This paper presents a local energy distribution based hyperparameter determination for stochastic simulated annealing (SSA). SSA is capable of solving combinatorial optimization problems faster than typical simulated annealing (SA), but requires a time-consuming hyperparameter search. The proposed method determines hyperparameters based on the local energy distributions of spins (probabilistic bits). The spin is a basic computing element of SSA and is graphically connected to other spins with its weights. The distribution of the local energy can be estimated based on the central limit theorem (CLT). The CLT-based normal distribution is used to determine the hyperparameters, which reduces the time complexity for hyperparameter search from O(n^3) of the conventional method to O(1). The performance of SSA with the determined hyperparameters is evaluated on the Gset and K2000 benchmarks for maximum-cut problems. The results show that the proposed method achieves mean cut values of approximately 98% of the best-known cut values. | ['Takahiro Hanyu', 'Duckgyu Shin', 'Kyo Kuroki', 'Naoya Onizawa'] | 2023-04-24 | null | null | null | null | ['combinatorial-optimization'] | ['methodology'] | [ 5.70868179e-02 -4.26950067e-01 -4.92453158e-01 -3.11190784e-01
-7.91473627e-01 -6.15814805e-01 4.73329902e-01 2.30683506e-01
-5.94759285e-01 1.10352898e+00 -2.97498971e-01 -3.15308511e-01
-3.52540225e-01 -8.73504817e-01 -4.53695655e-01 -1.43908811e+00
7.02936947e-02 1.00221968e+00 4.88748312e-01 -8.70882124e-02
8.76404345e-01 6.60820067e-01 -1.74224257e+00 -1.09358028e-01
1.05030775e+00 1.03978050e+00 7.18684718e-02 7.21734405e-01
-2.25685924e-01 -1.52706653e-01 -8.19739044e-01 -6.99478015e-02
3.52893062e-02 -6.44664586e-01 -9.40446198e-01 -4.90427136e-01
-4.56496738e-02 4.61727530e-01 3.42566520e-01 1.33311200e+00
3.36665511e-01 4.64595258e-01 8.46861303e-01 -1.12405539e+00
1.88538030e-01 7.36009240e-01 -7.38038659e-01 2.72662401e-01
1.62867159e-01 1.15362518e-01 7.27340400e-01 -1.78989977e-01
6.69579089e-01 1.20187426e+00 3.92447174e-01 6.58892393e-02
-1.32948744e+00 -6.44470930e-01 -4.43643808e-01 5.47975183e-01
-1.68877721e+00 -5.30488640e-02 6.28704429e-01 1.79280296e-01
1.11362052e+00 7.56471574e-01 8.54051232e-01 6.27873659e-01
6.80685222e-01 1.82224661e-01 1.47307825e+00 -7.85002053e-01
9.14500773e-01 -2.30107814e-01 6.24398291e-01 6.95351481e-01
6.67069733e-01 7.02775195e-02 -4.88343835e-01 -8.89498413e-01
4.43671405e-01 -6.72372758e-01 8.82824883e-02 -4.86122847e-01
-1.08074224e+00 8.57065797e-01 3.72644931e-01 1.85755994e-02
-2.10886046e-01 4.60884899e-01 3.25209051e-01 -2.22554192e-01
-1.19012199e-01 6.01018071e-01 -2.75635839e-01 -4.77760553e-01
-9.02650177e-01 3.57865542e-01 8.89790773e-01 5.94735920e-01
5.39713621e-01 -1.88618869e-01 -1.15854479e-01 7.98157692e-01
5.03325582e-01 9.29983735e-01 2.75533199e-01 -8.18518877e-01
1.60324574e-01 4.06084806e-01 3.68279368e-01 -6.11106277e-01
-5.58364868e-01 -4.19966340e-01 -8.02767277e-01 2.88224220e-01
5.25728464e-01 -7.41189495e-02 -1.02023482e+00 1.35246384e+00
5.82708538e-01 -2.39271313e-01 -1.14531457e-01 8.93409729e-01
3.47394913e-01 1.16641068e+00 2.58903876e-02 -6.25556886e-01
1.41593647e+00 -9.46409285e-01 -7.23103106e-01 9.40467417e-02
2.40790427e-01 -7.25860894e-01 9.77085948e-01 7.38405645e-01
-1.11572993e+00 7.77519494e-02 -1.48106396e+00 3.51336509e-01
-3.00138950e-01 -2.96318978e-01 8.46272945e-01 1.31606340e+00
-7.08568156e-01 4.17114735e-01 -1.12143338e+00 -1.87853217e-01
-1.59672245e-01 6.76364183e-01 3.87501627e-01 2.12502912e-01
-1.12035489e+00 8.26819181e-01 7.29879081e-01 1.55159935e-01
-5.31279683e-01 -2.60647744e-01 -4.19249207e-01 2.19273776e-01
3.29004645e-01 -6.29652441e-01 8.11657012e-01 -3.48700702e-01
-2.15138006e+00 2.52118379e-01 -2.04181820e-01 -1.68716103e-01
1.71091735e-01 4.73339230e-01 -1.98162720e-01 1.18082024e-01
-5.90979695e-01 2.41474167e-01 4.81116980e-01 -1.20076430e+00
-1.26732275e-01 -3.00138295e-01 -4.73826438e-01 2.54785329e-01
1.97875008e-01 6.54449984e-02 -5.42452872e-01 -1.56472027e-01
4.69137728e-01 -1.27444160e+00 -4.15073395e-01 -8.37071419e-01
-6.47635937e-01 -1.66242108e-01 2.02273250e-01 -5.79146817e-02
1.52695656e+00 -1.69795334e+00 2.05136895e-01 1.17924643e+00
-4.04142976e-01 2.48966441e-02 2.32342601e-01 5.00787258e-01
1.44636258e-01 7.85664283e-03 -3.99174631e-01 4.67018217e-01
1.09876089e-01 -3.46317776e-02 2.15613469e-01 6.35441422e-01
-7.05878913e-01 5.30336678e-01 -6.35009170e-01 -6.27201736e-01
8.34890381e-02 2.48530909e-01 -2.96697676e-01 -2.41575897e-01
-2.84553736e-01 -2.11781442e-01 -4.18850720e-01 4.13518816e-01
1.00234699e+00 -2.30469376e-01 6.48385346e-01 -2.07210690e-01
-8.55272859e-02 1.55819952e-01 -1.58581543e+00 1.32126212e+00
-1.86645120e-01 1.53649766e-02 -3.23147401e-02 -4.89337623e-01
1.15008247e+00 -1.78261548e-01 1.16452351e-01 -5.16743898e-01
4.10844296e-01 2.72566587e-01 5.56972958e-02 -2.25687847e-02
5.09853840e-01 5.30389659e-02 -3.92383844e-01 7.68154562e-01
-2.97362894e-01 -5.07934153e-01 3.89985919e-01 -2.70721433e-03
8.45628262e-01 -1.65212631e-01 4.32085693e-01 -8.75947475e-01
7.55694151e-01 4.42570031e-01 3.80256295e-01 8.60846937e-01
-4.82151248e-02 5.89549690e-02 6.34113729e-01 -4.13045675e-01
-9.87582684e-01 -1.03630912e+00 -5.20834029e-01 8.09438050e-01
7.01188922e-01 -4.55603242e-01 -1.21325207e+00 -2.72249848e-01
-2.28092745e-01 9.68252897e-01 -3.90825063e-01 -1.52994379e-01
-4.71937925e-01 -1.49934626e+00 2.24233419e-01 3.54527235e-02
5.22276342e-01 -9.75657701e-01 -7.48956561e-01 2.76364982e-02
-7.36549273e-02 -4.34207767e-01 -2.30225489e-01 5.39589167e-01
-1.05457222e+00 -8.32690001e-01 -1.49617702e-01 -3.00501078e-01
7.60625005e-01 -6.83848858e-02 8.88906538e-01 4.87813577e-02
-4.26467270e-01 -3.15776050e-01 -1.35942027e-01 9.17101651e-02
-4.07941937e-01 4.11024600e-01 -3.78365591e-02 -4.84317183e-01
2.82247335e-01 -2.30167568e-01 -5.39062321e-01 7.05862939e-01
-6.79282546e-01 -2.26765424e-01 5.08210540e-01 1.05107260e+00
9.33195233e-01 4.71941411e-01 1.62228748e-01 -7.39263415e-01
6.12020433e-01 -2.57888407e-01 -1.23376334e+00 5.06832123e-01
-1.01478875e+00 8.28002393e-01 3.91584277e-01 -2.71854013e-01
-1.34185338e+00 1.21604279e-01 3.92012537e-01 3.62093329e-01
1.22077689e-01 3.75643641e-01 -1.43515930e-01 -3.87024224e-01
6.23100877e-01 5.52931614e-02 -9.48875397e-02 -2.90832043e-01
1.49256647e-01 6.64967895e-01 4.81386036e-01 -1.10291123e+00
2.10742131e-01 2.07354441e-01 6.09585762e-01 -6.84288740e-01
-4.27283704e-01 -2.07013234e-01 -1.96151674e-01 -1.25826433e-01
5.99061608e-01 -7.13815987e-02 -1.29938877e+00 5.61369002e-01
-6.75646663e-01 -5.88282617e-03 7.25315213e-02 4.90211070e-01
-5.55327356e-01 4.73425567e-01 -4.33595657e-01 -9.66374278e-01
-4.50174451e-01 -1.49902129e+00 6.78794265e-01 4.98148382e-01
-1.51416540e-01 -7.00977862e-01 3.24088037e-01 5.63503504e-01
3.10820878e-01 1.67569458e-01 1.33600461e+00 -3.82765353e-01
-5.41303337e-01 -2.42309391e-01 1.13933720e-01 -3.91040176e-01
-4.82298195e-01 3.84201199e-01 -3.69630933e-01 -3.56593162e-01
6.97687566e-02 -9.14911181e-02 6.37995720e-01 7.24709809e-01
1.24678481e+00 2.50333846e-02 -5.18832147e-01 7.05278695e-01
1.84626400e+00 4.87189829e-01 8.34702373e-01 7.68169343e-01
-3.63361090e-02 8.26897398e-02 7.28736103e-01 6.08635128e-01
-4.95147109e-02 7.96316803e-01 3.59769344e-01 3.51444751e-01
2.60375857e-01 1.99803546e-01 2.50463694e-01 7.26275563e-01
-4.61536758e-02 -5.08334637e-01 -1.16821337e+00 1.15727238e-01
-1.74553180e+00 -7.52022564e-01 -4.80351418e-01 2.63687015e+00
1.05428064e+00 4.45282131e-01 1.00108147e-01 1.96347386e-01
1.13479400e+00 5.67173064e-02 -6.31319165e-01 -1.22465253e+00
8.49065781e-02 4.15410012e-01 1.06163955e+00 7.28619337e-01
-6.11092031e-01 7.11318135e-01 7.39368677e+00 1.47049105e+00
-7.26060927e-01 -1.81934670e-01 6.37872994e-01 -4.61938888e-01
-2.86116242e-01 1.40414253e-01 -9.68712151e-01 8.67824793e-01
1.23142397e+00 -1.77322015e-01 7.83519208e-01 4.75244612e-01
4.54742946e-02 -9.45464313e-01 -5.61288655e-01 1.08208323e+00
-1.84632197e-01 -1.21552169e+00 -4.93914187e-02 2.04547375e-01
1.04700923e+00 -2.36491695e-01 4.31538299e-02 -1.80370092e-01
4.71166819e-01 -1.01128519e+00 3.49038273e-01 5.09395659e-01
5.25587380e-01 -1.59734666e+00 1.07476950e+00 1.37895837e-01
-8.95834923e-01 4.51263227e-02 -4.75057095e-01 4.21641648e-01
1.53783396e-01 9.00122643e-01 -7.17402995e-01 3.87975276e-01
6.86030626e-01 -5.15478492e-01 -3.50771546e-01 1.41882777e+00
-4.85114567e-02 7.38390446e-01 -1.08728385e+00 -7.41955996e-01
3.25022727e-01 -8.58188212e-01 6.03992164e-01 7.55145550e-01
2.59095281e-01 1.28928170e-01 -1.06750801e-01 7.07536221e-01
2.68694639e-01 1.96226925e-01 3.45377266e-01 1.05037980e-05
8.78129482e-01 1.11318815e+00 -1.38345993e+00 -1.26152903e-01
4.19301003e-01 4.76171285e-01 -7.82807767e-02 8.52534622e-02
-8.96273255e-01 -9.06038582e-01 1.72135681e-01 -3.45078766e-01
2.75813013e-01 -3.02791655e-01 -6.99086726e-01 -4.29599166e-01
-2.56042510e-01 -8.70398045e-01 4.96599168e-01 -4.75971699e-01
-8.29604387e-01 5.45180082e-01 3.80890131e-01 -5.96749604e-01
-2.51833677e-01 -4.75271851e-01 -5.78070045e-01 7.76404738e-01
-8.16097796e-01 -5.05067647e-01 -1.70953393e-01 1.51738599e-01
2.83323843e-02 1.04924083e-01 7.25052059e-01 -4.50278848e-01
-8.13140154e-01 6.53181434e-01 8.63271773e-01 -7.95976937e-01
3.10539782e-01 -1.25529671e+00 2.39032552e-01 4.95995581e-01
-4.06959832e-01 5.03776848e-01 1.09984231e+00 -7.97776043e-01
-1.65910685e+00 4.46172059e-02 4.85285550e-01 2.46251654e-02
4.03075933e-01 -2.71656960e-01 -5.69748402e-01 -1.96607858e-01
2.22492605e-01 -2.96596318e-01 5.96656263e-01 2.36176729e-01
8.52230191e-02 -2.93453620e-03 -1.48053575e+00 4.45391357e-01
6.28951967e-01 1.19219795e-02 -1.63430795e-01 3.38353753e-01
1.06740154e-01 -4.80822593e-01 -8.60431254e-01 1.12219721e-01
5.73032737e-01 -9.91728663e-01 9.73180175e-01 -1.90662026e-01
-1.88480943e-01 -4.45293009e-01 -3.58204171e-02 -1.24259710e+00
-2.68365055e-01 -8.65227222e-01 -3.13227437e-02 8.31354320e-01
4.95075852e-01 -7.11493373e-01 1.00974131e+00 6.03962660e-01
2.44203761e-01 -9.37258482e-01 -1.21242785e+00 -9.86968994e-01
-1.46833688e-01 -9.79131609e-02 1.13135076e+00 6.26562953e-01
2.56091326e-01 7.74307996e-02 1.44732624e-01 2.13519558e-01
1.01030028e+00 3.24555814e-01 3.12747806e-01 -1.06600702e+00
-4.33982372e-01 -5.67563355e-01 -3.50256622e-01 -6.27623677e-01
-7.57988617e-02 -7.19234526e-01 1.67408749e-01 -1.18458092e+00
5.53301215e-01 -6.17597103e-01 -4.59337890e-01 7.00195432e-02
-5.72702847e-03 3.91607434e-02 -2.91748226e-01 -1.01867476e-02
-5.05046427e-01 4.64598835e-01 9.77190733e-01 1.06441729e-01
-3.77037287e-01 -9.99300256e-02 1.98966209e-02 5.30243158e-01
8.04907322e-01 -7.61655092e-01 -2.91061491e-01 1.82468548e-01
6.66240156e-01 1.95869818e-01 -1.89188078e-01 -9.83061194e-01
3.19215745e-01 -6.75437748e-01 2.55288780e-01 -9.70106304e-01
2.52208292e-01 -6.89858675e-01 5.59240639e-01 6.25812709e-01
-2.28962883e-01 1.20663367e-01 9.18895155e-02 4.79033619e-01
1.23185277e-01 -9.27807391e-01 1.03040767e+00 2.69401193e-01
-2.26157829e-01 -3.79761487e-01 -4.51552302e-01 -1.90559894e-01
1.16970170e+00 -4.12133753e-01 -5.65758467e-01 -2.82204170e-02
-4.51837659e-01 1.47162765e-01 7.42228746e-01 -3.54537994e-01
2.32737169e-01 -1.19635427e+00 -4.14037332e-02 1.81634545e-01
-2.03251481e-01 -2.02955559e-01 2.16539264e-01 7.03506052e-01
-1.26515830e+00 5.08621633e-01 -2.76946664e-01 -5.98547280e-01
-1.23802352e+00 4.50185806e-01 3.37122738e-01 -3.04342121e-01
-2.24935859e-01 6.90813780e-01 -6.29010677e-01 -2.18774796e-01
-3.75132225e-02 -1.89697698e-01 4.13003862e-02 -2.83620477e-01
2.55508274e-01 1.07853687e+00 2.54150152e-01 -2.68983424e-01
-5.88909626e-01 7.06231177e-01 7.40004238e-03 -4.92104352e-01
1.20209467e+00 8.99071321e-02 -5.95325947e-01 1.09754741e-01
9.34436738e-01 8.37516189e-02 -7.01774299e-01 2.79881448e-01
2.31357571e-02 -4.69630986e-01 6.10004246e-01 -1.10628521e+00
-7.59038627e-01 3.64164561e-01 8.34773958e-01 1.79099768e-01
8.58770847e-01 -3.07390004e-01 9.40174162e-01 6.86288714e-01
6.57835960e-01 -1.50712061e+00 -5.62154830e-01 3.96932960e-01
3.75659317e-01 -5.86078346e-01 6.71036839e-01 -4.65313733e-01
-5.65248609e-01 1.30554259e+00 3.65132302e-01 1.88420657e-02
5.28272092e-01 3.25006455e-01 -3.53751034e-01 -3.74888986e-01
-6.51120722e-01 3.46186042e-01 1.47537664e-01 7.57201612e-02
-8.66701826e-02 3.27979028e-01 -9.41569448e-01 2.88963914e-01
-4.05380100e-01 -4.20248777e-01 5.78640163e-01 9.15514290e-01
-7.44947135e-01 -1.37896359e+00 -8.82417023e-01 2.67037302e-01
-2.03298613e-01 9.25418437e-02 -2.48861834e-01 5.35815954e-01
-2.79685944e-01 9.38209057e-01 1.58084288e-01 7.08808098e-03
-2.80318797e-01 1.51518822e-01 6.71599329e-01 -3.76345008e-03
-6.31879628e-01 -5.65700345e-02 2.51386672e-01 -5.71622849e-01
-8.56117308e-02 -6.55290782e-01 -1.56470346e+00 -5.95854223e-01
-7.25032330e-01 8.38723540e-01 1.36394954e+00 7.01804340e-01
2.36943871e-01 1.62713379e-01 6.34649158e-01 -7.79851973e-01
-7.61013091e-01 -4.29652870e-01 -6.87518358e-01 -1.75110430e-01
-4.29205835e-01 -9.68274772e-01 -6.16090417e-01 -6.59945428e-01] | [6.188405990600586, 4.102367401123047] |
14050aa1-84f4-408e-9768-a7f075a48872 | towards-open-scenario-semi-supervised-medical | 2304.04059 | null | https://arxiv.org/abs/2304.04059v1 | https://arxiv.org/pdf/2304.04059v1.pdf | Towards Open-Scenario Semi-supervised Medical Image Classification | Semi-supervised learning (SSL) has attracted much attention since it reduces the expensive costs of collecting adequate well-labeled training data, especially for deep learning methods. However, traditional SSL is built upon an assumption that labeled and unlabeled data should be from the same distribution e.g., classes and domains. However, in practical scenarios, unlabeled data would be from unseen classes or unseen domains, and it is still challenging to exploit them by existing SSL methods. Therefore, in this paper, we proposed a unified framework to leverage these unseen unlabeled data for open-scenario semi-supervised medical image classification. We first design a novel scoring mechanism, called dual-path outliers estimation, to identify samples from unseen classes. Meanwhile, to extract unseen-domain samples, we then apply an effective variational autoencoder (VAE) pre-training. After that, we conduct domain adaptation to fully exploit the value of the detected unseen-domain samples to boost semi-supervised training. We evaluated our proposed framework on dermatology and ophthalmology tasks. Extensive experiments demonstrate our model can achieve superior classification performance in various medical SSL scenarios. | ['ZongYuan Ge', 'Zhuoting Zhu', 'Lin Wang', 'Zhen Yu', 'Wei Feng', 'Yicheng Wu', 'Lie Ju'] | 2023-04-08 | null | null | null | null | ['semi-supervised-medical-image-classification'] | ['medical'] | [ 1.44183353e-01 6.60125762e-02 -3.94422263e-01 -6.10323131e-01
-1.04185772e+00 -3.18334788e-01 1.05879731e-01 1.12584546e-01
-2.96663702e-01 8.01532507e-01 1.03684686e-01 -7.97062069e-02
8.80147442e-02 -4.03661102e-01 -6.13310516e-01 -9.73437607e-01
7.36774683e-01 5.61748564e-01 5.75529449e-02 3.18183124e-01
-1.78460792e-01 2.22650260e-01 -1.32314169e+00 1.78640008e-01
1.20859838e+00 9.08037901e-01 2.23118261e-01 1.36057585e-01
-1.76730916e-01 6.71879947e-01 -4.17091310e-01 -1.63736209e-01
1.47879925e-02 -4.56691384e-01 -5.02132475e-01 6.46379650e-01
1.58781871e-01 -4.40724820e-01 -1.59280017e-01 1.19035137e+00
5.58983028e-01 1.33530840e-01 8.15598071e-01 -1.09843838e+00
-7.68588781e-01 6.15587644e-03 -7.18214631e-01 -3.78174931e-02
-8.73401761e-02 1.22297730e-03 7.17882216e-01 -1.13431132e+00
6.52589321e-01 8.18050861e-01 3.71641219e-01 8.34930956e-01
-1.03246021e+00 -5.41630983e-01 4.99182791e-02 7.36359181e-03
-1.16899157e+00 -4.23057288e-01 1.03858030e+00 -4.73034292e-01
4.50128019e-02 6.99581672e-03 3.48358542e-01 1.28836346e+00
-2.46032253e-01 1.20875311e+00 9.54932392e-01 -3.10365617e-01
4.68190700e-01 5.23631454e-01 1.70543939e-01 5.71078658e-01
2.95196652e-01 -1.16549164e-01 -2.87844360e-01 -3.40177804e-01
6.59758210e-01 4.17919368e-01 -3.29087496e-01 -6.46258652e-01
-1.06792378e+00 7.84010410e-01 3.26054811e-01 6.48533702e-02
-4.86179739e-01 -5.45844853e-01 4.15627837e-01 -1.64340377e-01
6.81402087e-01 -1.60743624e-01 -3.84657592e-01 3.59807849e-01
-9.51482117e-01 -2.73846388e-01 3.52603287e-01 7.68916726e-01
7.03772128e-01 -1.91730648e-01 -1.31581649e-01 1.21875155e+00
5.33400595e-01 4.44591254e-01 7.57836699e-01 -6.22890353e-01
3.72397035e-01 8.25587571e-01 1.80018529e-01 -5.96414089e-01
-1.86131015e-01 -6.48814917e-01 -9.80533183e-01 -2.13874429e-01
4.18464154e-01 -2.01158524e-01 -1.15783346e+00 1.48935843e+00
7.83534944e-01 4.81935233e-01 2.52786607e-01 1.10429132e+00
9.12919164e-01 4.58024472e-01 2.18267720e-02 -3.90223265e-01
1.09415185e+00 -9.14829791e-01 -7.36125767e-01 -1.02338232e-01
6.96972311e-01 -6.54722691e-01 1.06291652e+00 3.67502242e-01
-7.22565114e-01 -4.79641229e-01 -9.28681672e-01 -8.52942467e-02
1.40545279e-01 4.95468706e-01 3.05749327e-01 4.28177327e-01
-3.67320031e-01 3.23882729e-01 -1.13317132e+00 -4.80998345e-02
8.60154033e-01 1.62242219e-01 -3.66468459e-01 -3.69699836e-01
-8.90862048e-01 1.99133113e-01 4.31698829e-01 2.52575904e-01
-1.04311633e+00 -5.05841017e-01 -1.01736379e+00 -1.89193130e-01
4.33285654e-01 -4.70891207e-01 1.10147274e+00 -1.06651449e+00
-1.07914639e+00 1.10691535e+00 -4.54988182e-01 -2.15830788e-01
4.70634192e-01 -8.53813812e-02 -4.44514573e-01 3.12553734e-01
2.75671571e-01 3.89332592e-01 1.04754674e+00 -1.33237541e+00
-4.89756584e-01 -7.51526773e-01 -5.08599401e-01 9.74839181e-02
-6.58874512e-01 -3.99913520e-01 -4.26520854e-01 -6.70571566e-01
4.95418042e-01 -9.10195470e-01 -2.60473281e-01 1.12573087e-01
-5.98537385e-01 -3.23454142e-01 7.75911152e-01 -6.10082388e-01
1.00419998e+00 -2.44861865e+00 -1.61621556e-01 1.69034466e-01
5.22600353e-01 4.55483258e-01 8.38374197e-02 -5.15678339e-02
-2.19366159e-02 -3.08109075e-01 -3.89089644e-01 -4.34812218e-01
-2.65280366e-01 4.80342478e-01 -2.37577081e-01 6.14110172e-01
4.00749236e-01 6.73020184e-01 -1.05344653e+00 -7.87035584e-01
2.56374359e-01 3.99509549e-01 -5.04536450e-01 2.56806374e-01
-1.99712321e-01 9.29032207e-01 -6.76651537e-01 8.43046963e-01
9.30878460e-01 -7.13947237e-01 9.92217213e-02 -1.10902645e-01
4.88881737e-01 -1.15100600e-01 -1.10523248e+00 1.71758008e+00
-2.69973963e-01 2.77248085e-01 -9.98442322e-02 -1.25311387e+00
8.53203714e-01 3.75785470e-01 5.32591283e-01 -3.75117213e-01
1.21577151e-01 5.37879348e-01 -3.61001521e-01 -6.79708004e-01
-1.26659460e-02 -3.95097405e-01 2.12854907e-01 1.30456805e-01
1.08529754e-01 3.77655983e-01 -1.65715083e-01 2.13682085e-01
6.50416434e-01 5.19223772e-02 3.09828460e-01 1.20246150e-01
5.82897782e-01 1.03685163e-01 1.05358458e+00 4.48194712e-01
-6.17888629e-01 8.64704788e-01 3.52277577e-01 -2.76558191e-01
-8.31512511e-01 -1.27813637e+00 -4.68519449e-01 7.72034943e-01
3.87787133e-01 3.89508270e-02 -6.85036898e-01 -1.10239971e+00
-1.47828430e-01 4.14791882e-01 -4.59400564e-01 -2.40935162e-01
-2.00333506e-01 -7.88457394e-01 1.76618353e-01 5.31862676e-01
2.84360349e-01 -9.53133702e-01 -1.75352037e-01 2.18661934e-01
-2.73550898e-01 -1.11511838e+00 -4.51267213e-01 -8.68750140e-02
-1.07033622e+00 -1.19147623e+00 -1.08098686e+00 -1.10625088e+00
1.05353653e+00 3.34994614e-01 8.72408986e-01 -1.98736459e-01
-3.53032202e-01 2.64072064e-02 -4.22368944e-01 -4.49029982e-01
-4.03610945e-01 -1.70189515e-01 1.01191156e-01 5.23940623e-01
7.61880815e-01 -2.83556640e-01 -8.51368785e-01 4.41684276e-01
-9.76676226e-01 -2.49979705e-01 6.72979295e-01 1.29949212e+00
1.14641142e+00 1.04857609e-01 7.46319294e-01 -1.11488867e+00
2.03802854e-01 -6.78494275e-01 -3.60385358e-01 3.16920102e-01
-4.50523257e-01 1.42726833e-02 7.19164014e-01 -5.29888332e-01
-1.14800143e+00 1.27877414e-01 -2.02624142e-01 -8.68888855e-01
-4.38515484e-01 5.15597701e-01 -3.19137037e-01 3.19959313e-01
5.72562277e-01 4.07000273e-01 4.94408086e-02 -5.78994274e-01
4.32580002e-02 1.14147294e+00 5.24904251e-01 -4.34593201e-01
6.05401158e-01 9.36711550e-01 -3.90174538e-01 -8.53396952e-01
-1.23683977e+00 -7.44439423e-01 -4.33450669e-01 6.16548061e-02
7.86800385e-01 -1.14199460e+00 -1.90634534e-01 4.33513969e-01
-7.27298498e-01 6.15280047e-02 -3.01999152e-01 8.36215198e-01
-1.71701148e-01 6.03663981e-01 -4.34010208e-01 -7.55956709e-01
-3.39611262e-01 -1.26046753e+00 1.26223683e+00 3.90538245e-01
3.14810947e-02 -9.72409606e-01 1.25335651e-02 7.51764655e-01
-1.50719866e-01 2.79474854e-01 6.80830240e-01 -1.04891324e+00
-3.77614379e-01 -4.06720161e-01 -2.94456571e-01 7.19274163e-01
2.86044538e-01 -2.85702735e-01 -1.02643943e+00 -4.81084168e-01
2.55010813e-01 -7.21382260e-01 8.76050472e-01 5.00965536e-01
1.35388207e+00 3.37311104e-02 -4.45863396e-01 6.63636148e-01
1.13735545e+00 -4.78282012e-02 3.01286519e-01 -5.74018620e-02
7.98019946e-01 5.70650935e-01 8.40825140e-01 6.90850437e-01
1.73737258e-01 1.07549258e-01 2.44075045e-01 -2.13141873e-01
1.79365978e-01 -3.90807450e-01 -4.07146895e-03 8.17055941e-01
3.03138435e-01 -2.61103250e-02 -7.94711649e-01 7.23768890e-01
-1.80355704e+00 -5.19937873e-01 -9.14750770e-02 2.19120431e+00
7.40744889e-01 1.35942250e-01 -7.96626508e-02 7.01558366e-02
9.16366458e-01 -1.04452342e-01 -1.12577415e+00 4.68990594e-01
-8.59416556e-03 1.67337805e-01 2.12209225e-01 2.41086148e-02
-1.15432823e+00 6.24407530e-01 4.67467022e+00 1.09238565e+00
-1.15734601e+00 3.25240850e-01 7.18617022e-01 -2.04441652e-01
-2.77207553e-01 -2.05930680e-01 -6.42248333e-01 6.39332294e-01
5.07104695e-01 2.18889624e-01 -2.25750953e-01 1.10946488e+00
2.64135659e-01 4.80174534e-02 -8.96777570e-01 1.28180337e+00
8.64591375e-02 -1.03309143e+00 -1.09252088e-01 8.65650699e-02
7.68293381e-01 -8.19162354e-02 1.77368894e-01 3.13733131e-01
2.70371437e-02 -6.04786813e-01 1.13182113e-01 2.55735457e-01
7.63952374e-01 -7.20241904e-01 7.88754940e-01 6.84145570e-01
-7.26678371e-01 -3.22127342e-02 -5.77250659e-01 4.53611314e-01
-1.08786233e-01 1.06919134e+00 -7.45497525e-01 4.90502119e-01
4.87587571e-01 9.94726777e-01 -2.47876361e-01 9.40942049e-01
-2.73036927e-01 8.08364749e-01 -1.81832746e-01 2.37024829e-01
2.09846601e-01 -2.52570421e-01 3.49667460e-01 5.96305728e-01
1.94701508e-01 6.97098598e-02 2.81586319e-01 6.31017327e-01
-2.33868450e-01 2.08851546e-01 -3.29458565e-01 -1.83034763e-02
4.70921934e-01 1.15213251e+00 -6.01795435e-01 -3.42096716e-01
-5.82351208e-01 1.14660192e+00 3.15021247e-01 5.16579211e-01
-8.15104365e-01 -1.26879126e-01 3.37649465e-01 8.92903358e-02
9.80002135e-02 2.10719272e-01 -1.25056654e-01 -1.69417071e+00
3.62026095e-01 -6.25373304e-01 6.59800708e-01 -5.41378200e-01
-1.64078784e+00 4.59853053e-01 -5.15325546e-01 -1.67549133e+00
5.93931824e-02 -5.24526179e-01 -3.97627771e-01 6.72255397e-01
-1.57790172e+00 -1.14940107e+00 -4.73789960e-01 7.47324824e-01
5.68689823e-01 -2.09543616e-01 5.13428271e-01 3.83076102e-01
-9.18261707e-01 7.12134421e-01 6.36033297e-01 3.98289502e-01
9.91006255e-01 -1.16233325e+00 -2.88565934e-01 6.99590862e-01
1.17491432e-01 6.40157580e-01 2.47742504e-01 -6.16522014e-01
-8.98889184e-01 -1.29962659e+00 5.08696675e-01 -2.18652874e-01
3.87349010e-01 -1.84710860e-01 -1.17869866e+00 6.33790135e-01
-2.83464164e-01 4.77671921e-01 1.21972775e+00 -1.14255428e-01
-1.05542347e-01 -3.57584395e-02 -1.19910908e+00 2.88803428e-01
7.23121166e-01 -4.74282056e-01 -6.76326871e-01 6.44195557e-01
5.88432074e-01 -4.35495347e-01 -7.94607043e-01 5.30679166e-01
3.46768647e-01 -8.88425827e-01 8.93808544e-01 -7.10381627e-01
5.32192707e-01 -3.12897444e-01 -1.41110510e-01 -1.22521794e+00
1.55271381e-01 -1.21704794e-01 -2.54112214e-01 1.13753486e+00
1.60107091e-01 -6.42831266e-01 1.21640825e+00 5.95942795e-01
-2.18168035e-01 -8.38035703e-01 -7.04726934e-01 -6.36572897e-01
-6.35373443e-02 -2.07462743e-01 3.07145447e-01 1.14995718e+00
-2.25277841e-01 2.99931616e-01 -4.04356271e-01 3.58851403e-01
9.64092851e-01 2.97658056e-01 5.31040668e-01 -1.24955654e+00
-3.67225915e-01 1.76123783e-01 -2.69790292e-01 -9.67086792e-01
3.20808619e-01 -9.29463744e-01 6.48805499e-03 -1.39140368e+00
3.64030391e-01 -5.43278039e-01 -5.72594106e-01 3.12084585e-01
-6.22828245e-01 2.76216179e-01 -3.74985904e-01 6.40399814e-01
-7.47949123e-01 7.89968431e-01 1.28364539e+00 -1.07987048e-02
-3.40255409e-01 1.79208696e-01 -6.29604757e-01 8.90615582e-01
6.69979990e-01 -5.31726301e-01 -6.97667181e-01 -4.11404341e-01
-2.97763556e-01 1.87398016e-01 4.23813790e-01 -9.36543882e-01
2.18282901e-02 -1.59695461e-01 5.71951389e-01 -7.99199581e-01
6.50090352e-02 -8.37217927e-01 -3.66226405e-01 3.41110289e-01
-2.69221663e-01 -8.59700322e-01 -1.69479430e-01 9.87585247e-01
-4.77573663e-01 -2.35063747e-01 9.63427186e-01 -1.89731037e-03
-5.87023854e-01 5.21776378e-01 7.08776861e-02 2.78898865e-01
1.24487126e+00 -1.36747092e-01 -2.22821869e-02 -5.33653200e-02
-1.14868057e+00 4.34135169e-01 4.50229585e-01 1.26369908e-01
8.79263937e-01 -1.25747752e+00 -5.93843102e-01 4.43014354e-01
5.23605764e-01 5.81632435e-01 6.59597993e-01 8.72953892e-01
-4.22476143e-01 6.51492476e-02 1.71552047e-01 -8.74257147e-01
-1.02948308e+00 9.43276346e-01 1.12485707e-01 -1.79691896e-01
-3.87436777e-01 8.12346101e-01 4.06776845e-01 -6.68717861e-01
2.15895325e-01 1.07673397e-02 -3.62988183e-04 7.59916157e-02
5.10750651e-01 1.53255627e-01 6.44437149e-02 -4.17036504e-01
-3.86907399e-01 3.69365603e-01 -3.70588958e-01 1.68919995e-01
1.09757733e+00 -1.25376761e-01 1.04811162e-01 3.64318877e-01
1.46571016e+00 -1.53507128e-01 -1.23523760e+00 -6.93729281e-01
-1.05218790e-01 -3.48623067e-01 1.04845099e-01 -3.31487536e-01
-1.00808740e+00 1.16012466e+00 6.61546350e-01 -2.59883195e-01
1.22071218e+00 1.85962111e-01 1.18857288e+00 2.67220080e-01
2.15818435e-01 -1.16431320e+00 2.56077856e-01 -1.43768802e-01
2.27680430e-01 -1.77855515e+00 -2.37349078e-01 -4.70218986e-01
-8.62645090e-01 7.99554884e-01 6.32554650e-01 -4.72988337e-02
6.96540952e-01 -2.64275253e-01 3.42296869e-01 -1.40405983e-01
-3.32404554e-01 -1.98605821e-01 2.38451615e-01 5.39252400e-01
2.02585712e-01 1.19081035e-01 -3.95226367e-02 7.50324130e-01
4.40140545e-01 3.48637938e-01 2.95852929e-01 9.50429022e-01
-3.95304650e-01 -1.02843928e+00 -4.52191710e-01 7.29260206e-01
-4.60116655e-01 9.75910202e-02 -8.64560455e-02 4.88070399e-01
1.34910524e-01 8.98338437e-01 -1.03345975e-01 -1.79898202e-01
8.91198590e-02 3.25853415e-02 1.38369679e-01 -8.67320478e-01
1.11535944e-01 4.86970782e-01 -4.70005274e-01 -2.49497920e-01
-4.74586189e-01 -5.79688847e-01 -1.36473179e+00 3.15556407e-01
-4.43601876e-01 2.66765684e-01 3.67986679e-01 1.01221156e+00
3.66852194e-01 4.42084521e-01 8.43940318e-01 -3.38281363e-01
-7.90985525e-01 -7.70967722e-01 -8.60878646e-01 8.01008165e-01
5.60171187e-01 -6.67169511e-01 -4.24001992e-01 1.13363512e-01] | [14.725704193115234, -1.9796679019927979] |
445d9927-2206-41e0-9f98-ae5e9d8e8ea4 | deberta-decoding-enhanced-bert-with | 2006.03654 | null | https://arxiv.org/abs/2006.03654v6 | https://arxiv.org/pdf/2006.03654v6.pdf | DeBERTa: Decoding-enhanced BERT with Disentangled Attention | Recent progress in pre-trained neural language models has significantly improved the performance of many natural language processing (NLP) tasks. In this paper we propose a new model architecture DeBERTa (Decoding-enhanced BERT with disentangled attention) that improves the BERT and RoBERTa models using two novel techniques. The first is the disentangled attention mechanism, where each word is represented using two vectors that encode its content and position, respectively, and the attention weights among words are computed using disentangled matrices on their contents and relative positions, respectively. Second, an enhanced mask decoder is used to incorporate absolute positions in the decoding layer to predict the masked tokens in model pre-training. In addition, a new virtual adversarial training method is used for fine-tuning to improve models' generalization. We show that these techniques significantly improve the efficiency of model pre-training and the performance of both natural language understanding (NLU) and natural langauge generation (NLG) downstream tasks. Compared to RoBERTa-Large, a DeBERTa model trained on half of the training data performs consistently better on a wide range of NLP tasks, achieving improvements on MNLI by +0.9% (90.2% vs. 91.1%), on SQuAD v2.0 by +2.3% (88.4% vs. 90.7%) and RACE by +3.6% (83.2% vs. 86.8%). Notably, we scale up DeBERTa by training a larger version that consists of 48 Transform layers with 1.5 billion parameters. The significant performance boost makes the single DeBERTa model surpass the human performance on the SuperGLUE benchmark (Wang et al., 2019a) for the first time in terms of macro-average score (89.9 versus 89.8), and the ensemble DeBERTa model sits atop the SuperGLUE leaderboard as of January 6, 2021, out performing the human baseline by a decent margin (90.3 versus 89.8). | ['Weizhu Chen', 'Xiaodong Liu', 'Jianfeng Gao', 'Pengcheng He'] | 2020-06-05 | null | https://openreview.net/forum?id=XPZIaotutsD | https://openreview.net/pdf?id=XPZIaotutsD | iclr-2021-1 | ['linguistic-acceptability'] | ['natural-language-processing'] | [ 1.98697612e-01 4.01842147e-01 -2.94829402e-02 -3.13424736e-01
-1.18756998e+00 -7.89685667e-01 9.42731202e-01 -2.40970686e-01
-6.46543622e-01 7.75471032e-01 4.91717130e-01 -5.66785097e-01
4.12964791e-01 -7.09751725e-01 -1.01151049e+00 -6.19309604e-01
7.52808526e-02 5.93603790e-01 -2.52425849e-01 -5.05937338e-01
-1.89770892e-01 9.08061713e-02 -1.01304352e+00 6.04808331e-01
8.99840355e-01 6.48958325e-01 7.12999105e-02 8.08685184e-01
5.86437546e-02 6.05643630e-01 -6.42681062e-01 -6.62608862e-01
3.84427965e-01 -3.31411481e-01 -6.67398155e-01 -6.15373433e-01
4.09349918e-01 -3.56342494e-01 -4.27234590e-01 6.34512007e-01
5.50303638e-01 4.33555990e-02 6.49147332e-01 -9.61450279e-01
-8.15102220e-01 1.12150741e+00 -6.76559746e-01 -2.06733476e-02
-6.68416098e-02 3.43740433e-01 1.37038648e+00 -1.01300240e+00
3.47311080e-01 1.23030782e+00 5.30687511e-01 9.48664665e-01
-1.20979786e+00 -1.02583718e+00 2.87857354e-01 -7.94490576e-02
-1.14405620e+00 -7.37427354e-01 1.93654954e-01 -2.98758656e-01
1.58763289e+00 3.55204135e-01 3.17366689e-01 1.39816284e+00
2.74821460e-01 9.33315516e-01 1.00245929e+00 -3.90486002e-01
9.84289777e-03 -1.56691670e-01 2.60212302e-01 8.17697763e-01
2.76214302e-01 3.17775667e-01 -7.34359384e-01 -2.33964603e-02
2.20950633e-01 -5.11773169e-01 -2.11218670e-01 8.64262134e-02
-1.41743135e+00 7.98073113e-01 4.17145491e-01 1.49597198e-01
-1.83159485e-01 3.56836468e-01 3.01282912e-01 2.23752663e-01
6.64586902e-01 8.64574432e-01 -7.61605024e-01 -1.33523479e-01
-9.37012196e-01 3.41348767e-01 7.97760427e-01 8.30123544e-01
5.57439923e-01 3.46832842e-01 -3.95080864e-01 7.40945697e-01
3.84890079e-01 6.19042039e-01 6.08308911e-01 -7.04393506e-01
8.78526807e-01 2.56812334e-01 -1.28113434e-01 -5.22813737e-01
-2.61047482e-01 -7.27558076e-01 -1.02222157e+00 9.22452062e-02
5.50027609e-01 -4.56130475e-01 -1.21107912e+00 2.13329196e+00
-2.26609394e-01 8.49219114e-02 3.68469298e-01 7.10649252e-01
5.40932715e-01 1.37372780e+00 6.77842274e-02 2.82222629e-01
1.38161421e+00 -1.18220329e+00 -4.16301250e-01 -8.00570309e-01
8.47352862e-01 -6.86714470e-01 1.03260398e+00 5.46346784e-01
-1.35126376e+00 -4.98142272e-01 -1.30178022e+00 -3.78315777e-01
-3.36677849e-01 7.05369562e-02 5.78899801e-01 6.44651711e-01
-1.21961486e+00 4.18115944e-01 -8.88041377e-01 1.09565839e-01
3.46216381e-01 3.92428279e-01 -4.20513809e-01 -3.38302791e-01
-1.59517670e+00 9.66553152e-01 3.23571682e-01 6.08846918e-02
-9.65672553e-01 -9.63083327e-01 -9.66114998e-01 2.83124000e-01
7.19870850e-02 -8.66116524e-01 1.31513929e+00 -5.98529160e-01
-1.59196818e+00 9.04113770e-01 -2.77866006e-01 -8.38786840e-01
5.08153975e-01 -5.37247956e-01 -2.55708486e-01 -3.21307987e-01
-7.21436460e-03 8.99764419e-01 4.82834518e-01 -1.00264478e+00
-3.64375561e-01 -2.67679155e-01 -1.46642849e-01 2.80698925e-01
-8.38075131e-02 -4.24511656e-02 -3.30034584e-01 -7.87082911e-01
-1.00236423e-01 -8.56127977e-01 -1.49756894e-01 -4.43356395e-01
-5.79012811e-01 5.82436612e-03 3.44188273e-01 -1.07766020e+00
9.21535969e-01 -2.01034641e+00 4.13984537e-01 -6.06673658e-02
5.29039085e-01 4.69945073e-01 -5.64356387e-01 4.70315218e-01
-1.67185947e-01 4.61917847e-01 -4.13378477e-01 -8.83385301e-01
1.46614745e-01 1.41252398e-01 -5.77025890e-01 2.37077087e-01
3.68078887e-01 1.22885764e+00 -8.11585844e-01 1.68853298e-01
-1.87442273e-01 4.75221813e-01 -8.06157768e-01 1.72711790e-01
-3.70469570e-01 3.75411808e-01 -1.07554093e-01 1.47330970e-01
4.81967062e-01 -6.72304556e-02 7.04531521e-02 1.48651823e-01
-2.00277522e-01 7.10005641e-01 -5.67287147e-01 1.85866833e+00
-8.07698727e-01 8.66681993e-01 2.54954360e-02 -6.67422414e-01
8.30000222e-01 5.44540703e-01 4.73868614e-03 -7.22119987e-01
5.63615076e-02 1.47446394e-01 4.53692228e-01 -1.18496679e-01
5.97288430e-01 -1.11083217e-01 -2.97863215e-01 7.51379132e-01
7.15893731e-02 -9.55816060e-02 3.55904102e-02 6.19936407e-01
1.14224756e+00 1.00202654e-02 1.39540985e-01 -3.86188984e-01
2.15025544e-01 -1.24005735e-01 4.76922005e-01 8.09847713e-01
1.16693005e-02 6.88618064e-01 7.15295315e-01 -2.34013453e-01
-1.09733450e+00 -1.18934011e+00 3.30375671e-01 1.34428763e+00
-3.99834961e-01 -4.43712384e-01 -8.89690638e-01 -5.00140011e-01
-2.78312434e-02 1.27892101e+00 -7.32357502e-01 -5.00828564e-01
-7.62591183e-01 -9.18554187e-01 9.64885473e-01 5.60961366e-01
5.31333923e-01 -8.94746006e-01 -3.99674214e-02 2.55098082e-02
-2.45085090e-01 -1.00097609e+00 -6.58836663e-01 1.35262847e-01
-4.42308396e-01 -5.52860558e-01 -6.86482310e-01 -6.18789315e-01
3.85009438e-01 9.35069695e-02 1.25957370e+00 -1.87531233e-01
6.45418912e-02 -3.51805270e-01 -2.88765669e-01 -3.68342638e-01
-7.28501201e-01 3.98425996e-01 1.16301827e-01 3.65561768e-02
4.02657986e-02 -3.42238486e-01 -4.49780583e-01 6.34069592e-02
-6.62906408e-01 6.24727428e-01 7.98552513e-01 1.07104492e+00
1.47859380e-01 -6.62859619e-01 6.66864991e-01 -9.82140362e-01
5.18172383e-01 -5.42343795e-01 -5.60180902e-01 1.65516168e-01
-7.00810432e-01 4.40018922e-01 6.78083301e-01 -3.02859724e-01
-1.04778993e+00 -2.06694245e-01 -3.05354357e-01 -1.23425378e-02
7.15485886e-02 4.11135912e-01 -3.51528376e-01 3.87641221e-01
7.21503079e-01 2.20061928e-01 -2.17932872e-02 -4.51402247e-01
7.70318210e-01 6.74604297e-01 4.51459706e-01 -4.42130893e-01
8.63600254e-01 1.06492385e-01 -5.07717550e-01 -3.22299957e-01
-1.09539604e+00 -1.23895265e-01 -3.31611097e-01 3.63151491e-01
8.71937454e-01 -1.22502220e+00 -3.38923484e-01 5.96530139e-01
-1.46375704e+00 -6.64510071e-01 -5.68380356e-02 5.24456859e-01
-3.22413325e-01 -2.46909652e-02 -8.78739834e-01 -4.37489986e-01
-7.32730269e-01 -1.20740628e+00 1.00568688e+00 -1.96890742e-01
-4.63373542e-01 -6.87628567e-01 8.80082697e-02 6.76997781e-01
5.23792982e-01 2.33248211e-02 1.14378881e+00 -8.28994811e-01
-4.98834997e-01 -1.67600021e-01 -2.93124288e-01 6.02098763e-01
-1.71761543e-01 -2.68525451e-01 -1.21341550e+00 -3.66122305e-01
-1.65906161e-01 -2.58376867e-01 1.17732060e+00 6.04071692e-02
8.76586139e-01 -4.03334767e-01 -2.37633288e-01 7.57392764e-01
9.52175438e-01 1.30814627e-01 5.89923441e-01 2.64902562e-02
7.59594679e-01 4.15465593e-01 -7.89896771e-02 4.13233750e-02
3.14761698e-01 4.76799816e-01 3.86610270e-01 1.34005751e-02
-3.99266571e-01 -4.49882925e-01 8.03115427e-01 1.27595496e+00
8.66313800e-02 -7.19192147e-01 -1.16912138e+00 4.94516701e-01
-1.57536948e+00 -7.12883949e-01 -1.55409932e-01 2.10784602e+00
1.07180786e+00 2.75059670e-01 -4.51735198e-01 -1.03584431e-01
4.53623801e-01 3.08968395e-01 -6.03747368e-01 -6.35007024e-01
-2.86140561e-01 4.58414048e-01 5.67245305e-01 9.58188891e-01
-7.89566636e-01 1.28707600e+00 5.82978773e+00 7.69121647e-01
-8.82211983e-01 2.22876489e-01 8.87374043e-01 -3.30831707e-01
-4.13167208e-01 -1.53798074e-01 -7.87531972e-01 3.31431478e-01
1.23433006e+00 -3.78663182e-01 6.82384551e-01 4.11417633e-01
5.12783751e-02 2.39561915e-01 -1.25538862e+00 7.58571863e-01
1.83954492e-01 -1.34184742e+00 2.19129845e-01 1.12345695e-01
9.75394487e-01 4.53284323e-01 3.76128227e-01 7.23327875e-01
6.94849133e-01 -1.49773335e+00 7.78765559e-01 2.43663073e-01
9.01398540e-01 -7.72473454e-01 6.28720582e-01 5.98610342e-01
-6.20932519e-01 2.07405925e-01 -1.29426450e-01 -2.74494559e-01
2.90905654e-01 4.89522457e-01 -9.40744340e-01 3.77157748e-01
2.66643763e-01 4.41867650e-01 -2.46674955e-01 3.71269286e-01
-6.74104095e-01 1.13971448e+00 -1.76400870e-01 -7.20384493e-02
5.23343801e-01 6.71450272e-02 5.82514405e-01 1.23698425e+00
1.35151625e-01 5.49004935e-02 -4.96782631e-01 1.01403034e+00
-6.34795845e-01 -4.05546762e-02 -4.02481139e-01 -5.23966029e-02
2.45035425e-01 1.07335889e+00 -1.23013277e-02 -4.70937610e-01
-3.13706696e-01 1.14272678e+00 5.20712078e-01 5.76724410e-01
-9.78812754e-01 -4.68408167e-01 7.84789741e-01 -1.02359176e-01
4.59663011e-02 -3.06254059e-01 -3.28750044e-01 -1.28724265e+00
-2.04024166e-01 -1.13532782e+00 8.17064643e-02 -8.71756673e-01
-1.19708276e+00 1.03122020e+00 -1.95279658e-01 -6.50120735e-01
-5.30825078e-01 -8.09336185e-01 -4.72506195e-01 1.42818868e+00
-1.30360818e+00 -1.02952552e+00 1.04186617e-01 9.74621773e-02
5.22043765e-01 -3.76868993e-01 1.02881396e+00 1.39612958e-01
-7.96387315e-01 8.91083002e-01 3.20434988e-01 4.07007009e-01
7.32870162e-01 -1.22365177e+00 1.34690034e+00 1.01540577e+00
4.23471838e-01 7.54693449e-01 4.71727252e-01 -3.56294811e-01
-1.29547250e+00 -1.13192797e+00 1.34713066e+00 -6.12862229e-01
7.69006133e-01 -1.05726290e+00 -8.60618651e-01 1.02109575e+00
5.76605618e-01 -2.25004300e-01 5.50082386e-01 2.38706127e-01
-8.78154457e-01 6.89770803e-02 -8.35239887e-01 8.37592661e-01
8.85797322e-01 -4.56077933e-01 -5.55148721e-01 3.30888718e-01
1.21976030e+00 -5.35912216e-01 -4.97541577e-01 2.86593318e-01
5.38303435e-01 -7.28435159e-01 7.59991050e-01 -9.48202610e-01
5.66108823e-01 9.39014554e-02 -2.46690169e-01 -1.62436414e+00
-3.68301392e-01 -8.95505667e-01 -2.07349196e-01 9.32815552e-01
9.89563942e-01 -9.89007652e-01 7.01351821e-01 5.51817715e-01
-3.32558364e-01 -9.83613014e-01 -8.64952683e-01 -5.96963048e-01
6.31857574e-01 -5.86428821e-01 5.60945630e-01 8.44733834e-01
-1.23125076e-01 8.54637027e-01 -4.00107175e-01 1.28655061e-01
2.93086708e-01 -1.14950433e-01 5.21504045e-01 -8.12762916e-01
-6.37828410e-01 -4.76249218e-01 1.52258854e-02 -1.28990519e+00
2.95106649e-01 -1.35543811e+00 1.34630008e-02 -1.49928546e+00
1.61292344e-01 -3.67820323e-01 -3.49624574e-01 7.40836382e-01
-2.92990744e-01 1.52302235e-01 4.41051692e-01 5.16773164e-02
-1.09912559e-01 7.15365291e-01 1.01348531e+00 -2.52115935e-01
6.06279224e-02 -1.05788283e-01 -1.01639986e+00 4.66416568e-01
1.20216274e+00 -4.51013744e-01 -2.23581776e-01 -1.22734880e+00
3.23875189e-01 -1.06233329e-01 2.83456743e-01 -7.22512722e-01
7.30521604e-02 3.29157770e-01 1.24386139e-01 -3.18252504e-01
4.96169180e-01 -1.51246279e-01 -1.60464019e-01 7.07866609e-01
-5.96587837e-01 2.56932646e-01 5.12719214e-01 3.78597260e-01
-5.54662477e-03 4.25390489e-02 6.26714468e-01 2.42172945e-02
-4.12926018e-01 2.40688935e-01 -5.02483666e-01 2.81455010e-01
7.33720243e-01 2.00619712e-01 -5.30137360e-01 -5.15145600e-01
-5.08774817e-01 2.84811586e-01 -7.35554844e-02 5.89067340e-01
3.59375477e-01 -9.69513953e-01 -1.27398944e+00 3.06927741e-01
-1.63554206e-01 1.35611385e-01 1.83563918e-01 4.08671170e-01
-5.41956186e-01 5.96309781e-01 1.25734448e-01 -7.29693621e-02
-9.13409412e-01 2.26363629e-01 4.29225862e-01 -6.69048548e-01
-2.63419867e-01 1.24413800e+00 4.88590658e-01 -6.28042817e-01
2.91398843e-03 -4.05172974e-01 2.98008621e-01 -1.82123601e-01
6.00973248e-01 2.14853272e-01 1.76638156e-01 -6.15076303e-01
-2.58090228e-01 2.23054126e-01 -3.83489937e-01 -4.71810043e-01
1.24176979e+00 2.70286888e-01 -1.12360731e-01 3.18791687e-01
1.39523911e+00 2.95916289e-01 -1.12328279e+00 -2.90173411e-01
-2.89814323e-01 1.85219958e-01 2.33009942e-02 -1.43867922e+00
-1.08502948e+00 1.25946534e+00 8.32461417e-02 -2.04568863e-01
7.68860579e-01 -4.08039987e-02 9.86503303e-01 3.14447701e-01
1.86982900e-01 -4.89232093e-01 -1.05395921e-01 9.49830830e-01
1.09179270e+00 -9.28304553e-01 -2.80218065e-01 -2.41271108e-01
-7.18143880e-01 5.92750609e-01 5.24271667e-01 -4.10197563e-02
2.10729524e-01 5.04668772e-01 1.35057703e-01 1.96612775e-01
-1.17523801e+00 1.95918709e-01 3.37256521e-01 2.43638411e-01
5.85992217e-01 2.47272775e-01 -4.01745364e-02 9.31783974e-01
-5.21928370e-01 -5.26965797e-01 1.97065815e-01 3.24627459e-01
-2.69594312e-01 -1.09712088e+00 -1.51747718e-01 2.78909653e-01
-4.22750771e-01 -8.09127331e-01 -2.49572083e-01 7.95548260e-01
-6.25208244e-02 9.24143732e-01 1.47585422e-01 -4.45638984e-01
2.16892138e-01 3.70924741e-01 2.15363845e-01 -7.59114385e-01
-8.01067054e-01 -2.66099125e-01 2.43537873e-01 -5.00239015e-01
3.21149498e-01 -4.85065609e-01 -1.33948684e+00 -4.42828149e-01
-1.48210287e-01 1.26796737e-01 6.70766413e-01 8.05938125e-01
6.32505238e-01 6.60113633e-01 2.92690456e-01 -5.93664587e-01
-8.92150760e-01 -1.10227334e+00 -1.52309880e-01 3.99842113e-02
3.55061054e-01 -7.11903274e-02 -3.48323166e-01 -2.68337391e-02] | [10.972209930419922, 7.682807922363281] |
497947e2-dc7a-474d-b1bc-225526463b25 | graph-anomaly-detection-with-graph-neural | 2209.14930 | null | https://arxiv.org/abs/2209.14930v2 | https://arxiv.org/pdf/2209.14930v2.pdf | Graph Anomaly Detection with Graph Neural Networks: Current Status and Challenges | Graphs are used widely to model complex systems, and detecting anomalies in a graph is an important task in the analysis of complex systems. Graph anomalies are patterns in a graph that do not conform to normal patterns expected of the attributes and/or structures of the graph. In recent years, graph neural networks (GNNs) have been studied extensively and have successfully performed difficult machine learning tasks in node classification, link prediction, and graph classification thanks to the highly expressive capability via message passing in effectively learning graph representations. To solve the graph anomaly detection problem, GNN-based methods leverage information about the graph attributes (or features) and/or structures to learn to score anomalies appropriately. In this survey, we review the recent advances made in detecting graph anomalies using GNN models. Specifically, we summarize GNN-based methods according to the graph type (i.e., static and dynamic), the anomaly type (i.e., node, edge, subgraph, and whole graph), and the network architecture (e.g., graph autoencoder, graph convolutional network). To the best of our knowledge, this survey is the first comprehensive review of graph anomaly detection methods based on GNNs. | ['Sungsu Lim', 'Won-Yong Shin', 'Byung Suk Lee', 'Hwan Kim'] | 2022-09-29 | null | null | null | null | ['graph-anomaly-detection'] | ['graphs'] | [ 8.03242475e-02 2.65098572e-01 2.36670226e-02 -5.06784283e-02
6.08952880e-01 -4.13177609e-01 5.06672680e-01 8.84969652e-01
3.19343567e-01 2.62969673e-01 -3.02785426e-01 -4.37465787e-01
-3.34887475e-01 -1.26340759e+00 -6.91137493e-01 -5.61665595e-01
-7.75921047e-01 4.56768721e-01 2.24233285e-01 -4.09393817e-01
1.11982375e-01 8.09087098e-01 -1.31589174e+00 -6.96145743e-02
7.84305990e-01 9.57001209e-01 -5.72825372e-01 8.79957557e-01
-3.74898076e-01 1.02852571e+00 -6.15964234e-01 -2.70504057e-01
9.71707981e-03 -5.31456470e-01 -5.56711912e-01 2.45497495e-01
4.42138225e-01 3.49007733e-02 -1.07983172e+00 1.46794569e+00
5.38910143e-02 -3.74094918e-02 6.19821250e-01 -1.93371856e+00
-7.53146410e-01 5.87726533e-01 -3.89322013e-01 5.47340810e-01
4.33306575e-01 1.20644430e-02 1.16775942e+00 -5.22133768e-01
4.48394775e-01 1.00997424e+00 8.20611477e-01 2.34014511e-01
-1.04293954e+00 -3.07763040e-01 5.93369424e-01 4.59792972e-01
-1.11379409e+00 6.38025478e-02 1.18176675e+00 -5.67638099e-01
9.93630230e-01 2.95261502e-01 1.03404844e+00 9.77620840e-01
5.99137127e-01 5.20650625e-01 3.88575196e-01 -2.68199474e-01
1.74490541e-01 -4.98574555e-01 2.97200501e-01 1.19940162e+00
7.21233964e-01 -1.32345799e-02 -1.64859101e-01 -2.65490204e-01
4.32364851e-01 3.89350295e-01 -3.67673039e-02 -4.62050200e-01
-8.47124279e-01 8.87954712e-01 8.49537849e-01 4.72892851e-01
-5.65375507e-01 4.45325732e-01 6.48176312e-01 7.27127612e-01
4.73708630e-01 4.47595060e-01 -2.77852029e-01 2.69600004e-01
-1.92171723e-01 -1.10216282e-01 1.03700578e+00 8.18231702e-01
7.34188259e-01 7.63601005e-01 5.14625721e-02 5.55066824e-01
3.97016257e-01 2.95517981e-01 4.14776951e-01 -1.59816861e-01
5.30778803e-02 1.33483648e+00 -9.33670998e-01 -1.63772380e+00
-6.81101322e-01 -4.29616094e-01 -1.35245287e+00 -8.70303214e-02
2.40960851e-01 1.45273909e-01 -1.04646397e+00 1.45328879e+00
2.48966336e-01 4.36239213e-01 -2.29986221e-01 2.86436528e-01
1.03845263e+00 3.30622196e-01 -1.24225736e-01 1.18797429e-01
9.14688587e-01 -6.29115522e-01 -6.71340883e-01 -3.73488635e-01
8.14887643e-01 -1.97883248e-01 6.25111341e-01 1.63183957e-01
-5.10933638e-01 -1.92795634e-01 -1.05069065e+00 4.66783941e-01
-8.37852836e-01 -5.60043693e-01 7.62665153e-01 5.77878714e-01
-1.11310112e+00 8.05781126e-01 -9.23623681e-01 -6.25336289e-01
3.41024250e-01 4.11635607e-01 -4.79135305e-01 -1.40380859e-01
-1.10765052e+00 3.95238727e-01 5.70126295e-01 1.87987298e-01
-7.23820329e-01 -2.53490239e-01 -1.28700364e+00 2.96786755e-01
4.34526950e-01 -2.44966730e-01 7.35777199e-01 -1.10012019e+00
-8.31705451e-01 5.46582043e-01 3.41117620e-01 -6.40511453e-01
-7.09109083e-02 2.69918233e-01 -1.11688793e+00 4.59874794e-02
-2.05952466e-01 -9.65027735e-02 8.35108757e-01 -7.67511725e-01
-3.15961361e-01 -4.37472612e-01 -3.40457223e-02 -3.15684438e-01
-4.95706558e-01 -3.07191283e-01 1.00624487e-01 -5.38929284e-01
4.52923417e-01 -8.54324222e-01 -2.73950011e-01 -1.88985303e-01
-8.50865960e-01 -3.83864135e-01 1.20082951e+00 -5.42029083e-01
1.45098305e+00 -2.03246570e+00 -2.16914132e-01 9.26388979e-01
1.01754999e+00 2.17323244e-01 -1.33644298e-01 7.71146119e-01
-4.96711135e-01 2.10876137e-01 -2.14032024e-01 3.29201192e-01
-1.36653095e-01 5.54069459e-01 1.90977771e-02 6.95081651e-01
2.55069017e-01 1.02460778e+00 -9.79081333e-01 -8.02450031e-02
2.15073466e-01 1.40370071e-01 -3.13109338e-01 2.74906635e-01
-2.96259791e-01 2.43403569e-01 -4.92071390e-01 8.76611173e-01
2.62475103e-01 -5.90999722e-01 4.45964336e-01 -2.69909180e-03
3.84305298e-01 -1.05418175e-01 -9.17682588e-01 7.16962397e-01
2.49581784e-02 8.32141876e-01 -3.24698389e-01 -1.47321486e+00
1.07546997e+00 2.55631834e-01 6.75364733e-01 -6.58034682e-01
3.29516381e-02 1.26531497e-01 6.54936731e-01 -2.89688259e-01
5.66893406e-02 5.02819419e-01 9.10437759e-03 4.49031025e-01
2.27709636e-01 3.63141179e-01 5.27131140e-01 3.98442805e-01
1.72531807e+00 -7.09018230e-01 6.18427336e-01 -4.87381257e-02
7.54359484e-01 -2.10488856e-01 3.84648234e-01 8.14939916e-01
-1.90161124e-01 1.84512779e-01 1.10078657e+00 -1.10600257e+00
-8.92569602e-01 -9.49879110e-01 4.15210158e-01 9.01502252e-01
-6.27726782e-04 -6.36023402e-01 -4.68107462e-01 -1.03398871e+00
3.66778582e-01 1.92163691e-01 -6.98433995e-01 -8.89097512e-01
-5.95921874e-01 -5.71695626e-01 6.66602314e-01 5.54860413e-01
2.30402321e-01 -1.36813903e+00 2.36250669e-01 2.88189113e-01
4.05396849e-01 -1.15439141e+00 -1.58204675e-01 -4.03551422e-02
-9.59053159e-01 -1.66486526e+00 2.07625657e-01 -8.03534985e-01
1.11003768e+00 -6.08046306e-03 1.34773755e+00 8.26029420e-01
-3.93827915e-01 7.64025807e-01 -4.01772320e-01 -4.29563135e-01
-6.26124144e-01 1.21634025e-02 2.07410440e-01 2.96722859e-01
2.83968329e-01 -8.54087472e-01 -3.17272484e-01 1.42384008e-01
-1.00305057e+00 -5.79929769e-01 5.48619628e-01 7.15557158e-01
4.68977034e-01 4.17270064e-01 3.39826524e-01 -1.19664073e+00
9.81254458e-01 -8.71590018e-01 -6.05816841e-01 3.67138535e-01
-1.07989073e+00 -3.07327062e-02 1.07948697e+00 -2.91134059e-01
-4.57381904e-02 -3.14154893e-01 -1.03298709e-01 -5.15457690e-01
-3.24700862e-01 8.95380735e-01 -9.79052484e-02 -4.75606680e-01
5.26199222e-01 3.23502660e-01 1.98033378e-01 -1.07586280e-01
5.13312072e-02 1.39505833e-01 1.91136792e-01 -1.64764270e-01
8.90685678e-01 1.75467670e-01 4.80969489e-01 -9.40090001e-01
-3.48996907e-01 -4.04750705e-01 -5.80599427e-01 -5.64065278e-01
4.81082827e-01 -3.15054119e-01 -6.23519838e-01 6.40023530e-01
-6.89793825e-01 -1.64581522e-01 -3.11330885e-01 1.95414439e-01
-9.76896286e-02 6.50014758e-01 -7.39766061e-01 -5.22583961e-01
-4.50680345e-01 -6.58389926e-01 6.29771352e-01 1.47486001e-01
-1.42722577e-01 -1.56744456e+00 1.19815446e-01 -5.57866395e-01
5.68986773e-01 7.98496664e-01 1.17046309e+00 -1.34221292e+00
-4.84917700e-01 -7.96921670e-01 -1.96745545e-01 1.69515952e-01
3.69451553e-01 2.88448036e-01 -4.67276096e-01 -5.40283799e-01
-4.52160925e-01 2.76668370e-01 5.73820293e-01 3.07292074e-01
1.34302282e+00 -5.97416401e-01 -4.31681156e-01 6.12717807e-01
1.23136246e+00 2.17691272e-01 3.64600390e-01 3.96122187e-02
1.11038005e+00 1.75999641e-01 -7.47998208e-02 2.84102261e-01
1.66424453e-01 1.28353879e-01 9.14642096e-01 -2.47558188e-02
4.65393737e-02 -3.08849990e-01 3.36141884e-01 1.02961326e+00
-3.43802087e-02 -6.44695222e-01 -1.16980219e+00 2.27918267e-01
-1.81400251e+00 -6.96569860e-01 -3.84275287e-01 1.77401638e+00
-2.29701027e-01 4.08373654e-01 1.39800355e-01 2.66615182e-01
1.07773805e+00 5.19374251e-01 -7.60574937e-01 -5.96138954e-01
-7.55159035e-02 3.03003676e-02 3.24296802e-01 2.07131445e-01
-1.09477651e+00 7.37174690e-01 5.88537836e+00 2.73193628e-01
-1.01557529e+00 -4.26099747e-01 3.78566116e-01 6.23484731e-01
-2.03147382e-01 1.27074856e-03 -5.01219667e-02 3.50375682e-01
9.99666691e-01 -3.44430238e-01 6.16542876e-01 9.00830388e-01
-3.58047426e-01 4.58331227e-01 -1.15195048e+00 7.25270748e-01
1.96461439e-01 -1.07448161e+00 2.28469074e-01 1.84690222e-01
4.61511523e-01 2.22581521e-01 -7.72394463e-02 4.02916640e-01
4.34086949e-01 -1.11421418e+00 -5.97067848e-02 5.94569087e-01
5.27226031e-01 -6.21323586e-01 9.58904207e-01 4.53370176e-02
-1.57756913e+00 -2.57140994e-01 -1.35056123e-01 -1.63713172e-01
-1.39733225e-01 9.01221335e-01 -7.84416378e-01 7.34719694e-01
6.39866590e-01 1.17779446e+00 -8.43596399e-01 9.71906066e-01
-3.45328510e-01 8.11131656e-01 -1.94173768e-01 -4.10784874e-03
2.12531134e-01 -4.14336622e-01 8.61374497e-01 7.02312052e-01
2.46310472e-01 -2.83057421e-01 5.78967094e-01 6.60397768e-01
-2.45199889e-01 -1.04179969e-02 -1.10212040e+00 -6.38357341e-01
2.04571366e-01 1.26882887e+00 -9.44957018e-01 -1.23547226e-01
-6.29693568e-01 6.72913373e-01 3.83697331e-01 3.68527770e-01
-3.55185181e-01 -5.59510469e-01 7.00013697e-01 1.45774037e-01
9.67004430e-03 -2.68679172e-01 2.21700773e-01 -1.00969708e+00
4.05919217e-02 -8.41958225e-01 9.77262139e-01 -2.91892141e-01
-1.69361675e+00 8.78025353e-01 -3.47838044e-01 -1.15477788e+00
-3.28809202e-01 -8.71561885e-01 -1.17493975e+00 2.61454761e-01
-1.06340635e+00 -8.66936088e-01 -7.56015062e-01 8.27416480e-01
-4.57293093e-02 -5.50045013e-01 8.39485526e-01 1.05369739e-01
-7.67876744e-01 4.00557011e-01 1.48466779e-02 8.07166517e-01
1.00530796e-01 -1.52303517e+00 8.68982553e-01 9.84994709e-01
4.61949617e-01 2.47904539e-01 5.11915028e-01 -8.41292918e-01
-1.46502829e+00 -1.29690754e+00 4.65069652e-01 -1.01126045e-01
1.10603678e+00 -4.28216368e-01 -1.14668524e+00 1.06006181e+00
-2.08755136e-01 7.11206317e-01 5.31378984e-01 2.96262741e-01
-3.72177422e-01 -1.80437952e-01 -1.05605197e+00 5.75609803e-01
1.20992649e+00 -5.45562387e-01 5.49509712e-02 4.04026389e-01
5.05167902e-01 -4.17427242e-01 -8.85673940e-01 5.23431063e-01
2.78187305e-01 -9.72163737e-01 7.60353923e-01 -1.10078585e+00
9.93674099e-02 -1.97119534e-01 1.21796131e-01 -1.48612738e+00
-5.53638160e-01 -3.93582582e-01 -9.91783321e-01 7.72050560e-01
2.66846567e-01 -1.07529449e+00 1.02543342e+00 1.91633344e-01
-3.20160002e-01 -8.09294760e-01 -7.57330060e-01 -6.68629646e-01
-4.47183251e-01 -2.78949112e-01 8.43756378e-01 1.02079380e+00
-9.86399576e-02 1.80638015e-01 -1.07484192e-01 4.57507849e-01
5.10259151e-01 6.20423071e-02 6.81499898e-01 -1.85907149e+00
5.85086383e-02 -7.79219389e-01 -1.33045733e+00 -3.15445513e-01
2.90019065e-01 -1.18990946e+00 -4.08347458e-01 -1.52419901e+00
-3.41211289e-01 -1.73171222e-01 -4.93902296e-01 3.99481058e-01
-1.71784073e-01 -4.83671390e-02 -2.16221675e-01 -9.52154677e-03
-7.39500105e-01 4.24700946e-01 8.41441929e-01 -3.99333835e-01
-1.16738841e-01 2.93111295e-01 -4.14165288e-01 8.10294628e-01
1.13385665e+00 -3.48049194e-01 -3.66526335e-01 -1.87652111e-02
5.49414575e-01 -1.08900614e-01 4.08459634e-01 -1.11233532e+00
3.72797430e-01 1.88750550e-01 3.54821503e-01 -2.50035107e-01
-2.22030014e-01 -1.00739133e+00 2.52170078e-02 9.01325166e-01
-5.17026745e-02 7.41422713e-01 -5.03118709e-02 1.09835863e+00
-5.18084705e-01 1.67912871e-01 3.80124629e-01 -3.87358814e-02
-9.72765863e-01 1.01657820e+00 -5.79331517e-01 3.32140215e-02
1.06454897e+00 -1.65795714e-01 -3.83377671e-01 -7.24119604e-01
-8.74873817e-01 2.50029117e-01 8.10441226e-02 5.82628310e-01
8.24714661e-01 -1.41457462e+00 -5.32665372e-01 6.16986036e-01
5.07623315e-01 -1.28909320e-01 1.95569426e-01 8.29452276e-01
-6.98692501e-01 9.82575119e-02 -2.91125566e-01 -6.90121233e-01
-1.17917168e+00 6.13555431e-01 6.26843691e-01 -6.55411124e-01
-9.19301033e-01 4.25547123e-01 -3.98379192e-02 -5.76562226e-01
1.10721305e-01 -1.17563449e-01 -4.62018132e-01 -3.06964070e-01
9.61302146e-02 3.55954021e-01 2.50228852e-01 -4.21506107e-01
-3.89985919e-01 1.78961664e-01 -1.84760228e-01 9.47491944e-01
1.10669565e+00 2.52238840e-01 -5.38327634e-01 3.94115359e-01
9.21863496e-01 -2.10604444e-01 -5.76449513e-01 -3.38331491e-01
4.29986566e-01 -1.78458557e-01 -2.44356453e-01 -2.66050696e-01
-1.57707953e+00 5.51869869e-01 3.52564722e-01 1.16242218e+00
9.85616326e-01 1.33120999e-01 6.46293998e-01 6.52978480e-01
4.29666862e-02 -7.43888557e-01 3.20084542e-01 7.57411897e-01
8.37564290e-01 -1.11219037e+00 -9.36150923e-02 -4.94720191e-01
-1.29616216e-01 1.46140027e+00 9.70904827e-01 -5.44125140e-01
1.16970670e+00 -8.00677910e-02 -2.71260321e-01 -8.84673595e-01
-5.88730872e-01 -1.01448804e-01 5.18959284e-01 7.39455581e-01
2.61457920e-01 1.06497444e-01 2.66498089e-01 9.82260182e-02
-8.51300433e-02 -7.76317179e-01 6.56628489e-01 6.46066070e-01
-3.87374431e-01 -8.15014780e-01 -2.80936132e-03 1.21616471e+00
-3.07361364e-01 -3.27118337e-02 -8.35082769e-01 9.33730304e-01
-3.49121243e-01 7.81577289e-01 3.10806751e-01 -6.11339033e-01
4.91803139e-01 -6.59531876e-02 1.24689534e-01 -4.99332905e-01
-3.09173673e-01 -7.72993684e-01 1.69518311e-02 -8.02347124e-01
5.44664860e-02 -2.33013555e-01 -1.08836019e+00 -5.82359850e-01
-3.19018573e-01 1.31380960e-01 3.11801970e-01 9.12190318e-01
5.12097716e-01 9.49386299e-01 7.02153563e-01 -1.96373016e-01
-4.06701379e-02 -9.73907411e-01 -9.31536913e-01 5.91729879e-01
4.50954676e-01 -6.14923596e-01 -5.84649086e-01 -6.19910419e-01] | [6.655348777770996, 5.8279619216918945] |
ee646f8e-b95c-46fe-8393-ce3582d27c70 | unsupervised-explanation-generation-via | 2211.11160 | null | https://arxiv.org/abs/2211.11160v1 | https://arxiv.org/pdf/2211.11160v1.pdf | Unsupervised Explanation Generation via Correct Instantiations | While large pre-trained language models (PLM) have shown their great skills at solving discriminative tasks, a significant gap remains when compared with humans for explanation-related tasks. Among them, explaining the reason why a statement is wrong (e.g., against commonsense) is incredibly challenging. The major difficulty is finding the conflict point, where the statement contradicts our real world. This paper proposes Neon, a two-phrase, unsupervised explanation generation framework. Neon first generates corrected instantiations of the statement (phase I), then uses them to prompt large PLMs to find the conflict point and complete the explanation (phase II). We conduct extensive experiments on two standard explanation benchmarks, i.e., ComVE and e-SNLI. According to both automatic and human evaluations, Neon outperforms baselines, even for those with human-annotated instantiations. In addition to explaining a negative prediction, we further demonstrate that Neon remains effective when generalizing to different scenarios. | ['Lingpeng Kong', 'Yang Liu', 'Zhixing Li', 'Jiangjie Chen', 'Zhiyong Wu', 'Sijie Cheng'] | 2022-11-21 | null | null | null | null | ['explanation-generation'] | ['natural-language-processing'] | [ 5.48168898e-01 6.76822662e-01 -2.74677217e-01 -4.53845471e-01
-8.98078620e-01 -5.54761887e-01 6.73639655e-01 3.70040894e-01
4.55865301e-02 8.64101648e-01 4.55203086e-01 -8.57425869e-01
1.46704406e-01 -3.13492835e-01 -5.51797152e-01 -7.58827701e-02
4.12043750e-01 8.54705215e-01 -1.96074903e-01 -1.47439674e-01
6.89760685e-01 -1.00999296e-01 -1.24233103e+00 5.35158932e-01
1.15866840e+00 3.79230738e-01 7.72177279e-02 3.67887944e-01
-4.77292240e-01 9.83642578e-01 -6.71602786e-01 -9.52739954e-01
-2.26730153e-01 -7.55249500e-01 -1.23688138e+00 5.64256944e-02
3.60595465e-01 -2.50908405e-01 3.52102727e-01 9.97393489e-01
1.16252385e-01 1.44358322e-01 7.90426850e-01 -1.59444547e+00
-9.45537865e-01 1.24124825e+00 -4.75635529e-01 1.32656798e-01
6.29599750e-01 3.14147145e-01 1.44220042e+00 -1.24807799e+00
5.34531653e-01 1.42666519e+00 5.96771240e-01 9.82132554e-01
-1.18918550e+00 -7.00302124e-01 2.64158010e-01 4.67121959e-01
-9.33144927e-01 -2.66549587e-01 7.09141552e-01 -1.38880953e-01
1.49719453e+00 4.42368478e-01 3.47802669e-01 1.46132445e+00
2.52272308e-01 7.66358197e-01 9.64276969e-01 -4.28853452e-01
1.90043524e-01 1.37242928e-01 3.42916995e-01 4.58166271e-01
5.97329795e-01 -2.57318228e-01 -6.10388041e-01 -1.03381433e-01
3.11586916e-01 -8.96362960e-02 -2.74179339e-01 2.35854194e-01
-1.30682826e+00 9.53479826e-01 3.08105111e-01 1.17613755e-01
-5.07568538e-01 2.17150971e-01 2.53909618e-01 5.78048360e-03
1.55308768e-01 1.01407051e+00 -6.25593007e-01 -2.62016237e-01
-8.41115892e-01 5.55803061e-01 1.01282287e+00 1.00459683e+00
5.88940620e-01 1.03702337e-01 -2.57096469e-01 5.75249791e-01
2.82821059e-01 2.41656229e-01 4.43804622e-01 -9.93514419e-01
7.47821212e-01 7.20655084e-01 1.86405793e-01 -1.04519165e+00
-3.73555750e-01 -7.00205863e-01 -7.52523959e-01 -2.18401864e-01
2.37014547e-01 -6.30213916e-02 -5.87212622e-01 1.88722658e+00
-3.11879106e-02 2.22612336e-01 3.67349923e-01 1.00556469e+00
1.14662051e+00 6.27874136e-01 3.82033259e-01 -2.25247979e-01
1.32511044e+00 -1.32649291e+00 -7.32756555e-01 -1.17667377e+00
5.68235099e-01 -5.78049898e-01 1.51803827e+00 3.28146964e-01
-1.02718186e+00 -3.35251421e-01 -8.83875489e-01 -3.65749955e-01
-1.19178168e-01 2.38271765e-02 9.77582693e-01 -7.38364905e-02
-7.10288227e-01 3.14593405e-01 -4.00943875e-01 -2.66466945e-01
2.10022569e-01 1.62582442e-01 -1.80841565e-01 -1.18053459e-01
-1.31425261e+00 1.06911492e+00 5.45737386e-01 -1.29846320e-01
-6.59070909e-01 -7.02502906e-01 -1.17664874e+00 4.61933374e-01
5.15213847e-01 -9.63821769e-01 1.52302730e+00 -9.15176034e-01
-9.23988879e-01 9.15024281e-01 -6.39241457e-01 -7.73416877e-01
4.31681603e-01 -4.20992613e-01 -3.65560293e-01 -3.50056648e-01
6.29721284e-01 9.60682392e-01 5.50104856e-01 -1.36012232e+00
-3.47256422e-01 3.90376821e-02 1.97668880e-01 1.49610806e-02
9.40986276e-02 5.89915849e-02 -1.80495486e-01 -5.56616724e-01
3.37205797e-01 -8.33183646e-01 -3.59581769e-01 -4.95839238e-01
-9.44279492e-01 -6.77226007e-01 5.80781162e-01 -5.69567382e-01
1.34999073e+00 -1.76183963e+00 1.16304323e-01 -8.18049535e-02
3.14931750e-01 4.67099436e-02 -2.22743109e-01 4.62041765e-01
-4.82024074e-01 6.42920613e-01 -4.23874676e-01 -5.79794049e-01
2.86981434e-01 3.07059258e-01 -9.11220074e-01 -3.84864241e-01
4.28133965e-01 1.15486479e+00 -1.08273709e+00 -5.36672950e-01
-4.74779569e-02 1.65430635e-01 -7.56392062e-01 2.12866440e-01
-5.45446336e-01 3.06580991e-01 -3.10644299e-01 4.87873524e-01
3.27591032e-01 -9.63617563e-01 1.12932548e-01 6.41133031e-03
2.70005822e-01 9.11967635e-01 -8.90121639e-01 1.45979619e+00
-4.74956959e-01 7.61684358e-01 -7.16255128e-01 -5.60038626e-01
7.61863112e-01 1.57596201e-01 -4.25783843e-01 -4.10717577e-01
-1.27984256e-01 3.52942288e-01 1.64024442e-01 -5.45224607e-01
6.05238020e-01 -3.77174765e-01 -5.38896695e-02 7.04152465e-01
-1.87401026e-01 -3.25374514e-01 1.97651178e-01 7.28311718e-01
1.10318112e+00 6.76446408e-02 9.00750756e-01 -7.76822818e-03
5.41897118e-01 2.86328435e-01 6.35899901e-01 9.20083940e-01
-1.57531593e-02 6.69777691e-01 8.31775963e-01 -5.51054239e-01
-7.27877259e-01 -8.08585942e-01 3.78670663e-01 6.40977383e-01
3.56967151e-01 -8.65488350e-01 -4.77985144e-01 -1.04699636e+00
-9.75693539e-02 1.87796938e+00 -6.01405203e-01 -1.82402834e-01
-4.59594458e-01 -3.06503028e-01 1.79531977e-01 6.39415562e-01
5.83472610e-01 -1.36411273e+00 -7.81618357e-01 1.15780935e-01
-7.14621186e-01 -1.27769756e+00 -3.10212582e-01 2.63371885e-01
-7.16940463e-01 -1.02520585e+00 1.71559989e-01 -4.21154022e-01
8.71968210e-01 3.90254825e-01 1.74980283e+00 6.66523337e-01
1.86131760e-01 1.19644791e-01 -2.41966590e-01 -6.66398287e-01
-5.20139158e-01 -1.04620606e-01 -2.59245783e-01 -6.37723029e-01
6.13199711e-01 -5.22243619e-01 -3.18027377e-01 1.66440427e-01
-7.39902020e-01 6.13845348e-01 6.24922514e-01 7.83707261e-01
5.49886227e-01 -9.12624821e-02 5.48272252e-01 -1.16608775e+00
9.48032141e-01 -6.61430180e-01 -6.63961768e-02 2.91087121e-01
-8.17402959e-01 2.93700993e-01 7.02871025e-01 -2.73638457e-01
-9.04441535e-01 -3.44681323e-01 -4.19113263e-02 -2.74065020e-03
-2.13789642e-01 6.55461788e-01 -1.40381202e-01 7.10110545e-01
7.38545060e-01 1.62346706e-01 -5.27482808e-01 -5.31718619e-02
2.78676718e-01 2.19562232e-01 7.12668300e-01 -6.69064462e-01
9.87894237e-01 4.41846699e-02 -2.71398902e-01 -1.50756434e-01
-1.53151047e+00 -1.36876866e-01 -1.81954116e-01 6.87798485e-02
7.60990441e-01 -6.41500533e-01 -5.81700444e-01 -3.92272025e-01
-1.85884619e+00 -3.21942747e-01 -2.22350612e-01 2.34878734e-01
-4.50478196e-01 2.83713788e-01 -3.08110535e-01 -8.31739128e-01
-4.34025586e-01 -9.86032844e-01 1.04259014e+00 2.88701534e-01
-1.07514620e+00 -9.86664355e-01 -8.99468437e-02 5.88635147e-01
3.94665182e-01 8.85741115e-02 1.35499537e+00 -1.14264810e+00
-4.01947975e-01 -2.60073561e-02 -3.11908990e-01 -5.37775569e-02
-4.97755036e-02 -7.77641013e-02 -6.42597437e-01 3.39193195e-01
-8.95985588e-02 -5.18634677e-01 8.36263955e-01 -7.73845911e-02
1.48494077e+00 -6.83183372e-01 -3.04846227e-01 2.06608117e-01
1.04001951e+00 -1.25116721e-01 6.12546384e-01 5.06949484e-01
3.88935953e-01 4.67900783e-01 6.83113396e-01 2.75033712e-01
7.51438439e-01 6.03112876e-01 6.64880157e-01 -6.81905821e-02
-1.45242646e-01 -7.25866377e-01 3.70935291e-01 3.18588048e-01
1.56177253e-01 -5.41187942e-01 -1.17710078e+00 4.99630809e-01
-2.07628202e+00 -1.07019079e+00 -4.78723764e-01 1.80519056e+00
7.90561736e-01 4.11164880e-01 -4.15517569e-01 1.13555260e-01
3.37945789e-01 1.38861790e-01 -6.97461545e-01 -7.30028868e-01
-2.77919322e-01 1.63361069e-03 -3.37401360e-01 7.07862616e-01
-7.32092917e-01 1.21271086e+00 6.11083364e+00 5.11714160e-01
-7.31528103e-01 -5.35643939e-03 6.29007578e-01 3.04404259e-01
-9.51504648e-01 4.37464744e-01 -5.93911111e-01 1.76889300e-01
6.98321402e-01 -1.89834461e-01 3.24079663e-01 1.00451791e+00
1.49126932e-01 1.04873598e-01 -1.43964231e+00 8.43608677e-01
4.38289404e-01 -1.31414187e+00 6.15834951e-01 -3.40774387e-01
6.87377095e-01 -3.42001945e-01 -5.99628128e-02 6.66370571e-01
2.04958156e-01 -1.28348577e+00 7.88067758e-01 1.14384685e-02
3.84187818e-01 -5.42762399e-01 9.77301657e-01 6.84615552e-01
-7.30786443e-01 4.78132926e-02 -3.21586609e-01 -4.29023147e-01
2.43291244e-01 5.71231961e-01 -1.13431966e+00 2.60552645e-01
2.80789405e-01 7.29939878e-01 -6.88189387e-01 5.49173653e-01
-1.40310061e+00 6.75427198e-01 1.49387449e-01 -7.14540407e-02
2.50084609e-01 9.06095430e-02 6.10633373e-01 1.31545293e+00
3.14161539e-01 4.25691694e-01 -8.38242471e-02 1.66928256e+00
-2.96380550e-01 -8.22557285e-02 -5.07481515e-01 -2.90616333e-01
4.56490517e-01 1.32336354e+00 -4.75358844e-01 -6.26612604e-01
-2.97400504e-01 1.24077511e+00 4.86018151e-01 4.45718199e-01
-1.23528600e+00 3.82267754e-03 3.88507307e-01 -1.34563133e-01
-1.51701421e-01 -4.59454581e-02 -7.67972589e-01 -1.38908696e+00
-7.68814087e-02 -1.17176986e+00 4.41196352e-01 -1.36489534e+00
-1.19190633e+00 6.84644520e-01 -1.35571867e-01 -9.20570672e-01
-6.04661465e-01 -3.76006693e-01 -1.15116787e+00 8.23465407e-01
-1.59442973e+00 -1.06399608e+00 -4.35109317e-01 7.24437460e-02
1.12139952e+00 2.00336412e-01 9.25885320e-01 -2.93549985e-01
-5.77694297e-01 4.28812951e-01 -1.12008250e+00 -1.74408227e-01
5.45701265e-01 -1.49133861e+00 7.54109144e-01 1.03738165e+00
4.52483743e-01 1.08935273e+00 1.30913055e+00 -8.01638007e-01
-1.10205793e+00 -9.42837358e-01 1.66644943e+00 -7.19766557e-01
5.50433278e-01 -1.19722836e-01 -1.17403185e+00 9.96381581e-01
3.94482195e-01 -4.46675301e-01 7.89951384e-01 3.26738894e-01
-7.57530153e-01 6.24207377e-01 -1.01641464e+00 9.60412264e-01
1.10323358e+00 -3.68501276e-01 -1.18574095e+00 7.71002412e-01
9.74879920e-01 -5.26307583e-01 1.38767436e-02 2.45963573e-01
2.63597846e-01 -9.85097945e-01 7.80204177e-01 -1.27197433e+00
1.38265574e+00 -3.03900778e-01 -1.04631625e-01 -1.42654371e+00
-2.62544990e-01 -6.82211161e-01 -1.45458102e-01 1.24959147e+00
1.06910741e+00 -4.89486188e-01 4.60133284e-01 1.15142047e+00
-2.83025086e-01 -7.44797051e-01 -5.43897271e-01 -5.88814557e-01
-2.11199015e-01 -6.13833547e-01 6.47993922e-01 9.96540844e-01
3.24622661e-01 8.35561812e-01 -3.24482232e-01 2.91061908e-01
6.13045156e-01 4.85565454e-01 9.57063913e-01 -9.43976998e-01
-4.72920388e-01 -5.04499912e-01 1.29714578e-01 -1.03096938e+00
7.40142226e-01 -1.03053248e+00 2.54102379e-01 -1.99713683e+00
7.27139413e-01 2.18553394e-02 2.62494422e-02 9.23888445e-01
-8.27475309e-01 -1.25082135e-01 2.73151040e-01 2.05540955e-01
-8.42554927e-01 3.14868003e-01 9.58443880e-01 -1.49674729e-01
2.07443330e-02 -2.55767941e-01 -1.22546017e+00 1.11069477e+00
8.73668313e-01 -5.63574016e-01 -3.54458004e-01 -7.35758603e-01
6.43847048e-01 -4.07157056e-02 6.38697863e-01 -6.98799729e-01
1.84095904e-01 -5.55712104e-01 1.28477126e-01 -5.47872186e-01
3.24833170e-02 -6.18308783e-01 3.79928239e-02 4.94715273e-01
-8.86389554e-01 4.82432067e-01 3.22231591e-01 3.91420364e-01
-2.06207618e-01 -2.68290937e-01 3.94637227e-01 -4.17313054e-02
-7.38419056e-01 -1.90673128e-01 3.11026424e-02 3.75419617e-01
6.20469332e-01 -7.11498857e-02 -6.82650745e-01 -7.67678142e-01
-4.04184490e-01 4.98838127e-01 2.95311540e-01 4.49696302e-01
9.49359417e-01 -1.11961544e+00 -8.32476795e-01 -1.93028405e-01
3.21296006e-01 -5.77349849e-02 6.56019002e-02 7.23469973e-01
2.13187393e-02 5.92762649e-01 1.69813722e-01 -2.58010447e-01
-1.18439531e+00 4.62222248e-01 2.57385105e-01 -4.81459588e-01
-6.86682105e-01 9.75382209e-01 3.47820580e-01 -3.32493365e-01
1.46434695e-01 -5.20335436e-01 -1.52375832e-01 -4.29768264e-01
7.12662280e-01 -3.56246792e-02 -2.00300992e-01 -2.16068953e-01
-6.41089261e-01 1.78486392e-01 -9.44252010e-04 -7.88283050e-02
1.36107743e+00 -8.08732235e-04 -2.64700532e-01 4.32851642e-01
5.97455263e-01 1.45925790e-01 -7.23120868e-01 -1.52179420e-01
2.76617765e-01 -4.13475484e-01 -4.46291506e-01 -1.38727784e+00
-5.79245567e-01 1.00316560e+00 -3.97071451e-01 1.08230248e-01
7.64238358e-01 2.63204753e-01 8.39015663e-01 4.95568573e-01
2.01023042e-01 -5.71732700e-01 2.78042227e-01 6.11593425e-01
1.52060592e+00 -1.56793773e+00 -4.22217734e-02 -5.23381054e-01
-1.03976786e+00 1.10744369e+00 1.10455823e+00 2.42900550e-01
-2.92676359e-01 -4.45536673e-02 -2.65177060e-02 -3.77739340e-01
-1.28140521e+00 1.77681252e-01 3.38578105e-01 2.63966024e-01
8.59645247e-01 2.02806950e-01 -4.30736452e-01 1.09425759e+00
-5.81150651e-01 -3.42948049e-01 7.16803074e-01 2.73480922e-01
-4.57334280e-01 -8.00017178e-01 -1.78378046e-01 3.09077859e-01
-2.95537293e-01 -5.81672132e-01 -1.09315979e+00 8.77513409e-01
1.20408181e-03 1.22211444e+00 -3.23773980e-01 -1.90388754e-01
1.95288047e-01 1.36444852e-01 9.28709805e-02 -9.98943686e-01
-7.10748911e-01 -4.71956283e-01 4.42124009e-01 -7.23506033e-01
-1.82175711e-01 -4.73589510e-01 -1.97699726e+00 -2.37090915e-01
-1.09413736e-01 4.77565855e-01 5.36817670e-01 1.33233380e+00
3.09685439e-01 5.43868661e-01 8.99998099e-02 -2.93821007e-01
-7.17552125e-01 -7.93547809e-01 1.35439828e-01 6.35703564e-01
2.17502296e-01 -4.49276268e-01 -7.62639821e-01 -1.23068802e-01] | [9.58715534210205, 6.943080425262451] |
49107e5a-0f75-4973-8531-8371adbb2e4d | general-adversarial-defense-against-black-box | 2212.05387 | null | https://arxiv.org/abs/2212.05387v1 | https://arxiv.org/pdf/2212.05387v1.pdf | General Adversarial Defense Against Black-box Attacks via Pixel Level and Feature Level Distribution Alignments | Deep Neural Networks (DNNs) are vulnerable to the black-box adversarial attack that is highly transferable. This threat comes from the distribution gap between adversarial and clean samples in feature space of the target DNNs. In this paper, we use Deep Generative Networks (DGNs) with a novel training mechanism to eliminate the distribution gap. The trained DGNs align the distribution of adversarial samples with clean ones for the target DNNs by translating pixel values. Different from previous work, we propose a more effective pixel level training constraint to make this achievable, thus enhancing robustness on adversarial samples. Further, a class-aware feature-level constraint is formulated for integrated distribution alignment. Our approach is general and applicable to multiple tasks, including image classification, semantic segmentation, and object detection. We conduct extensive experiments on different datasets. Our strategy demonstrates its unique effectiveness and generality against black-box attacks. | ['Jiaya Jia', 'Philip Torr', 'Hengshuang Zhao', 'Xiaogang Xu'] | 2022-12-11 | null | null | null | null | ['adversarial-defense'] | ['adversarial'] | [ 5.78271508e-01 1.93285808e-01 -6.37046471e-02 -2.67060220e-01
-7.15815425e-01 -1.10203731e+00 5.22215605e-01 -4.66432810e-01
-4.53904122e-01 7.50062048e-01 -3.52449149e-01 -2.80926079e-01
2.62595952e-01 -1.14573514e+00 -1.16137457e+00 -1.07258642e+00
3.57692510e-01 -5.73504381e-02 4.60160464e-01 -7.97419101e-02
4.16517854e-02 8.21590185e-01 -9.06168699e-01 6.82673091e-03
1.08747768e+00 9.93486106e-01 -1.75715387e-01 6.46203876e-01
-2.00118423e-02 6.37707949e-01 -1.18249023e+00 -6.51065171e-01
8.48447442e-01 -5.27375519e-01 -4.40961570e-01 -8.54692012e-02
7.39804924e-01 -4.98237520e-01 -6.81533277e-01 1.89725471e+00
4.90230441e-01 -7.95130655e-02 6.59649134e-01 -1.56503713e+00
-1.03048468e+00 5.24203777e-01 -5.95632076e-01 -1.86276045e-02
-2.77482808e-01 4.12803978e-01 4.15436834e-01 -2.85519809e-01
4.28743601e-01 1.32815063e+00 5.41396618e-01 1.03228319e+00
-8.96704853e-01 -1.01190019e+00 1.88718960e-01 1.16686242e-04
-1.11994815e+00 -1.01593837e-01 8.62450898e-01 -2.70843238e-01
1.80110142e-01 3.90805662e-01 2.52781630e-01 1.51579332e+00
3.17899644e-01 7.67795026e-01 1.13014174e+00 -2.91673481e-01
3.65196049e-01 1.53350219e-01 -1.97826967e-01 3.13988537e-01
5.47528446e-01 3.02420676e-01 7.70660862e-02 -8.64556208e-02
9.85797942e-01 1.38061330e-01 -2.63545245e-01 -4.61600363e-01
-6.13853514e-01 8.51530790e-01 6.94112062e-01 2.36231685e-02
9.91210416e-02 2.55327255e-01 2.99008459e-01 2.05322489e-01
2.41841957e-01 3.17684919e-01 -3.78673017e-01 4.27119076e-01
-4.72006172e-01 2.55506665e-01 6.57182932e-01 1.14868617e+00
5.83407402e-01 4.33966637e-01 -3.59457254e-01 5.60461700e-01
2.24648908e-01 9.30779278e-01 3.21311325e-01 -8.43382895e-01
5.03510714e-01 2.80783296e-01 -1.87485710e-01 -1.05355918e+00
1.41737223e-01 -4.94332194e-01 -9.61563766e-01 8.09804142e-01
5.87825000e-01 -5.34746468e-01 -1.35570514e+00 1.95249534e+00
5.18093050e-01 2.47006908e-01 2.91530788e-01 8.46252084e-01
5.63744545e-01 5.22277772e-01 -4.97347042e-02 2.22069666e-01
9.49774444e-01 -7.88006783e-01 -6.62776828e-01 -4.20841098e-01
1.63042009e-01 -5.83534956e-01 8.21796298e-01 1.48327172e-01
-6.45691931e-01 -5.33807456e-01 -1.26140702e+00 4.20479476e-02
-5.14742434e-01 -2.30657786e-01 4.29306626e-01 1.26666915e+00
-5.24878681e-01 5.74413419e-01 -8.15279961e-01 7.88936317e-02
1.07577169e+00 3.44430953e-01 -3.15090150e-01 -5.63516915e-02
-1.30011475e+00 5.53176641e-01 4.99616444e-01 9.15684104e-02
-1.06088328e+00 -7.21587121e-01 -9.38850343e-01 -8.76811743e-02
2.32384294e-01 -5.30244112e-01 9.12590802e-01 -1.40547144e+00
-1.58948207e+00 7.34776199e-01 4.61791247e-01 -6.39115989e-01
9.13902223e-01 -2.43743733e-01 -3.11819226e-01 1.69074088e-01
-2.62347255e-02 6.25704825e-01 1.24970126e+00 -1.47661948e+00
-4.57927734e-01 -4.90178257e-01 9.55240130e-02 -1.71602100e-01
-5.46746790e-01 3.35463299e-03 -2.17553407e-01 -1.09195840e+00
-1.23651020e-01 -8.64911079e-01 -3.22440922e-01 3.36235851e-01
-8.03324580e-01 3.63581240e-01 1.16079342e+00 -4.62205023e-01
6.45734847e-01 -2.33279562e+00 -1.67860627e-01 3.50615114e-01
1.61820009e-01 7.58794248e-01 -2.65664309e-01 -6.43285364e-02
-1.32347003e-01 1.35160550e-01 -6.28457129e-01 -5.70480004e-02
1.29367262e-01 4.10517752e-01 -8.08852494e-01 6.73930764e-01
3.87566328e-01 9.70538795e-01 -8.02846730e-01 -2.47648075e-01
1.78850725e-01 5.46876907e-01 -5.29833972e-01 1.69224650e-01
-2.17935517e-01 5.36004305e-01 -5.10874212e-01 6.51782215e-01
1.37504268e+00 3.23222280e-01 -2.05692813e-01 -1.19989395e-01
4.53102201e-01 -3.50137889e-01 -9.25867617e-01 1.21388447e+00
-7.18927309e-02 5.32643974e-01 1.08430192e-01 -9.22792435e-01
1.04538119e+00 -9.99277309e-02 3.13525237e-02 -5.80527782e-01
4.43999559e-01 1.19866394e-01 2.06886068e-01 -1.22604750e-01
1.75395563e-01 -5.14082611e-02 -1.98584735e-01 1.69411182e-01
2.77472436e-02 -2.88755566e-01 -3.01155061e-01 9.28703994e-02
9.30549204e-01 1.48346545e-02 1.53704407e-02 -2.97843069e-01
4.33654457e-01 -3.16664308e-01 8.66087079e-01 9.54729915e-01
-5.03326118e-01 7.74906933e-01 6.66345119e-01 -2.49396726e-01
-1.07896066e+00 -1.32757449e+00 -1.66115940e-01 5.83643854e-01
4.22934979e-01 2.71127284e-01 -1.24775875e+00 -1.19361889e+00
9.32639614e-02 4.80496824e-01 -7.51192868e-01 -4.71329361e-01
-5.34450233e-01 -7.17598736e-01 9.90562677e-01 6.62799299e-01
8.80734622e-01 -9.17110085e-01 -1.70035809e-01 -1.13462217e-01
2.75646389e-01 -1.35793984e+00 -5.20271540e-01 6.85102046e-02
-6.25037253e-01 -1.00526643e+00 -7.66532242e-01 -6.97323620e-01
1.05403078e+00 -4.60500829e-02 5.77593446e-01 -2.56051868e-01
-2.99340963e-01 8.10190365e-02 -2.91838229e-01 -8.24678719e-01
-7.35296667e-01 -8.41118395e-02 -6.47053421e-02 6.98383227e-02
7.75827914e-02 -5.74839056e-01 -5.49271643e-01 3.91619325e-01
-1.35799181e+00 -3.49051982e-01 4.57918018e-01 7.61610150e-01
6.19743407e-01 2.10098356e-01 5.54009974e-01 -1.08069026e+00
3.55307281e-01 -2.93170571e-01 -9.13434029e-01 8.79291669e-02
-1.09469309e-01 -1.18544407e-01 1.10872638e+00 -7.14912236e-01
-8.96432400e-01 1.01249382e-01 -3.29835504e-01 -6.95394933e-01
-3.86501282e-01 -3.53371888e-01 -9.29792821e-01 -5.10418057e-01
7.38575459e-01 1.23879565e-02 -2.82355770e-02 -1.82409346e-01
4.93936300e-01 4.21590626e-01 8.95197332e-01 -5.83793461e-01
1.42401040e+00 7.69168198e-01 1.76030248e-01 -5.13144076e-01
-7.60704100e-01 3.42096090e-01 -4.63362634e-01 1.67038992e-01
8.35619152e-01 -5.88816464e-01 -4.27513003e-01 1.11599076e+00
-1.07175839e+00 -4.38663363e-01 -4.38179612e-01 2.34916076e-01
-4.03893948e-01 4.38179016e-01 -5.82521319e-01 -6.37384295e-01
-2.93933779e-01 -1.20269108e+00 7.43127346e-01 5.20710230e-01
4.15445298e-01 -1.01358163e+00 -1.27182692e-01 1.45203397e-01
2.66881704e-01 8.61031651e-01 5.81942558e-01 -8.71486783e-01
-7.31996834e-01 -3.89352232e-01 -3.28721762e-01 1.00966990e+00
4.06765103e-01 1.49800867e-01 -1.16542351e+00 -4.22009319e-01
3.27694684e-01 -1.43945470e-01 9.38996732e-01 3.77002418e-01
1.46576250e+00 -4.76567775e-01 -2.74053186e-01 1.02608168e+00
1.38405836e+00 1.21288061e-01 9.61745322e-01 4.18079048e-01
1.16430032e+00 3.05444837e-01 4.76585835e-01 1.68823320e-02
-1.63744837e-01 4.26006764e-01 9.15566921e-01 -3.72494459e-01
-1.62780210e-01 -1.97694421e-01 3.82590503e-01 5.81940934e-02
4.29421306e-01 -5.65348029e-01 -5.55299461e-01 3.22172940e-01
-1.70069730e+00 -8.70148480e-01 1.24892771e-01 1.97345972e+00
8.79579663e-01 3.81706238e-01 -1.13042518e-01 6.74164444e-02
9.49274898e-01 2.71394372e-01 -9.21074688e-01 -4.30848509e-01
-3.13880742e-01 3.56864482e-01 1.05015838e+00 4.11727071e-01
-1.42831492e+00 9.86074746e-01 6.14419842e+00 1.22271729e+00
-1.10575879e+00 -9.55727100e-02 7.55058765e-01 6.69896081e-02
-4.22117412e-01 -3.49468201e-01 -8.39179814e-01 6.68694317e-01
3.42336059e-01 4.59700897e-02 1.95607275e-01 1.02144349e+00
-2.67343700e-01 3.64954263e-01 -8.33598971e-01 6.28298938e-01
1.78081039e-02 -1.12708712e+00 1.92560971e-01 1.05791010e-01
1.00602269e+00 -1.93167135e-01 5.40674448e-01 7.11882263e-02
6.44943297e-01 -1.03357208e+00 6.98645473e-01 1.78476289e-01
8.25317562e-01 -1.03930390e+00 6.98126972e-01 2.20800802e-01
-6.93864048e-01 -1.09038129e-01 -4.78186220e-01 3.45788807e-01
-5.13643436e-02 5.64969122e-01 -4.59201276e-01 6.04676664e-01
5.70990741e-01 4.61494952e-01 -4.98063803e-01 6.97766900e-01
-5.57522953e-01 5.08353829e-01 -2.27846041e-01 2.73610771e-01
2.07820982e-01 -7.64467195e-02 7.28426278e-01 1.01197779e+00
1.04717843e-01 -3.41803282e-01 -1.65864583e-02 1.16857934e+00
-2.72307128e-01 -2.59340972e-01 -7.15056896e-01 1.49117738e-01
5.82571328e-01 1.23787367e+00 -7.98644543e-01 -9.71956030e-02
-9.03222412e-02 1.21951091e+00 1.10887296e-01 5.27778685e-01
-1.19491243e+00 -7.45035887e-01 1.06001174e+00 -2.26962581e-01
4.57005382e-01 -1.54095953e-02 -6.02695107e-01 -9.85477507e-01
8.38441327e-02 -9.69777763e-01 1.75133988e-01 -3.13027382e-01
-1.71856081e+00 5.67999899e-01 -1.17987946e-01 -1.20283937e+00
3.34930643e-02 -9.13068652e-01 -8.55568349e-01 8.59248579e-01
-1.50946820e+00 -1.27283621e+00 -1.12916045e-01 8.55815351e-01
8.37835595e-02 -3.07631999e-01 5.81764162e-01 2.64635682e-01
-7.51703858e-01 1.11061752e+00 1.61885813e-01 7.49356627e-01
7.00851321e-01 -1.26750851e+00 6.99350893e-01 1.47434044e+00
-1.63202360e-02 5.34136117e-01 6.59057200e-01 -6.47470176e-01
-1.03043771e+00 -1.51562858e+00 -1.26539633e-01 -3.94593626e-01
6.00864768e-01 -6.05437994e-01 -9.37278688e-01 6.82026744e-01
1.74714282e-01 3.54370028e-01 5.75728416e-01 -6.46410406e-01
-6.77068651e-01 -2.28511974e-01 -1.63141274e+00 7.73556709e-01
9.91509855e-01 -4.20243323e-01 -3.56076926e-01 2.43308306e-01
9.53346074e-01 -7.87358880e-01 -5.20558596e-01 5.98197699e-01
3.27355832e-01 -7.73082674e-01 1.18513203e+00 -4.94411975e-01
3.38824093e-01 -4.66889769e-01 -2.11457685e-01 -1.25711274e+00
5.23517057e-02 -7.85461068e-01 -4.74436842e-02 1.33846295e+00
7.81417713e-02 -1.03145719e+00 7.97384381e-01 5.56979358e-01
-1.45945683e-01 -3.28791648e-01 -9.19362009e-01 -1.03108346e+00
5.58515668e-01 -1.20447852e-01 8.04055750e-01 9.39563274e-01
-7.29530394e-01 -3.55562776e-01 -3.58692348e-01 6.60815835e-01
9.22035635e-01 -1.30837277e-01 9.27711129e-01 -7.95768499e-01
-4.09198225e-01 -4.75156099e-01 -5.83581686e-01 -8.89057934e-01
3.34528506e-01 -6.62083447e-01 6.61291108e-02 -1.05600464e+00
-1.12132169e-01 -3.46368343e-01 -3.18782479e-01 5.03656566e-01
-4.01135594e-01 5.42629957e-01 3.97463024e-01 -1.70137957e-01
-7.65614212e-02 4.81374890e-01 1.54667175e+00 -1.92792282e-01
9.39010382e-02 1.54446512e-01 -8.42906177e-01 7.66740680e-01
1.07189071e+00 -7.52546549e-01 -5.18598616e-01 -4.85454351e-01
-2.71830112e-01 -6.62743211e-01 6.44694388e-01 -9.93597507e-01
2.22676080e-02 -2.85347313e-01 6.34012163e-01 -3.18610579e-01
2.38282084e-02 -9.98612702e-01 -3.72288190e-02 5.23789883e-01
-1.30977795e-01 -5.47018051e-01 4.57169086e-01 5.49070656e-01
-9.54758376e-02 -2.79592335e-01 1.24895966e+00 8.84190425e-02
-5.18828511e-01 5.96073627e-01 -1.48093952e-02 1.65453315e-01
1.33471966e+00 -2.54348785e-01 -6.95800185e-01 -1.98995486e-01
-4.12905276e-01 -1.32388230e-02 5.27294457e-01 4.31322783e-01
4.98514235e-01 -1.46324909e+00 -4.58074301e-01 5.59834063e-01
-2.13885427e-01 3.41846973e-01 2.99994081e-01 2.13397622e-01
-5.74925363e-01 -1.72643989e-01 -5.92632353e-01 -4.13475782e-01
-1.08232284e+00 7.58879602e-01 6.14136219e-01 -3.64988260e-02
-5.30845881e-01 9.73230243e-01 6.19011104e-01 -5.27506888e-01
2.64821291e-01 -2.39112556e-01 1.72869101e-01 -4.33386207e-01
5.93870878e-01 1.71613470e-01 -3.18723083e-01 -3.92625630e-01
-2.28712961e-01 4.46791142e-01 -1.72334015e-01 1.28687054e-01
9.86291409e-01 9.33990255e-02 1.35844667e-02 -1.86508685e-01
1.27064586e+00 2.85032719e-01 -1.72924721e+00 -1.12285465e-01
-5.52215278e-01 -6.36155546e-01 -2.62102872e-01 -6.25836313e-01
-1.52647972e+00 8.59826148e-01 6.01298809e-01 3.28913212e-01
1.14558542e+00 -2.06339389e-01 1.02548349e+00 -1.31639605e-03
3.24576232e-03 -8.89664352e-01 3.55961733e-02 4.19261485e-01
6.37528837e-01 -1.03366303e+00 -2.94218093e-01 -5.85932314e-01
-4.30323660e-01 9.21416223e-01 9.19259191e-01 -5.55562794e-01
4.81317192e-01 5.31889498e-01 3.77549589e-01 2.58441329e-01
7.43560493e-02 6.37581944e-02 1.26181424e-01 1.13220572e+00
-3.38847607e-01 -3.05550434e-02 1.48708746e-01 5.47759593e-01
-1.90765426e-01 -3.40071470e-01 4.46141183e-01 9.16865826e-01
-2.96500713e-01 -1.28777587e+00 -5.84806323e-01 -1.67453848e-02
-7.73872972e-01 -2.85168923e-03 -4.22378302e-01 7.49138653e-01
2.37064108e-01 6.72178984e-01 3.40050906e-02 -2.83036172e-01
2.22391173e-01 -3.71285915e-01 5.50172508e-01 -3.94871205e-01
-3.67763340e-01 -1.81616023e-01 -4.93060440e-01 -5.62466204e-01
-1.41329587e-01 -3.82942885e-01 -1.23394585e+00 -2.78636634e-01
-1.74223810e-01 -2.45380178e-01 5.06323576e-01 8.03211451e-01
1.64179415e-01 7.39495933e-01 9.56379414e-01 -5.33274174e-01
-7.52921999e-01 -5.35936296e-01 -6.55807674e-01 4.44062114e-01
5.27608812e-01 -3.92112494e-01 -5.49878895e-01 -3.89445201e-02] | [5.5108866691589355, 7.974135875701904] |
5efd8b57-7354-4ca5-85aa-cf3c1d7fe75f | who-sides-with-whom-towards-computational | null | null | https://aclanthology.org/P19-1273 | https://aclanthology.org/P19-1273.pdf | Who Sides with Whom? Towards Computational Construction of Discourse Networks for Political Debates | Understanding the structures of political debates (which actors make what claims) is essential for understanding democratic political decision making. The vision of computational construction of such discourse networks from newspaper reports brings together political science and natural language processing. This paper presents three contributions towards this goal: (a) a requirements analysis, linking the task to knowledge base population; (b) an annotated pilot corpus of migration claims based on German newspaper reports; (c) initial modeling results. | ["Sebastian Pad{\\'o}", 'Jonas Kuhn', 'Sebastian Haunss', 'Erenay Dayanik', 'Andre Blessing', 'Nico Blokker'] | 2019-07-01 | null | null | null | acl-2019-7 | ['knowledge-base-population'] | ['natural-language-processing'] | [ 4.41426903e-01 1.33068597e+00 -7.94664264e-01 -5.16364813e-01
-8.14056873e-01 -1.08104885e+00 1.48199379e+00 7.03791499e-01
-4.20890331e-01 1.08574998e+00 1.60868537e+00 -1.33028936e+00
-1.65187091e-01 -9.51225460e-01 -5.60446858e-01 1.40335798e-01
1.39276460e-01 6.54741645e-01 1.19374551e-01 -6.60007179e-01
4.60808545e-01 1.55974124e-02 -8.66757989e-01 7.60283768e-01
5.92843831e-01 4.59418774e-01 -4.03192818e-01 5.60881495e-01
-6.21773303e-01 1.87142146e+00 -4.74746138e-01 -8.64680111e-01
1.90639086e-02 -5.05272269e-01 -1.62467206e+00 1.40512697e-02
3.84876102e-01 3.09379339e-01 -9.27818790e-02 9.18731749e-01
1.83955640e-01 -1.86478913e-01 5.30116618e-01 -4.49783564e-01
-6.28812551e-01 1.60967469e+00 -2.09731847e-01 6.10430360e-01
5.52746713e-01 -1.58136636e-01 1.05661488e+00 -3.83122951e-01
1.40901887e+00 1.27212131e+00 6.66789711e-01 2.22334906e-01
-1.02895522e+00 5.75584210e-02 3.59856576e-01 -3.95076945e-02
-7.45173216e-01 -9.65071023e-01 6.69739664e-01 -1.10894048e+00
1.04841971e+00 6.49395704e-01 7.80918717e-01 1.15654457e+00
-9.81907845e-02 2.29349121e-01 1.58651209e+00 -9.35878336e-01
-1.14629373e-01 4.38717544e-01 7.87056267e-01 2.10259229e-01
4.18210417e-01 -1.01226024e-01 -3.79739463e-01 -7.38254726e-01
3.12282681e-01 -1.17464817e+00 1.16547078e-01 3.70269924e-01
-1.14776111e+00 1.26032758e+00 4.97367382e-02 9.53059256e-01
-6.30993783e-01 1.46506643e-02 3.90697509e-01 6.23685479e-01
7.58065462e-01 3.88683885e-01 -2.30054513e-01 -3.76271874e-01
-6.75836742e-01 4.98001754e-01 1.55704057e+00 6.07674360e-01
3.66156369e-01 -3.75104219e-01 -6.46006688e-02 6.51230872e-01
7.04746842e-01 6.46308482e-01 -2.40332112e-01 -8.63836706e-01
1.20543385e+00 6.90722823e-01 4.89186317e-01 -1.66324294e+00
-5.29452145e-01 9.30305421e-02 -2.10196331e-01 -5.38513660e-01
7.11208105e-01 -6.14554644e-01 -4.34935391e-01 1.52042150e+00
4.13224578e-01 -8.16409469e-01 5.41812360e-01 5.04838109e-01
1.16276383e+00 6.43901110e-01 3.20288241e-01 -6.85069203e-01
1.85831010e+00 -3.63223135e-01 -1.01507318e+00 -5.83785236e-01
3.52952272e-01 -1.11032450e+00 3.76576781e-01 -4.39060062e-01
-1.25491321e+00 1.64929211e-01 -4.91459012e-01 -3.07775915e-01
-1.43207178e-01 7.92335123e-02 8.36768687e-01 7.81744123e-01
-6.22937083e-01 -3.06320608e-01 -4.83088553e-01 -6.89194560e-01
4.03322011e-01 -3.88415933e-01 8.59431401e-02 4.22586739e-01
-1.58724618e+00 1.32932758e+00 5.06070197e-01 1.54974952e-01
-1.11430675e-01 -2.94006735e-01 -9.15766478e-01 -5.73784530e-01
5.78163564e-01 -6.85885191e-01 1.22577453e+00 -1.07025802e+00
-1.27741194e+00 1.57202590e+00 -1.41165137e-01 -6.74923420e-01
6.27886355e-01 -1.92754399e-02 -7.64567912e-01 -6.55704215e-02
5.22616923e-01 -2.46191204e-01 3.09621215e-01 -1.19349492e+00
-7.69188106e-01 -9.49848443e-02 5.68139195e-01 1.98186025e-01
8.26559961e-02 8.89204323e-01 1.50708154e-01 -6.55560434e-01
2.04030909e-02 -7.66084254e-01 -2.98934013e-01 -8.59382927e-01
-5.37468970e-01 -3.19907039e-01 1.49493009e-01 -1.11405551e+00
1.66239798e+00 -1.56453037e+00 -9.35691670e-02 4.18105751e-01
9.08347368e-02 -5.72082028e-02 3.82321060e-01 1.17644417e+00
3.30948504e-03 5.05491078e-01 3.49328332e-02 8.04344833e-01
1.04801729e-01 2.21248046e-01 -1.06667876e+00 5.35143197e-01
2.98198797e-02 8.75918865e-01 -6.24076784e-01 -7.60132670e-01
-1.40239954e-01 -1.22497588e-01 -2.84840375e-01 -7.34050512e-01
-4.49705780e-01 2.65762627e-01 -8.66273582e-01 4.16857928e-01
5.62428404e-03 -4.80721407e-02 1.15468895e+00 -2.95518845e-01
-9.33316469e-01 1.00541961e+00 -8.18688750e-01 1.11452258e+00
-9.77328513e-03 1.20831800e+00 5.58123350e-01 -9.23824728e-01
6.69556081e-01 5.43067336e-01 2.11709917e-01 -6.52743220e-01
4.98088747e-01 -6.89438656e-02 3.21150571e-02 -9.33324754e-01
6.35473013e-01 -4.77608830e-01 -9.75868940e-01 7.91544139e-01
-5.47379732e-01 -1.64507344e-01 4.63901103e-01 2.48794228e-01
8.08041751e-01 -1.25843644e-01 7.62142479e-01 -7.37759769e-01
5.43860257e-01 1.06770062e+00 7.44699359e-01 5.40311456e-01
-6.14072429e-04 -2.27425143e-01 7.99564958e-01 -1.26846611e+00
-1.05789685e+00 -3.01336557e-01 -3.66590470e-01 8.42420042e-01
-9.09053609e-02 -5.27245164e-01 -5.25344729e-01 -4.92935270e-01
-2.48132199e-01 1.08400047e+00 -8.16517949e-01 1.07393181e+00
-1.24918997e+00 -9.22752202e-01 7.67542720e-01 -6.87301904e-02
2.43793532e-01 -8.98665965e-01 -7.86989093e-01 3.55661988e-01
-1.03454709e+00 -1.34125578e+00 2.59064615e-01 -4.99469280e-01
-3.18487912e-01 -1.76768661e+00 1.73663467e-01 -7.15725362e-01
6.29842043e-01 -2.89504796e-01 1.23549592e+00 -1.04049869e-01
1.56462654e-01 3.32868159e-01 -2.16723785e-01 -8.79312456e-01
-1.11147583e+00 4.94583398e-02 -2.51534730e-01 -5.20899415e-01
3.56909961e-01 -2.88134068e-01 9.54067782e-02 1.12843560e-02
-7.49079049e-01 4.02141124e-01 1.59654319e-01 3.68324518e-01
-5.17171994e-02 -2.67393917e-01 2.41964862e-01 -1.38788736e+00
1.25943112e+00 -5.51298440e-01 -5.84696651e-01 5.53754628e-01
-2.69063264e-01 -3.54449451e-01 -1.34830624e-01 2.12078974e-01
-1.55788636e+00 -7.68098235e-01 -9.10920277e-02 1.20514619e+00
1.32974729e-01 1.25177038e+00 4.07627851e-01 6.33822501e-01
1.35883117e+00 -5.13176024e-01 7.80509114e-02 -8.37966353e-02
8.26418519e-01 5.68389654e-01 3.79226267e-01 -1.16609597e+00
6.30918086e-01 6.06294870e-01 -6.05059505e-01 -1.19109976e+00
-1.04816842e+00 -5.31523488e-02 -6.05802357e-01 -5.63512981e-01
8.48272741e-01 -1.05173993e+00 -3.96736473e-01 -3.06050718e-01
-1.28386116e+00 -3.83479327e-01 -5.72700918e-01 6.36278927e-01
-4.55284417e-01 1.12936713e-01 -7.22059071e-01 -1.06439066e+00
-1.79703936e-01 -3.98750603e-01 1.36059180e-01 7.32775554e-02
-9.35459375e-01 -1.37751818e+00 6.84431493e-01 9.47140217e-01
1.68748558e-01 1.06296289e+00 1.10158956e+00 -4.76692528e-01
1.26576722e-01 1.94775119e-01 -2.48935997e-01 -2.81496525e-01
-4.19321805e-02 3.64759654e-01 -5.74899197e-01 1.93686917e-01
1.46784738e-01 -4.56013173e-01 4.00958985e-01 4.78547513e-01
-4.53548908e-01 -1.21407282e+00 -4.10985023e-01 -3.69978249e-01
1.30736446e+00 7.19525944e-03 6.65182173e-01 8.94771516e-01
1.53370509e-02 1.30654418e+00 5.47434688e-01 4.08197865e-02
9.55380619e-01 4.84922290e-01 -3.69813472e-01 1.63616389e-01
-2.21072078e-01 -2.15685785e-01 4.05414283e-01 1.09157813e+00
-6.35541201e-01 1.93199813e-01 -1.65100980e+00 9.85700727e-01
-2.25380683e+00 -1.46587896e+00 -6.57820582e-01 1.15397859e+00
1.10062861e+00 2.41692439e-01 -1.62187666e-02 -2.88970470e-01
9.21335340e-01 6.78507030e-01 1.86446965e-01 -5.80858648e-01
-4.96540159e-01 -1.86798468e-01 4.40780640e-01 9.96843219e-01
-1.21278620e+00 1.00403738e+00 7.16827154e+00 4.39131528e-01
-3.52507442e-01 4.45396960e-01 4.48181242e-01 2.92954803e-01
-8.18423748e-01 6.13135457e-01 -5.76239169e-01 -6.74077822e-03
9.58214402e-01 -6.72751546e-01 1.35781556e-01 3.21229845e-01
4.22370374e-01 -2.06493109e-01 -3.56397033e-01 1.89184681e-01
6.12938963e-02 -2.31332755e+00 -8.77757072e-02 1.95588723e-01
9.36598003e-01 2.19980761e-01 -6.74722373e-01 4.54307348e-02
1.39881539e+00 -6.14445150e-01 1.42317557e+00 6.26911342e-01
6.09300435e-01 -1.75359771e-01 7.76277244e-01 5.16546190e-01
-6.69670165e-01 -1.81661561e-01 -1.62768692e-01 -6.38789833e-01
9.07005608e-01 7.27580965e-01 -7.73924649e-01 8.93927693e-01
2.62386918e-01 3.24514687e-01 -1.58378668e-03 -3.93341668e-02
-8.06035817e-01 1.03099012e+00 -2.87054002e-01 -4.26671132e-02
4.79640633e-01 -1.94963947e-01 8.89220893e-01 1.37272942e+00
-2.92226017e-01 7.31140137e-01 1.53407559e-01 5.42654216e-01
-9.80431810e-02 1.87353864e-01 -6.90398574e-01 -3.41816992e-01
4.99731183e-01 9.99876142e-01 -9.21649098e-01 -3.10504377e-01
-3.65521073e-01 -2.70010918e-01 1.87060475e-01 2.51746207e-01
-5.94760478e-01 2.40192980e-01 2.87457049e-01 4.21222270e-01
-1.57639325e-01 -1.82815269e-01 -1.89312637e-01 -9.81844962e-01
-9.45214182e-02 -1.12729096e+00 6.28499448e-01 -2.19000310e-01
-1.27138603e+00 7.20529139e-01 3.48612249e-01 -4.94071305e-01
-2.52830654e-01 2.06877273e-02 -4.16474998e-01 7.30322182e-01
-1.14626813e+00 -1.52775836e+00 2.05084875e-01 2.16427036e-02
2.17543080e-01 -3.16705443e-02 1.00223053e+00 1.31202355e-01
-1.88517392e-01 -4.94349062e-01 -4.25422907e-01 5.71028054e-01
3.97125125e-01 -7.61833012e-01 6.91983342e-01 8.09783101e-01
1.83428392e-01 7.07242250e-01 8.76396358e-01 -9.79954660e-01
-1.00780880e+00 -6.74923539e-01 1.70627642e+00 -1.05601215e+00
1.13628912e+00 3.93231995e-02 -1.77771673e-01 8.26658189e-01
8.92914414e-01 -7.38580167e-01 1.20844197e+00 5.45924902e-01
-1.20943964e-01 5.40470004e-01 -9.83000755e-01 5.70814431e-01
9.72977400e-01 -5.57038844e-01 -1.76078653e+00 4.85523880e-01
4.20680851e-01 -3.64509583e-01 -8.42218757e-01 7.58870617e-02
7.30365336e-01 -3.56780767e-01 6.53570414e-01 -9.64909196e-01
4.61425036e-01 -4.30413663e-01 -1.82294354e-01 -7.19274700e-01
-3.84408981e-01 -8.42860103e-01 3.05558354e-01 1.32093668e+00
9.89498436e-01 -6.08851671e-01 3.17199916e-01 9.57551003e-01
1.26686648e-01 -7.20099509e-02 -1.13853371e+00 1.55604571e-01
9.58145335e-02 -6.97367311e-01 2.14485615e-01 1.89763367e+00
5.64537764e-01 7.96784163e-01 -7.90528357e-02 2.20648739e-02
3.50965381e-01 4.35873955e-01 8.22445810e-01 -1.49014521e+00
1.34921223e-01 -3.86492550e-01 1.19107492e-01 -6.63961947e-01
-3.12615931e-02 -7.29154468e-01 -3.22768897e-01 -2.18761253e+00
7.07958266e-02 -4.70513761e-01 7.99056768e-01 3.79704773e-01
3.99514467e-01 -7.96237811e-02 1.09239802e-01 6.23739779e-01
-2.43131563e-01 -1.13150172e-01 8.12026322e-01 -3.14282358e-01
-3.02910179e-01 -2.41104752e-01 -1.55683005e+00 1.27571321e+00
7.00834334e-01 -6.65650904e-01 -6.05095699e-02 -7.14184403e-01
1.45689785e+00 2.09487960e-01 -3.40684922e-03 -2.03601912e-01
4.79616910e-01 -6.29804492e-01 -2.60141104e-01 -5.24075687e-01
-3.48807067e-01 -4.51688230e-01 5.17418504e-01 4.80534762e-01
-5.88373244e-01 1.43014967e-01 2.51765430e-01 3.90366912e-01
-3.90802354e-01 -2.00625747e-01 6.42660186e-02 -2.24520296e-01
-5.24704933e-01 -6.68067336e-01 -1.02668095e+00 4.92957026e-01
1.03367078e+00 -1.76414326e-01 -1.27472699e+00 -4.07067567e-01
-1.08893001e+00 2.00712919e-01 2.88050532e-01 2.48742089e-01
-5.63192181e-02 -1.26417327e+00 -1.42174089e+00 -5.06888926e-01
-1.53306350e-01 -4.04559761e-01 -1.69197381e-01 8.71652246e-01
-8.45345199e-01 7.85182893e-01 1.02826841e-01 3.38388741e-01
-1.33448553e+00 5.13954870e-02 -1.84686050e-01 -6.38793111e-01
-4.70537484e-01 4.68338460e-01 -4.25389498e-01 -4.59356457e-01
-5.45574427e-01 -1.86161265e-01 -8.57231617e-01 7.92661846e-01
5.63140631e-01 1.80750579e-01 -4.60729003e-01 -1.24230969e+00
-4.51586366e-01 2.69133478e-01 2.11247832e-01 -7.48714387e-01
1.64134610e+00 -6.18543506e-01 -7.35582411e-01 6.24155343e-01
1.44706607e-01 9.75837290e-01 -2.94076711e-01 -3.93057972e-01
6.15474224e-01 -2.79469311e-01 -2.41468266e-01 -9.57927525e-01
-2.45871365e-01 -1.72642857e-01 -3.31236959e-01 1.13144529e+00
1.96783066e-01 3.66088718e-01 1.16043150e-01 4.62541372e-01
3.89086515e-01 -1.68046260e+00 -6.13012731e-01 7.36995459e-01
1.41088057e+00 -7.80370355e-01 5.89929283e-01 -7.09280312e-01
-8.94902349e-01 1.19600987e+00 -1.93551287e-01 2.06590697e-01
8.05889845e-01 2.86709517e-01 6.24874711e-01 -1.14678407e+00
-7.60955572e-01 -1.24929003e-01 7.19749406e-02 2.84959853e-01
6.04219258e-01 2.53978252e-01 -1.41734171e+00 6.43430650e-01
-5.22122741e-01 -5.19950129e-02 5.98084033e-01 1.11892843e+00
-4.92615789e-01 -1.15244222e+00 -6.52965128e-01 3.18797827e-01
-8.32932413e-01 -2.13023975e-01 -1.19677007e+00 9.49400663e-01
1.59208521e-01 1.36200583e+00 1.39126524e-01 1.34305507e-01
5.50131679e-01 -5.38443923e-02 2.22233430e-01 -6.67899370e-01
-9.85726953e-01 -1.17940590e-01 1.86704957e+00 -1.46425515e-01
-1.50862241e+00 -7.60640383e-01 -7.48865485e-01 -6.78419828e-01
-3.39061111e-01 8.82951796e-01 5.80752492e-01 8.66708815e-01
2.04912290e-01 3.38636369e-01 1.29039019e-01 8.53481889e-02
-1.18887378e-02 -1.09528399e+00 1.27176031e-01 1.99037775e-01
-1.72729895e-01 -1.00119166e-01 -3.49870250e-02 5.91975987e-01] | [9.082672119140625, 9.817211151123047] |
2ec38969-501d-48d2-b248-c8f6cd36d916 | trigger-free-event-detection-via-derangement | 2208.09659 | null | https://arxiv.org/abs/2208.09659v1 | https://arxiv.org/pdf/2208.09659v1.pdf | Trigger-free Event Detection via Derangement Reading Comprehension | Event detection (ED), aiming to detect events from texts and categorize them, is vital to understanding actual happenings in real life. However, mainstream event detection models require high-quality expert human annotations of triggers, which are often costly and thus deter the application of ED to new domains. Therefore, in this paper, we focus on low-resource ED without triggers and aim to tackle the following formidable challenges: multi-label classification, insufficient clues, and imbalanced events distribution. We propose a novel trigger-free ED method via Derangement mechanism on a machine Reading Comprehension (DRC) framework. More specifically, we treat the input text as Context and concatenate it with all event type tokens that are deemed as Answers with an omitted default question. So we can leverage the self-attention in pre-trained language models to absorb semantic relations between input text and the event types. Moreover, we design a simple yet effective event derangement module (EDM) to prevent major events from being excessively learned so as to yield a more balanced training process. The experiment results show that our proposed trigger-free ED model is remarkably competitive to mainstream trigger-based models, showing its strong performance on low-source event detection. | ['Haiqin Yang', 'Jiachen Zhao'] | 2022-08-20 | null | null | null | null | ['machine-reading-comprehension'] | ['natural-language-processing'] | [ 0.36089382 0.02801965 -0.02338756 -0.48894516 -0.9566152 -0.5059604
0.67351115 0.73692644 -0.68880576 0.42927012 0.54657817 -0.34520498
0.08530479 -0.81397253 -0.65738136 -0.31246743 0.3979181 0.07919145
0.418652 -0.15475759 0.11333106 -0.05640946 -1.6768565 0.5774322
1.208531 0.90920794 0.35537958 0.36347565 -0.3491349 1.3242347
-0.8322721 -0.44290835 -0.35737097 -0.7003444 -0.91379434 -0.1574562
0.00949984 -0.26472577 -0.26488256 0.83107865 0.76150525 0.2895723
0.63309336 -1.0394875 -0.43363613 0.88131624 -0.41101444 0.63094825
0.605936 -0.02657166 1.2041726 -0.9462488 0.49279165 1.1405689
0.42754957 0.5081706 -0.6998077 -0.56669086 0.59883916 0.6494755
-0.98012894 -0.39602655 0.78651565 -0.18863174 1.005693 0.409625
0.24466865 1.3579328 -0.09324937 0.9904323 0.6952461 -0.62359726
0.3536396 -0.1454493 0.31621584 0.38032007 -0.0478929 -0.45650753
-0.7076321 0.11805586 0.28187314 0.11373742 -0.41113925 0.4945405
-1.3290415 0.6725739 0.27920678 0.280318 -0.7149534 -0.15358226
0.6759809 0.02713449 0.55385 0.20766526 -0.63992625 -0.12812352
-0.5913888 0.24913119 0.49716753 0.8215789 0.34577414 -0.27457196
-0.76462024 0.89886487 0.15756595 0.20868118 0.7069659 -0.49894252
0.857204 0.9672285 0.10871843 -1.1031101 -0.71890306 -0.52150434
-0.6554941 -0.5557198 0.23933066 -0.23917733 -0.5737913 1.9845501
0.56818587 0.3219359 0.05429868 0.80665106 1.0567584 0.82468444
0.530288 -0.3829275 1.7061269 -0.82739747 -0.9933308 -0.5126403
0.7581442 -0.6957565 1.4415523 0.3052492 -0.8834838 -0.4046031
-0.89168 -0.36245745 -0.18907984 0.21492925 0.40502042 0.08513816
-0.14850396 0.24212699 -0.7058621 -0.26219764 0.32644093 -0.27518174
-0.01706184 -0.07189736 -1.6026864 0.8700385 0.7188302 -0.03568488
-0.7126161 -0.6046659 -0.8713399 0.17490618 0.78010976 -0.3297559
1.6511052 -0.7656893 -1.0049505 0.7411876 -0.4669212 -0.2532006
0.21567409 -0.48712698 -0.67011636 0.2902055 0.3461208 0.17462818
0.6232868 -0.8139869 -0.8636892 -0.07053802 0.05498462 0.40195224
-0.50386894 0.2878694 -0.20497032 -0.8892771 0.11082154 -0.23048863
0.04806142 -0.48558882 -0.46704543 -0.8285502 0.40272436 -0.7445178
1.754446 -2.1733103 -0.1586734 -0.32489225 0.15315858 0.10646016
-0.08564958 0.39155224 -0.12918694 -0.13981272 -0.08615813 -0.25186372
0.05845974 0.20700897 -0.6417103 0.03263739 0.63822407 0.76201224
-1.233412 -0.7069312 0.1591005 -0.00925404 -0.45965675 0.52875984
-0.59071136 0.29024076 -0.5295204 0.490088 0.41438323 -0.36921814
-0.09376625 -0.14198695 -0.14135632 0.99508053 -1.2194904 1.4288334
-0.53441614 0.05799258 -0.3758017 -1.0595212 0.6950071 0.5852169
0.1345789 -0.93466485 0.21237765 0.21179713 -0.18215524 -0.8079051
0.2911079 -0.11842485 -0.2902128 0.47734264 0.10354266 0.49828684
0.34240076 0.47429636 1.2798121 -0.09064431 0.45326853 0.0735862
0.5836137 -0.03156297 0.9306111 0.67624617 -0.21323992 0.570227
0.7315496 -0.1424161 -0.5258017 -0.73281884 0.04037785 1.5095112
0.16479145 -0.6913052 -0.75901467 -0.9526244 -0.45290437 1.1311418
-0.4417796 -0.3316187 -0.76937634 -1.0130796 0.5826526 0.7058373
0.5805625 -1.246398 -0.67943823 0.6892013 -0.9237572 -1.0146904
-0.5647176 0.39666253 -0.29604658 -1.1084918 -0.4133333 -0.9172634
0.3862609 0.25356734 1.2836668 -0.03236846 -0.01386864 0.08566017
-0.7651332 -0.6721768 -0.24342346 0.11634278 -0.24146314 0.15927568
0.8258472 -0.3738922 -0.69875574 0.29091787 -1.1122748 0.17465028
0.3598813 0.72996575 0.49026766 0.26805663 1.1970118 -0.8420494
0.55807763 -0.98989296 -0.19665968 0.36292094 -0.35775018 -0.16461714
0.98851335 -0.680169 -1.497153 -0.24885608 -0.39862737 0.10936093
-0.27140749 0.546577 -0.59261703 0.9058589 0.7036578 0.3105197
-0.6578743 -0.4403624 0.27652106 0.7972407 0.726007 -0.61492056
0.33534738 0.15080814 -0.5584802 -0.31364846 -1.6298071 -0.6250424
-0.13461146 -0.17175744 0.84651697 -1.0525562 -0.44710648 0.5430312
-1.353009 -0.13633138 -0.16131376 0.43329716 -0.18989639 0.12697491
-0.63478965 -0.6918409 -0.2838185 -0.6380095 1.0959694 0.4554461
-0.37774745 -0.8149381 0.03729369 0.29049328 0.20511347 0.12190129
1.1029481 -1.0536329 -0.35858753 -0.10077636 -0.21143514 0.11186929
0.04505992 -0.3674387 -1.0328745 0.28966278 0.25070086 -0.4514664
0.96565515 -0.08929804 1.2507547 -0.373791 -0.2172777 0.10215557
1.0429063 0.13685055 0.42607468 0.27491444 0.62549317 0.7213509
0.782654 0.6424633 0.8321358 0.3497965 0.27708966 -0.04085328
-0.09912866 -0.57560486 0.40355918 1.0086094 0.3626816 -0.68991977
-0.7373214 0.7543253 -1.8836119 -1.0003861 -0.32716584 1.892866
1.33387 0.26772165 -0.13055024 0.41933268 0.85084605 0.22267348
-0.5544683 -0.050347 -0.2679641 0.15090679 -0.16186939 0.06844525
-1.2350551 0.70251215 4.9086785 1.0584441 -0.8757722 0.3416202
0.5456107 -0.15055297 -0.46174937 -0.222328 -0.9615798 0.7525424
0.95874363 -0.11631479 0.00754945 0.5617378 0.3307779 -0.18016338
-1.1835973 0.77743286 0.04997604 -0.9833926 -0.07675976 -0.59722435
0.39439142 -0.2580282 -0.40445903 0.8162808 0.06287316 -0.5752964
0.9195825 0.38638392 0.35684386 -0.6773262 0.6825321 0.6559114
-1.1051636 -0.06446242 -0.25533944 -0.10710954 0.35477918 1.0325353
-0.54173774 0.6773219 0.4814553 0.72569436 -0.5478962 0.82290816
-0.5567364 1.2322799 -0.36164495 -0.10820699 0.06693367 0.43344843
0.46215937 1.3228403 0.18397023 0.46419132 0.21942689 0.6881516
-0.34941417 0.33310965 -0.15408793 0.14534801 0.6970796 1.2173973
-0.80838233 -0.51498353 -0.61559474 1.0147686 0.54219645 0.15866344
-1.121338 -0.5548327 0.18461363 -0.15098885 0.15068135 0.40114737
-0.24996717 -1.3873833 0.37674478 -0.82852644 0.8487114 -0.6700334
-1.5985377 0.43704534 -0.12879862 -1.1452975 -0.10241157 -0.21668075
-1.0113363 0.67640483 -1.6299647 -0.70021963 -0.33078447 0.56898975
0.786911 0.3237913 0.6629262 0.8124033 -1.0053242 0.65762645
-0.36352533 0.13233118 0.96763545 -1.3559504 0.14989267 1.1285156
-0.03813483 0.46190384 0.56683713 -0.74698144 -0.98787546 -1.40084
1.3614299 -0.39977852 0.49733287 -0.3208282 -1.3120486 0.6469159
0.20315847 -0.20212847 0.6830991 0.11139926 -0.462669 0.00948931
-0.8752567 0.5561402 1.0570328 -0.48252717 -1.1565388 0.47762793
1.0627009 -0.31497738 -0.38624096 0.37285975 -0.18001424 -0.5993361
0.7896483 -0.63867176 0.6126438 -0.3762919 -0.03600199 -1.210585
-0.18771243 -0.53190875 -0.3780465 1.8218162 0.3083382 -0.46341774
0.10114512 0.44262698 -0.37248874 -0.6003003 -0.6771661 -0.5257266
-0.14970197 -0.66702414 0.66790456 1.0530413 0.1940269 0.7514559
-0.15584348 0.37819272 0.2428415 0.03084015 0.15652846 -1.1074948
-0.12925535 -0.31159726 0.26677346 -1.1112218 0.05099866 -0.89455396
0.37037295 -1.4809974 0.28174648 -0.3819805 -0.55551004 0.65646785
-0.9520341 -0.15860546 -0.21346544 0.00935727 -0.98919207 0.7170308
0.9558489 0.20639345 -0.09742496 -0.08191592 -0.8680725 0.90554637
0.8306027 -0.62834734 -0.39039406 -0.46323198 0.54442394 -0.03088538
0.40920043 -0.729105 0.30464774 -0.27571383 0.23197308 -0.8127633
-0.22039704 -0.40879723 -0.39487308 0.06995478 -0.66968006 0.01266355
0.03579362 0.564583 -0.295601 -0.48433498 0.44352916 -0.05354492
-0.74513644 -0.1294816 -0.46257085 0.68030703 0.9331073 0.4132922
-0.52815515 -0.08556077 -0.45122313 0.49856856 -0.07423154 0.5205394
0.3979299 -1.2685982 -0.85022205 0.05727218 0.43105182 0.40428782
0.5011252 0.7360604 -0.08373704 0.20584768 0.39535835 -0.27500892
-0.87851495 0.5551326 0.03540747 -0.32240745 -0.73119146 0.90696555
0.20582785 -0.37498906 0.528064 -0.5063971 -0.39041403 0.41006055
0.89435226 0.32673252 0.2848014 -0.1122182 -0.4701295 -0.03844374
-0.23068784 0.11646157 1.118373 -0.27553457 -0.08273093 0.5427125
0.8356914 0.01918702 -1.1415801 -0.46827388 0.26855272 0.03968593
0.07648603 -1.0908453 -0.595273 0.83885217 0.11727138 0.30362177
1.3861877 0.23955221 1.2233264 0.22339839 -0.06355757 -1.1220112
0.24596275 0.64784914 0.62392855 -1.1645948 -0.49244922 -0.39390376
-0.66900414 0.8103121 0.84261566 0.22776563 0.21623513 0.04418122
0.03710046 -0.14407608 -0.9973867 -0.35797247 0.253661 0.1192136
0.40390068 -0.05067569 -0.51229614 1.1554315 0.10026184 -0.14011447
0.38716492 0.9185818 -0.6372205 -0.7659609 -0.25822553 0.50038767
-0.7563294 -0.27711993 -0.09059853 0.34639528 0.49024543 1.1987299
0.05350976 0.01945482 0.5754476 0.45192257 0.09264909 -0.692876
-0.7051553 -0.01354146 0.1609268 -0.38524616 -0.32324237 -0.70617115
-1.587658 0.19000578 -0.45590487 0.05356707 0.24867614 1.2154238
0.46380523 0.986305 0.6648667 -0.07987768 -0.52023995 -1.0471605
-0.04549875 0.64129955 0.23533337 -0.70576906 -0.38237584 0.15583587] | [9.113868713378906, 9.174711227416992] |
c9c50180-7ff7-444b-a979-b21df764266a | fast-online-object-tracking-and-segmentation | 1812.05050 | null | https://arxiv.org/abs/1812.05050v2 | https://arxiv.org/pdf/1812.05050v2.pdf | Fast Online Object Tracking and Segmentation: A Unifying Approach | In this paper we illustrate how to perform both visual object tracking and semi-supervised video object segmentation, in real-time, with a single simple approach. Our method, dubbed SiamMask, improves the offline training procedure of popular fully-convolutional Siamese approaches for object tracking by augmenting their loss with a binary segmentation task. Once trained, SiamMask solely relies on a single bounding box initialisation and operates online, producing class-agnostic object segmentation masks and rotated bounding boxes at 55 frames per second. Despite its simplicity, versatility and fast speed, our strategy allows us to establish a new state of the art among real-time trackers on VOT-2018, while at the same time demonstrating competitive performance and the best speed for the semi-supervised video object segmentation task on DAVIS-2016 and DAVIS-2017. The project website is http://www.robots.ox.ac.uk/~qwang/SiamMask. | ['Qiang Wang', 'Li Zhang', 'Philip H. S. Torr', 'Weiming Hu', 'Luca Bertinetto'] | 2018-12-12 | fast-online-object-tracking-and-segmentation-1 | http://openaccess.thecvf.com/content_CVPR_2019/html/Wang_Fast_Online_Object_Tracking_and_Segmentation_A_Unifying_Approach_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Wang_Fast_Online_Object_Tracking_and_Segmentation_A_Unifying_Approach_CVPR_2019_paper.pdf | cvpr-2019-6 | ['real-time-visual-tracking'] | ['computer-vision'] | [ 1.73783563e-02 -9.48030129e-02 -3.46331149e-01 -1.74483702e-01
-7.65839398e-01 -1.03828621e+00 6.70133591e-01 -2.66571462e-01
-8.65633011e-01 5.38475871e-01 -4.88928258e-01 -1.96024224e-01
2.65386701e-01 -3.03904247e-02 -1.09631312e+00 -5.27624190e-01
-2.97336876e-01 7.17361212e-01 6.95713937e-01 2.07341388e-01
-1.69651099e-02 6.32519007e-01 -1.49612045e+00 -1.59022227e-01
6.48789644e-01 1.03085041e+00 2.26838365e-01 1.08200300e+00
1.12328693e-01 5.08455932e-01 -4.95900631e-01 -5.47467291e-01
6.61078751e-01 -9.75985602e-02 -9.04775739e-01 1.96499318e-01
1.20402598e+00 -3.05606395e-01 -5.82656026e-01 1.17705250e+00
1.88465223e-01 4.64576520e-02 4.49987531e-01 -1.61847079e+00
-4.78572160e-01 4.43022460e-01 -6.66625321e-01 3.28149557e-01
3.54506485e-02 4.67804998e-01 8.08171451e-01 -6.23433650e-01
9.00350809e-01 1.10966253e+00 9.52131212e-01 7.89751172e-01
-1.35967803e+00 -8.83044720e-01 2.73773283e-01 6.87368140e-02
-1.26787627e+00 -4.65355784e-01 3.78339410e-01 -6.89523578e-01
8.56511891e-01 -9.70728789e-03 7.13510394e-01 1.12519121e+00
-2.63988972e-01 1.14570975e+00 8.19432318e-01 -6.51457906e-02
9.11455899e-02 6.78142160e-02 5.73355928e-02 8.65371943e-01
4.74368274e-01 3.67863864e-01 -2.13696718e-01 1.69286624e-01
7.57314026e-01 -2.05667898e-01 -8.65946859e-02 -1.05688787e+00
-1.51107597e+00 5.55461228e-01 6.13135099e-01 1.80574867e-03
9.31042805e-02 7.87015915e-01 7.70549059e-01 -3.57149355e-02
1.90108240e-01 2.58146197e-01 -5.90839565e-01 -3.02566797e-01
-1.35457134e+00 4.21493322e-01 5.80171943e-01 1.33218944e+00
5.82227707e-01 1.58517063e-01 -2.59775102e-01 3.55248243e-01
4.06818122e-01 7.76871741e-01 1.67353123e-01 -1.23743379e+00
3.79416525e-01 1.76142648e-01 4.73604411e-01 -4.66332078e-01
-3.74372989e-01 -3.98364723e-01 -1.17368005e-01 4.83675510e-01
9.26422596e-01 -1.54970974e-01 -1.34216225e+00 1.74883783e+00
5.26062906e-01 3.39902818e-01 -2.12617099e-01 1.06851411e+00
5.14127254e-01 2.79113173e-01 3.86573553e-01 9.33745280e-02
1.35326862e+00 -1.40623701e+00 -5.58407307e-01 -3.04990888e-01
6.32628560e-01 -5.93324065e-01 8.07965100e-01 2.86175430e-01
-9.09647167e-01 -6.51607335e-01 -1.07323217e+00 -3.43546346e-02
-4.58126873e-01 5.64953208e-01 7.15531409e-01 7.99571991e-01
-1.09042823e+00 6.90877676e-01 -1.32537484e+00 -4.74747956e-01
9.75811124e-01 4.63014901e-01 -3.56660187e-01 2.88332939e-01
-7.17019439e-01 8.40214014e-01 5.85459948e-01 7.56818801e-03
-1.28645003e+00 -8.15997720e-01 -8.36311519e-01 -3.21153194e-01
6.86680198e-01 -2.20079347e-01 1.63766670e+00 -8.82140458e-01
-1.34707785e+00 1.11987138e+00 1.48166597e-01 -1.00418448e+00
1.16815901e+00 -5.69074452e-01 -1.68582618e-01 3.08811843e-01
1.06024034e-01 1.24215841e+00 1.10305440e+00 -1.14016140e+00
-8.13017726e-01 -1.02914780e-01 -1.08010240e-01 -8.95661190e-02
2.38494612e-02 2.15183064e-01 -8.89944196e-01 -5.92420042e-01
-4.53325897e-01 -1.11914420e+00 -2.10061952e-01 6.95506394e-01
-3.57365340e-01 -2.68329293e-01 1.19360614e+00 -6.16852522e-01
7.46218622e-01 -1.98851430e+00 1.11741401e-01 -1.11595809e-01
3.32795560e-01 6.76166534e-01 -1.68332696e-01 -1.02024429e-01
5.29056713e-02 -1.88020542e-01 -5.30581951e-01 -5.90908110e-01
2.11461738e-01 1.02068007e-01 -2.79690951e-01 9.34017181e-01
3.06508273e-01 1.23042881e+00 -1.12176716e+00 -6.80387914e-01
5.09727895e-01 4.37926710e-01 -4.26360309e-01 -5.41341230e-02
-5.09811342e-01 3.60365570e-01 -4.27776203e-02 8.74769628e-01
7.22929418e-01 -2.20352173e-01 -9.55914259e-02 -1.57452390e-01
-2.91652888e-01 -1.37183100e-01 -1.23853099e+00 1.87810338e+00
1.26868650e-01 1.06447554e+00 3.54479820e-01 -9.72866118e-01
4.69457150e-01 -4.59896177e-02 5.77252388e-01 -4.56025511e-01
2.53499776e-01 3.06033790e-01 -1.56849816e-01 -1.98429614e-01
5.11355519e-01 3.77724230e-01 1.11887537e-01 1.53687790e-01
5.16475022e-01 5.53305596e-02 6.87231064e-01 3.25806975e-01
8.39658737e-01 8.80928874e-01 1.28205009e-02 -4.08095479e-01
3.15694511e-01 4.14282769e-01 3.54191482e-01 7.83746183e-01
-8.32977772e-01 4.85289037e-01 3.85653913e-01 -2.07577392e-01
-1.25591505e+00 -1.13868475e+00 -2.43007019e-01 1.16363037e+00
4.04654711e-01 -2.94675916e-01 -8.64827812e-01 -1.16532564e+00
3.87774229e-01 3.04713964e-01 -7.35963225e-01 1.07461274e-01
-6.95560336e-01 -3.07945102e-01 9.06902909e-01 7.36167371e-01
4.16764528e-01 -1.00691605e+00 -8.06074381e-01 1.47228792e-01
2.78522432e-01 -1.47472537e+00 -7.04219401e-01 4.11056668e-01
-6.69132948e-01 -1.35650873e+00 -1.03194141e+00 -6.03596091e-01
5.52094102e-01 2.83996344e-01 9.07497942e-01 -2.13768706e-01
-7.52241075e-01 4.72562045e-01 -1.61139920e-01 -4.08341736e-01
-3.17258328e-01 8.22670981e-02 1.18729599e-01 -2.56115854e-01
1.34541348e-01 -2.90399753e-02 -5.93595505e-01 3.39597583e-01
-6.03353262e-01 -2.66636491e-01 4.56987739e-01 4.57542896e-01
5.23967743e-01 -5.42380631e-01 1.87066182e-01 -6.48515999e-01
-2.64294654e-01 -1.44810140e-01 -1.37287593e+00 9.39311460e-02
-4.88020390e-01 -1.74269110e-01 4.22551602e-01 -5.72890818e-01
-7.08138645e-01 6.16578639e-01 1.14958085e-01 -8.54252338e-01
-2.67689943e-01 -3.67872864e-01 1.26730204e-01 -6.67201042e-01
5.99030375e-01 6.29826933e-02 2.08737656e-01 -5.34826100e-01
7.18358338e-01 2.66341627e-01 9.57123756e-01 -4.30707186e-01
1.30264008e+00 7.77224243e-01 -2.15947762e-01 -5.49488187e-01
-8.56847703e-01 -6.97862387e-01 -8.20334256e-01 -5.32647967e-01
1.22621047e+00 -1.01256490e+00 -7.25140393e-01 6.35761619e-01
-8.12931776e-01 -7.28640199e-01 -3.70158315e-01 4.04906809e-01
-7.09859192e-01 3.82241428e-01 -3.83677036e-01 -7.03323305e-01
-1.08892336e-01 -1.01149619e+00 1.23785639e+00 4.12024677e-01
-5.56483865e-02 -7.45522618e-01 -6.06140755e-02 3.82530689e-01
2.45159328e-01 3.10390830e-01 -2.23923966e-01 -7.10059404e-01
-9.27238107e-01 -1.80694759e-01 -5.82815886e-01 4.47452724e-01
-2.09631115e-01 3.03201377e-01 -9.55607474e-01 -6.03294075e-01
-7.66057432e-01 -4.14815962e-01 9.65274453e-01 3.94778043e-01
1.00111067e+00 6.22386113e-02 -5.61286032e-01 9.11083281e-01
1.35996985e+00 7.32059032e-02 8.66568312e-02 4.75725681e-01
9.33395326e-01 7.61819705e-02 8.29883456e-01 7.85471946e-02
1.80099100e-01 7.71473825e-01 5.97861350e-01 -1.14722699e-01
-4.65079606e-01 -1.58621952e-01 4.29216057e-01 6.06604666e-02
4.47944067e-02 -2.39289086e-02 -1.02338505e+00 9.06883836e-01
-1.89357769e+00 -8.41868103e-01 -2.30863973e-01 2.12603879e+00
6.32510483e-01 3.20788622e-01 6.39058709e-01 -2.03367934e-01
8.31438959e-01 2.25127071e-01 -7.26840854e-01 6.07170016e-02
-7.42632314e-04 1.39659151e-01 1.21100926e+00 2.67049342e-01
-1.76476216e+00 1.32822394e+00 6.01908827e+00 7.30842054e-01
-9.44437206e-01 2.21071348e-01 7.62863904e-02 -3.35207433e-01
6.17172897e-01 -7.31283352e-02 -1.12786257e+00 5.08299708e-01
8.05255055e-01 1.15354536e-02 4.24123406e-01 1.09597707e+00
-1.64744258e-01 -1.58872843e-01 -1.05725861e+00 8.95458221e-01
-1.04417115e-01 -1.43895388e+00 -3.54443133e-01 -7.16798902e-02
7.66771972e-01 6.09824598e-01 3.08287446e-04 3.86813760e-01
3.46737951e-01 -7.43364394e-01 1.22639692e+00 1.42426074e-01
6.98258877e-01 -5.08464813e-01 3.91613364e-01 -1.15095712e-02
-1.46800649e+00 2.03750432e-02 -8.46792534e-02 2.28786021e-01
2.54782468e-01 3.15164998e-02 -5.83511353e-01 3.03491086e-01
9.47890222e-01 8.57967138e-01 -7.63867199e-01 1.58606005e+00
-2.32062086e-01 6.41460240e-01 -5.50666153e-01 1.43523216e-01
6.35672092e-01 -1.88100524e-02 7.97501147e-01 1.55729282e+00
-9.64461043e-02 -5.02061129e-01 4.00050968e-01 8.12253535e-01
-1.30867869e-01 -4.32857484e-01 -2.87427396e-01 -2.90107310e-01
4.33212757e-01 1.33741403e+00 -1.18864870e+00 -4.57821578e-01
-2.03640357e-01 8.75209689e-01 2.56876409e-01 3.37506235e-01
-1.44468248e+00 -4.38383967e-01 8.93387914e-01 -1.88832626e-01
9.37304318e-01 -5.34250259e-01 -1.43017739e-01 -9.85022068e-01
-1.22640058e-01 -6.23536229e-01 3.45233381e-01 -5.73080420e-01
-7.68345296e-01 2.60047078e-01 1.42559662e-01 -1.24074721e+00
8.09304193e-02 -9.36058342e-01 -4.39466089e-01 3.26202333e-01
-1.44472134e+00 -1.24389005e+00 -3.21359098e-01 3.64686579e-01
4.84976828e-01 5.56618012e-02 2.04669565e-01 5.19638777e-01
-6.95084631e-01 6.43364668e-01 2.27390781e-01 4.69821453e-01
7.28593588e-01 -1.57311010e+00 8.29402983e-01 1.08841860e+00
3.31455559e-01 2.10240081e-01 8.82952571e-01 -6.03085160e-01
-1.33378422e+00 -1.30985355e+00 2.81707972e-01 -6.27818942e-01
1.05669129e+00 -6.22680843e-01 -7.22472966e-01 1.01277328e+00
5.99563606e-02 3.74936312e-01 1.02841578e-01 -2.69121528e-01
-3.51303726e-01 8.92331526e-02 -1.05191052e+00 4.98555273e-01
1.32120192e+00 -3.30334306e-01 -5.39260149e-01 6.08844995e-01
9.02101338e-01 -8.35202277e-01 -5.31339943e-01 3.49868983e-01
5.13156891e-01 -4.97561961e-01 1.22401857e+00 -6.12649381e-01
-1.57161921e-01 -7.74296999e-01 1.52507514e-01 -6.73476934e-01
-6.39921725e-02 -9.99979794e-01 -5.63760519e-01 1.11881018e+00
2.00035989e-01 -4.37108040e-01 1.04459810e+00 2.29309440e-01
-1.59437627e-01 -4.85095710e-01 -1.00855041e+00 -1.24389946e+00
-2.38315448e-01 -5.21413028e-01 1.30733907e-01 5.73009849e-01
-5.85002363e-01 -3.37544054e-01 -3.37670863e-01 2.40709782e-01
1.21141469e+00 5.27446456e-02 8.57263982e-01 -1.04615498e+00
-4.08586860e-02 -6.72750831e-01 -6.36965990e-01 -1.06691396e+00
1.43590003e-01 -8.27218533e-01 3.05681676e-01 -1.16255248e+00
-1.72393024e-02 -2.93994367e-01 -2.46092990e-01 5.90283990e-01
-1.14687972e-01 8.07274222e-01 4.81136888e-01 2.17964530e-01
-1.14963567e+00 3.88761342e-01 1.07672215e+00 -1.26753524e-01
-4.78338003e-02 4.27079126e-02 -1.77433848e-01 6.45645916e-01
8.17644000e-01 -6.91926837e-01 -2.71932743e-02 -3.50438356e-01
-4.14969653e-01 -5.26090622e-01 8.90516400e-01 -1.27883267e+00
3.38017464e-01 1.78687617e-01 3.82259279e-01 -8.07766497e-01
3.48855823e-01 -7.51676679e-01 -1.71006285e-02 7.74305165e-01
-1.37923643e-01 -1.05276980e-01 4.99333501e-01 7.01893210e-01
1.08351760e-01 -1.48023337e-01 1.17650628e+00 8.52596462e-02
-1.22050011e+00 5.15069604e-01 -1.97299972e-01 3.36236417e-01
1.38122904e+00 -3.83875847e-01 -3.72681916e-01 3.29530209e-01
-7.79787064e-01 6.55609787e-01 5.98328412e-01 6.79365575e-01
2.82253958e-02 -1.21320951e+00 -3.42315346e-01 -2.66014915e-02
8.63309056e-02 -1.96114033e-02 7.76727274e-02 9.42875564e-01
-8.57165813e-01 5.18786848e-01 -3.29251558e-01 -9.49117780e-01
-1.32524216e+00 6.37380600e-01 4.27372307e-01 3.13193016e-02
-7.80040503e-01 1.24187386e+00 -1.05548523e-01 -2.88365394e-01
7.32824862e-01 -3.07794780e-01 1.31835923e-01 9.06632021e-02
3.43689501e-01 4.60933149e-01 -1.69168457e-01 -6.60130918e-01
-6.59729302e-01 6.22634470e-01 -6.64196014e-02 -3.03792432e-02
1.04769635e+00 4.11270820e-02 3.18921477e-01 2.17804626e-01
1.13023317e+00 -2.77033985e-01 -2.10028267e+00 -1.04068622e-01
3.78769428e-01 -4.70386416e-01 7.62847438e-02 -7.92967081e-01
-1.13439548e+00 5.78200340e-01 7.37371385e-01 2.57331673e-02
6.11629903e-01 2.70950407e-01 7.36695588e-01 4.84982520e-01
2.46072069e-01 -1.24159408e+00 -4.32672314e-02 4.89967316e-01
5.33568025e-01 -1.41077065e+00 -3.12691852e-02 -3.09742242e-01
-5.83553553e-01 9.24620509e-01 8.71011794e-01 -4.87389594e-01
1.82393745e-01 4.33155954e-01 1.98526800e-01 -9.88155082e-02
-3.08750749e-01 -5.95253825e-01 4.07285631e-01 7.38760352e-01
-1.76936593e-02 -6.33841679e-02 1.61801070e-01 2.17007548e-02
-6.65850416e-02 -2.54068039e-02 2.21141487e-01 1.08652079e+00
-4.61993754e-01 -6.95748687e-01 -3.04607064e-01 2.31027752e-01
-3.84982616e-01 2.87102610e-01 -3.92593920e-01 1.28133619e+00
9.12079960e-02 4.31410253e-01 4.69555967e-02 -1.97991338e-02
1.69975400e-01 7.67638907e-02 6.57167971e-01 -3.40552926e-01
-5.78276038e-01 -5.39923348e-02 -2.35265028e-03 -1.00371718e+00
-5.34546971e-01 -9.50319707e-01 -1.40100157e+00 -6.02463745e-02
-4.15468097e-01 6.19227113e-03 8.98352742e-01 7.74759293e-01
1.06282979e-01 6.01694584e-01 4.38094661e-02 -1.42864394e+00
-6.04859948e-01 -6.84524953e-01 -2.12206066e-01 1.75535113e-01
7.12832391e-01 -9.48465347e-01 -1.97721004e-01 2.35373825e-01] | [6.511534690856934, -1.873499870300293] |
040f9838-7f97-4c33-86c0-a00073be4e6c | deep-amended-gradient-descent-for-efficient | 2108.05547 | null | https://arxiv.org/abs/2108.05547v1 | https://arxiv.org/pdf/2108.05547v1.pdf | Deep Amended Gradient Descent for Efficient Spectral Reconstruction from Single RGB Images | This paper investigates the problem of recovering hyperspectral (HS) images from single RGB images. To tackle such a severely ill-posed problem, we propose a physically-interpretable, compact, efficient, and end-to-end learning-based framework, namely AGD-Net. Precisely, by taking advantage of the imaging process, we first formulate the problem explicitly based on the classic gradient descent algorithm. Then, we design a lightweight neural network with a multi-stage architecture to mimic the formed amended gradient descent process, in which efficient convolution and novel spectral zero-mean normalization are proposed to effectively extract spatial-spectral features for regressing an initialization, a basic gradient, and an incremental gradient. Besides, based on the approximate low-rank property of HS images, we propose a novel rank loss to promote the similarity between the global structures of reconstructed and ground-truth HS images, which is optimized with our singular value weighting strategy during training. Moreover, AGD-Net, a single network after one-time training, is flexible to handle the reconstruction with various spectral response functions. Extensive experiments over three commonly-used benchmark datasets demonstrate that AGD-Net can improve the reconstruction quality by more than 1.0 dB on average while saving 67$\times$ parameters and 32$\times$ FLOPs, compared with state-of-the-art methods. The code will be publicly available at https://github.com/zbzhzhy/GD-Net. | ['Qingfu Zhang', 'Sen Jia', 'Junhui Hou', 'Hui Liu', 'Zhiyu Zhu'] | 2021-08-12 | null | null | null | null | ['spectral-reconstruction'] | ['computer-vision'] | [ 3.06260586e-01 -3.80558223e-01 2.55484253e-01 -4.34305489e-01
-7.62751877e-01 -4.11883555e-02 7.30384588e-02 -3.53079408e-01
-4.25041080e-01 5.71556389e-01 -6.20182417e-03 -3.22684973e-01
-4.43304628e-01 -8.05408180e-01 -7.00561285e-01 -1.09437180e+00
1.44920781e-01 -1.91989377e-01 -9.88669470e-02 -1.66772172e-01
-4.48031500e-02 5.94303548e-01 -1.25856769e+00 -1.25559732e-01
1.12903059e+00 1.42837656e+00 4.56459939e-01 1.97733015e-01
1.11822084e-01 6.59851551e-01 1.62625611e-01 -1.11131772e-01
5.79777420e-01 -4.21010882e-01 -5.57917476e-01 3.16930175e-01
3.87829304e-01 -5.67283690e-01 -7.19076633e-01 1.42545104e+00
8.88175011e-01 3.44474882e-01 3.98231834e-01 -6.35363698e-01
-5.98590434e-01 2.53181338e-01 -8.59061182e-01 -1.30915076e-01
-5.86666577e-02 3.71214598e-01 9.61159587e-01 -1.24053204e+00
1.29837990e-01 7.09164381e-01 8.59134972e-01 1.12570204e-01
-1.19464648e+00 -6.23355150e-01 6.14809394e-02 3.09881538e-01
-1.63449323e+00 -4.36219871e-01 1.03327751e+00 -1.70981750e-01
4.80315149e-01 3.32992256e-01 5.90753019e-01 5.99622488e-01
-4.07831281e-01 5.74702144e-01 1.15283585e+00 -3.20531398e-01
-8.65599662e-02 -5.16301215e-01 -3.28972302e-02 9.52701390e-01
3.13410431e-01 7.21220896e-02 -3.73085111e-01 3.46321128e-02
7.68504202e-01 1.65503815e-01 -8.81308556e-01 -2.08911195e-01
-1.01324785e+00 6.48868263e-01 8.64006221e-01 5.22242449e-02
-5.88283241e-01 2.72147115e-02 1.50989160e-01 5.92606552e-02
5.94929814e-01 7.88830072e-02 -2.97088772e-01 4.53761727e-01
-8.69737267e-01 -1.67238325e-01 4.28806454e-01 5.34072042e-01
1.05489087e+00 2.89784044e-01 -1.92538261e-01 1.01553273e+00
4.49543864e-01 7.60548413e-01 2.82019824e-01 -8.38641942e-01
4.05836612e-01 4.89985645e-01 2.92050205e-02 -1.12969208e+00
-4.92912114e-01 -8.19784522e-01 -1.39441478e+00 9.00962055e-02
-2.07759403e-02 -1.48669228e-01 -7.39400506e-01 1.50784957e+00
4.18970555e-01 4.09542829e-01 4.06075902e-02 1.32153571e+00
9.03492928e-01 7.69947112e-01 -3.23837340e-01 -4.31748897e-01
1.00883138e+00 -1.04392338e+00 -4.78633970e-01 -1.99003682e-01
2.71762669e-01 -6.74488664e-01 1.26784384e+00 3.61005723e-01
-9.78468478e-01 -4.57792014e-01 -1.17867827e+00 -1.37668878e-01
5.87048791e-02 5.46575487e-01 6.75355315e-01 2.89417475e-01
-7.73424268e-01 8.74545634e-01 -6.30986154e-01 6.25830069e-02
4.26541865e-01 2.06067756e-01 -1.61127448e-01 -3.31285477e-01
-9.62474585e-01 3.54350984e-01 4.20684278e-01 6.63957655e-01
-6.72622740e-01 -6.82671428e-01 -7.00727582e-01 2.62723118e-02
5.24778485e-01 -6.57639503e-01 9.63934243e-01 -8.31986964e-01
-1.74931550e+00 6.39249623e-01 -1.05083950e-01 5.07566035e-02
3.98169726e-01 -2.12047294e-01 -5.40774822e-01 4.09519523e-01
-7.97234178e-02 -3.24944742e-02 9.13963318e-01 -1.18876302e+00
-2.92093813e-01 -4.57209319e-01 -1.54992625e-01 4.32999909e-01
-5.99963725e-01 -2.03056112e-01 -5.39420962e-01 -5.84229648e-01
6.23783171e-01 -8.03925335e-01 -2.91989118e-01 3.16990793e-01
-5.07106304e-01 3.65550190e-01 4.87426043e-01 -1.00587487e+00
1.04906213e+00 -2.33096170e+00 1.10154182e-01 2.89404154e-01
2.99670815e-01 3.87717217e-01 -2.25186333e-01 1.74685776e-01
-2.42286891e-01 -1.70519561e-01 -7.27926373e-01 -2.81396002e-01
-7.04945549e-02 -1.54149398e-01 -2.00063542e-01 8.81993055e-01
2.50887941e-03 6.82478726e-01 -8.08391213e-01 -8.14430863e-02
2.82642365e-01 6.01003289e-01 -1.15817368e-01 3.41352105e-01
8.04236066e-03 5.31600058e-01 -5.04388094e-01 7.16017604e-01
1.22853994e+00 -4.41485018e-01 8.19233283e-02 -7.20785916e-01
-4.58711386e-01 3.03080082e-02 -1.37385297e+00 1.87853336e+00
-4.55424935e-01 2.61199236e-01 5.27650416e-01 -1.28560078e+00
8.78803492e-01 4.66100387e-02 5.10560870e-01 -8.24074924e-01
9.32436213e-02 5.47333181e-01 -3.70356530e-01 -5.22906423e-01
8.09833854e-02 -3.27069461e-01 5.09732187e-01 2.46453822e-01
-1.21462442e-01 -1.35680839e-01 -1.22583978e-01 -1.46539435e-01
7.84816802e-01 1.62333325e-01 7.29244128e-02 -2.06521913e-01
8.31305921e-01 -2.00150102e-01 6.47943377e-01 4.83434588e-01
9.72840115e-02 8.29723835e-01 -1.87642444e-02 -4.84713197e-01
-7.44996071e-01 -7.90775418e-01 -2.60068923e-01 8.63083720e-01
4.00165021e-01 -3.02354228e-02 -4.87743229e-01 -2.85793453e-01
-1.81717291e-01 3.62560123e-01 -1.86540112e-01 -1.94399446e-01
-2.74829865e-01 -1.27518058e+00 2.19529867e-01 1.69759542e-01
1.08906877e+00 -6.84854865e-01 -1.95378318e-01 2.01864213e-01
-2.46118069e-01 -9.77779984e-01 -2.54291564e-01 1.44471467e-01
-8.62124979e-01 -8.45299721e-01 -7.33268738e-01 -6.58276260e-01
6.03001535e-01 7.41553128e-01 8.48603189e-01 1.81903690e-01
-3.54249269e-01 5.91838323e-02 -3.20632517e-01 -5.39298281e-02
2.75265068e-01 -1.20514140e-01 -1.56856045e-01 4.32985604e-01
-1.13020442e-01 -9.04875278e-01 -9.46919680e-01 1.68964341e-01
-1.11692524e+00 3.99788797e-01 8.38891029e-01 1.06426764e+00
8.62195730e-01 2.85837561e-01 2.64960397e-02 -5.70907772e-01
3.16514730e-01 -3.70555490e-01 -7.75910735e-01 2.63196707e-01
-7.19718516e-01 -1.56229287e-02 7.86577940e-01 -2.14542434e-01
-1.18049121e+00 2.23662823e-01 -2.06963181e-01 -5.68296969e-01
1.39375612e-01 9.32612360e-01 -1.71906948e-01 -6.79224610e-01
5.96420467e-01 6.17332101e-01 3.89092937e-02 -6.28205955e-01
3.22908610e-01 5.94191551e-01 6.45091116e-01 -5.88294327e-01
1.29331255e+00 6.95476472e-01 2.70144314e-01 -9.59538102e-01
-1.18669069e+00 -6.07636809e-01 -3.56799752e-01 -2.00514734e-01
4.94737923e-01 -1.25810003e+00 -7.37379551e-01 9.00702834e-01
-8.86906862e-01 -6.01127207e-01 -1.26469731e-01 6.11154795e-01
-2.59287685e-01 7.47154236e-01 -6.32862747e-01 -6.61537170e-01
-6.80934310e-01 -8.87211680e-01 9.65560079e-01 3.29384655e-01
8.53287578e-01 -7.61596620e-01 -2.06159055e-01 3.81386220e-01
5.05740047e-01 2.90486455e-01 6.71539187e-01 1.19741499e-01
-7.42270172e-01 9.15466174e-02 -7.75324464e-01 6.95732772e-01
6.11900091e-02 -2.96154082e-01 -1.03405035e+00 -5.30841172e-01
3.74552518e-01 -4.50222909e-01 1.14021933e+00 3.37996334e-01
1.57473528e+00 -2.91990250e-01 1.87505096e-01 1.47302365e+00
1.92394137e+00 -2.50585109e-01 6.98688328e-01 3.90675813e-01
9.57095206e-01 4.19736594e-01 4.66590792e-01 7.20947266e-01
1.89466953e-01 3.06309253e-01 8.79341841e-01 -5.39274275e-01
-4.73905504e-02 5.26372977e-02 1.64093301e-01 9.75777984e-01
-3.77812743e-01 5.57106873e-03 -6.53567672e-01 1.40919179e-01
-1.81994545e+00 -7.54467666e-01 -3.27859551e-01 2.32525897e+00
7.54360378e-01 -2.59567887e-01 -3.37455332e-01 9.70802158e-02
6.34040713e-01 5.54089904e-01 -8.74312282e-01 5.05467176e-01
-3.73584807e-01 2.92050421e-01 7.26917505e-01 4.41725075e-01
-1.13657618e+00 6.67033553e-01 4.63573647e+00 1.00658774e+00
-1.52955234e+00 1.05017744e-01 5.92787564e-01 -5.15569234e-03
-2.52216190e-01 -8.50396976e-02 -2.89104283e-01 3.02827567e-01
3.84612888e-01 8.61228406e-02 9.25378382e-01 5.26492536e-01
4.45796400e-01 1.72932714e-01 -3.62106353e-01 1.22387791e+00
-3.12014157e-03 -1.19928014e+00 -1.60268620e-01 -2.27726214e-02
5.93074143e-01 2.93911755e-01 1.03433758e-01 -8.12196955e-02
-9.96900126e-02 -8.59227121e-01 6.81645274e-01 6.86895788e-01
1.10095048e+00 -6.47335112e-01 6.34211302e-01 1.82824939e-01
-1.39892757e+00 -1.80509642e-01 -6.91832423e-01 -5.33960499e-02
-3.19166929e-02 1.17360568e+00 -2.20048763e-02 1.22357380e+00
7.95300245e-01 8.54335070e-01 -3.56343359e-01 1.15477431e+00
-4.77957040e-01 5.90608418e-01 -2.80158222e-01 3.08356822e-01
3.70279551e-01 -8.80506039e-01 4.14376676e-01 9.63380694e-01
6.52936697e-01 5.83842754e-01 2.87306190e-01 7.70276785e-01
-9.36684608e-02 2.59265900e-01 -2.52931178e-01 1.15230255e-01
1.97839484e-01 1.65905499e+00 -3.98592651e-01 3.84568870e-02
-5.27287841e-01 1.15650880e+00 1.70249313e-01 5.90478361e-01
-7.91782558e-01 -5.89670181e-01 5.21000862e-01 -2.56971922e-02
3.52817982e-01 -2.80648649e-01 -8.10545981e-02 -1.30173945e+00
3.80768031e-01 -8.58536839e-01 1.59548610e-01 -1.04259491e+00
-1.35766697e+00 6.27196789e-01 -4.55513746e-01 -1.35022593e+00
5.14326096e-01 -8.58530581e-01 -6.34275436e-01 1.04186630e+00
-2.14632773e+00 -1.16432512e+00 -1.05182910e+00 6.48841798e-01
1.08382940e-01 6.68551475e-02 6.58074737e-01 3.92746955e-01
-9.85794008e-01 1.08711295e-01 5.01672745e-01 2.64849681e-02
4.50818002e-01 -9.68400896e-01 -8.26181993e-02 1.06566036e+00
-2.38983765e-01 4.41448659e-01 2.97022879e-01 -2.81250745e-01
-1.60095346e+00 -1.23719418e+00 2.95053571e-01 5.84080040e-01
7.61512816e-01 4.27207388e-02 -1.07080865e+00 2.07793906e-01
2.59872470e-02 4.54653949e-01 6.03049397e-01 -2.43751898e-01
-3.72679263e-01 -6.53418779e-01 -9.33213651e-01 4.51826245e-01
1.28403914e+00 -6.36852801e-01 -1.57752708e-02 6.69027746e-01
7.73687184e-01 -5.45924842e-01 -7.53919363e-01 7.23490179e-01
2.97925234e-01 -1.15208113e+00 1.19513762e+00 -1.12609804e-01
4.59287435e-01 -6.08010888e-01 -3.42386335e-01 -1.17876112e+00
-6.16072655e-01 -6.68089330e-01 -1.14736125e-01 8.70701969e-01
2.77610242e-01 -7.91362584e-01 5.19406855e-01 2.71219283e-01
-5.26912808e-01 -8.87049079e-01 -6.44578278e-01 -7.58013606e-01
-3.53754938e-01 -2.33036652e-01 4.90180433e-01 9.85050738e-01
-6.19684279e-01 2.47129381e-01 -4.88433540e-01 7.40608931e-01
9.98083711e-01 4.44567502e-01 4.16366428e-01 -1.11843121e+00
-4.31628853e-01 -3.58731657e-01 -2.86368951e-02 -1.14039338e+00
9.48203951e-02 -8.29634488e-01 2.24028364e-01 -1.56588149e+00
2.45665655e-01 -3.60004097e-01 -5.17339766e-01 5.17874837e-01
-2.39265397e-01 3.07734072e-01 7.42783174e-02 3.40170622e-01
-3.70023012e-01 9.72584724e-01 1.31326377e+00 -2.40399420e-01
-4.12037149e-02 -9.76157114e-02 -7.35056639e-01 6.72356248e-01
8.76054943e-01 -2.65415728e-01 -3.35780889e-01 -7.03390419e-01
3.04497659e-01 -9.70110968e-02 5.53440273e-01 -1.21179032e+00
2.38900259e-01 -2.98613846e-01 4.14499998e-01 -3.64187002e-01
4.77847010e-01 -8.10066879e-01 1.81578919e-01 3.73986721e-01
7.42118731e-02 -4.61188197e-01 4.66645658e-02 4.64425921e-01
-2.48423412e-01 -4.42746311e-01 8.77157807e-01 -1.27021357e-01
-7.73692191e-01 7.51867294e-01 2.33679473e-01 -3.15417379e-01
6.07212365e-01 2.75947433e-02 -2.36958563e-01 -2.36162424e-01
-4.39491540e-01 1.23689957e-01 4.46586579e-01 -4.77632821e-01
7.78250515e-01 -1.27255607e+00 -8.62547219e-01 1.30599439e-01
1.00445496e-02 3.21469396e-01 5.98670959e-01 1.03266370e+00
-7.77382195e-01 -1.25861600e-01 4.03406061e-02 -4.28827375e-01
-8.51264179e-01 1.79685295e-01 5.77156246e-01 -1.76226303e-01
-7.13468313e-01 9.46410537e-01 1.00238279e-01 -6.98209763e-01
2.16722135e-02 -6.08709780e-03 1.05244830e-01 -1.42770723e-01
5.01104653e-01 4.47937518e-01 2.34258875e-01 -4.51691270e-01
-8.21624324e-02 6.93553030e-01 4.28396553e-01 1.23222098e-01
1.85968840e+00 -1.26460344e-01 -4.97850865e-01 1.76883683e-01
1.26005888e+00 -2.11016890e-02 -1.59506118e+00 -5.36015511e-01
-3.97818804e-01 -5.90542257e-01 6.39127851e-01 -6.17992043e-01
-1.50068212e+00 8.31112981e-01 6.95062816e-01 1.03764705e-01
1.72472179e+00 -4.67727602e-01 7.92637825e-01 6.24252200e-01
2.05020919e-01 -8.19800258e-01 8.59637633e-02 5.12088895e-01
9.68976200e-01 -1.24865103e+00 3.44300598e-01 -5.77286839e-01
-2.72548199e-01 1.18688846e+00 4.33000833e-01 9.74271372e-02
5.27885795e-01 -2.89959103e-01 -9.48782545e-03 -2.72601932e-01
6.42414540e-02 -3.67584139e-01 3.10447514e-01 2.58125007e-01
1.98080033e-01 4.64211330e-02 -1.36138886e-01 4.13015544e-01
1.98638484e-01 -9.36286245e-03 2.29645252e-01 6.87807262e-01
-4.81564164e-01 -6.75762713e-01 -2.73565441e-01 4.59665596e-01
-1.54070228e-01 -4.60775644e-01 4.76906374e-02 3.86339545e-01
5.62651120e-02 6.81802273e-01 -4.25270110e-01 -4.37745541e-01
3.95230174e-01 -2.13715956e-01 2.83397794e-01 -4.20878083e-01
-4.90684696e-02 1.60302490e-01 -1.10163897e-01 -6.17443144e-01
-5.71165681e-01 -4.57956433e-01 -1.18388343e+00 -2.70346701e-01
-4.61023211e-01 -5.98342568e-02 6.99593186e-01 7.14271247e-01
3.87899041e-01 3.85878175e-01 1.06024432e+00 -1.17337751e+00
-8.04440081e-01 -9.57066178e-01 -9.72953141e-01 2.34972358e-01
4.59938437e-01 -4.95221227e-01 -6.45129859e-01 -1.79988042e-01] | [10.343132019042969, -2.052926540374756] |
2ad97161-ef13-4b21-b380-d6818d588f77 | improving-model-based-reinforcement-learning | 2102.05599 | null | https://arxiv.org/abs/2102.05599v1 | https://arxiv.org/pdf/2102.05599v1.pdf | Improving Model-Based Reinforcement Learning with Internal State Representations through Self-Supervision | Using a model of the environment, reinforcement learning agents can plan their future moves and achieve superhuman performance in board games like Chess, Shogi, and Go, while remaining relatively sample-efficient. As demonstrated by the MuZero Algorithm, the environment model can even be learned dynamically, generalizing the agent to many more tasks while at the same time achieving state-of-the-art performance. Notably, MuZero uses internal state representations derived from real environment states for its predictions. In this paper, we bind the model's predicted internal state representation to the environment state via two additional terms: a reconstruction model loss and a simpler consistency loss, both of which work independently and unsupervised, acting as constraints to stabilize the learning process. Our experiments show that this new integration of reconstruction model loss and simpler consistency loss provide a significant performance increase in OpenAI Gym environments. Our modifications also enable self-supervised pretraining for MuZero, so the algorithm can learn about environment dynamics before a goal is made available. | ['Stefan Wermter', 'Muhammad Burhan Hafez', 'Cornelius Weber', 'Julien Scholz'] | 2021-02-10 | null | null | null | null | ['board-games'] | ['playing-games'] | [-1.89247847e-01 3.13367933e-01 -2.89544910e-01 -2.03384206e-01
-3.82679760e-01 -4.36660528e-01 4.83490556e-01 2.79081494e-01
-8.16237450e-01 9.00059164e-01 4.37016599e-02 -3.21930391e-03
-1.00459464e-01 -7.95797586e-01 -7.99165606e-01 -6.19626939e-01
-4.76969898e-01 6.17397964e-01 4.99986857e-01 -5.45866013e-01
1.68227673e-01 3.08919638e-01 -1.78857636e+00 -1.27002802e-02
8.30569863e-01 5.00628352e-01 4.19086248e-01 8.34484100e-01
3.35337073e-01 1.10287988e+00 -5.14449477e-01 1.93190023e-01
4.41554219e-01 -4.72425312e-01 -8.72826397e-01 -1.07958205e-01
8.16312730e-02 -3.24638963e-01 -7.67403185e-01 6.34630799e-01
4.21678543e-01 7.89572358e-01 2.50013053e-01 -1.11775017e+00
-1.38790205e-01 6.14544749e-01 -6.46442920e-02 1.61542952e-01
3.77795577e-01 6.83309972e-01 8.74567926e-01 -2.88324943e-03
6.42247260e-01 1.03436720e+00 6.58848822e-01 6.76736712e-01
-1.45506883e+00 -5.37850320e-01 5.02796173e-01 1.53855830e-01
-1.11948419e+00 -3.37350041e-01 5.60465157e-01 -1.08358540e-01
1.25356233e+00 5.04001006e-02 1.05637360e+00 9.34597135e-01
5.83450615e-01 9.17521596e-01 1.07915521e+00 -3.32471400e-01
7.29902267e-01 -2.82349735e-01 -1.72667176e-01 8.86462510e-01
-2.48800125e-02 6.04513764e-01 -6.61870718e-01 -1.37993172e-02
8.59583616e-01 -1.95834011e-01 -1.32651150e-01 -9.81504381e-01
-1.03647554e+00 4.60421979e-01 5.59589207e-01 -1.35857075e-01
-2.94401169e-01 5.97004712e-01 1.96945429e-01 4.74679321e-01
-9.14801955e-02 8.29447150e-01 -4.38462466e-01 -4.89452124e-01
-4.56394672e-01 7.96876073e-01 7.88177669e-01 7.15587139e-01
7.39953756e-01 2.69275397e-01 2.56166831e-02 4.99002188e-01
2.03350052e-01 4.70412672e-01 6.75614238e-01 -1.34232080e+00
3.34866971e-01 3.30426812e-01 4.67845887e-01 -6.89335585e-01
-7.09981561e-01 -6.26154006e-01 -1.90231323e-01 6.79794371e-01
5.43137074e-01 -4.05691504e-01 -9.03678834e-01 2.18918896e+00
4.22090441e-01 5.12567401e-01 1.03536725e-01 8.66929531e-01
2.98804436e-02 4.58595008e-01 1.61753237e-01 9.68651026e-02
8.31150055e-01 -1.00859678e+00 -4.59146738e-01 -7.44296849e-01
8.08691561e-01 -1.11967452e-01 1.17480993e+00 4.68931943e-01
-1.36988592e+00 -6.36464000e-01 -1.15026355e+00 1.82899311e-01
-7.79896751e-02 -2.14183390e-01 9.91304755e-01 4.02012855e-01
-1.09856808e+00 1.16760039e+00 -1.60472310e+00 -2.14002833e-01
6.51732311e-02 6.25406086e-01 -3.46801370e-01 3.01212430e-01
-9.60170627e-01 1.14073980e+00 6.76581800e-01 -1.31091624e-01
-1.22875178e+00 -3.90064895e-01 -1.06951940e+00 1.72238573e-01
4.29030895e-01 -6.01333678e-01 1.58921742e+00 -7.19964147e-01
-2.06303191e+00 3.99199247e-01 2.64660209e-01 -7.63030410e-01
4.14424539e-01 -2.25201175e-01 -1.13110341e-01 -5.23668379e-02
2.58101858e-02 7.67593145e-01 5.25543928e-01 -1.03320944e+00
-6.34598911e-01 -2.91524112e-01 2.61134416e-01 7.67192960e-01
-1.73248067e-01 -5.87729871e-01 -4.25906330e-01 -3.47855657e-01
1.95878997e-01 -1.18548334e+00 -6.55021727e-01 -7.55277351e-02
2.44770676e-01 -1.51022837e-01 2.62475610e-01 -3.43884856e-01
1.10739839e+00 -1.98416030e+00 4.32623625e-01 3.26913625e-01
-7.30967848e-04 -1.56151904e-02 -4.03479815e-01 3.46724331e-01
1.11486614e-01 -3.73927563e-01 1.34262852e-02 -3.10390294e-01
6.07086457e-02 5.04047990e-01 -5.30457236e-02 4.57074761e-01
-8.29656795e-02 8.23661625e-01 -1.35592663e+00 -1.03148095e-01
2.84084588e-01 4.73261438e-02 -1.06575787e+00 1.50855541e-01
-3.50141943e-01 5.26972711e-01 -4.51668531e-01 6.89203152e-03
-2.09157504e-02 -6.86684623e-02 5.80489814e-01 4.80998367e-01
3.75574194e-02 5.46060443e-01 -1.42906582e+00 2.05793691e+00
-4.93098140e-01 2.45260537e-01 2.40151281e-03 -8.25665057e-01
7.91089714e-01 8.38734135e-02 5.60076773e-01 -8.08882713e-01
1.65174857e-01 -8.33703429e-02 3.38886768e-01 -2.32702941e-01
6.29719794e-01 -1.85323372e-01 -1.27225667e-01 3.71128529e-01
2.00324863e-01 -3.99896115e-01 1.98255554e-01 -1.54372957e-02
1.26171541e+00 9.00006950e-01 8.64369646e-02 -3.57630849e-01
8.69932845e-02 2.78346688e-01 7.40973711e-01 1.07642007e+00
-3.62357378e-01 1.64350882e-01 2.86264896e-01 -5.42230904e-01
-9.87775147e-01 -1.12013948e+00 2.09572047e-01 1.10962093e+00
4.15114373e-01 -5.85003495e-01 -6.51993155e-01 -3.89630556e-01
8.10489357e-02 8.08525324e-01 -6.79084361e-01 -6.59730077e-01
-9.25128818e-01 -4.51112151e-01 3.82344216e-01 8.15690219e-01
4.51588511e-01 -1.16469920e+00 -1.20747066e+00 4.86803621e-01
-1.03806153e-01 -6.45939469e-01 -2.34687135e-01 6.64482057e-01
-9.96889710e-01 -8.85555029e-01 -1.22076832e-01 -6.39066875e-01
4.19048429e-01 -8.99950042e-02 1.01105428e+00 1.53955117e-01
-2.30329886e-01 5.20701170e-01 -5.01210801e-02 -2.23010734e-01
-3.21721345e-01 1.22172989e-01 5.12476027e-01 -6.98901534e-01
5.83742261e-02 -8.22234333e-01 -5.08578122e-01 1.24914542e-01
-5.17133534e-01 1.42174214e-01 2.26942241e-01 1.07044971e+00
5.49305558e-01 1.62977964e-01 2.47761264e-01 -4.98496622e-01
5.10202408e-01 -1.68935314e-01 -7.11175084e-01 -7.67526701e-02
-6.10909581e-01 5.46557307e-01 5.95737934e-01 -5.84188461e-01
-1.02842963e+00 2.06974879e-01 -1.77423298e-01 -2.49087811e-01
-5.78189380e-02 3.48923981e-01 1.24078818e-01 -6.18837448e-03
9.73213434e-01 2.63573080e-01 3.20158273e-01 -3.45278442e-01
1.81327879e-01 -6.88231736e-02 6.50337875e-01 -1.03678834e+00
7.77489245e-01 1.96004972e-01 -9.56698060e-02 -3.54547799e-01
-7.21255720e-01 -2.40150511e-01 -3.43405485e-01 -1.19584866e-01
6.42719865e-01 -9.41875458e-01 -1.25547695e+00 4.80695784e-01
-3.74836206e-01 -1.22715855e+00 -6.27794385e-01 5.54320514e-01
-9.99709249e-01 1.18169896e-01 -8.65784883e-01 -7.00894058e-01
3.26384634e-01 -1.06643128e+00 6.22322083e-01 4.05382127e-01
-1.58340678e-01 -8.64397943e-01 4.51991111e-01 5.14357798e-02
3.33494455e-01 3.86060365e-02 6.37169838e-01 -3.45358491e-01
-3.91062558e-01 4.54233475e-02 5.29079318e-01 8.74386951e-02
-1.00489911e-02 -5.24137557e-01 -6.33694649e-01 -6.13552988e-01
-1.72622874e-01 -5.15014768e-01 6.34424090e-01 2.51080304e-01
1.03492081e+00 -1.30650565e-01 -2.64440686e-01 5.14057398e-01
1.17739177e+00 2.33340055e-01 6.53494716e-01 7.73857534e-01
2.56955594e-01 2.43691877e-01 6.78119361e-01 6.27559364e-01
5.62376261e-01 7.35688090e-01 5.35898387e-01 1.93849728e-01
-1.77441873e-02 -5.91839671e-01 4.68353420e-01 5.37942767e-01
-1.41710415e-01 6.28391951e-02 -9.72423851e-01 4.14426833e-01
-2.13246727e+00 -1.13883662e+00 4.95483518e-01 2.38897848e+00
9.04776633e-01 5.07376969e-01 2.42202312e-01 -3.93264405e-02
1.23329632e-01 1.21243142e-01 -9.05149579e-01 -2.53100514e-01
3.00422281e-01 1.87839985e-01 5.62737167e-01 8.38039577e-01
-1.04397631e+00 1.25964248e+00 7.53018522e+00 4.20068115e-01
-1.03719258e+00 -2.80289322e-01 4.31305617e-01 -3.52146983e-01
-1.00037128e-01 1.16541848e-01 -6.33882880e-01 2.41970569e-01
1.02279639e+00 -7.96016827e-02 9.43900883e-01 1.12309420e+00
1.66885972e-01 -2.70369262e-01 -1.13454223e+00 7.13653743e-01
-1.78327888e-01 -1.23562205e+00 -4.63489085e-01 -1.18625294e-02
5.94399869e-01 1.80009320e-01 -6.27948791e-02 9.15777028e-01
9.97537851e-01 -8.39052558e-01 9.69665468e-01 3.83791894e-01
2.53892541e-01 -8.43714595e-01 3.73639226e-01 7.10595965e-01
-9.80412841e-01 -4.33985889e-01 -2.47779101e-01 -7.13366926e-01
-1.53230414e-01 -3.98655057e-01 -7.40470529e-01 2.38405466e-01
6.52818441e-01 4.05853033e-01 -4.72668201e-01 1.07159436e+00
-2.50939846e-01 5.22965372e-01 -4.62088376e-01 -9.73274708e-02
3.35777879e-01 -1.67629942e-01 4.90058988e-01 5.54876685e-01
-2.87134238e-02 2.97457010e-01 7.70484209e-01 5.92998385e-01
3.65815043e-01 -2.59846330e-01 -4.27957684e-01 2.44715124e-01
4.55426514e-01 7.88030326e-01 -6.57107353e-01 -2.51773357e-01
2.10683003e-01 9.17594731e-01 7.25398421e-01 3.30547839e-01
-8.89033258e-01 -2.41861597e-01 9.57120776e-01 3.67417559e-02
2.49903053e-01 -5.57681501e-01 -1.65665805e-01 -1.37224138e+00
-2.29738459e-01 -1.00052035e+00 1.51572913e-01 -6.50795162e-01
-6.93527281e-01 3.58183175e-01 6.94555119e-02 -1.20263171e+00
-5.93809366e-01 -4.98012304e-01 -4.32578713e-01 5.55847704e-01
-1.26052642e+00 -5.98577142e-01 -1.29255280e-01 4.66102570e-01
4.85483944e-01 -1.56972781e-01 1.01315761e+00 -1.32503077e-01
-5.16316891e-01 4.91703957e-01 1.28422529e-01 -2.42301542e-02
5.01932919e-01 -1.54502678e+00 4.31077927e-01 6.55634105e-01
2.01595612e-02 7.13782310e-01 1.06848109e+00 -6.40296161e-01
-1.39873755e+00 -5.10085464e-01 1.63271502e-01 -4.63899314e-01
6.62711978e-01 -2.74469078e-01 -6.63099825e-01 1.00636280e+00
-1.04748592e-01 -4.35667783e-02 3.82830173e-01 5.61436236e-01
-1.50029659e-01 -3.37639935e-02 -1.01546288e+00 1.01800859e+00
1.24518466e+00 -3.07126075e-01 -7.85014033e-01 1.18139073e-01
4.56017494e-01 -1.00714958e+00 -7.85195351e-01 7.02072456e-02
6.50058329e-01 -7.84620464e-01 9.60602343e-01 -1.07053399e+00
2.08204567e-01 -4.41973090e-01 9.77487788e-02 -1.55729783e+00
-6.54486120e-01 -6.45722806e-01 -3.11559737e-01 5.39139748e-01
2.22915664e-01 -5.10949433e-01 1.20138276e+00 9.23339546e-01
-3.11372131e-01 -6.32917106e-01 -8.95831466e-01 -9.11095202e-01
2.07445309e-01 -3.60506505e-01 6.10771537e-01 6.00489616e-01
4.23639894e-01 2.28614733e-01 -5.06801486e-01 2.03123987e-01
5.41913509e-01 -1.88326575e-02 9.44722891e-01 -1.13071918e+00
-8.89921665e-01 -4.09780055e-01 -3.47159714e-01 -1.37545705e+00
2.38779679e-01 -6.63594186e-01 4.55765337e-01 -1.41101813e+00
-8.56191069e-02 -7.68729389e-01 -4.96484399e-01 8.98081720e-01
-1.31843403e-01 -1.22574151e-01 4.01886821e-01 4.45982516e-02
-7.82576442e-01 7.72048593e-01 1.33231866e+00 -2.60334127e-02
-6.29097760e-01 -1.04232701e-02 -5.29192686e-01 7.99639046e-01
9.63209629e-01 -4.50860918e-01 -7.70157397e-01 -5.68369746e-01
6.80396855e-02 2.69176543e-01 1.94631696e-01 -1.51734364e+00
2.97877163e-01 -5.84207356e-01 4.66830105e-01 1.00551182e-02
6.50770187e-01 -6.25382721e-01 -1.60994697e-02 9.77038205e-01
-5.69223940e-01 1.87071294e-01 4.50923383e-01 7.64946878e-01
1.71864942e-01 -1.06162131e-01 6.55447543e-01 -2.29754061e-01
-8.93427074e-01 2.02165712e-02 -7.22885787e-01 6.67694956e-02
1.00960076e+00 -2.64265984e-01 -1.20130964e-01 -4.12012696e-01
-1.04515684e+00 6.05433583e-01 6.02495670e-01 4.07464921e-01
2.89000213e-01 -1.10182381e+00 -4.07965302e-01 2.17264861e-01
-1.15589567e-01 1.31159145e-02 3.07108045e-01 4.72553641e-01
-6.17482305e-01 -2.29819752e-02 -6.52721286e-01 -4.15908962e-01
-9.69774663e-01 4.03617591e-01 5.98638833e-01 -6.16767704e-01
-9.34363365e-01 6.88476384e-01 -6.39242455e-02 -9.08769667e-01
3.84013742e-01 -3.93968314e-01 2.83487681e-02 -6.60264254e-01
5.47768176e-01 1.46117777e-01 -1.30580187e-01 -4.37761135e-02
-2.20801592e-01 -4.51082252e-02 -8.42877626e-02 -3.58953387e-01
1.43665552e+00 1.10866122e-01 3.48900467e-01 3.45439851e-01
2.51807213e-01 -2.05206946e-01 -1.85415411e+00 -8.20787922e-02
6.39235228e-03 -4.05330420e-01 8.70284438e-02 -9.35119092e-01
-6.38219595e-01 4.44528699e-01 7.14807451e-01 -2.56152719e-01
8.71768892e-01 -3.77659768e-01 5.93821049e-01 9.62309361e-01
8.83875072e-01 -1.38777173e+00 3.65481853e-01 7.80566275e-01
4.77004021e-01 -1.01238775e+00 -3.79758179e-02 1.89315423e-01
-7.34792769e-01 8.01800311e-01 1.08679676e+00 -5.52828014e-01
3.18711609e-01 3.84398013e-01 -8.77920538e-02 2.28312597e-01
-1.07750618e+00 -1.82863340e-01 -7.89460540e-02 7.41516590e-01
1.00107759e-01 1.19933218e-01 -8.23689178e-02 4.39124584e-01
-4.23313171e-01 -9.40792486e-02 5.18500149e-01 1.17611039e+00
-7.49992013e-01 -1.30234742e+00 -2.07982451e-01 2.69382358e-01
-3.36433873e-02 2.48827502e-01 1.97726399e-01 8.60177040e-01
-1.68701887e-01 6.37507498e-01 1.23930372e-01 -3.68726224e-01
4.62636948e-01 9.47115272e-02 8.22005928e-01 -6.29153788e-01
-8.53387296e-01 -8.18374455e-02 1.02298476e-01 -9.66686130e-01
1.04348101e-01 -7.20875978e-01 -1.61316776e+00 -2.96757013e-01
-1.96950600e-01 2.83903301e-01 3.41343701e-01 7.99026310e-01
3.54670554e-01 7.33080029e-01 3.49543929e-01 -1.01521957e+00
-1.01893806e+00 -6.13979638e-01 -4.70144451e-01 3.85087162e-01
2.11873248e-01 -9.17848885e-01 6.48894683e-02 -2.21874624e-01] | [4.180840492248535, 1.5734302997589111] |
ad490ff1-c224-4e07-b598-e8b8961cc56e | the-dlcc-node-classification-benchmark-for | 2207.06014 | null | https://arxiv.org/abs/2207.06014v1 | https://arxiv.org/pdf/2207.06014v1.pdf | The DLCC Node Classification Benchmark for Analyzing Knowledge Graph Embeddings | Knowledge graph embedding is a representation learning technique that projects entities and relations in a knowledge graph to continuous vector spaces. Embeddings have gained a lot of uptake and have been heavily used in link prediction and other downstream prediction tasks. Most approaches are evaluated on a single task or a single group of tasks to determine their overall performance. The evaluation is then assessed in terms of how well the embedding approach performs on the task at hand. Still, it is hardly evaluated (and often not even deeply understood) what information the embedding approaches are actually learning to represent. To fill this gap, we present the DLCC (Description Logic Class Constructors) benchmark, a resource to analyze embedding approaches in terms of which kinds of classes they can represent. Two gold standards are presented, one based on the real-world knowledge graph DBpedia and one synthetic gold standard. In addition, an evaluation framework is provided that implements an experiment protocol so that researchers can directly use the gold standard. To demonstrate the use of DLCC, we compare multiple embedding approaches using the gold standards. We find that many DL constructors on DBpedia are actually learned by recognizing different correlated patterns than those defined in the gold standard and that specific DL constructors, such as cardinality constraints, are particularly hard to be learned for most embedding approaches. | ['Heiko Paulheim', 'Jan Portisch'] | 2022-07-13 | null | null | null | null | ['knowledge-graph-embeddings', 'knowledge-graph-embeddings'] | ['graphs', 'methodology'] | [-2.46304572e-01 6.03477776e-01 -5.36235452e-01 -2.55733520e-01
-9.81773213e-02 -6.86150193e-01 8.03237557e-01 7.61546075e-01
-2.82898247e-01 7.39093781e-01 3.32297593e-01 -3.20084453e-01
-5.07763386e-01 -1.34011304e+00 -6.32722139e-01 -2.57337868e-01
-5.46441376e-01 8.13757837e-01 3.93703729e-01 -5.44841468e-01
-2.54831046e-01 1.73746869e-01 -1.51756489e+00 4.19600725e-01
5.50913990e-01 6.47574723e-01 -2.86812901e-01 2.21006930e-01
-2.47141317e-01 7.69426703e-01 -5.34200728e-01 -1.06658137e+00
7.65179619e-02 5.91175677e-03 -1.15301478e+00 -6.86149538e-01
4.09946829e-01 2.84354985e-01 -6.06663883e-01 1.08322465e+00
3.56724650e-01 -8.82232860e-02 6.68920577e-01 -1.43099010e+00
-1.09704649e+00 9.91248190e-01 2.47143939e-01 1.42352074e-01
6.99839056e-01 -1.50390789e-01 1.73097241e+00 -7.68666446e-01
1.15461028e+00 1.20138407e+00 8.08818817e-01 3.84437025e-01
-1.42189312e+00 -2.43816122e-01 -3.05404956e-03 5.60322821e-01
-1.24141276e+00 -1.18333928e-03 6.42942905e-01 -6.20777905e-01
1.42159152e+00 3.17988127e-01 6.84601367e-01 1.03688908e+00
-7.30494559e-02 3.97634864e-01 7.66357720e-01 -5.86649776e-01
9.08458754e-02 5.02390981e-01 6.75971627e-01 8.84338975e-01
7.54940927e-01 1.94121286e-01 -3.02748352e-01 -3.43636632e-01
2.04251766e-01 -3.05673361e-01 -6.65205717e-01 -7.93280840e-01
-1.30035853e+00 1.09340310e+00 6.29626930e-01 4.83234137e-01
3.74442292e-03 3.49556446e-01 7.06549823e-01 6.09094679e-01
2.86216080e-01 8.50633025e-01 -5.46079993e-01 5.69173284e-02
-2.11284339e-01 3.38774502e-01 1.26879883e+00 1.07670915e+00
7.28731036e-01 -2.37135217e-01 -1.61316037e-01 8.29607606e-01
3.77035141e-01 8.69310796e-02 2.43204072e-01 -4.73980933e-01
6.01524293e-01 9.89586592e-01 -5.02300933e-02 -1.32855344e+00
-3.13804030e-01 -2.21654594e-01 -1.37859583e-01 6.85680732e-02
4.62877661e-01 5.36172278e-02 -4.67716634e-01 1.62098622e+00
1.97762586e-02 1.55263767e-01 4.46056932e-01 7.21379578e-01
1.14712548e+00 5.53501785e-01 5.58043011e-02 2.02161700e-01
1.58293116e+00 -9.09401834e-01 -5.91651738e-01 -9.86977145e-02
1.11839449e+00 -2.82046765e-01 1.03954530e+00 4.51538563e-02
-5.94525874e-01 -3.00916106e-01 -1.41053116e+00 -6.81006685e-02
-1.18389130e+00 -3.44810724e-01 9.75920498e-01 6.87283397e-01
-8.31950128e-01 7.73967564e-01 -5.79050839e-01 -4.61004019e-01
1.28661871e-01 2.65502810e-01 -7.86506236e-01 -2.45598525e-01
-1.79540253e+00 1.26561391e+00 9.23283756e-01 -2.98196107e-01
-5.59933305e-01 -7.10096121e-01 -1.10669041e+00 2.34562054e-01
3.59427094e-01 -6.55741990e-01 5.41983247e-01 -4.92985785e-01
-8.36837888e-01 8.75687480e-01 5.05251050e-01 -5.72726488e-01
1.20709561e-01 1.15168475e-01 -7.94303298e-01 -1.33593485e-01
-1.14918210e-01 2.85589188e-01 2.70918697e-01 -1.28898418e+00
-5.10673165e-01 -1.91088632e-01 6.57935500e-01 -6.59293756e-02
-5.13448894e-01 -1.57373890e-01 -4.60357606e-01 -4.04960364e-01
-2.73908675e-01 -8.73894811e-01 1.54964924e-01 -1.25317276e-01
-5.04265189e-01 -4.74891931e-01 7.34572649e-01 -4.09754187e-01
1.42167854e+00 -1.87876797e+00 2.67871886e-01 4.20152426e-01
4.39291239e-01 3.56403142e-01 -1.03555471e-01 8.99065852e-01
-4.37305242e-01 4.26775903e-01 -8.48798230e-02 1.25680640e-01
4.96520817e-01 5.76079071e-01 -1.55827507e-01 2.58620054e-01
2.61833668e-01 1.01881266e+00 -1.30736709e+00 -2.34646022e-01
-7.33916312e-02 5.65118372e-01 -5.49744487e-01 -9.15008038e-02
-4.22308981e-01 -4.12979245e-01 -1.77569568e-01 5.46838164e-01
1.76506102e-01 -2.75353074e-01 7.38682091e-01 -6.45383775e-01
4.14265722e-01 2.18323022e-01 -1.15937948e+00 1.34403074e+00
-2.02638701e-01 8.38582039e-01 -6.56058252e-01 -1.13953853e+00
8.98651063e-01 3.71136934e-01 2.55507171e-01 -4.31960851e-01
-3.69665682e-01 1.81248307e-01 2.15300590e-01 -6.53046072e-01
5.24904311e-01 -2.05149986e-02 -4.62814048e-02 1.86986536e-01
4.80069667e-01 1.82089761e-01 4.59313035e-01 3.03780019e-01
1.49743223e+00 -1.56296849e-01 3.95026773e-01 -1.94455013e-01
3.66432428e-01 1.67999685e-01 5.41049540e-01 3.29828680e-01
2.29597129e-02 1.80224001e-01 1.03864348e+00 -6.42335296e-01
-8.91707122e-01 -1.19783688e+00 -3.32864016e-01 8.48704994e-01
1.54965565e-01 -1.09475136e+00 -1.50775671e-01 -1.24842250e+00
4.86764878e-01 8.57449770e-01 -8.66525471e-01 -3.10413092e-01
-1.28092304e-01 -6.81854308e-01 9.50807154e-01 5.53145885e-01
-2.47015730e-02 -1.00233090e+00 -1.32366240e-01 1.22153319e-01
1.71501130e-01 -1.11108327e+00 2.47437172e-02 1.98097602e-01
-5.45273423e-01 -1.59813225e+00 -3.56864855e-02 -9.61229980e-01
6.14345074e-01 -1.98082209e-01 1.60921955e+00 2.40302354e-01
1.54464291e-02 6.31826282e-01 -6.48127437e-01 -2.66683936e-01
-4.14606839e-01 -9.07707438e-02 2.55831555e-02 -3.53990287e-01
7.45963514e-01 -4.38316584e-01 -2.28363752e-01 2.10158303e-01
-8.20586205e-01 -4.11642432e-01 3.18107396e-01 1.05773473e+00
4.06794637e-01 1.87551409e-01 6.14271164e-01 -1.35724974e+00
1.06673706e+00 -7.07904875e-01 -4.37722266e-01 6.27504945e-01
-1.08144248e+00 3.87912840e-01 3.95446360e-01 -1.69846520e-01
-4.62349653e-01 -3.05034995e-01 7.37144276e-02 -2.50503093e-01
1.68845743e-01 1.18274808e+00 -2.05229089e-01 -5.84635250e-02
9.62335289e-01 -1.54258281e-01 -7.85067230e-02 -3.95829856e-01
7.04621255e-01 4.06601578e-01 2.38061175e-01 -9.16333437e-01
7.51523137e-01 8.01717341e-02 -1.44179374e-01 -5.06562948e-01
-5.75075626e-01 -2.41603777e-01 -4.91644144e-01 1.74951062e-01
8.41668069e-01 -6.66823626e-01 -4.51946795e-01 -3.84552300e-01
-1.20618618e+00 -1.73528329e-01 -4.71769035e-01 5.41792870e-01
-3.73005867e-01 2.42979854e-01 -5.16412258e-01 -1.64173692e-01
-1.02022439e-01 -9.63946998e-01 4.88899440e-01 -1.78089619e-01
-2.33605772e-01 -1.70025301e+00 3.75624806e-01 3.40441205e-02
2.60449141e-01 4.76253897e-01 1.60718763e+00 -9.94182289e-01
-3.11887711e-01 -3.61351132e-01 -2.14412212e-01 3.92079562e-01
8.63032639e-02 1.61557198e-01 -8.27775955e-01 -2.03838855e-01
-8.28802884e-01 -1.69567302e-01 7.33197689e-01 -3.09919596e-01
8.89801502e-01 -3.61235887e-01 -6.40035629e-01 3.99046689e-01
1.82649767e+00 -4.52417284e-02 8.86322737e-01 4.93826479e-01
6.22319162e-01 6.55739844e-01 4.93656665e-01 -1.33899242e-01
6.39704645e-01 1.00094378e+00 5.48591793e-01 2.16298595e-01
-1.30822763e-01 -4.29077983e-01 1.96132034e-01 7.50751674e-01
-3.89190644e-01 -3.06516200e-01 -1.20593381e+00 5.90576112e-01
-1.92268074e+00 -1.11224055e+00 -2.92496711e-01 2.05122900e+00
1.02259767e+00 2.52892971e-01 1.71237290e-02 1.89744428e-01
4.97878969e-01 -2.70154625e-02 1.20482534e-01 -5.31295896e-01
-1.57029405e-01 2.75798976e-01 4.08744276e-01 5.49283266e-01
-1.03478920e+00 8.49946380e-01 6.08202076e+00 4.71052021e-01
-8.00582051e-01 8.09859931e-02 -1.72511980e-01 4.08479989e-01
-6.10299408e-01 2.89949507e-01 -5.74821234e-01 5.05793035e-01
8.73629987e-01 -4.42320108e-01 2.67921329e-01 8.78758967e-01
-4.96602029e-01 4.57074374e-01 -1.74970996e+00 9.33908880e-01
-9.88016054e-02 -1.61375749e+00 1.59953997e-01 4.91155796e-02
4.22731608e-01 -4.88854945e-03 -3.26266021e-01 8.84478331e-01
4.73709911e-01 -1.26928830e+00 2.66943216e-01 6.35469794e-01
6.64060712e-01 -4.67909634e-01 1.07985377e+00 -1.60225421e-01
-1.17749763e+00 -4.01934125e-02 -5.00382304e-01 1.27368718e-01
-1.77713949e-02 5.49734533e-01 -9.86711025e-01 9.90532279e-01
5.37297368e-01 9.34669793e-01 -8.64910841e-01 9.41962540e-01
-6.82708621e-01 7.13700533e-01 -4.92289178e-02 -5.12257628e-02
1.04885943e-01 -4.19969782e-02 4.53989416e-01 1.57891023e+00
1.44680545e-01 -3.69527698e-01 6.34322781e-03 8.06393027e-01
-2.58169830e-01 6.49336800e-02 -1.13147175e+00 -4.25266057e-01
6.00949764e-01 1.12001038e+00 -1.67000204e-01 -1.86894268e-01
-8.34655404e-01 5.62683403e-01 5.84054053e-01 4.21505362e-01
-9.16677475e-01 -7.21149325e-01 9.14618075e-01 9.15393233e-02
1.77586108e-01 -1.05308130e-01 2.45094866e-01 -1.23054194e+00
8.98624212e-02 -6.91316903e-01 7.15154052e-01 -6.16770685e-01
-1.63677919e+00 5.83915293e-01 2.68324792e-01 -1.01574874e+00
-2.04549983e-01 -9.98893321e-01 -5.14502108e-01 6.00685716e-01
-1.58034098e+00 -9.86290514e-01 -2.91101992e-01 4.22276586e-01
-2.85657346e-01 -4.03579235e-01 1.39882827e+00 6.77354455e-01
-4.71267223e-01 7.24334002e-01 -1.31026983e-01 5.44738114e-01
6.39894664e-01 -1.58011734e+00 1.26435965e-01 3.12382340e-01
5.50602019e-01 8.63039732e-01 7.50593781e-01 -4.99497831e-01
-1.48843408e+00 -1.13026798e+00 9.42319334e-01 -8.51868629e-01
1.01097620e+00 -2.95916706e-01 -1.14312279e+00 1.06742120e+00
1.89294145e-01 5.29033065e-01 9.42568719e-01 6.88166797e-01
-7.72824526e-01 -2.52596498e-01 -1.00495839e+00 3.55907261e-01
1.23444545e+00 -6.08418286e-01 -1.03199184e+00 4.60896879e-01
7.96915412e-01 -8.91694054e-02 -1.51559913e+00 4.35957670e-01
4.47138101e-01 -7.20970452e-01 1.11305189e+00 -1.31529498e+00
5.12780845e-01 -4.57440585e-01 -4.53374982e-01 -1.68109584e+00
-3.45389009e-01 -4.15489636e-02 -6.24456763e-01 1.33914566e+00
9.16565657e-01 -8.97852004e-01 7.83451855e-01 4.43408102e-01
2.06950195e-02 -9.72236216e-01 -7.08970308e-01 -1.18785846e+00
6.34850338e-02 -3.24649543e-01 6.56157851e-01 1.64163256e+00
5.93068957e-01 6.85108006e-01 1.66440606e-02 2.14527920e-01
3.54138225e-01 1.13545984e-01 5.96236467e-01 -1.46104300e+00
-2.67700672e-01 -3.34990084e-01 -1.35479605e+00 -2.24486366e-01
2.31774390e-01 -1.71271145e+00 -6.77903354e-01 -1.91227794e+00
-1.06848776e-01 -6.94371164e-01 -5.04351556e-01 8.36648762e-01
-3.76415215e-02 -6.65788874e-02 7.40972087e-02 3.17319296e-02
-5.03538609e-01 4.14642096e-01 8.26649845e-01 -6.03599250e-01
4.65114117e-02 -4.35082793e-01 -7.54751146e-01 4.99941319e-01
6.49325252e-01 -5.11549115e-01 -7.53431439e-01 -4.84086215e-01
8.87199819e-01 -3.56768280e-01 4.52707857e-01 -8.57829869e-01
9.93136019e-02 6.61745593e-02 -1.07423827e-01 1.64305702e-01
2.74249047e-01 -9.73194003e-01 2.96396852e-01 3.35905612e-01
-2.14465857e-01 1.28327832e-01 7.48651922e-02 8.04668665e-01
-5.29489100e-01 -5.68851769e-01 2.53457993e-01 1.07959248e-01
-1.21519458e+00 1.77606449e-01 2.44181693e-01 3.79559875e-01
1.22471964e+00 -6.54770285e-02 -6.50797784e-01 -2.12079152e-01
-9.55939829e-01 2.73717016e-01 3.61697406e-01 5.67033172e-01
6.35471880e-01 -1.73138523e+00 -6.21534884e-01 -1.67885959e-01
8.33062112e-01 -3.75945091e-01 -5.81820488e-01 7.65824914e-01
-6.63032234e-01 4.47692573e-01 -5.11993878e-02 -2.32591033e-01
-9.44292068e-01 8.28449130e-01 4.07761812e-01 -5.06756961e-01
-7.43688285e-01 5.99284410e-01 -3.37195396e-01 -6.60240114e-01
1.86727405e-01 -4.38693106e-01 -4.11121160e-01 2.95894116e-01
2.11333722e-01 2.37981871e-01 1.51978031e-01 -2.71187901e-01
-4.95279521e-01 2.87530631e-01 1.55424088e-01 3.91663104e-01
1.49722910e+00 4.26434755e-01 -1.91576570e-01 3.34350318e-01
1.25683856e+00 2.25899145e-01 -5.08550048e-01 -2.52973527e-01
5.30855298e-01 -3.96232575e-01 -2.22807005e-01 -8.63498211e-01
-1.03346860e+00 6.54378176e-01 4.50589567e-01 7.87362278e-01
3.82894039e-01 1.62782907e-01 3.87733102e-01 7.49163032e-01
4.74692732e-01 -1.01497912e+00 -1.15371317e-01 6.66065633e-01
9.20107007e-01 -1.16181362e+00 6.83444217e-02 -6.71051264e-01
-5.55160284e-01 1.28454125e+00 4.46956933e-01 -7.43899718e-02
9.14353907e-01 1.25956625e-01 -1.92777291e-01 -7.13963270e-01
-8.68878841e-01 -3.93335044e-01 4.57597136e-01 9.16448236e-01
6.33546472e-01 1.94775835e-01 -5.19643843e-01 6.84945285e-01
-1.45922109e-01 -2.19136059e-01 3.93140286e-01 6.65316999e-01
-2.09238663e-01 -1.42557573e+00 2.13509992e-01 6.71996832e-01
-2.30213344e-01 -1.28252298e-01 -5.38355529e-01 1.23505068e+00
3.36135030e-01 5.80440104e-01 -2.19275817e-01 -7.15467095e-01
6.11736655e-01 2.36720160e-01 4.79304165e-01 -1.06499088e+00
-3.45050216e-01 -1.05845845e+00 8.60890210e-01 -5.17259836e-01
-2.53163487e-01 -2.58318007e-01 -1.21880007e+00 -3.16801220e-01
-3.36792290e-01 4.15513277e-01 4.39276218e-01 5.92337430e-01
3.11002314e-01 5.38851619e-01 -3.18707377e-02 -1.36434183e-01
-4.18706656e-01 -6.68508708e-01 -6.90055847e-01 7.77205348e-01
-2.13488340e-01 -1.06265211e+00 -3.09968114e-01 -2.67998457e-01] | [8.853761672973633, 7.742548942565918] |
c6ae109e-28bc-49aa-9965-d43b6644ae9a | tracking-objects-with-3d-representation-from | 2306.05416 | null | https://arxiv.org/abs/2306.05416v1 | https://arxiv.org/pdf/2306.05416v1.pdf | Tracking Objects with 3D Representation from Videos | Data association is a knotty problem for 2D Multiple Object Tracking due to the object occlusion. However, in 3D space, data association is not so hard. Only with a 3D Kalman Filter, the online object tracker can associate the detections from LiDAR. In this paper, we rethink the data association in 2D MOT and utilize the 3D object representation to separate each object in the feature space. Unlike the existing depth-based MOT methods, the 3D object representation can be jointly learned with the object association module. Besides, the object's 3D representation is learned from the video and supervised by the 2D tracking labels without additional manual annotations from LiDAR or pretrained depth estimator. With 3D object representation learning from Pseudo 3D object labels in monocular videos, we propose a new 2D MOT paradigm, called P3DTrack. Extensive experiments show the effectiveness of our method. We achieve new state-of-the-art performance on the large-scale Waymo Open Dataset. | ['Zhaoxiang Zhang', 'Naiyan Wang', 'Zehao Huang', 'Yuntao Chen', 'Yuqi Wang', 'Lue Fan', 'JiaWei He'] | 2023-06-08 | null | null | null | null | ['object-tracking', 'multiple-object-tracking'] | ['computer-vision', 'computer-vision'] | [-2.62790889e-01 -1.03515573e-01 -4.60213155e-01 -2.21682638e-01
-5.43238580e-01 -5.05380332e-01 3.73045444e-01 -1.59267247e-01
-4.33930188e-01 5.81496298e-01 -3.86399895e-01 6.25536144e-02
-9.12627429e-02 -4.32369500e-01 -9.46097076e-01 -6.25149608e-01
-6.76157251e-02 8.96032989e-01 9.49313879e-01 4.28381622e-01
8.37517306e-02 5.68052769e-01 -1.84938097e+00 -2.52529204e-01
4.45052624e-01 1.30463052e+00 4.96242583e-01 7.42748320e-01
-3.24960321e-01 5.00625908e-01 -3.12797904e-01 2.76986241e-01
6.68445408e-01 1.81595594e-01 -1.47375509e-01 3.85424793e-01
1.00657022e+00 -6.92181110e-01 -5.68985701e-01 9.90442514e-01
2.63859987e-01 -7.27031054e-03 6.17198467e-01 -1.72679031e+00
-3.41057718e-01 -8.42706710e-02 -7.91709781e-01 8.34169313e-02
3.26787114e-01 1.64708629e-01 7.64154255e-01 -1.02945924e+00
6.13781512e-01 1.61720419e+00 4.78715867e-01 6.03213251e-01
-1.06261289e+00 -8.87704670e-01 5.30469596e-01 3.41922551e-01
-1.52219164e+00 -3.12608480e-01 6.57732189e-01 -7.20311284e-01
6.54954195e-01 -3.68301198e-02 9.59131241e-01 7.02385068e-01
3.64687331e-02 9.97251630e-01 9.86380100e-01 5.05909659e-02
-4.81701978e-02 1.52145162e-01 1.90509394e-01 8.02489638e-01
6.51844203e-01 2.80486315e-01 -6.74721599e-01 -1.00650750e-01
7.65350878e-01 3.99217069e-01 9.21080858e-02 -1.21477520e+00
-1.20051634e+00 4.75785673e-01 3.52774382e-01 -3.87983680e-01
-8.68174508e-02 5.97589433e-01 5.93590364e-02 1.33271113e-01
3.41834605e-01 -2.40819037e-01 -5.43587446e-01 4.32367586e-02
-5.06704569e-01 3.00402582e-01 4.57645088e-01 1.70715153e+00
1.08456993e+00 -8.85392800e-02 4.72419709e-02 2.77607769e-01
9.82437730e-01 1.05972588e+00 7.76225515e-03 -9.81524706e-01
2.30057269e-01 8.00984502e-01 3.61084789e-01 -4.69048887e-01
-2.08457112e-01 -1.31822973e-01 -2.14313731e-01 5.83458185e-01
3.14475417e-01 1.56532124e-01 -8.88269544e-01 1.47139549e+00
1.07989979e+00 5.21022141e-01 -6.53123064e-03 9.61627126e-01
9.04631913e-01 2.32797116e-01 -2.46333182e-01 -2.13615701e-01
1.23081982e+00 -8.79569113e-01 -7.90300667e-01 -3.06602418e-01
5.59968770e-01 -5.44303894e-01 2.82694906e-01 2.65340626e-01
-7.06397057e-01 -6.92743719e-01 -1.00656140e+00 -1.91465691e-01
-2.46735618e-01 4.80649136e-02 5.61195195e-01 4.42364365e-01
-5.80993176e-01 4.39894237e-02 -1.09172869e+00 -5.59385002e-01
5.90713620e-01 7.57773101e-01 -4.98037457e-01 -2.59785056e-01
-4.86082017e-01 9.18236911e-01 6.15740001e-01 4.20175977e-02
-1.26982296e+00 -3.58996332e-01 -9.08681631e-01 -5.78239977e-01
7.52352953e-01 -6.89797878e-01 1.19922495e+00 3.84971164e-02
-1.16155159e+00 1.02986097e+00 -3.94905835e-01 -2.99668223e-01
5.05741596e-01 -6.15789831e-01 -1.31362513e-01 -1.43067032e-01
2.76171565e-01 9.45600092e-01 9.72730160e-01 -1.44555259e+00
-1.21861684e+00 -6.58370674e-01 7.38380775e-02 2.34293997e-01
1.19147211e-01 -3.85656714e-01 -7.14749396e-01 -1.91365872e-02
5.69035411e-01 -1.26133060e+00 -1.04543835e-01 7.93605626e-01
-1.90983936e-01 -6.16549432e-01 1.57195735e+00 1.54956104e-02
5.87609410e-01 -2.29376936e+00 1.42614365e-01 -2.44994521e-01
4.07940626e-01 1.56826875e-03 5.09402528e-02 -2.00138688e-01
4.95307416e-01 -3.52532238e-01 2.83444196e-01 -6.20583892e-01
2.32314140e-01 6.03915751e-01 -1.96012095e-01 8.32981884e-01
2.17628255e-01 7.78064013e-01 -1.13003981e+00 -8.13550532e-01
4.22415823e-01 3.15362960e-01 -5.81874311e-01 4.05771047e-01
-3.82092386e-01 4.96681929e-01 -6.99528694e-01 9.63102341e-01
9.49122429e-01 -1.27006635e-01 -1.64234519e-01 -2.60235876e-01
-3.47791851e-01 3.61536056e-01 -1.55397463e+00 2.07332873e+00
1.42642155e-01 5.29448032e-01 1.36655584e-01 -6.38468623e-01
1.02463126e+00 2.01690122e-01 8.76189530e-01 -1.84637576e-01
2.30192556e-03 3.38827372e-01 -1.62057877e-01 -5.36770344e-01
5.68385839e-01 1.66448206e-01 -6.05533943e-02 1.71513006e-01
1.95611253e-01 -3.77417922e-01 -6.82342947e-02 9.52149481e-02
1.05408430e+00 5.51926970e-01 2.09577963e-01 3.25742848e-02
3.41563433e-01 1.72062442e-01 9.70685720e-01 7.76564062e-01
-5.25237262e-01 1.90180123e-01 -2.49724358e-01 -2.35544667e-01
-5.89588583e-01 -1.37533593e+00 -3.91969353e-01 5.43432891e-01
8.25840533e-01 -3.14058959e-01 8.38669166e-02 -1.05213356e+00
6.34862959e-01 9.93911084e-03 -3.58465225e-01 -9.86468419e-03
-6.34652793e-01 -1.55245200e-01 6.88418299e-02 4.53301579e-01
3.42569113e-01 -3.46918523e-01 -8.22741628e-01 1.90243244e-01
7.34157711e-02 -1.51568103e+00 -4.85734224e-01 4.17394251e-01
-1.09739089e+00 -1.31622851e+00 -8.22270960e-02 -6.84536099e-01
5.32006681e-01 9.70360816e-01 6.20320797e-01 2.93327160e-02
-2.09785089e-01 7.74818957e-01 -2.74914682e-01 -7.35005081e-01
1.17003873e-01 -1.98708326e-01 7.35370874e-01 -2.03180477e-01
7.34729111e-01 -3.57335359e-01 -4.20563787e-01 4.83626962e-01
-3.80934119e-01 -1.59900188e-01 4.75379080e-01 4.09228951e-01
1.02301979e+00 -9.84700918e-02 1.70396194e-01 -2.95315653e-01
-6.70300663e-01 -3.58481020e-01 -9.55192804e-01 -2.19817147e-01
-2.42283762e-01 -8.30499232e-02 -2.61548877e-01 -7.68707871e-01
-5.48431337e-01 7.86298752e-01 5.74261248e-01 -1.15497637e+00
-2.24621564e-01 -1.83698043e-01 -3.33927602e-01 -3.70883673e-01
2.28491187e-01 -1.33924186e-01 -7.36379204e-03 -8.60856533e-01
3.74541730e-01 6.41766250e-01 4.69170183e-01 -4.82137501e-01
1.44771099e+00 9.41265643e-01 3.78347725e-01 -6.53314292e-01
-1.18416739e+00 -7.66391397e-01 -1.00632095e+00 -4.76630390e-01
1.03688800e+00 -1.56206417e+00 -9.16317523e-01 2.03997090e-01
-1.29063678e+00 1.18698264e-02 -4.96549815e-01 1.05929661e+00
-6.10507250e-01 2.57501215e-01 -1.23368032e-01 -1.11313415e+00
1.96681082e-01 -1.16100669e+00 1.40026724e+00 2.53729910e-01
2.85282731e-01 -5.17159522e-01 6.72393590e-02 2.42042944e-01
-2.52485067e-01 3.98096517e-02 2.41298094e-01 -5.20389080e-01
-1.50239491e+00 -1.76102057e-01 -2.46075764e-01 7.17688305e-03
3.03881794e-01 -1.51821360e-01 -1.01003778e+00 -3.97046864e-01
3.24078761e-02 -2.33451918e-01 8.07834208e-01 2.37483725e-01
9.03137207e-01 2.14391381e-01 -8.46109271e-01 5.65602303e-01
1.13530147e+00 1.16879247e-01 -8.38385075e-02 1.54995129e-01
9.76117790e-01 5.38760483e-01 1.36884725e+00 4.65426773e-01
7.78538525e-01 9.28850591e-01 1.02083135e+00 4.11297262e-01
-3.80481005e-01 -3.60690951e-01 5.22345662e-01 8.01382542e-01
2.66832978e-01 5.86962774e-02 -6.50631905e-01 4.77289826e-01
-2.14273715e+00 -5.56912541e-01 -4.86755252e-01 2.22494316e+00
3.73920321e-01 3.73728216e-01 1.56172037e-01 -2.64053941e-01
8.12823415e-01 -1.26121268e-01 -8.96254838e-01 6.76032066e-01
-4.21766527e-02 -4.82556403e-01 7.90136397e-01 4.67021078e-01
-1.30141890e+00 9.83401477e-01 5.80084181e+00 4.83730942e-01
-6.16224468e-01 2.00504944e-01 -6.05636775e-01 -4.21059400e-01
1.40744418e-01 3.00143689e-01 -1.74425638e+00 3.52122605e-01
3.56297344e-01 -1.03783384e-02 8.00542012e-02 9.61198509e-01
-4.32993136e-02 -1.83521435e-01 -1.53103900e+00 1.30710316e+00
2.21194569e-02 -1.08002675e+00 -2.06753746e-01 5.17455220e-01
4.95841473e-01 3.04030180e-01 -1.84672609e-01 4.27908212e-01
3.44519466e-01 -2.91763753e-01 9.64170575e-01 3.78195822e-01
6.55867517e-01 -1.78319767e-01 2.81062722e-01 5.43285906e-01
-1.83295715e+00 -1.75607398e-01 -5.87684155e-01 -1.06343091e-01
2.97714561e-01 3.87771398e-01 -7.92716146e-01 5.81496656e-01
9.18605626e-01 1.24975586e+00 -4.31389362e-01 1.50325084e+00
-1.55168697e-01 2.03909129e-02 -8.02742600e-01 7.86811709e-02
-4.80286777e-02 -2.96739470e-02 1.02686262e+00 6.24262512e-01
4.59461987e-01 1.49525449e-01 7.82574117e-01 5.77589691e-01
-4.09997534e-03 -4.97204393e-01 -9.51177955e-01 1.50527745e-01
8.01261544e-01 1.19414639e+00 -4.36556190e-01 -3.89494509e-01
-5.46009362e-01 4.63110864e-01 1.48154989e-01 3.46557833e-02
-9.23622966e-01 1.01727940e-01 1.06839168e+00 6.35647103e-02
8.43699157e-01 -6.73521101e-01 2.18500301e-01 -1.22479391e+00
9.21691060e-02 -2.22998202e-01 2.53955245e-01 -8.28851223e-01
-1.23487723e+00 -3.62494625e-02 1.54985860e-01 -1.83929181e+00
1.92462295e-01 -8.31620157e-01 -4.45406921e-02 3.82179737e-01
-1.78196645e+00 -1.09439504e+00 -5.12400866e-01 6.23352945e-01
4.35274839e-01 1.26465991e-01 3.69510472e-01 5.89315116e-01
-2.81055987e-01 8.85692909e-02 -3.86579931e-02 -3.82374786e-02
9.44987416e-01 -1.14489567e+00 9.89587158e-02 5.11237562e-01
2.26060659e-01 1.71657309e-01 5.25382638e-01 -1.01823187e+00
-2.03187299e+00 -1.30242419e+00 4.17160034e-01 -1.03957224e+00
6.76566958e-01 -6.31510079e-01 -6.70845330e-01 1.07129920e+00
-2.88412333e-01 5.05815446e-01 4.00192767e-01 -1.72226846e-01
-3.69131148e-01 -3.18432182e-01 -1.02590895e+00 1.44547552e-01
1.63746965e+00 -3.84115726e-01 -7.57006586e-01 4.85987902e-01
1.15822005e+00 -8.57602656e-01 -8.65840077e-01 6.01153731e-01
5.66385269e-01 -2.87806451e-01 1.16549063e+00 -5.07961392e-01
-5.31547308e-01 -1.11269295e+00 -6.14556789e-01 -6.23239100e-01
-1.16829313e-01 -3.87336433e-01 -9.05992806e-01 1.07322836e+00
-2.09987491e-01 -4.79727417e-01 1.05277014e+00 7.65315220e-02
-4.13084298e-01 -1.65045455e-01 -1.42180562e+00 -1.28271163e+00
-5.69233537e-01 -5.13333499e-01 4.59427625e-01 5.73278785e-01
-5.50873458e-01 4.32152212e-01 -4.88838255e-01 7.93726623e-01
1.32345116e+00 3.93828601e-01 1.47426510e+00 -1.87237465e+00
-1.15147084e-01 1.94914982e-01 -8.30229104e-01 -1.88195145e+00
2.95398265e-01 -9.22684729e-01 4.68421936e-01 -1.28758180e+00
1.88778162e-01 -7.22485900e-01 -3.16317618e-01 3.38770181e-01
-3.06424070e-02 1.02943882e-01 3.96623075e-01 3.75313759e-01
-1.19108474e+00 7.89903820e-01 1.27900290e+00 -3.26808661e-01
-6.79915398e-03 -1.19620979e-01 -1.17541134e-01 6.39084220e-01
2.31777877e-01 -9.34018016e-01 -1.62488297e-01 -7.11591423e-01
-2.54593223e-01 -7.06726089e-02 6.83622360e-01 -1.22103965e+00
3.71310085e-01 -3.40377659e-01 2.74394721e-01 -1.67824256e+00
1.06965196e+00 -1.54680824e+00 1.78367913e-01 5.58370948e-01
2.76977479e-01 -3.06583703e-01 1.42844096e-01 1.07552052e+00
2.79861301e-01 -2.80575473e-02 6.92291141e-01 -2.03250237e-02
-9.98127937e-01 9.17538106e-01 -8.75177756e-02 -2.02652261e-01
1.35057259e+00 -5.12169242e-01 -3.00844967e-01 2.68008828e-01
-5.80289483e-01 8.50402594e-01 6.78677857e-01 6.47446096e-01
6.38940215e-01 -1.70300889e+00 -2.96057910e-01 2.05751851e-01
4.53819543e-01 6.70398772e-01 1.03481017e-01 1.02651739e+00
1.56010494e-01 3.62610817e-01 -1.07137918e-01 -1.54391277e+00
-1.44562232e+00 7.74252892e-01 1.76358968e-01 3.67616534e-01
-7.91087925e-01 7.62022018e-01 4.89923418e-01 -5.18417776e-01
5.81093967e-01 -3.95692766e-01 -2.69094110e-02 -3.37239867e-03
3.98070484e-01 2.22474962e-01 -3.96284610e-01 -6.56958282e-01
-6.76058710e-01 1.02520263e+00 -1.29589394e-01 1.23956032e-01
1.10403764e+00 -5.24344385e-01 1.22499473e-01 9.39690888e-01
7.84825444e-01 -2.32903838e-01 -1.70547068e+00 -6.49887204e-01
-3.24280299e-02 -8.97709608e-01 -1.80923454e-02 -1.08562848e-02
-7.95437515e-01 8.39886725e-01 9.63364780e-01 -5.04977666e-02
4.67149913e-01 3.82254303e-01 5.91698945e-01 7.76131809e-01
7.62822866e-01 -8.95253479e-01 3.30596566e-01 6.53538525e-01
3.81131291e-01 -1.59075010e+00 1.93617076e-01 -4.29680079e-01
-7.12671578e-02 7.71374106e-01 1.14528489e+00 -1.72545612e-01
8.13182890e-01 1.88239723e-01 4.32027616e-02 -3.51983458e-01
-8.31603646e-01 -5.91527045e-01 7.83504620e-02 8.49013627e-01
-3.00284833e-01 -1.34896159e-01 2.70009041e-01 1.19968645e-01
2.88473666e-01 -1.46921948e-01 2.80763566e-01 1.19653070e+00
-9.00735199e-01 -1.13840878e+00 -5.86533487e-01 3.17999303e-01
2.06885606e-01 5.44715345e-01 -1.67395994e-01 9.71611023e-01
4.48256701e-01 8.63285542e-01 2.12120518e-01 -4.95212823e-01
3.98392528e-01 -2.89530400e-02 8.05275321e-01 -1.02347934e+00
7.99307674e-02 2.44761452e-01 -2.05937505e-01 -5.95734239e-01
-7.42203951e-01 -9.27997231e-01 -1.46780396e+00 -4.49450426e-02
-7.88373709e-01 6.20876513e-02 8.82384598e-01 9.62207019e-01
3.25011611e-01 3.47338349e-01 4.94124502e-01 -1.17070997e+00
-3.32737088e-01 -7.42403924e-01 -5.71628511e-01 -5.86179085e-02
8.31074238e-01 -1.50915670e+00 -3.11064929e-01 -1.08749650e-01] | [6.61021614074707, -2.2677292823791504] |
d602205b-ec11-4d66-ad1b-148a3d58fcc3 | bayesflow-can-reliably-detect-model | 2112.08866 | null | https://arxiv.org/abs/2112.08866v5 | https://arxiv.org/pdf/2112.08866v5.pdf | Detecting Model Misspecification in Amortized Bayesian Inference with Neural Networks | Neural density estimators have proven remarkably powerful in performing efficient simulation-based Bayesian inference in various research domains. In particular, the BayesFlow framework uses a two-step approach to enable amortized parameter estimation in settings where the likelihood function is implicitly defined by a simulation program. But how faithful is such inference when simulations are poor representations of reality? In this paper, we conceptualize the types of model misspecification arising in simulation-based inference and systematically investigate the performance of the BayesFlow framework under these misspecifications. We propose an augmented optimization objective which imposes a probabilistic structure on the latent data space and utilize maximum mean discrepancy (MMD) to detect potentially catastrophic misspecifications during inference undermining the validity of the obtained results. We verify our detection criterion on a number of artificial and realistic misspecifications, ranging from toy conjugate models to complex models of decision making and disease outbreak dynamics applied to real data. Further, we show that posterior inference errors increase as a function of the distance between the true data-generating distribution and the typical set of simulations in the latent summary space. Thus, we demonstrate the dual utility of MMD as a method for detecting model misspecification and as a proxy for verifying the faithfulness of amortized Bayesian inference. | ['Stefan T. Radev', 'Ullrich Köthe', 'Paul-Christian Bürkner', 'Marvin Schmitt'] | 2021-12-16 | null | null | null | null | ['probabilistic-deep-learning'] | ['computer-vision'] | [ 1.6098145e-01 -1.1718173e-01 -5.4377887e-02 -2.7205935e-01
-5.8303040e-01 -4.4529685e-01 8.1899476e-01 -1.1168682e-01
-5.1760685e-01 1.0894004e+00 -1.7437932e-01 -7.1693623e-01
-3.9240524e-01 -8.4558678e-01 -9.2076516e-01 -8.1554472e-01
-3.0900878e-01 7.4803412e-01 -8.2089260e-02 4.2635494e-01
1.9438265e-01 6.1064583e-01 -1.3426113e+00 -3.1617728e-01
1.0788684e+00 3.8734528e-01 8.4509887e-02 7.8296328e-01
4.3871045e-01 1.4856224e-01 -9.1813684e-01 -4.5031118e-01
-4.9772765e-02 -2.5713325e-01 -2.0865458e-01 -4.8437916e-02
-1.4804143e-01 -6.1780053e-01 2.3921331e-02 1.3202739e+00
4.3847758e-01 -2.6763076e-01 1.1322246e+00 -1.4695711e+00
1.4555452e-02 6.2508327e-01 -4.9362749e-01 3.4984362e-01
2.5756600e-01 5.2358210e-01 4.2949504e-01 -6.0531420e-01
3.1935936e-01 1.5441279e+00 8.1348741e-01 2.4915887e-03
-1.5096371e+00 -7.0309186e-01 -1.1061772e-01 -2.9586440e-01
-1.6005342e+00 -3.8057739e-01 2.1219221e-01 -6.4969152e-01
7.5448728e-01 6.6095494e-02 4.3770117e-01 1.5522943e+00
5.6722540e-01 2.4196565e-01 1.1508942e+00 -3.1029478e-01
7.2996151e-01 2.1058483e-01 2.0721745e-01 4.6607912e-01
1.0739446e+00 7.0376086e-01 -2.0612535e-01 -9.5233464e-01
1.0123301e+00 -3.2998383e-02 -2.8341806e-01 -3.3189315e-02
-1.0611036e+00 1.0240614e+00 -4.6544555e-01 -1.4732496e-01
-4.8628426e-01 3.0952874e-01 1.8083198e-01 6.8126343e-02
5.8915162e-01 2.7554941e-01 -3.2002333e-01 -6.0640357e-02
-1.0369563e+00 6.5225095e-01 8.9308631e-01 5.8880776e-01
2.6514181e-01 2.7898622e-01 -2.9949296e-01 3.1660792e-01
4.8635614e-01 1.1148232e+00 1.5455632e-01 -9.7665125e-01
2.0487423e-01 6.4922860e-03 6.8687940e-01 -8.9305127e-01
-2.5801784e-01 -5.5842578e-01 -8.7166262e-01 3.4791194e-03
6.9445866e-01 -4.0060624e-01 -7.5306386e-01 2.1401038e+00
5.1196790e-01 4.7634640e-01 -9.5053330e-02 5.5180871e-01
-1.7725340e-01 7.2604442e-01 1.3017648e-01 -7.1455067e-01
1.1930747e+00 -2.6769083e-02 -8.3757710e-01 1.6204095e-01
5.7646745e-01 -4.8508132e-01 1.0248936e+00 3.7068737e-01
-1.1179140e+00 -1.6317496e-01 -8.5743088e-01 9.7578609e-01
-3.3826120e-02 -1.8669796e-01 4.5457181e-01 8.9403939e-01
-8.0639625e-01 5.4172343e-01 -1.1726199e+00 -2.3182888e-01
1.3444582e-01 -2.9225124e-02 5.2302118e-02 7.8559041e-02
-1.2058026e+00 8.2218593e-01 3.9249438e-01 1.6460976e-01
-1.2742091e+00 -9.1612601e-01 -5.9303594e-01 2.7961609e-01
4.1605633e-01 -8.6146766e-01 1.1808629e+00 -5.0464755e-01
-1.1356430e+00 5.5521464e-01 -1.9949298e-01 -6.7620569e-01
8.7669086e-01 4.4314437e-02 -3.9416307e-01 1.9373037e-02
8.2410008e-02 -5.5722032e-02 1.0400951e+00 -1.0490600e+00
-1.8218522e-01 -3.5628021e-01 -1.4583123e-01 -3.3092713e-01
3.8739836e-01 6.7309111e-02 1.8619873e-01 -7.5547212e-01
-4.3899812e-02 -6.9851696e-01 -2.9481357e-01 -2.1052685e-01
-5.0774997e-01 3.4855697e-02 3.1871873e-01 -5.6623530e-01
1.3075542e+00 -1.7542896e+00 -2.8702837e-01 4.0263444e-01
-1.3271119e-01 9.5337722e-03 1.1231131e-01 5.8416069e-01
9.9781804e-02 1.1882738e-01 -7.9231620e-01 2.8563373e-02
2.9907858e-01 1.6043654e-01 -5.8522224e-01 8.2110929e-01
2.9229635e-01 5.6894177e-01 -7.8525174e-01 -3.4735698e-01
-2.0692360e-02 3.7029183e-01 -5.2412164e-01 4.1277897e-01
-2.4360052e-01 2.1466003e-01 -1.9447440e-01 3.9446384e-01
9.5529759e-01 -4.0179181e-01 3.8797623e-01 1.2603286e-01
-5.4364666e-02 2.3809102e-01 -1.2817886e+00 8.6699897e-01
-1.7892776e-01 2.0395979e-01 5.7330560e-02 -1.1748731e+00
4.1484833e-01 2.5443447e-01 -1.4000650e-02 -7.2791114e-02
8.8005036e-02 2.0515576e-02 9.9768847e-02 -3.8560525e-01
1.1576508e-01 -3.6350971e-01 -1.7796783e-01 8.8122541e-01
-1.6259361e-02 -1.3668005e-01 2.1898609e-02 7.0162997e-02
1.0223151e+00 -2.9855154e-02 5.3518265e-01 -5.2522296e-01
-1.6689138e-01 -1.3301566e-01 5.4783022e-01 1.4776222e+00
-8.0695987e-02 1.9789401e-01 8.8476896e-01 4.0038709e-02
-1.2198477e+00 -1.6228316e+00 -6.9099224e-01 4.9148643e-01
-2.7482468e-01 5.4718800e-02 -8.3814096e-01 -4.4772702e-01
1.3229229e-01 1.1147280e+00 -5.6328231e-01 -3.7625778e-01
-1.4596972e-01 -1.6361700e+00 7.9219061e-01 2.7653205e-01
1.5880297e-01 -5.5799794e-01 -9.7005737e-01 1.1446574e-01
4.2149585e-02 -9.0522337e-01 -5.3624272e-02 7.7415168e-02
-9.1859889e-01 -1.1663104e+00 -5.9237134e-01 1.9442199e-01
6.8166852e-01 -1.8798056e-01 1.1614375e+00 -3.3098823e-01
-1.5196690e-01 3.5657096e-01 1.8624711e-01 -2.0473623e-01
-8.7910038e-01 -5.3513640e-01 4.9361667e-01 -2.6828784e-01
2.1198218e-01 -5.1749545e-01 -3.5506931e-01 3.3535719e-01
-1.0335889e+00 -1.4626552e-01 3.4004143e-01 8.0978251e-01
1.3314703e-01 6.1303280e-02 6.3680816e-01 -7.7972174e-01
7.9616880e-01 -9.1942132e-01 -1.3267252e+00 4.9751443e-01
-5.1628119e-01 2.5880978e-01 2.6369083e-01 -6.6046423e-01
-1.1330909e+00 -5.1503468e-01 2.2785941e-01 -3.1357536e-01
-2.2274643e-01 7.9001856e-01 -6.5702759e-02 6.0158968e-01
3.7774298e-01 3.4827981e-02 7.6477461e-02 -3.4458396e-01
-1.2763044e-01 5.3314143e-01 4.3355694e-01 -8.2651603e-01
5.1948857e-01 4.5218816e-01 2.0515156e-01 -7.8122050e-01
-7.1317267e-01 5.0582763e-02 -1.0551669e-01 -1.7480242e-01
3.8463044e-01 -7.9301190e-01 -1.0122807e+00 6.0678774e-01
-1.2144802e+00 -3.7096256e-01 -1.7372765e-01 8.7579161e-01
-6.4290607e-01 3.3132428e-01 -3.0713755e-01 -1.3915751e+00
9.5717542e-02 -1.2095661e+00 9.5312136e-01 -8.4342904e-02
-1.9182050e-01 -1.1760340e+00 4.6024507e-01 -3.3642399e-01
2.4428298e-01 2.8534493e-01 1.0765152e+00 -5.7198524e-01
-2.2927111e-01 -1.8430711e-01 -3.0755094e-01 -4.1937515e-02
-3.2422505e-03 4.4413435e-01 -1.0767635e+00 -4.5406151e-01
3.5992181e-01 -1.4592255e-02 7.1330762e-01 9.4117242e-01
9.7499406e-01 -6.7314762e-01 -3.0475876e-01 5.1146936e-01
1.2819588e+00 9.6722767e-02 3.1512135e-01 -1.7727214e-01
-4.2202167e-02 6.8177581e-01 4.1261840e-01 9.2250955e-01
4.3806389e-02 5.5689949e-01 2.4896319e-01 3.2194597e-01
5.4200530e-01 -5.0359392e-01 3.5823846e-01 6.3532937e-01
3.6169785e-01 -4.1489846e-01 -1.0354395e+00 3.9574289e-01
-1.7020097e+00 -1.1428611e+00 8.3197489e-02 2.8423693e+00
9.8098791e-01 2.7310210e-01 3.1452170e-01 -1.4792836e-01
1.1684319e+00 -1.6042711e-01 -7.3886001e-01 -1.8946174e-01
-9.6156098e-02 -5.5442564e-02 5.3327364e-01 6.1879778e-01
-6.5450573e-01 3.5388637e-01 7.7203302e+00 8.6968404e-01
-7.0361996e-01 2.1773426e-01 6.9926643e-01 1.0825054e-01
-3.0559707e-01 2.6049532e-02 -6.8451595e-01 8.8077217e-01
1.4937918e+00 -4.1774505e-01 1.7145911e-01 4.9054912e-01
6.1850131e-01 -6.6393256e-01 -1.1495352e+00 5.8366191e-01
-3.8478786e-01 -1.1279625e+00 1.6128965e-02 2.4133657e-01
6.8228966e-01 -2.8908211e-01 2.6478356e-01 1.5775594e-01
7.3810971e-01 -1.0650064e+00 5.7354510e-01 8.5524535e-01
7.1615982e-01 -6.8563449e-01 8.1555116e-01 7.2081751e-01
-4.1299674e-01 2.8029016e-01 -3.2884243e-01 -2.5755525e-01
2.4007884e-01 1.0005119e+00 -1.1450014e+00 5.8325149e-02
2.9578021e-01 6.7305714e-03 -7.6128721e-02 1.0892889e+00
-8.3096117e-02 1.1036782e+00 -8.6128461e-01 5.8196146e-02
7.1132243e-02 -2.3633282e-01 7.3394936e-01 1.2267290e+00
4.5844540e-01 -2.2217092e-01 -2.7175158e-01 1.4669130e+00
3.5595575e-01 -6.4196634e-01 -6.9244868e-01 -1.9970571e-01
9.3445706e-01 5.2341825e-01 -7.6883924e-01 -3.4354889e-01
7.7443808e-02 4.1049078e-01 -2.2722255e-02 7.7034712e-01
-9.8940539e-01 -9.3677476e-02 5.7474750e-01 -4.4698507e-02
2.3976529e-01 -1.1487011e-01 -1.2900454e-01 -1.3176413e+00
-1.3480005e-01 -7.1371388e-01 3.4589618e-01 -5.4631341e-01
-1.3199146e+00 -1.1949037e-01 7.8117764e-01 -7.3457044e-01
-8.2091475e-01 -3.8223261e-01 -7.2690117e-01 9.9806857e-01
-1.0572450e+00 -4.2527163e-01 2.1944952e-01 1.5475583e-01
1.7883658e-01 1.4637753e-01 6.2506729e-01 -1.3673082e-01
-7.2635466e-01 5.2717561e-01 4.3818668e-01 -3.3700097e-01
3.0029470e-01 -9.7985709e-01 4.0754595e-01 9.9072945e-01
-3.4092495e-01 7.4862361e-01 1.2855377e+00 -1.0378913e+00
-1.0954552e+00 -9.6701407e-01 4.5504123e-01 -3.5158095e-01
7.7582014e-01 -4.2457199e-01 -9.4849616e-01 6.3130522e-01
-4.0335926e-01 -3.3204839e-01 2.3041601e-01 2.9607946e-03
-3.8248402e-01 2.7563030e-01 -1.4569123e+00 5.8836377e-01
6.5446544e-01 -4.0311778e-01 -6.1902261e-01 3.4703937e-01
5.6934178e-01 2.1095356e-01 -7.1040499e-01 5.7046467e-01
5.7575792e-01 -9.6086836e-01 8.6546922e-01 -8.3128953e-01
2.1843781e-01 -9.2094064e-02 -3.0207181e-01 -1.0525674e+00
1.8976739e-01 -7.1504891e-01 -3.6829200e-01 9.9922246e-01
1.5761174e-01 -8.8658023e-01 3.5756260e-01 3.7754267e-01
5.1459718e-01 -3.2986048e-01 -1.1596265e+00 -8.9890242e-01
2.2089878e-01 -7.6392698e-01 6.3048327e-01 7.7684927e-01
-3.3470157e-01 -1.6629300e-01 -3.3515918e-01 6.8453610e-01
1.2650037e+00 -1.1207204e-01 5.7443720e-01 -1.1213131e+00
-6.0015285e-01 -2.3991366e-01 -1.7488937e-01 -6.6251165e-01
2.2704266e-01 -2.1430059e-01 -1.2563046e-02 -7.2694659e-01
6.2057114e-01 -2.1543632e-01 3.9062276e-02 -2.1088286e-01
-2.7634495e-01 -2.9299077e-01 -3.0012259e-01 1.5533778e-01
-2.3385125e-01 5.0461829e-01 7.2245467e-01 2.8636158e-01
1.2089467e-01 3.8406971e-01 -1.3808534e-01 9.9748820e-01
5.4132891e-01 -8.2263279e-01 -3.8656533e-01 -9.8947007e-03
5.1311076e-01 6.7806327e-01 9.9013972e-01 -5.2271926e-01
1.9197636e-03 -5.1006997e-01 1.4965630e-01 -8.2071555e-01
1.2974784e-01 -5.2222884e-01 5.0694299e-01 8.0328691e-01
-2.9875982e-01 1.4364186e-01 2.1887094e-01 8.1270027e-01
2.7934611e-01 -4.1469610e-01 7.9681259e-01 -4.2543687e-02
1.7827278e-01 -8.1277207e-02 -7.9186231e-01 2.7373773e-01
6.6113895e-01 8.4515870e-02 -3.9068729e-01 -4.6572435e-01
-5.8999115e-01 3.7720002e-02 4.7191095e-01 -3.8658515e-01
5.0841320e-01 -7.9982030e-01 -8.8036561e-01 2.3608279e-01
-1.8955067e-01 -4.4624320e-01 2.4225818e-01 1.0092189e+00
-3.8014176e-01 3.8104445e-01 1.4341490e-01 -9.0378678e-01
-6.2349105e-01 4.7171015e-01 5.1190639e-01 -3.2560700e-01
-2.0595178e-01 3.3299518e-01 5.1778603e-01 -4.0963688e-01
2.3346657e-01 -4.8612237e-01 5.1061392e-01 -1.9686961e-01
5.9861189e-01 5.1395291e-01 -2.2286469e-01 -1.2134151e-01
-2.2846892e-01 4.2880856e-02 7.9405844e-02 -5.8559096e-01
9.4748253e-01 -2.2746250e-01 1.1293356e-01 8.1552505e-01
7.8656226e-01 -3.3990780e-01 -1.6107161e+00 3.2975774e-02
8.5611627e-02 -3.4079042e-01 9.8314844e-03 -9.2901379e-01
-3.0701438e-01 8.2834888e-01 4.8418725e-01 3.5059947e-01
6.3610798e-01 -1.5111211e-01 1.1164453e-02 3.4846243e-01
3.9393365e-01 -7.3929918e-01 -4.0285435e-01 1.2999624e-01
6.7427838e-01 -1.0689999e+00 1.3137892e-01 2.1994246e-02
-1.3703635e-01 6.2054324e-01 1.1194196e-01 -9.6032999e-02
7.6746660e-01 5.1755154e-01 -5.3219086e-01 -1.1383785e-01
-1.0081739e+00 1.5802395e-01 -3.0440986e-02 6.7308706e-01
-8.3422795e-02 2.5726408e-01 -2.2125116e-01 7.4413472e-01
-7.8065917e-02 3.6236972e-02 5.9668261e-01 6.4096761e-01
-4.2390811e-01 -2.5736910e-01 -6.3550395e-01 4.1636318e-01
-5.1856422e-01 -1.4855333e-01 1.4640492e-01 9.5990384e-01
-3.8979614e-01 8.5794246e-01 4.3771607e-01 2.6571792e-01
-1.4144975e-01 -1.8357877e-02 4.5586082e-01 -3.0417344e-01
-1.3375337e-01 1.3337338e-01 3.9876185e-02 -4.2134148e-01
-2.5795299e-01 -7.8763783e-01 -5.5486679e-01 -6.6496235e-01
-4.3364477e-01 2.5158221e-01 6.3947797e-01 1.2080988e+00
1.4488496e-01 1.8873829e-01 4.9252892e-01 -6.6241521e-01
-1.3728030e+00 -1.0649540e+00 -6.6195512e-01 -4.2607278e-02
4.0046829e-01 -8.1877381e-01 -1.0150579e+00 -2.7929336e-01] | [6.807300567626953, 3.9544498920440674] |
3e6c4b61-77c1-4da1-9008-47d3e5d11bde | bridging-the-gap-between-f-gans-and-1 | null | null | https://proceedings.icml.cc/static/paper_files/icml/2020/5361-Paper.pdf | https://proceedings.icml.cc/static/paper_files/icml/2020/5361-Paper.pdf | Bridging the Gap Between f-GANs and Wasserstein GANs | Generative adversarial networks (GANs) variants approximately minimize divergences between the model and the data distribution using a discriminator. Wasserstein GANs (WGANs) enjoy superior empirical performance, however, unlike in f-GANs, the discriminator does not provide an estimate for the ratio between model and data densities, which is useful in applications such as inverse reinforcement learning. To overcome this limitation, we propose an new training objective where we additionally optimize over a set of importance weights over the generated samples. By suitably constraining the feasible set of importance weights, we obtain a family of objectives which includes and generalizes the original f-GAN and WGAN objectives. We show that a natural extension outperforms WGANs while providing density ratios as in f-GAN, and demonstrate empirical success on distribution modeling, density ratio estimation and image generation, where we achieve state-of-the-art FID scores on CIFAR10 generation. | ['Jiaming Song', 'Stefano Ermon'] | null | null | https://proceedings.icml.cc/static/paper_files/icml/2020/5361-Paper.pdf | https://proceedings.icml.cc/static/paper_files/icml/2020/5361-Paper.pdf | icml-2020-1 | ['density-ratio-estimation'] | ['methodology'] | [ 2.97501624e-01 2.34686658e-01 -1.36012614e-01 -2.64731526e-01
-1.04633844e+00 -6.00641072e-01 8.83428216e-01 -3.98010671e-01
-2.83293486e-01 1.20704651e+00 1.42197251e-01 -1.59472108e-01
-8.27150196e-02 -9.61662352e-01 -8.90361607e-01 -8.55388463e-01
1.26751378e-01 8.33860338e-01 -4.26730961e-01 -3.28493677e-02
-1.18532896e-01 2.97592342e-01 -1.10991287e+00 -2.22655803e-01
1.21362305e+00 9.88643110e-01 -4.89345612e-03 8.48083317e-01
6.85159788e-02 7.34250486e-01 -1.03874815e+00 -5.62218726e-01
2.83064216e-01 -9.39254761e-01 -4.64873850e-01 -4.44712937e-02
5.02963722e-01 -6.01486683e-01 -3.63553554e-01 1.15956056e+00
4.24801350e-01 2.26460591e-01 1.37047720e+00 -1.44816446e+00
-1.13313043e+00 4.55693692e-01 -5.64753950e-01 -5.67725673e-02
-1.12167768e-01 4.90885191e-02 8.90746236e-01 -5.42610466e-01
2.42266566e-01 1.18805850e+00 6.84816301e-01 9.76984859e-01
-1.49717855e+00 -6.15708470e-01 2.81286947e-02 -3.58231902e-01
-1.25961733e+00 -2.91883528e-01 4.80751067e-01 -3.18475127e-01
4.80513632e-01 -3.34299705e-03 4.51978087e-01 1.47557628e+00
9.11526680e-02 8.65761817e-01 9.30784523e-01 -4.26083297e-01
5.93222797e-01 1.38428941e-01 -4.91469771e-01 3.44014794e-01
1.12930097e-01 2.93122977e-02 -2.45308906e-01 -1.69025749e-01
1.24631560e+00 -1.97477043e-01 -4.13607091e-01 -3.61383289e-01
-9.40385044e-01 1.17673779e+00 4.51362491e-01 -1.46229928e-02
-2.95646042e-01 6.45461857e-01 -1.21378221e-01 2.27148131e-01
7.40002394e-01 4.54247445e-01 -1.10518388e-01 -2.88132846e-01
-8.82606566e-01 4.89672422e-01 5.90735018e-01 1.13909388e+00
6.13226771e-01 6.52044058e-01 -4.89213109e-01 8.38869631e-01
2.54018098e-01 8.81906271e-01 4.80804443e-01 -1.15039682e+00
4.99405146e-01 -6.34292662e-02 3.05282950e-01 -2.90695339e-01
2.37775490e-01 -5.45263767e-01 -1.00910211e+00 3.06335628e-01
7.33664691e-01 -4.56875980e-01 -1.31331241e+00 2.14293838e+00
9.44505185e-02 2.67760336e-01 6.30382374e-02 6.75655007e-01
2.56085515e-01 6.41077042e-01 2.05940574e-01 5.43735325e-02
8.49949062e-01 -7.86975801e-01 -6.96247756e-01 -2.40140617e-01
1.94187284e-01 -5.61680377e-01 1.19344747e+00 2.43036479e-01
-1.35317826e+00 -5.29601455e-01 -9.28075314e-01 2.35765338e-01
-8.53327364e-02 8.11960995e-02 5.53800523e-01 9.24833715e-01
-1.21474743e+00 7.85029709e-01 -8.07041287e-01 1.11658819e-01
7.41367936e-01 3.62112105e-01 -1.10975593e-01 7.69745559e-02
-1.04824924e+00 7.23058522e-01 9.76508036e-02 -2.01912642e-01
-1.33190620e+00 -1.02313793e+00 -8.43152225e-01 1.06605224e-01
-8.09914693e-02 -8.29496801e-01 1.33021224e+00 -1.08008945e+00
-1.78659308e+00 4.09670413e-01 3.40702236e-02 -8.07978272e-01
7.88684607e-01 -2.11161941e-01 -2.38024712e-01 2.92720087e-02
-1.14071667e-01 1.06079996e+00 1.24144566e+00 -1.34039748e+00
-3.98824424e-01 1.76389664e-02 4.14601527e-02 8.14126283e-02
-3.57577324e-01 -5.63023329e-01 2.41760514e-03 -1.02585161e+00
-5.30012548e-01 -7.51758277e-01 -1.67558655e-01 4.95320857e-02
-4.05572236e-01 5.72508834e-02 8.20660651e-01 -5.27512193e-01
8.20113003e-01 -2.01435733e+00 1.42924726e-01 1.11580752e-01
3.54031980e-01 1.75357088e-01 -3.96754742e-01 9.96413603e-02
4.47254777e-02 2.21534654e-01 -5.89854658e-01 -6.67419612e-01
2.99006283e-01 3.64705205e-01 -5.89157403e-01 3.30685049e-01
4.54136193e-01 1.00651562e+00 -9.25154269e-01 -1.41899377e-01
1.19735740e-01 8.03776622e-01 -7.72044480e-01 5.47126949e-01
-5.16575515e-01 5.38513422e-01 -2.43437603e-01 2.43576512e-01
8.01268280e-01 -2.18578890e-01 -2.96102315e-02 1.65509269e-01
6.29157841e-01 3.46444063e-02 -6.89856887e-01 1.50066483e+00
-6.33005619e-01 5.18636823e-01 -1.84261665e-01 -9.99203265e-01
9.83937621e-01 1.73264891e-01 3.23289275e-01 -3.51822108e-01
-2.13525537e-02 -7.49478862e-03 4.48324047e-02 2.29687020e-01
4.78434503e-01 -5.09066582e-01 1.13725752e-01 4.81695682e-01
4.96027440e-01 -4.20087546e-01 5.32515161e-02 1.91935286e-01
7.72880197e-01 3.29520732e-01 5.29982932e-02 -2.69822568e-01
1.15808964e-01 -4.67472017e-01 3.33699971e-01 9.26753581e-01
8.70536789e-02 1.02122223e+00 8.64462137e-01 -1.43615743e-02
-1.27338171e+00 -1.71928501e+00 -1.05458528e-01 6.30636394e-01
-2.27560639e-01 -7.05437586e-02 -1.22313964e+00 -1.04391086e+00
5.75534329e-02 1.07300234e+00 -1.00364900e+00 -3.00742537e-01
-2.72876590e-01 -9.89577770e-01 6.22754276e-01 8.53493273e-01
4.41744506e-01 -7.79041290e-01 -9.99327451e-02 1.17544971e-01
1.22238426e-02 -9.70218599e-01 -8.67969036e-01 1.90795511e-01
-7.14647591e-01 -6.93313777e-01 -1.18479693e+00 -3.50333154e-01
7.99180210e-01 -4.10522133e-01 1.45550573e+00 -4.96017337e-01
7.00161234e-02 4.84096169e-01 -1.48640245e-01 -4.99400586e-01
-7.43949354e-01 1.35547951e-01 -4.62854765e-02 -8.21239948e-02
4.99346890e-02 -7.73818016e-01 -6.56674087e-01 2.04828963e-01
-9.30923045e-01 -1.88558608e-01 4.61375922e-01 1.10173380e+00
6.62678659e-01 -1.18606649e-02 1.02284539e+00 -1.04146814e+00
7.54317522e-01 -5.22269666e-01 -7.29390204e-01 1.56713352e-01
-5.93255043e-01 2.81639397e-01 9.65670228e-01 -6.28624201e-01
-1.26603496e+00 -4.80750591e-01 -2.85358489e-01 -7.28379846e-01
-8.90634432e-02 3.23698260e-02 -1.87343910e-01 1.75364330e-01
6.11139119e-01 1.42541617e-01 1.37682706e-01 -2.42930248e-01
5.23468673e-01 3.96362454e-01 7.06896305e-01 -7.65802979e-01
8.97307038e-01 1.92069516e-01 -6.24718191e-03 -5.68569958e-01
-9.18470442e-01 2.65720159e-01 -2.94897228e-01 6.73029050e-02
8.15689683e-01 -8.65468264e-01 -4.04766381e-01 4.39942867e-01
-8.70970070e-01 -8.11799347e-01 -1.08430517e+00 4.59015876e-01
-1.03271604e+00 -1.26503725e-02 -4.85688061e-01 -9.99202967e-01
-2.82519937e-01 -9.17400241e-01 1.06619978e+00 2.06661880e-01
-1.72953457e-01 -1.54071498e+00 1.13121197e-01 3.03232786e-03
7.14235365e-01 4.63116974e-01 1.07207382e+00 -4.26741898e-01
-3.51956040e-01 5.50262444e-02 -7.39835054e-02 7.92125940e-01
2.92298764e-01 -8.09135288e-02 -1.05866802e+00 -4.66338962e-01
-3.75442505e-02 -5.11527658e-01 8.95323396e-01 8.79367828e-01
1.32438934e+00 -4.27393675e-01 5.43961078e-02 8.54615092e-01
1.37252367e+00 2.38709718e-01 9.61238265e-01 -1.70413375e-01
6.35726154e-01 1.04377970e-01 1.93164587e-01 5.27011395e-01
7.39300773e-02 4.55819607e-01 4.38047290e-01 -1.89546391e-01
-2.43854865e-01 -6.95779800e-01 3.92718971e-01 3.34815830e-01
-9.85667780e-02 -7.23341763e-01 -4.48383301e-01 3.83742630e-01
-1.54305816e+00 -8.60504806e-01 4.73409563e-01 2.39289403e+00
9.43011701e-01 1.49649549e-02 3.43672723e-01 -1.02085076e-01
5.91988742e-01 1.93111971e-01 -8.95751297e-01 -1.01607054e-01
-1.20808855e-01 7.25476205e-01 5.30339897e-01 6.30556464e-01
-9.58281398e-01 6.84623837e-01 7.62373495e+00 1.10116887e+00
-7.48472869e-01 1.19364001e-01 1.12876546e+00 -1.52006775e-01
-8.26842546e-01 -3.53843123e-01 -6.09151244e-01 6.31714940e-01
9.95584309e-01 -2.39922017e-01 7.22114623e-01 8.53488207e-01
-2.56374300e-01 1.40473813e-01 -1.29076684e+00 8.12892795e-01
-5.40310144e-02 -1.19912100e+00 2.59452373e-01 4.90131021e-01
1.14621115e+00 -1.96907427e-02 6.34000957e-01 4.10012871e-01
8.30841541e-01 -1.50720966e+00 7.42580473e-01 5.55024803e-01
1.17699230e+00 -1.10841835e+00 6.65782392e-01 1.00906432e-01
-7.31949270e-01 2.13136598e-01 -5.35404027e-01 2.88009852e-01
1.87517703e-01 6.88200593e-01 -6.98946476e-01 2.80374944e-01
2.07957461e-01 5.63143671e-01 -1.74467117e-01 6.33096278e-01
-2.39435732e-01 6.95661306e-01 -1.63682416e-01 2.12452605e-01
1.47168934e-01 -4.42924619e-01 4.10787463e-01 8.64637554e-01
5.72689950e-01 -3.84400845e-01 -2.34035254e-01 1.40685248e+00
-6.14170790e-01 -4.01685268e-01 -4.27508503e-01 -2.56822467e-01
4.43522245e-01 9.85647500e-01 -4.29389745e-01 -2.15794727e-01
-5.98392189e-02 1.02897060e+00 4.06138211e-01 7.29728162e-01
-1.16743314e+00 -2.84599185e-01 9.19352889e-01 1.06132917e-01
4.44927275e-01 -6.44444674e-02 -6.20877109e-02 -1.16217196e+00
-2.15711847e-01 -8.85179996e-01 3.52729894e-02 -5.85332274e-01
-1.67796350e+00 5.90765655e-01 7.85693973e-02 -1.11249149e+00
-7.58616567e-01 -5.17847776e-01 -7.43369222e-01 1.21373880e+00
-1.43816519e+00 -9.59954977e-01 -1.89022765e-01 4.72897112e-01
3.58094394e-01 -3.84902537e-01 9.32926178e-01 1.25589877e-01
-2.74901271e-01 8.85916293e-01 4.13427711e-01 4.18018783e-03
5.88007867e-01 -1.83716559e+00 7.40914881e-01 6.62069499e-01
2.28027403e-01 3.39598805e-01 6.69779956e-01 -4.88885671e-01
-8.74930978e-01 -1.26790893e+00 -7.38126831e-03 -4.03919190e-01
5.11828721e-01 -4.14151043e-01 -5.60510457e-01 8.98164153e-01
3.13917816e-01 1.40545055e-01 6.59067154e-01 -1.71694979e-01
-2.14217290e-01 8.50604996e-02 -1.48480105e+00 4.24655616e-01
8.89840543e-01 -3.81852210e-01 -1.47741809e-02 2.79132932e-01
5.08829236e-01 -4.70777303e-01 -8.98123026e-01 2.05987230e-01
5.03366828e-01 -9.56576586e-01 9.73103166e-01 -6.15695953e-01
6.89476788e-01 -3.84947434e-02 -2.50548124e-01 -1.87790167e+00
-1.35724217e-01 -7.08958268e-01 -4.64798003e-01 1.41576827e+00
2.39911303e-01 -7.68957794e-01 9.85034943e-01 4.32450324e-01
8.92994702e-02 -7.73161948e-01 -7.59982347e-01 -9.44499135e-01
5.18866897e-01 -1.86779171e-01 7.93530047e-01 5.87688744e-01
-6.62748277e-01 1.09687924e-01 -5.49804926e-01 -2.70401359e-01
1.02359474e+00 -3.71490628e-01 7.33552337e-01 -8.72974932e-01
-6.96707726e-01 -4.85323578e-01 -1.55993357e-01 -1.37808895e+00
3.06316227e-01 -6.30642831e-01 5.78262173e-02 -1.31677699e+00
-4.59212437e-02 -5.82075775e-01 -1.68788061e-01 1.64312482e-01
-2.08593115e-01 4.56314087e-01 1.07995957e-01 2.42411136e-03
-1.58780396e-01 1.06287515e+00 1.44042873e+00 -1.83097348e-01
8.27458650e-02 1.93473145e-01 -9.13012326e-01 5.71522295e-01
6.91694438e-01 -3.74813765e-01 -9.78776515e-01 -4.42048222e-01
-1.99077293e-01 -5.66427782e-02 2.74833173e-01 -9.64482784e-01
-3.78727227e-01 -1.96316078e-01 8.69370580e-01 -1.80641130e-01
4.84769285e-01 -4.78106976e-01 2.32363179e-01 1.53959140e-01
-4.24822718e-01 -1.53225183e-01 1.29694017e-02 7.11056709e-01
-2.06917748e-01 -2.47171938e-01 1.03410888e+00 1.38933873e-02
1.71754248e-02 4.99095440e-01 -1.97589323e-01 6.00696802e-01
8.43670547e-01 4.55922782e-02 -4.92224634e-01 -9.17596340e-01
-5.09313643e-01 -5.78453182e-04 5.01110077e-01 2.57469006e-02
5.52162707e-01 -1.58306444e+00 -7.91883171e-01 3.25554401e-01
-2.86807716e-01 2.94729084e-01 1.71293184e-01 3.38985533e-01
-4.29996103e-01 -2.33531203e-02 -1.93326578e-01 -4.27467048e-01
-4.46903288e-01 2.28948995e-01 6.43068433e-01 -4.89084005e-01
-2.76374131e-01 9.67179239e-01 6.57465756e-01 -2.75886774e-01
8.97515416e-02 -1.91452682e-01 1.06056318e-01 -2.91878045e-01
5.28412521e-01 2.83463418e-01 -1.78921252e-01 -2.96476573e-01
5.80056496e-02 2.75780588e-01 -8.81297663e-02 -3.09328794e-01
1.31495786e+00 6.36138543e-02 3.76003742e-01 1.71507776e-01
1.11874092e+00 -2.56968122e-02 -2.03286099e+00 1.01614296e-01
-5.60578525e-01 -5.94939470e-01 3.03156767e-02 -7.55733728e-01
-1.34957767e+00 8.83659959e-01 4.38319981e-01 2.84053564e-01
1.12688029e+00 -1.07126616e-01 7.65173435e-01 -1.46222919e-01
1.18161969e-01 -7.39529967e-01 2.23371521e-01 3.55001956e-01
7.88799584e-01 -1.01869392e+00 -2.67395169e-01 2.32046680e-03
-7.52315402e-01 8.20491850e-01 6.33532703e-01 -4.32137638e-01
5.70749581e-01 6.27519250e-01 -2.50206664e-02 3.03033799e-01
-5.56568503e-01 2.74252743e-02 3.28265488e-01 1.17276084e+00
3.61564666e-01 4.82740365e-02 2.30209500e-01 3.51897091e-01
-5.02429903e-01 -1.58941194e-01 4.00403529e-01 3.90785336e-01
-1.04059137e-01 -1.25092065e+00 -8.82611424e-02 6.51547194e-01
-5.08736372e-01 -2.39977062e-01 3.22803669e-02 8.23621511e-01
-8.38154703e-02 6.49508893e-01 4.67933655e-01 -1.61725298e-01
3.27181928e-02 -6.68383539e-02 8.48642111e-01 -3.60625654e-01
-2.03508511e-01 7.89766666e-03 -2.18389496e-01 -7.07214102e-02
-3.21517467e-01 -3.97397697e-01 -7.57084966e-01 -5.26414692e-01
-4.61284369e-01 3.76787722e-01 4.27273840e-01 8.53450894e-01
1.68656632e-02 7.29673743e-01 6.68345630e-01 -7.87035763e-01
-8.57437968e-01 -1.15877998e+00 -9.31687474e-01 3.65100145e-01
4.79295045e-01 -4.81430203e-01 -5.02456725e-01 1.17259033e-01] | [11.557414054870605, -0.08200128376483917] |
0513ba78-621d-48b1-8780-4d62cb6438b3 | injecting-word-embeddings-with-another | null | null | https://aclanthology.org/I17-2020 | https://aclanthology.org/I17-2020.pdf | Injecting Word Embeddings with Another Language's Resource : An Application of Bilingual Embeddings | Word embeddings learned from text corpus can be improved by injecting knowledge from external resources, while at the same time also specializing them for similarity or relatedness. These knowledge resources (like WordNet, Paraphrase Database) may not exist for all languages. In this work we introduce a method to inject word embeddings of a language with knowledge resource of another language by leveraging bilingual embeddings. First we improve word embeddings of German, Italian, French and Spanish using resources of English and test them on variety of word similarity tasks. Then we demonstrate the utility of our method by creating improved embeddings for Urdu and Telugu languages using Hindi WordNet, beating the previously established baseline for Urdu. | ['P', 'Manish Shrivastava', 'Vikram Pudi', 'Prakhar ey'] | 2017-11-01 | injecting-word-embeddings-with-another-1 | https://aclanthology.org/I17-2020 | https://aclanthology.org/I17-2020.pdf | ijcnlp-2017-11 | ['learning-word-embeddings'] | ['methodology'] | [-6.50603712e-01 -1.94852039e-01 -3.88470054e-01 -2.88496703e-01
-5.44381738e-01 -1.03501546e+00 7.19886541e-01 2.85091281e-01
-1.11410177e+00 9.40412223e-01 8.07626963e-01 -4.22492385e-01
2.61993021e-01 -1.02156126e+00 -3.34660828e-01 -9.96851847e-02
2.71159738e-01 5.85875034e-01 1.09468915e-01 -6.88508689e-01
8.81840587e-02 4.06438500e-01 -6.74352944e-01 -7.50203803e-03
9.96209741e-01 1.61855936e-01 2.64280200e-01 3.22209269e-01
-3.00027639e-01 5.03188789e-01 -4.95427489e-01 -8.49984109e-01
2.95657635e-01 -1.72324359e-01 -1.04606187e+00 -4.64542389e-01
3.15227240e-01 -3.33276033e-01 -7.14659691e-01 1.03936088e+00
7.17630088e-01 2.51023054e-01 5.91414452e-01 -7.59281456e-01
-1.75615919e+00 9.90952611e-01 -1.32418036e-01 4.90322232e-01
3.24903160e-01 -5.91759011e-02 1.54929698e+00 -1.17735946e+00
9.37063158e-01 1.22817564e+00 5.81162810e-01 4.78105396e-01
-1.02520823e+00 -6.16843402e-01 -1.49787977e-01 5.24393916e-01
-1.47628677e+00 -3.52510549e-02 6.43615186e-01 -1.93993181e-01
1.59733689e+00 -2.79949546e-01 3.99247587e-01 1.34461319e+00
-1.27191842e-01 5.88450551e-01 6.96052849e-01 -6.13303483e-01
-2.96976089e-01 4.82091248e-01 5.85459054e-01 6.28545582e-01
3.44721705e-01 -4.93725352e-02 -4.58697140e-01 -1.29507780e-01
5.83621085e-01 -9.79117230e-02 -1.81635782e-01 -2.75464535e-01
-1.24490142e+00 1.12403226e+00 3.30163807e-01 7.07938731e-01
-1.18194878e-01 -6.78892732e-02 8.29747438e-01 7.24356234e-01
5.61247349e-01 8.36742818e-01 -8.55884552e-01 -1.51389852e-01
-4.47668850e-01 -2.20455006e-02 8.25190663e-01 1.06466234e+00
1.13241661e+00 -1.90389641e-02 -4.26351428e-02 1.35031915e+00
2.84509659e-01 6.20923698e-01 1.24899769e+00 -3.44875485e-01
5.47946811e-01 4.57454890e-01 -1.05474994e-01 -7.19564259e-01
1.02448657e-01 -8.59826803e-02 -2.81844378e-01 -4.15927410e-01
1.52082741e-01 -3.02167386e-01 -8.35506678e-01 1.80134571e+00
4.58091721e-02 1.84644669e-01 4.97057885e-01 6.50442541e-01
1.04474485e+00 6.15943491e-01 -8.70354548e-02 2.54478782e-01
1.30351627e+00 -1.15524101e+00 -6.64985955e-01 -2.73053288e-01
8.92247319e-01 -8.86029720e-01 1.36554933e+00 -6.03820197e-02
-7.15923667e-01 -7.62414873e-01 -1.07409024e+00 -4.93020028e-01
-1.04063308e+00 -1.29039690e-01 3.32015097e-01 7.59466290e-01
-1.00727272e+00 5.27678788e-01 -6.28258049e-01 -7.70021617e-01
-7.97418691e-03 1.66409284e-01 -7.84655213e-01 -4.20361042e-01
-1.79428732e+00 1.48382354e+00 8.55778813e-01 -5.08052409e-01
-7.47057319e-01 -9.35859978e-01 -1.29588509e+00 -2.38761827e-01
-2.99874872e-01 -3.96799684e-01 7.90908277e-01 -6.84851050e-01
-1.32854092e+00 1.09136903e+00 7.64254630e-02 -4.40123707e-01
4.44507506e-03 -3.77991110e-01 -6.72587276e-01 -1.52529255e-01
2.41481021e-01 5.86662233e-01 2.31290221e-01 -7.13949919e-01
-2.22714856e-01 -2.93694258e-01 4.24395055e-01 2.00089350e-01
-1.01288378e+00 1.98908418e-01 -2.59160012e-01 -8.59962344e-01
-4.57053334e-01 -5.65221250e-01 -1.50538877e-01 -4.78200048e-01
6.53457940e-02 -5.30726790e-01 7.14734018e-01 -1.19295180e+00
1.33494091e+00 -1.92937791e+00 1.16643809e-01 8.77232179e-02
-1.16001591e-01 6.02272451e-01 -7.41732299e-01 9.66316879e-01
4.27974872e-02 4.07206625e-01 2.41536759e-02 -3.62382010e-02
1.03327833e-01 9.12093818e-01 -3.06441814e-01 2.91085422e-01
4.70601439e-01 1.18664074e+00 -1.36733162e+00 -4.46611166e-01
2.39738733e-01 4.60991383e-01 -7.97465146e-01 1.73494011e-01
2.59260029e-01 2.36990359e-02 -1.46467015e-01 4.09025282e-01
4.03318346e-01 4.07895267e-01 3.76353115e-01 -2.70878613e-01
6.44134656e-02 7.23341823e-01 -8.61647069e-01 1.85705590e+00
-1.19398880e+00 6.16214752e-01 -5.02437830e-01 -1.07117260e+00
1.10140109e+00 4.68435675e-01 1.16013721e-01 -6.66788995e-01
1.00895554e-01 5.09485185e-01 2.11569026e-01 -5.62451065e-01
6.96910918e-01 -2.45275438e-01 -1.77442506e-01 6.11096263e-01
1.02367222e+00 -2.92333126e-01 4.09296721e-01 1.42598972e-01
1.10279393e+00 1.02431580e-01 6.42338097e-01 -5.13080239e-01
6.78666353e-01 -1.81007862e-01 3.08053166e-01 4.44468588e-01
-2.05218717e-01 4.04270113e-01 1.16803385e-02 -3.33394051e-01
-1.40234029e+00 -1.41022015e+00 -4.98724252e-01 1.33612931e+00
-7.70217329e-02 -7.59890020e-01 -1.97290376e-01 -1.02385271e+00
1.80494681e-01 9.58508730e-01 -6.07710063e-01 -3.00871372e-01
-8.38447154e-01 -5.67531586e-01 8.99579406e-01 7.10122526e-01
2.92155623e-01 -1.03202736e+00 6.57818168e-02 3.58270645e-01
-8.41952208e-03 -1.26803958e+00 -8.22275639e-01 8.45668763e-02
-6.22539818e-01 -9.25348878e-01 -7.39195704e-01 -1.07546997e+00
3.35768074e-01 6.35274053e-02 1.55273259e+00 -8.98701027e-02
-1.89388886e-01 5.51094294e-01 -7.48577416e-01 -9.85527262e-02
-4.60318804e-01 3.40848446e-01 3.40653569e-01 -4.82282311e-01
9.92315888e-01 -6.11173928e-01 -1.20554462e-01 8.05807635e-02
-9.99981880e-01 -6.37684166e-01 1.36240497e-01 1.19246256e+00
1.31397024e-01 -6.46261454e-01 8.32080126e-01 -8.67418051e-01
8.63282323e-01 -8.62533033e-01 -1.73196465e-01 3.57459038e-01
-3.30270052e-01 3.05224508e-01 6.83383405e-01 -5.12954056e-01
-8.02930355e-01 -3.28587860e-01 -4.83103126e-01 -1.15247220e-01
7.45557323e-02 7.07287610e-01 -2.16507241e-01 5.67128621e-02
8.49293411e-01 1.76786557e-01 -3.71701270e-01 -7.53282607e-01
1.11730111e+00 1.03543532e+00 2.26433560e-01 -9.24911380e-01
7.88207233e-01 8.10613707e-02 -7.48591721e-01 -9.21300292e-01
-6.92577481e-01 -6.69865966e-01 -1.10754848e+00 4.22099799e-01
7.97906816e-01 -9.70449388e-01 1.76508471e-01 -4.34221588e-02
-1.54531896e+00 5.80373816e-02 -4.30834591e-01 7.05319524e-01
-2.05534194e-02 5.64415336e-01 -7.26150632e-01 9.32315737e-03
-2.66541153e-01 -7.99684405e-01 6.60446763e-01 -8.42382908e-02
-3.36355746e-01 -1.74361515e+00 8.22907984e-01 9.44941416e-02
5.47826469e-01 -3.18552583e-01 1.13381350e+00 -1.35530329e+00
1.93690389e-01 -1.75697088e-01 -1.98709756e-01 1.04190135e+00
5.69490910e-01 -2.08675966e-01 -6.54275954e-01 -3.13311785e-01
-2.78920740e-01 -5.53106248e-01 7.06574082e-01 -3.73665363e-01
3.90912056e-01 -2.40006670e-01 6.08169474e-02 4.96862471e-01
1.64942360e+00 -1.49347395e-01 4.36257541e-01 4.57962483e-01
6.95544302e-01 4.08264667e-01 1.50782093e-01 8.58096257e-02
4.82878834e-01 5.03674328e-01 -2.06079558e-01 6.27400726e-02
-4.05191272e-01 -3.81270558e-01 5.79537392e-01 1.66280937e+00
2.44899001e-02 1.96542200e-02 -1.10804915e+00 1.37268388e+00
-1.51017845e+00 -6.77485108e-01 1.92147002e-01 2.01347089e+00
1.40052104e+00 -2.84873635e-01 -1.41905874e-01 -3.43952030e-01
4.54203397e-01 1.33414835e-01 -1.40055060e-01 -8.51697087e-01
-3.84616315e-01 1.10837746e+00 6.16710722e-01 8.55257452e-01
-7.01894581e-01 1.52409506e+00 6.29151487e+00 7.02320099e-01
-8.08527589e-01 7.38653004e-01 -6.37545958e-02 3.04479927e-01
-7.65453160e-01 2.15334505e-01 -8.80547166e-01 1.78165495e-01
9.04047966e-01 -7.34578550e-01 3.53007525e-01 6.66746855e-01
-3.43956620e-01 5.96527100e-01 -1.26732421e+00 7.72258878e-01
4.47452098e-01 -1.08916426e+00 3.25654268e-01 -3.72049212e-01
9.42304254e-01 6.89843416e-01 -3.48861545e-01 7.00978279e-01
9.40288424e-01 -1.09504747e+00 8.96610022e-02 -1.18457757e-01
6.80478871e-01 -7.47243941e-01 1.01977432e+00 -7.69019201e-02
-1.01599467e+00 3.92547935e-01 -9.02248263e-01 1.69636693e-03
1.86640218e-01 3.24637473e-01 -8.01305830e-01 7.15168417e-01
4.29245323e-01 1.02676046e+00 -8.01238239e-01 5.53183317e-01
-7.83530653e-01 6.67192936e-01 -1.74709454e-01 -1.11944228e-01
6.00735426e-01 -2.56975830e-01 2.82634944e-01 1.66032219e+00
4.67302054e-01 -3.45744342e-01 4.76245321e-02 6.38447702e-01
-3.50241780e-01 7.22548246e-01 -1.11903107e+00 -2.91879028e-01
5.33153355e-01 1.09497583e+00 8.24953765e-02 -4.05648559e-01
-9.47164893e-01 1.27071214e+00 6.65531039e-01 3.66200715e-01
-7.10828781e-01 -8.09754193e-01 1.03863025e+00 -3.27104300e-01
2.48615101e-01 -6.77767634e-01 1.11186042e-01 -1.56494951e+00
-1.88029379e-01 -7.10650086e-01 4.40232784e-01 -5.84140480e-01
-1.94627953e+00 7.48818398e-01 -4.25324924e-02 -8.26122284e-01
-2.73873031e-01 -1.19773161e+00 -4.44552064e-01 1.16583192e+00
-1.65766025e+00 -1.31383753e+00 2.86488384e-01 7.66074121e-01
6.12419426e-01 -7.26164818e-01 1.10387552e+00 5.42406261e-01
-1.97692186e-01 7.80492604e-01 3.45674276e-01 6.25384808e-01
1.15970159e+00 -1.30132103e+00 6.76725090e-01 5.48369825e-01
8.20585370e-01 9.77855921e-01 4.00356948e-01 -5.90222180e-01
-1.16478813e+00 -9.59123433e-01 1.49629283e+00 -7.31306016e-01
1.54995489e+00 -4.70681429e-01 -9.90550220e-01 9.45770562e-01
9.22861576e-01 4.89995405e-02 1.02402008e+00 4.52118069e-01
-8.57581615e-01 7.78866187e-02 -9.10237968e-01 7.89428413e-01
1.14064610e+00 -9.26689267e-01 -1.44254041e+00 5.24468005e-01
9.59012508e-01 1.91627085e-01 -1.14841783e+00 1.05956674e-01
3.58102858e-01 -3.54901791e-01 1.17022550e+00 -1.34335136e+00
4.78379756e-01 3.33917153e-04 -6.27563000e-01 -1.86477709e+00
-3.30091327e-01 -2.23753139e-01 1.58133209e-01 1.54735899e+00
5.38536668e-01 -9.21881735e-01 2.44166151e-01 -2.57448778e-02
-1.13288864e-01 -1.80324063e-01 -1.05994606e+00 -1.29943895e+00
9.42724705e-01 -2.46589392e-01 6.22442365e-01 1.44751215e+00
3.76229703e-01 6.91065729e-01 -1.95294529e-01 -3.98579985e-02
1.16698809e-01 -3.59594017e-01 4.45894629e-01 -7.87748039e-01
-1.86875880e-01 -1.69902787e-01 -6.17746949e-01 -8.96926403e-01
8.66810620e-01 -1.66841948e+00 -3.05732846e-01 -1.50468242e+00
1.10183828e-01 -1.18776664e-01 -7.07885563e-01 4.95300621e-01
-3.37154061e-01 5.46299577e-01 3.28799188e-02 -1.57027870e-01
-1.53535306e-01 7.12574959e-01 8.11110556e-01 -1.84446812e-01
2.11558700e-01 -9.08517659e-01 -5.62835455e-01 5.48518360e-01
8.74410689e-01 -4.95730162e-01 -3.16474646e-01 -1.10465455e+00
3.28813285e-01 -6.57670081e-01 2.04295758e-02 -5.26836157e-01
-5.22510335e-02 -6.31878674e-02 -1.23580180e-01 -1.13921635e-01
6.63160905e-02 -7.37944484e-01 -4.55871671e-01 2.60483056e-01
-1.61628649e-01 6.20882750e-01 3.21654469e-01 1.93753868e-01
-6.53761744e-01 -7.92817712e-01 7.17706025e-01 -3.00306171e-01
-9.74995613e-01 5.72298542e-02 -3.55787545e-01 5.87038159e-01
7.48460770e-01 9.30440575e-02 -3.28006327e-01 -2.83381399e-02
-4.70535129e-01 1.71756491e-01 4.27904099e-01 9.76191044e-01
4.34853345e-01 -1.79147160e+00 -1.10769594e+00 3.82322639e-01
5.25694788e-01 -8.50956202e-01 -3.41843843e-01 2.73285568e-01
-6.43339753e-01 3.69229555e-01 -5.02109587e-01 1.20460331e-01
-1.01022959e+00 5.55821657e-01 1.80474207e-01 -4.03963804e-01
-7.85456300e-02 6.93052471e-01 -3.02597452e-02 -1.27341616e+00
-1.97341487e-01 -1.08724833e-01 -4.00637865e-01 2.01553449e-01
2.31529668e-01 2.31889784e-01 1.25535697e-01 -7.93673396e-01
-4.88554984e-01 6.52616441e-01 -1.63516492e-01 -3.07113171e-01
1.43996251e+00 -6.52796552e-02 -3.55159611e-01 3.38485062e-01
1.58766139e+00 4.42242533e-01 -1.76197901e-01 -5.30519545e-01
3.30813706e-01 -5.49006760e-01 -2.05354944e-01 -5.42575896e-01
-9.02163863e-01 1.01552355e+00 4.01062578e-01 -2.27789000e-01
5.26544511e-01 9.05067995e-02 1.07215178e+00 6.83073342e-01
3.10665816e-01 -1.34147906e+00 2.07793012e-01 1.16666150e+00
7.35802412e-01 -1.22401083e+00 -3.92286569e-01 1.02980912e-01
-6.74784541e-01 1.27050567e+00 4.60601538e-01 -6.29822552e-01
9.50750172e-01 7.05973059e-02 3.48056138e-01 5.27741797e-02
-4.77378696e-01 -4.09219027e-01 3.23043555e-01 8.12233865e-01
9.04897571e-01 -1.48251638e-01 -9.89718318e-01 7.31502831e-01
-2.51824677e-01 -2.10811287e-01 3.98771554e-01 5.80175579e-01
-2.35776827e-01 -1.84397173e+00 1.11669660e-01 1.75100431e-01
-5.33282280e-01 -8.65429997e-01 -3.80560756e-01 8.67812872e-01
2.75976032e-01 6.67309105e-01 7.01934248e-02 -1.24839619e-01
3.74363959e-01 3.30794871e-01 5.46437919e-01 -1.12752593e+00
-6.81349635e-01 -5.04913330e-01 4.21629220e-01 -1.45542860e-01
-2.17029735e-01 -3.66816372e-01 -6.62837386e-01 -2.94727943e-05
-1.74329028e-01 2.98886091e-01 3.85181487e-01 8.47163439e-01
9.88726243e-02 7.59600252e-02 5.23091316e-01 -4.03300047e-01
-6.08058393e-01 -1.24566543e+00 -4.36220288e-01 8.15854073e-01
-1.81555822e-01 -4.27354574e-01 -2.72694081e-01 -2.75027994e-02] | [10.946320533752441, 9.907103538513184] |
6e4ed74b-b442-4110-b562-ea8506a8428a | self-supervised-image-representation-learning-1 | 2301.09299 | null | https://arxiv.org/abs/2301.09299v1 | https://arxiv.org/pdf/2301.09299v1.pdf | Self-Supervised Image Representation Learning: Transcending Masking with Paired Image Overlay | Self-supervised learning has become a popular approach in recent years for its ability to learn meaningful representations without the need for data annotation. This paper proposes a novel image augmentation technique, overlaying images, which has not been widely applied in self-supervised learning. This method is designed to provide better guidance for the model to understand underlying information, resulting in more useful representations. The proposed method is evaluated using contrastive learning, a widely used self-supervised learning method that has shown solid performance in downstream tasks. The results demonstrate the effectiveness of the proposed augmentation technique in improving the performance of self-supervised models. | ['Shaofei Wang', 'Han Ding', 'Yinheng Li'] | 2023-01-23 | null | null | null | null | ['image-augmentation'] | ['computer-vision'] | [ 5.67418337e-01 5.15564382e-01 -4.08264846e-01 -4.20709133e-01
-2.26271927e-01 -7.56736696e-02 8.90810370e-01 4.32612032e-01
-4.11879897e-01 7.04104185e-01 3.15443397e-01 -4.96978238e-02
1.83994070e-01 -4.59402740e-01 -3.52175981e-01 -6.27857029e-01
3.17148585e-03 1.31296441e-01 2.05263376e-01 -1.86967328e-01
2.00215340e-01 3.94367188e-01 -1.70231700e+00 1.75602779e-01
8.47067773e-01 6.38776600e-01 2.35443622e-01 1.44300491e-01
-4.35180038e-01 8.36110532e-01 -6.42200291e-01 -5.11521380e-03
6.72027767e-02 -5.14989853e-01 -7.07106769e-01 3.39847118e-01
4.00314897e-01 -3.34342048e-02 -2.18607351e-01 8.70116711e-01
1.37207523e-01 2.25389153e-01 8.42693150e-01 -1.12222028e+00
-4.15922523e-01 4.43289518e-01 -6.86442018e-01 4.41532940e-01
-3.48830856e-02 -1.67850032e-01 7.04734325e-01 -6.43278539e-01
4.63730872e-01 1.18130469e+00 3.67920935e-01 4.73329484e-01
-1.03500187e+00 -8.87566209e-01 1.67604834e-01 1.21558972e-01
-1.05624688e+00 -3.74240398e-01 1.02658927e+00 -3.22709799e-01
7.84841359e-01 -1.12953164e-01 6.02853954e-01 7.73872495e-01
-1.02196380e-01 8.90565932e-01 1.40315235e+00 -9.28827822e-01
1.84417933e-01 5.73768198e-01 2.94602096e-01 7.92671561e-01
2.79730737e-01 2.82025158e-01 -5.91926336e-01 1.17912546e-01
8.16666842e-01 2.35099286e-01 -7.44200731e-03 -5.01248538e-01
-9.35109675e-01 7.68464148e-01 6.84219718e-01 6.89167917e-01
-1.47673473e-01 3.36949490e-02 4.67707753e-01 1.61216274e-01
9.85502064e-01 4.60178196e-01 -2.09067523e-01 3.44176069e-02
-7.67326117e-01 -9.97994691e-02 2.02789590e-01 5.87877154e-01
6.49623036e-01 3.68717939e-01 -4.47985940e-02 7.82227337e-01
4.64795619e-01 6.40052631e-02 9.58380163e-01 -4.40504491e-01
2.22294822e-01 1.17476237e+00 -1.17397614e-01 -7.65177011e-01
-4.35248762e-01 -7.52711296e-01 -7.89464056e-01 4.00554597e-01
6.58822060e-02 1.37655716e-02 -1.11507559e+00 1.57586169e+00
1.85276493e-01 4.25101519e-01 4.65735078e-01 5.66840291e-01
9.29346383e-01 6.62547946e-01 5.02661884e-01 -2.46683404e-01
9.48698282e-01 -1.30989528e+00 -9.30708110e-01 -4.20008749e-01
7.74591863e-01 -5.88814676e-01 1.00762856e+00 2.65813291e-01
-5.82013071e-01 -8.67785037e-01 -1.25140226e+00 3.19039643e-01
-5.55357575e-01 2.79292852e-01 8.69347155e-01 5.70305884e-01
-7.73980439e-01 5.10947764e-01 -6.57600343e-01 -7.08474338e-01
8.21801722e-01 4.09570009e-01 -5.37343800e-01 2.08138958e-01
-8.65801275e-01 1.05121052e+00 6.82622612e-01 -2.44277060e-01
-7.10933566e-01 -3.54521066e-01 -9.33545828e-01 6.73746839e-02
2.13917270e-01 -1.28751040e-01 1.22357261e+00 -1.11207056e+00
-1.20941961e+00 9.94919538e-01 -1.58931926e-01 -7.14898050e-01
2.89706618e-01 -3.42693269e-01 -4.20956045e-01 1.83669642e-01
2.30093465e-05 8.80062103e-01 1.00509036e+00 -1.39855230e+00
-4.48855817e-01 -4.45545703e-01 -2.39426881e-01 5.48908591e-01
-9.87032473e-01 -2.23873824e-01 -2.14717463e-01 -9.43135738e-01
3.17200512e-01 -8.23486090e-01 -1.73219875e-01 -3.23061235e-02
-1.38840541e-01 -3.77427131e-01 1.32478786e+00 -2.99077213e-01
9.08035755e-01 -2.22470403e+00 -3.52244496e-01 6.60619065e-02
5.97468764e-02 7.33703613e-01 8.80164187e-03 4.91690964e-01
-3.86336237e-01 3.76573950e-02 -4.08657968e-01 -5.31804085e-01
-6.77830756e-01 3.42605591e-01 -1.14737682e-01 1.69338882e-01
4.48111653e-01 8.30494225e-01 -8.39487553e-01 -4.86367255e-01
5.08356631e-01 3.71260613e-01 2.88908351e-02 4.48285341e-01
-4.04412858e-02 6.95144117e-01 -2.56998152e-01 2.75338084e-01
3.67461711e-01 -2.64473975e-01 7.35838041e-02 -1.15777403e-01
-1.15576684e-01 1.75530557e-02 -9.19603288e-01 1.56204700e+00
-2.68887937e-01 6.74801826e-01 -5.00491858e-01 -1.29228461e+00
1.31597495e+00 2.34429047e-01 3.35038781e-01 -6.71396792e-01
2.46561766e-01 -1.75856069e-01 -2.25847680e-02 -5.46220183e-01
1.14897653e-01 -2.49758273e-01 3.66108507e-01 4.38859731e-01
2.37164229e-01 -1.57353934e-02 1.86570585e-01 3.18323344e-01
6.95416808e-01 2.61191010e-01 6.27141893e-01 -2.56089538e-01
6.02239370e-01 2.92751133e-01 3.53663653e-01 3.65356147e-01
-6.90012053e-02 3.26868206e-01 -9.12149325e-02 -3.12725753e-01
-9.26881850e-01 -7.57722616e-01 -3.76440026e-02 8.08145761e-01
4.06018347e-02 -3.00333679e-01 -5.96453547e-01 -8.84682834e-01
-1.25878587e-01 5.31529546e-01 -7.64967084e-01 -4.54949856e-01
-1.78522855e-01 -5.76018274e-01 2.75662273e-01 7.44187236e-01
1.07631457e+00 -1.20231926e+00 -3.72492373e-01 -2.48885546e-02
1.71948195e-01 -9.63080108e-01 -1.68928001e-02 3.93542111e-01
-1.59429610e+00 -1.12149918e+00 -8.29535723e-01 -1.16703117e+00
1.29737449e+00 7.47765720e-01 5.59362292e-01 2.33545274e-01
-2.00019658e-01 3.36835980e-01 -6.32712424e-01 -7.50977755e-01
-5.85740983e-01 1.03404664e-01 4.65531573e-02 1.13521315e-01
2.94417053e-01 -3.57620358e-01 -3.16441208e-01 2.28950843e-01
-1.13757753e+00 3.15869659e-01 8.60921502e-01 9.29136395e-01
4.58391935e-01 2.47649744e-01 1.00183499e+00 -1.47602248e+00
6.32635534e-01 -2.91323751e-01 -3.62024963e-01 6.24258853e-02
-9.82471883e-01 1.67441472e-01 5.45014322e-01 -3.38454753e-01
-1.39088356e+00 2.47890100e-01 5.88658899e-02 -2.58798629e-01
-4.22558814e-01 6.84724927e-01 6.16299920e-02 -3.22625607e-01
8.36513102e-01 2.91227132e-01 4.49007452e-01 -7.60225594e-01
2.54818708e-01 8.22861612e-01 4.79266971e-01 -5.04299253e-03
8.38725865e-01 4.66120273e-01 2.20109634e-02 -8.28603506e-01
-1.03274596e+00 -8.65457177e-01 -9.43877578e-01 -1.25099406e-01
5.84787190e-01 -9.79591906e-01 -5.35684265e-02 3.75071853e-01
-7.36895978e-01 -2.02495769e-01 -4.29942161e-01 5.08727670e-01
-3.16665143e-01 4.05911475e-01 -1.02342673e-01 -9.68515337e-01
-4.44251895e-01 -6.37708783e-01 5.04819632e-01 5.08019209e-01
-2.18839183e-01 -1.14181483e+00 3.72252278e-02 5.18402576e-01
3.95834714e-01 7.86266476e-02 7.42245793e-01 -1.08682001e+00
-9.66648012e-02 -4.07869369e-01 -2.50557095e-01 6.42692626e-01
6.14288509e-01 -2.41878882e-01 -1.17594910e+00 -2.83310384e-01
-1.77688822e-01 -4.43266153e-01 1.01570499e+00 -4.20131609e-02
1.07324696e+00 -2.72868037e-01 -4.15020496e-01 1.90076694e-01
1.23850262e+00 3.51926267e-01 5.76497555e-01 5.70212066e-01
4.76111382e-01 6.68945432e-01 9.46571887e-01 1.62135735e-01
2.31418535e-02 3.26986909e-01 4.48084772e-01 -7.35144973e-01
-2.66723633e-01 -4.01124299e-01 -9.00066178e-03 6.83745444e-01
1.83175523e-02 1.89010739e-01 -8.76621962e-01 4.39279705e-01
-2.02901316e+00 -7.00917661e-01 -5.00309840e-03 1.99494731e+00
6.25529230e-01 3.39132279e-01 6.85866624e-02 4.72527742e-01
5.72504640e-01 4.07746732e-02 -6.24630809e-01 -3.23579043e-01
-4.10159267e-02 2.92237014e-01 2.86664426e-01 1.10919915e-01
-1.26882136e+00 1.15146124e+00 7.15686941e+00 5.01626790e-01
-1.10066724e+00 7.91886002e-02 6.66026473e-01 5.61420918e-01
1.16405644e-01 -5.02946973e-02 -6.75683439e-01 1.81542322e-01
6.68964207e-01 -1.63771108e-01 -1.95054963e-01 1.09030998e+00
4.21198726e-01 -4.18725550e-01 -8.71310949e-01 8.36902022e-01
2.99158663e-01 -1.36721969e+00 2.69120872e-01 -2.10132822e-01
8.99463236e-01 -3.29590768e-01 8.16358328e-02 3.46714497e-01
1.88361928e-02 -8.15523267e-01 8.60179812e-02 2.98959166e-01
5.23567617e-01 -6.79210603e-01 8.70974541e-01 4.57433879e-01
-9.12931144e-01 -1.54244199e-01 -1.84975624e-01 -3.25283736e-01
-1.81532294e-01 2.26844162e-01 -1.36743736e+00 2.30504528e-01
5.09490728e-01 9.09256756e-01 -8.83211613e-01 1.16170716e+00
-4.14277911e-01 9.05721605e-01 9.23833698e-02 7.81031922e-02
1.39551178e-01 -3.55860149e-03 1.60395578e-01 1.16064239e+00
-6.95519969e-02 4.08374220e-02 2.70996064e-01 4.56969649e-01
-6.79072440e-02 4.88302588e-01 -9.67408478e-01 -3.61943483e-01
1.98046699e-01 1.23122096e+00 -9.38028574e-01 -5.61531782e-01
-5.55086374e-01 8.28046262e-01 2.63328314e-01 1.91109419e-01
-3.12871099e-01 -1.22665122e-01 8.11406896e-02 3.81545126e-01
-1.18192673e-01 -2.64735729e-01 -3.14729154e-01 -8.28488708e-01
-3.79361212e-01 -6.20383084e-01 5.40232360e-01 -8.81215274e-01
-1.09865510e+00 5.18456936e-01 2.12940991e-01 -1.33103347e+00
-3.15534353e-01 -4.34628040e-01 -7.31118739e-01 6.59957767e-01
-1.68027329e+00 -1.45442450e+00 -5.82126677e-01 5.91767907e-01
6.97902977e-01 -6.71746671e-01 1.00164330e+00 6.61990717e-02
-6.05507374e-01 5.09358943e-01 1.79150090e-01 8.44236929e-03
8.26461017e-01 -1.17291760e+00 1.21635631e-01 8.43504846e-01
3.79514158e-01 6.00053489e-01 6.13277137e-01 -8.58616054e-01
-8.52249503e-01 -1.12493026e+00 4.13722843e-01 2.88847834e-01
2.70816535e-01 -4.87593263e-02 -1.05507708e+00 6.10280633e-01
2.89355665e-01 4.77547161e-02 1.00713968e+00 4.35493737e-02
-2.09810451e-01 -3.01061809e-01 -1.27672970e+00 3.49760592e-01
7.16204047e-01 -2.19147190e-01 -7.84970999e-01 1.86658606e-01
5.60977519e-01 3.61565389e-02 -5.00414014e-01 6.15134299e-01
1.12907268e-01 -6.33017004e-01 6.48023129e-01 -5.22032201e-01
3.10780078e-01 -2.44110823e-01 3.72973889e-01 -1.49293113e+00
-2.23349392e-01 -2.92498052e-01 -3.24111909e-01 1.33609879e+00
3.65965098e-01 -6.90695286e-01 1.23512089e+00 2.37818584e-01
-1.16156049e-01 -6.90567017e-01 -5.28983116e-01 -7.75384009e-01
-3.51471066e-01 -2.70536244e-02 1.28898084e-01 1.15393901e+00
2.17006415e-01 5.35448909e-01 -2.14410469e-01 -2.00158402e-01
5.20600319e-01 -1.70223996e-01 8.60679507e-01 -1.46864069e+00
8.88529122e-02 -1.28994539e-01 -6.27175272e-01 -7.26043820e-01
6.00257777e-02 -9.61697996e-01 -1.16546318e-01 -1.80667043e+00
3.13106418e-01 -4.64018434e-01 -5.54446220e-01 7.69675791e-01
-2.83882350e-01 5.16814649e-01 1.38023064e-01 5.67598164e-01
-4.41285372e-01 6.86323106e-01 1.28882158e+00 -1.91784307e-01
-3.67079675e-01 -1.38416626e-02 -9.23499107e-01 8.76092315e-01
1.21780574e+00 -3.32150578e-01 -8.14630032e-01 -3.22784781e-02
-4.40283179e-01 -4.56286401e-01 1.06631957e-01 -1.28078222e+00
1.12497717e-01 6.76014721e-02 6.13519549e-01 -5.11875212e-01
2.47318879e-01 -9.04909849e-01 -2.25027293e-01 5.05475938e-01
-6.03457868e-01 -7.84048736e-02 5.40410817e-01 6.67297721e-01
-5.45842111e-01 -4.49626178e-01 8.57337713e-01 -1.46521047e-01
-1.15049219e+00 4.87969518e-02 -1.87194079e-01 -4.03364241e-01
1.28698230e+00 -3.66529733e-01 -1.14839219e-01 -3.85321647e-01
-9.45917010e-01 5.96838817e-02 1.98135838e-01 6.24343991e-01
8.21216941e-01 -1.25891590e+00 -3.45240414e-01 2.78545678e-01
5.44326365e-01 -1.02981411e-01 1.56425815e-02 4.42992687e-01
-3.00726116e-01 3.89961272e-01 -4.38643873e-01 -5.33427119e-01
-1.58300292e+00 3.80225301e-01 -1.35918468e-01 -3.15905333e-01
-6.89860106e-01 6.08501732e-01 1.34577617e-01 -1.42259911e-01
3.99955511e-01 2.24981457e-02 -9.36529934e-01 -8.60112999e-03
5.62085092e-01 2.29639232e-01 4.98189405e-02 -6.37335956e-01
7.17940554e-02 3.09426278e-01 -4.25803632e-01 5.83859719e-03
1.32396805e+00 -8.23136792e-02 1.48970500e-01 6.50974631e-01
1.02464759e+00 -4.58182216e-01 -1.16540837e+00 -5.25831878e-01
2.34369531e-01 -4.68365967e-01 3.43337595e-01 -8.46244335e-01
-9.70768094e-01 9.36189890e-01 1.13943791e+00 4.22820672e-02
1.05630648e+00 -4.27979082e-02 5.07559419e-01 4.36206818e-01
1.32674754e-01 -1.07503390e+00 5.50109804e-01 2.23603979e-01
8.47099781e-01 -1.75739324e+00 1.71954751e-01 -7.02485740e-01
-7.16627657e-01 1.10341978e+00 1.01448524e+00 -3.33145298e-02
6.32476449e-01 -2.16678008e-02 2.98394799e-01 -1.63551345e-01
-3.91026676e-01 -4.82433051e-01 3.17652851e-01 8.63880992e-01
6.33425236e-01 -2.78468698e-01 -3.16527814e-01 1.41729921e-01
2.09102601e-01 8.42468143e-02 2.47432113e-01 1.15352046e+00
-6.47007942e-01 -1.31834519e+00 -2.34331161e-01 7.72807300e-01
-2.03310981e-01 7.38038048e-02 -4.64095443e-01 8.45571935e-01
5.46531603e-02 8.71888280e-01 4.49469127e-02 -2.51613051e-01
1.32884428e-01 9.64817107e-02 2.27423340e-01 -1.14620876e+00
-4.41703498e-01 8.90461877e-02 -5.28067723e-02 4.29084003e-02
-9.28093374e-01 -2.88689792e-01 -1.45595884e+00 3.33057791e-01
-4.49140191e-01 2.67153770e-01 8.42331052e-01 9.10507023e-01
2.71524787e-01 6.21650815e-01 6.80711985e-01 -6.85264766e-01
-2.38324240e-01 -1.26683724e+00 -3.39041919e-01 6.60415471e-01
2.01592930e-02 -8.81388485e-01 -1.30939528e-01 2.01616302e-01] | [9.460387229919434, 2.464905023574829] |
04bb6247-c20b-4138-bc38-18fac5f78d25 | lapnet-automatic-balanced-loss-and-optimal | 1911.01149 | null | https://arxiv.org/abs/1911.01149v2 | https://arxiv.org/pdf/1911.01149v2.pdf | LapNet : Automatic Balanced Loss and Optimal Assignment for Real-Time Dense Object Detection | Real-time single-stage object detectors based on deep learning still remain less accurate than more complex ones. The trade-off between model performance and computational speed is a major challenge. In this paper, we propose a new way to efficiently learn a single-shot detector which offers a very good compromise between these two objectives. To this end, we introduce LapNet, an anchor based detector, trained end-to-end without any sampling strategy. Our approach aims to overcome two important problems encountered in training an anchor based detector: (1) ambiguity in the assignment of anchor to ground truth and (2) class and object size imbalance. To address the first limitation, we propose a soft positive/negative anchor assignment procedure based on a new overlapping function called "Per-Object Normalized Overlap" (PONO). This soft assignment can be self-corrected by the network itself to avoid ambiguity between close objects. To cope with the second limitation, we propose to learn additional weights, that are not used at inference, to efficiently manage sample imbalance. These two contributions make the detector learning more generic whatever the training dataset. Various experiments show the effectiveness of the proposed approach. | ['Mohamed Chaouch', 'Quoc-Cuong Pham', 'Florian Chabot'] | 2019-11-04 | null | null | null | null | ['dense-object-detection'] | ['computer-vision'] | [ 1.18515015e-01 2.02930838e-01 -2.21137211e-01 -4.92693871e-01
-5.81286728e-01 -6.32441193e-02 3.56038421e-01 5.26786208e-01
-5.81481695e-01 6.62019908e-01 -4.43478584e-01 2.38924697e-01
-2.11910725e-01 -8.25611055e-01 -7.19799638e-01 -6.92756951e-01
2.20160764e-02 6.45034015e-01 8.46178710e-01 2.19012555e-02
3.28265548e-01 5.50337493e-01 -1.83397841e+00 1.60497397e-01
8.76656115e-01 1.24616027e+00 3.59966516e-01 3.65391910e-01
-3.07719797e-01 6.36552334e-01 -5.93729138e-01 -3.29655319e-01
3.91744107e-01 -2.03342304e-01 -5.03928244e-01 1.90726757e-01
3.28872800e-01 -3.99283051e-01 2.25408629e-01 1.15630472e+00
7.57719517e-01 4.55870740e-02 7.38474309e-01 -1.38251650e+00
-1.68548465e-01 7.66263783e-01 -8.07379067e-01 4.61762518e-01
-1.58041686e-01 1.39761880e-01 1.01203465e+00 -1.03908646e+00
4.17449892e-01 1.07821131e+00 9.79918361e-01 4.43754226e-01
-9.91672218e-01 -4.45590913e-01 2.37830639e-01 5.04629731e-01
-1.55117226e+00 -4.12254006e-01 8.04314137e-01 -3.65922332e-01
6.17647588e-01 1.02865919e-01 6.15380466e-01 8.02459955e-01
-4.89790663e-02 8.27369571e-01 7.66093254e-01 -7.38472700e-01
5.84991872e-01 5.33252299e-01 1.94121450e-01 5.65680623e-01
6.04492605e-01 -2.25076824e-01 -3.78750801e-01 -1.13359086e-01
5.48029602e-01 -1.31637752e-01 -7.32516944e-02 -8.73754799e-01
-7.53552377e-01 8.04297209e-01 5.38881958e-01 5.04332900e-01
-1.54503688e-01 8.67708400e-02 4.64881569e-01 1.56094013e-02
5.26796699e-01 3.84112567e-01 -3.06669563e-01 3.15808654e-01
-1.13177752e+00 3.32490474e-01 6.92357540e-01 6.64593935e-01
7.75620639e-01 -1.22952685e-01 -3.60963583e-01 9.43744183e-01
3.68989676e-01 1.66261345e-01 4.35830921e-01 -3.17708433e-01
3.94420266e-01 7.73366451e-01 2.32717231e-01 -1.06524169e+00
-6.65091515e-01 -8.65200877e-01 -7.03338742e-01 5.08884490e-01
5.56803048e-01 9.80062187e-02 -8.82409692e-01 1.57804692e+00
6.36736214e-01 2.03065991e-01 -2.50966758e-01 9.61962342e-01
6.26353621e-01 3.23907465e-01 1.26067534e-01 -1.86077282e-01
1.36955655e+00 -9.46863949e-01 -7.73073196e-01 -3.19602698e-01
5.88860512e-01 -6.09048247e-01 8.68196964e-01 2.22790778e-01
-1.03067422e+00 -5.75083256e-01 -1.40105236e+00 2.02127323e-01
-4.98746991e-01 4.50909555e-01 4.41479832e-01 6.73965991e-01
-6.54926479e-01 6.77259922e-01 -7.90701389e-01 -3.51074994e-01
5.84956884e-01 5.76862812e-01 6.09551705e-02 2.47271448e-01
-1.02664065e+00 1.09640479e+00 7.78559506e-01 1.18228488e-01
-4.85870719e-01 -6.10206723e-01 -5.36463857e-01 3.78330141e-01
7.11446583e-01 -5.04950464e-01 1.23913705e+00 -1.21678936e+00
-1.36196506e+00 9.54289615e-01 2.42926836e-01 -7.44293928e-01
7.02326715e-01 -6.25909790e-02 -1.39752060e-01 -1.00378610e-01
1.00953996e-01 5.30998051e-01 8.62041533e-01 -1.29479933e+00
-8.14722419e-01 -4.39399451e-01 -9.92630795e-02 7.50719234e-02
-5.37511051e-01 5.43436147e-02 -2.89407939e-01 -5.69378495e-01
2.85299327e-02 -4.83895004e-01 -1.93220869e-01 4.23244506e-01
-2.91336834e-01 -5.01770258e-01 7.90340483e-01 -1.79609120e-01
1.20523024e+00 -2.00148582e+00 -2.15329275e-01 2.11509600e-01
1.85419485e-01 7.73589194e-01 -2.89705601e-02 1.43802404e-01
-3.98586020e-02 -3.13082516e-01 -3.11248839e-01 -5.72979391e-01
-1.20673487e-02 2.23341584e-01 -7.83538222e-02 5.63375533e-01
5.43235898e-01 5.82810700e-01 -8.30055177e-01 -7.37778187e-01
1.82932168e-01 2.68201262e-01 -4.14999485e-01 1.69213519e-01
-2.34627336e-01 -5.16118407e-02 -2.95292258e-01 6.31876707e-01
1.05630827e+00 -7.73076117e-02 5.32262735e-02 -3.22392702e-01
-2.53964454e-01 -7.96296373e-02 -1.92164421e+00 1.46825850e+00
-7.65404552e-02 2.69913882e-01 1.04887426e-01 -1.39302337e+00
1.16186500e+00 8.43278617e-02 4.17196304e-01 -5.98805904e-01
3.88778239e-01 4.84570086e-01 -1.78371165e-02 -4.69016135e-01
4.08084273e-01 -1.50155351e-01 2.49739140e-01 1.19352438e-01
3.08054209e-01 4.24512684e-01 3.26699346e-01 -1.67759612e-01
9.27886307e-01 7.86770731e-02 5.81762195e-01 -3.89511555e-01
5.95375121e-01 -2.15836570e-01 8.69832158e-01 9.50993061e-01
-3.88164818e-01 7.35364318e-01 6.06682062e-01 -7.03655660e-01
-1.09505665e+00 -8.03517938e-01 -2.12490469e-01 1.07837760e+00
1.83717534e-01 -2.53266916e-02 -8.17477405e-01 -7.41772354e-01
1.14360467e-01 3.45189035e-01 -6.24805093e-01 -1.69031948e-01
-7.14822650e-01 -1.06638646e+00 3.95862341e-01 6.95956171e-01
3.93092394e-01 -7.90919900e-01 -9.99526441e-01 2.89861113e-01
-9.26063675e-03 -9.53813076e-01 -1.56174824e-01 5.00227630e-01
-8.44761014e-01 -1.01200974e+00 -7.41476893e-01 -6.66143537e-01
7.66519785e-01 1.93262056e-01 1.03705168e+00 2.39386722e-01
-5.61086237e-01 -2.24335231e-02 -4.33729053e-01 -8.33488166e-01
-1.72955707e-01 3.83290470e-01 1.29520193e-01 2.44058609e-01
4.45261508e-01 -5.39525092e-01 -5.45755565e-01 3.85497004e-01
-1.02718925e+00 -2.03063115e-01 8.31530988e-01 8.32027078e-01
5.91866791e-01 -1.45934567e-01 8.15934062e-01 -7.81518400e-01
2.57342905e-01 -3.09607029e-01 -9.21409428e-01 3.67359281e-01
-5.92210412e-01 2.21363425e-01 5.13726473e-01 -5.45820653e-01
-7.61485875e-01 3.52492183e-01 -2.50115544e-01 -3.71829301e-01
5.33885024e-02 1.37033880e-01 -3.48725826e-01 -2.54743308e-01
8.51552844e-01 -1.15657337e-01 -5.50882891e-02 -7.13462114e-01
1.23387732e-01 6.77557945e-01 4.96555626e-01 -2.95493603e-01
7.90932238e-01 6.15878940e-01 5.78032285e-02 -4.84093755e-01
-1.03992641e+00 -6.90266669e-01 -8.32506895e-01 -2.49993742e-01
6.07434928e-01 -7.54546463e-01 -6.08081937e-01 4.07017589e-01
-1.17706716e+00 -7.44818524e-02 -5.94940186e-01 3.86728585e-01
-4.98729110e-01 3.53328377e-01 -1.57173648e-01 -9.59828973e-01
-4.75536138e-01 -7.59097695e-01 8.63374770e-01 3.87085646e-01
1.10749774e-01 -6.02174342e-01 1.01478890e-01 -4.26071286e-02
3.58803481e-01 3.01894218e-01 4.63826060e-01 -9.51466560e-01
-5.61794698e-01 -3.51584226e-01 -5.37135363e-01 4.33814466e-01
-1.76378042e-01 -2.75337826e-02 -9.76406932e-01 -3.47579569e-01
1.36602402e-01 -2.67557681e-01 9.42625284e-01 2.22903579e-01
1.10037696e+00 -4.29803170e-02 -4.15710896e-01 4.01092201e-01
1.60123146e+00 -1.33705959e-01 5.93905985e-01 5.71976483e-01
3.72545034e-01 4.29075122e-01 8.85446429e-01 7.02728510e-01
1.49284035e-01 9.93266463e-01 6.57292962e-01 -1.02962099e-01
-2.94217438e-01 1.23467222e-02 -2.85023102e-03 4.52857137e-01
-1.90603826e-02 -2.92766660e-01 -9.46594417e-01 5.92640400e-01
-2.06260633e+00 -8.60877156e-01 -2.20537663e-01 2.31407809e+00
5.99182785e-01 5.95856428e-01 2.42210269e-01 4.03400481e-01
8.08375895e-01 -5.93110770e-02 -4.69722837e-01 -8.10323060e-02
-9.59392413e-02 -3.80783621e-03 6.49579585e-01 2.73269892e-01
-1.22418809e+00 5.95876276e-01 5.30655146e+00 9.66138124e-01
-1.14624989e+00 3.22650999e-01 3.67780715e-01 3.17899287e-02
2.05182895e-01 -8.53169188e-02 -1.21671176e+00 5.18789887e-01
5.64525485e-01 1.26738265e-01 -1.68216377e-01 1.19356024e+00
-3.24345753e-02 -2.96833932e-01 -1.11843371e+00 7.78545260e-01
2.46277198e-01 -1.21207404e+00 -1.77063450e-01 -2.29299992e-01
4.06493485e-01 -1.37891173e-01 -2.02852547e-01 3.51137489e-01
-3.07848424e-01 -4.71740723e-01 9.53299999e-01 5.03009558e-01
3.84072602e-01 -7.96651006e-01 9.78966951e-01 5.36801636e-01
-1.09580457e+00 -3.20020437e-01 -7.85242736e-01 -2.27495898e-02
-5.76052535e-03 9.21305120e-01 -1.00576758e+00 4.99667019e-01
5.67276001e-01 2.30149552e-01 -5.18381715e-01 1.61716235e+00
-3.46978195e-02 2.44058043e-01 -5.72605252e-01 -1.59164786e-01
1.62656248e-01 3.93901408e-01 7.39956737e-01 1.24671841e+00
2.15623617e-01 -2.24462777e-01 2.12863103e-01 9.85692382e-01
6.57057613e-02 2.07036331e-01 -3.35828334e-01 4.73136276e-01
3.78994316e-01 1.37255359e+00 -1.25426197e+00 -3.53003561e-01
-1.51781470e-01 7.19895124e-01 5.52641153e-01 -2.30275616e-01
-1.03769863e+00 -3.95647168e-01 1.81222945e-01 3.77906412e-01
5.60617805e-01 3.47153209e-02 -3.53292882e-01 -1.00129676e+00
3.29071760e-01 -4.82521594e-01 3.21356088e-01 -3.54753554e-01
-1.36819565e+00 3.30877960e-01 9.71043482e-02 -1.43095624e+00
4.63382620e-03 -7.24244833e-01 -5.42203486e-01 4.93148059e-01
-1.71294034e+00 -1.04996979e+00 -4.21583325e-01 2.34471604e-01
5.40211916e-01 3.39357071e-02 5.74899256e-01 8.14144969e-01
-6.07198954e-01 9.00548935e-01 5.29681556e-02 4.12065722e-02
7.95970380e-01 -1.25644040e+00 1.84218377e-01 8.56493413e-01
-1.43756783e-02 1.19482890e-01 8.66786718e-01 -5.03955245e-01
-8.67602170e-01 -9.96720672e-01 7.97286928e-01 -2.55227655e-01
3.17493051e-01 -4.75826949e-01 -1.17353916e+00 2.92328030e-01
-2.60273337e-01 3.87937039e-01 4.34210330e-01 5.20852655e-02
-2.86476612e-01 -5.48714101e-01 -1.29966140e+00 1.62806064e-01
8.21036339e-01 1.45964727e-01 -5.85133612e-01 3.79594624e-01
4.61667448e-01 -4.20211077e-01 -4.74450678e-01 6.78402245e-01
3.67406547e-01 -1.13507545e+00 9.43463027e-01 -2.77654141e-01
4.35244516e-02 -4.53630328e-01 8.51810351e-02 -1.01037765e+00
-3.37650329e-01 -1.14525817e-01 -2.71517426e-01 1.35395336e+00
1.85809016e-01 -4.85170990e-01 8.77078712e-01 2.34984547e-01
-1.69959843e-01 -9.32014167e-01 -1.23640001e+00 -9.70027030e-01
-3.04074347e-01 -1.68654606e-01 4.37870622e-01 8.04491997e-01
-2.72502273e-01 2.72313088e-01 -4.55015272e-01 2.24085763e-01
6.92644358e-01 -9.49062258e-02 6.37447298e-01 -1.63590908e+00
-3.61322165e-01 -4.03492421e-01 -7.17557967e-01 -7.57225096e-01
-3.74159187e-01 -5.66493392e-01 3.83131802e-01 -1.35487282e+00
3.14084500e-01 -7.29312241e-01 -4.34596986e-01 4.53437030e-01
-1.11383781e-01 3.15774828e-01 2.15600923e-01 2.19040766e-01
-8.52598429e-01 5.56784213e-01 7.87949741e-01 1.34996772e-01
-1.68884933e-01 2.54614949e-01 -3.30158263e-01 9.08990383e-01
7.41060317e-01 -8.56124341e-01 -1.65406123e-01 -3.16596866e-01
2.95692861e-01 -2.94764787e-01 3.89849275e-01 -1.48349500e+00
4.19696718e-01 1.45093560e-01 3.44836444e-01 -8.33609283e-01
1.83620095e-01 -8.88506353e-01 -3.65649790e-01 5.76033652e-01
-1.32174402e-01 -2.20515415e-01 1.63136255e-02 6.27620578e-01
-9.71891955e-02 -7.96217561e-01 1.13163865e+00 -1.26741529e-01
-6.04392886e-01 4.87586632e-02 -7.59192184e-02 -1.66006207e-01
1.29047215e+00 -3.49754035e-01 -8.96879509e-02 1.80553019e-01
-6.39062226e-01 3.05605829e-01 1.86613724e-01 4.02785927e-01
2.44021013e-01 -1.26974797e+00 -5.75448215e-01 1.27729565e-01
1.91049233e-01 1.02971144e-01 2.76250660e-01 9.36265409e-01
-4.13108706e-01 1.90402642e-01 -3.96789342e-01 -6.93119943e-01
-1.17638588e+00 5.95536530e-01 4.10166651e-01 -5.85847497e-01
-5.15124500e-01 9.15133297e-01 -2.48981640e-01 -2.90429771e-01
6.82304323e-01 -3.47021341e-01 -2.91569561e-01 4.51920688e-01
7.50248194e-01 4.28848803e-01 2.80040294e-01 -2.79243410e-01
-4.27848011e-01 4.67203647e-01 -1.34766623e-01 2.85776168e-01
1.42311645e+00 -1.46867707e-01 -3.35627757e-02 5.52966416e-01
8.53794992e-01 -2.47896239e-01 -1.21127927e+00 -3.20251018e-01
2.30211779e-01 -4.84054029e-01 -9.47294012e-02 -5.76664150e-01
-9.12925541e-01 8.39024484e-01 1.11164606e+00 4.09804910e-01
1.02434146e+00 -1.47148430e-01 5.94107509e-01 3.92385364e-01
8.91623050e-02 -1.48116004e+00 2.67230541e-01 2.43414521e-01
6.16873503e-01 -1.30399919e+00 1.29880875e-01 -4.46936786e-01
-1.95518479e-01 1.22872972e+00 9.16151106e-01 -1.86656326e-01
4.19061720e-01 3.31767708e-01 -1.74274832e-01 -1.00237332e-01
-4.46400881e-01 -4.53138769e-01 2.05648705e-01 4.74822938e-01
8.36143568e-02 -2.56322831e-01 -7.58220971e-01 5.59939146e-01
4.38770860e-01 2.85901308e-01 3.32497358e-01 9.62947071e-01
-9.56485152e-01 -1.08487654e+00 -5.87259293e-01 2.62181520e-01
-4.70331818e-01 3.10517192e-01 -6.50877133e-02 7.78963089e-01
6.94302917e-01 5.05491436e-01 2.39461750e-01 -3.00642964e-03
4.67277735e-01 6.55386001e-02 3.52384686e-01 -4.98370796e-01
-4.72332537e-01 -1.33106858e-01 -2.81858146e-01 -3.09136569e-01
-4.92461801e-01 -4.09694046e-01 -1.06429482e+00 2.19208032e-01
-7.44845986e-01 -9.33450088e-02 8.06921124e-01 9.37402964e-01
1.05721764e-01 6.63290977e-01 5.12531519e-01 -1.00514591e+00
-1.09303236e+00 -9.42554057e-01 -4.72979873e-01 1.76365584e-01
2.18915045e-01 -9.75046277e-01 -2.95337111e-01 -2.54021555e-01] | [9.082897186279297, 1.1942384243011475] |
4260ed8b-7bc4-440b-8dae-764b8dfafb00 | efficient-algorithms-for-sparse-moment | 2207.13008 | null | https://arxiv.org/abs/2207.13008v1 | https://arxiv.org/pdf/2207.13008v1.pdf | Efficient Algorithms for Sparse Moment Problems without Separation | We consider the sparse moment problem of learning a $k$-spike mixture in high dimensional space from its noisy moment information in any dimension. We measure the accuracy of the learned mixtures using transportation distance. Previous algorithms either assume certain separation assumptions, use more recovery moments, or run in (super) exponential time. Our algorithm for the 1-dimension problem (also called the sparse Hausdorff moment problem) is a robust version of the classic Prony's method, and our contribution mainly lies in the analysis. We adopt a global and much tighter analysis than previous work (which analyzes the perturbation of the intermediate results of Prony's method). A useful technical ingredient is a connection between the linear system defined by the Vandermonde matrix and the Schur polynomial, which allows us to provide tight perturbation bound independent of the separation and may be useful in other contexts. To tackle the high dimensional problem, we first solve the 2-dimensional problem by extending the 1-dimension algorithm and analysis to complex numbers. Our algorithm for the high dimensional case determines the coordinates of each spike by aligning a 1-d projection of the mixture to a random vector and a set of 2d-projections of the mixture. Our results have applications to learning topic models and Gaussian mixtures, implying improved sample complexity results or running time over prior work. | ['Jian Li', 'Zhiyuan Fan'] | 2022-07-26 | null | null | null | null | ['topic-models'] | ['natural-language-processing'] | [-1.22506864e-01 -1.39962863e-02 1.33462682e-01 -5.03662713e-02
-1.24424338e+00 -6.21127307e-01 4.29712415e-01 -6.89941570e-02
-5.04194736e-01 5.55228531e-01 -3.09503190e-02 -8.93194377e-02
-5.10397136e-01 -2.98230022e-01 -8.02211404e-01 -1.24547589e+00
-4.00664777e-01 7.98991442e-01 1.45818228e-02 8.51959065e-02
4.75811750e-01 8.48553836e-01 -1.20887339e+00 -3.72314215e-01
4.34811622e-01 1.03538942e+00 -1.24534421e-01 8.97891462e-01
-2.59171993e-01 2.02985659e-01 -3.44098270e-01 -2.43402705e-01
3.84534478e-01 -4.72360402e-01 -7.24592745e-01 1.35036111e-01
2.36213192e-01 2.49839246e-01 -2.69778837e-02 1.39793384e+00
6.00730300e-01 3.16593377e-03 1.16857505e+00 -1.51919568e+00
-2.86070466e-01 6.94430947e-01 -7.81953633e-01 3.68196219e-01
1.33242849e-02 -3.60534340e-01 4.53845561e-01 -8.49491060e-01
2.34102950e-01 1.13373363e+00 9.02142823e-01 4.74385470e-01
-1.57897937e+00 -5.60455859e-01 -1.61843151e-01 -2.80663334e-02
-1.53683662e+00 -7.01020181e-01 8.64591241e-01 -7.62486100e-01
5.98226190e-01 2.04416245e-01 5.36159694e-01 7.16443121e-01
1.62351839e-02 8.08626175e-01 1.01662743e+00 -4.44029301e-01
4.76482064e-01 1.84541613e-01 3.92448753e-01 6.29738510e-01
3.95234078e-01 -4.29943740e-01 3.93478274e-02 -4.25168812e-01
7.36447155e-01 8.54221359e-03 -2.80820459e-01 -6.93719029e-01
-1.27696025e+00 7.70294964e-01 -2.73279428e-01 5.75474739e-01
-1.99070454e-01 1.11376330e-01 -2.88270731e-02 1.74918756e-01
5.71212053e-01 2.29249790e-01 -2.90499955e-01 -3.60068262e-01
-1.23818910e+00 5.94978034e-01 1.22104943e+00 1.06541288e+00
6.30947769e-01 -2.56808430e-01 4.34852958e-01 6.73774660e-01
3.77526939e-01 7.71297395e-01 3.49012464e-01 -1.25766301e+00
2.65875369e-01 -1.38734639e-01 1.17538311e-01 -7.63774335e-01
-4.99582052e-01 -3.88357639e-01 -9.92952585e-01 -8.73175934e-02
8.49316239e-01 -1.59348816e-01 -4.03284520e-01 1.99560046e+00
1.14283532e-01 5.29658079e-01 1.82518438e-01 4.35841471e-01
-2.72816028e-02 4.42046583e-01 -5.97669423e-01 -5.11892200e-01
8.96785915e-01 -3.26558322e-01 -4.77919161e-01 1.12586007e-01
6.11214817e-01 -7.78491139e-01 5.37139177e-01 5.01630008e-01
-1.42031896e+00 2.60450728e-02 -8.82293582e-01 1.98036000e-01
-2.97295719e-01 -1.10223643e-01 2.31863648e-01 5.29392838e-01
-1.27183628e+00 8.08293939e-01 -9.41797137e-01 -3.02496344e-01
2.74512500e-01 3.96998048e-01 -2.08980009e-01 -1.05569690e-01
-7.39867568e-01 6.47723556e-01 -1.56041890e-01 -7.90963173e-02
-2.91725487e-01 -5.99293053e-01 -7.72343636e-01 -1.74651191e-01
-1.66405886e-01 -4.63330716e-01 1.02150369e+00 -1.40867829e-01
-1.40089953e+00 6.18326068e-01 -2.54696637e-01 -4.04645711e-01
4.26569492e-01 3.13920528e-02 -5.15354201e-02 2.28464469e-01
1.94353908e-01 2.42457703e-01 8.77324879e-01 -9.64338899e-01
-2.04965457e-01 -7.20524967e-01 -4.28936183e-01 -1.50885567e-01
2.17308521e-01 -2.94796646e-01 -2.12663919e-01 -5.45187950e-01
5.74276388e-01 -1.15810323e+00 -4.04150754e-01 -4.95519012e-01
-2.75502592e-01 -1.31926104e-01 6.81550145e-01 -2.30749622e-01
8.82886350e-01 -2.27070856e+00 4.02464241e-01 3.47779632e-01
3.78301561e-01 -1.88452885e-01 -7.15362877e-02 5.09092450e-01
-3.54451120e-01 4.54358570e-02 -6.35738194e-01 -7.56624579e-01
2.43664846e-01 -4.66977321e-02 -4.64886814e-01 1.25158536e+00
2.17712969e-02 4.28854108e-01 -6.64930403e-01 -4.01022702e-01
-1.28289178e-01 5.12451231e-01 -6.43971860e-01 -2.22577244e-01
1.89522833e-01 2.34493509e-01 -1.84701949e-01 2.40084574e-01
9.58128273e-01 -2.15275407e-01 -5.13599932e-01 -1.74653426e-01
-1.54058160e-02 -1.05554886e-01 -1.67732477e+00 1.64319074e+00
-2.46644109e-01 5.51353395e-01 3.54021043e-01 -1.29209816e+00
4.76139665e-01 1.84921354e-01 1.01988089e+00 -4.29601297e-02
1.28943324e-01 4.46219176e-01 -1.78953260e-01 -3.96301597e-01
7.31550753e-02 -5.45493662e-01 -1.40509203e-01 6.21030450e-01
2.09179759e-01 -2.86267638e-01 1.87721699e-01 3.86014789e-01
1.00899875e+00 -2.79763013e-01 -1.05713330e-01 -6.13861084e-01
1.96353763e-01 -4.19508189e-01 2.72016376e-01 5.82946002e-01
-2.87627816e-01 5.70063770e-01 8.22096705e-01 2.33782873e-01
-1.08914721e+00 -1.32151794e+00 -5.65706909e-01 3.46904188e-01
-6.28727227e-02 -2.71441549e-01 -1.04408848e+00 -2.17380062e-01
-2.80066393e-02 6.41172588e-01 -5.44103801e-01 -1.27887696e-01
-3.19540739e-01 -1.00718021e+00 6.52928889e-01 3.24547142e-01
3.11099470e-01 -1.73741117e-01 -1.96643561e-01 6.99429587e-02
-1.55024529e-01 -1.02531326e+00 -6.67501926e-01 4.04124618e-01
-9.39556837e-01 -9.47804689e-01 -1.14019883e+00 -6.35056198e-01
4.54325408e-01 4.66721244e-02 6.05565906e-01 -9.21075106e-01
1.19903861e-02 7.92772949e-01 1.28758773e-01 -3.86776596e-01
-6.98335245e-02 -1.69141307e-01 6.10281229e-01 6.62500039e-02
6.72689736e-01 -8.44636023e-01 -2.81758308e-01 2.30538517e-01
-1.05732453e+00 -5.11950850e-01 3.32467526e-01 5.25866568e-01
6.32610381e-01 4.03169781e-01 5.23361623e-01 -5.14875591e-01
6.94724143e-01 -6.71029806e-01 -7.43022263e-01 -2.68112928e-01
-4.79323447e-01 6.85611308e-01 4.16313678e-01 -5.45958877e-01
-2.26626575e-01 1.72777519e-01 -8.19117352e-02 -5.57667315e-01
-1.10198736e-01 1.43045425e-01 -2.80645370e-01 -1.67718381e-01
4.63407040e-01 2.51153648e-01 1.38181344e-01 -6.96789801e-01
4.32210535e-01 5.82495928e-01 5.90433359e-01 -7.29322493e-01
7.29036391e-01 6.59176409e-01 3.91639650e-01 -1.19348371e+00
-5.95348358e-01 -8.36822510e-01 -7.87105262e-01 1.98688433e-01
6.53914750e-01 -5.75886369e-01 -8.70154142e-01 5.55773973e-01
-1.08987749e+00 -2.84139782e-01 -5.17630696e-01 8.79362404e-01
-1.16949797e+00 5.85283875e-01 -7.02417791e-01 -1.21233046e+00
2.59637713e-01 -1.01549911e+00 1.27969587e+00 -1.41402453e-01
4.06788522e-03 -9.97203052e-01 7.37238765e-01 -1.66344732e-01
2.66017884e-01 1.05268419e-01 9.53497767e-01 -7.89479792e-01
-1.99271336e-01 -4.56379205e-01 1.04932405e-01 3.68911177e-01
-3.19597065e-01 -9.83358473e-02 -7.51179039e-01 -1.02879308e-01
8.63241076e-01 1.62830800e-01 9.02939379e-01 8.33217025e-01
8.92435014e-01 -4.26213771e-01 -3.57724696e-01 7.11567581e-01
1.27688587e+00 -1.40293568e-01 3.38465929e-01 5.73008806e-02
4.37675834e-01 5.74651301e-01 -4.03678305e-02 4.11061734e-01
2.95853674e-01 3.88880283e-01 -6.13733195e-03 4.40539807e-01
3.33048224e-01 1.66874856e-01 4.05441791e-01 1.07240665e+00
7.93996006e-02 2.28365377e-01 -8.14244270e-01 4.96183604e-01
-1.82064605e+00 -1.22713304e+00 -3.41968425e-02 2.53601313e+00
6.99682891e-01 6.49634376e-02 7.40248084e-01 5.38800776e-01
7.59532809e-01 -2.13753924e-01 -5.95793128e-01 -1.51356980e-01
-3.06705743e-01 1.74343854e-01 8.07331085e-01 7.91812420e-01
-8.93817782e-01 4.12925988e-01 6.86636972e+00 8.15594554e-01
-8.02887321e-01 2.40581706e-02 2.73625672e-01 -3.16061527e-01
-1.82149068e-01 -3.07992756e-01 -8.88927042e-01 4.53071564e-01
1.26910079e+00 -3.35813969e-01 5.35925686e-01 8.20403218e-01
-1.11475512e-01 -1.24226011e-01 -1.37127030e+00 1.47652161e+00
2.16407612e-01 -1.05996621e+00 -4.89370912e-01 6.53926909e-01
5.13409913e-01 9.56167206e-02 4.44426298e-01 7.41800070e-02
1.66551054e-01 -8.64168942e-01 5.61412990e-01 7.42986023e-01
5.54273188e-01 -1.00837433e+00 4.66914415e-01 7.09179103e-01
-1.03721750e+00 1.72362044e-01 -4.95463550e-01 5.00597060e-02
2.33945489e-01 1.01385772e+00 -4.58895922e-01 -2.36409381e-02
3.16412598e-01 6.46576524e-01 -3.14615786e-01 1.17567801e+00
7.76928782e-01 4.64771777e-01 -9.37282860e-01 1.03267267e-01
2.11137816e-01 -6.88542366e-01 8.36142182e-01 1.35549223e+00
7.01595962e-01 2.87103713e-01 -1.88358605e-01 6.87831163e-01
7.00545013e-02 7.11541697e-02 -8.45746756e-01 -3.39886844e-01
2.22089261e-01 9.93474901e-01 -9.23821270e-01 -3.20691884e-01
-1.98122129e-01 8.52072120e-01 7.13591203e-02 4.61663842e-01
-5.14922857e-01 -7.02654421e-01 8.69814217e-01 1.77639708e-01
5.68988979e-01 -5.71402848e-01 -2.56602138e-01 -1.26036513e+00
1.90577224e-01 -4.67657954e-01 8.31017792e-02 -3.11937213e-01
-1.56516743e+00 5.55006683e-01 3.02860647e-01 -1.04698730e+00
-6.15039349e-01 -6.00495040e-01 -2.59822637e-01 9.62300837e-01
-8.89335871e-01 -4.02908742e-01 6.07349813e-01 7.54937291e-01
1.43688709e-01 -6.32271962e-03 7.60022104e-01 2.47334853e-01
-4.63952243e-01 3.87846380e-01 6.97126210e-01 -1.29841380e-02
5.00873208e-01 -1.48679495e+00 2.38015831e-01 7.49981046e-01
1.78633928e-01 5.33647656e-01 1.04727161e+00 -4.11110610e-01
-1.42829204e+00 -8.02956283e-01 7.83377171e-01 -7.48979330e-01
9.57084596e-01 -4.38809484e-01 -7.35980153e-01 7.34154344e-01
-3.75396699e-01 -8.56831521e-02 1.03614819e+00 -1.13943681e-01
-2.38585919e-01 1.19735181e-01 -1.31243753e+00 4.62754995e-01
8.32741559e-01 -5.93313277e-01 -5.86514652e-01 4.28886056e-01
4.02357310e-01 -4.56083715e-02 -8.75214219e-01 7.05919117e-02
4.68056768e-01 -8.15184534e-01 9.04969633e-01 -7.50722110e-01
-1.30344005e-02 -1.98449150e-01 -7.60182202e-01 -1.21707261e+00
-3.65922332e-01 -8.31015646e-01 -3.30715090e-01 1.01718664e+00
2.86898583e-01 -5.49304605e-01 6.18155062e-01 7.24830508e-01
8.47174451e-02 -9.97504592e-01 -1.06692719e+00 -1.02525711e+00
5.30181110e-01 -8.89333367e-01 4.07849908e-01 8.14032555e-01
4.84595895e-01 2.98699200e-01 -1.06498420e-01 3.60375553e-01
1.23409438e+00 -1.15357593e-01 5.25941193e-01 -1.37933803e+00
-5.13378501e-01 -8.88023019e-01 -7.20334411e-01 -1.36966014e+00
4.14074659e-01 -9.33680773e-01 2.10183322e-01 -1.10838079e+00
-3.02556716e-03 -3.77479166e-01 -1.31703258e-01 -3.96311134e-01
3.36015493e-01 9.79282409e-02 6.36391193e-02 3.16474169e-01
-5.39001942e-01 3.52665335e-01 6.48710132e-01 6.37244135e-02
-2.80043960e-01 4.86280710e-01 -7.36208439e-01 1.01147294e+00
6.04879022e-01 -5.66811383e-01 -2.64027625e-01 -1.74498796e-01
1.98747709e-01 7.51933604e-02 1.61531761e-01 -1.24383473e+00
4.93250787e-01 1.61848187e-01 2.78294563e-01 -6.55081272e-01
6.00237966e-01 -6.95436299e-01 -2.34527066e-02 1.76507473e-01
-3.70207578e-01 1.55718878e-01 -6.43407255e-02 8.17847490e-01
1.73700467e-01 -4.35822427e-01 1.07911682e+00 -1.74018706e-03
3.37916672e-01 5.23529112e-01 -4.11020517e-01 3.92271101e-01
8.14387262e-01 -8.80815983e-02 -1.76896842e-03 -4.90454793e-01
-1.07579815e+00 -1.46209911e-01 5.33010960e-01 -4.49629039e-01
4.49133933e-01 -1.53747964e+00 -5.59935868e-01 3.27141970e-01
-2.18402311e-01 1.02035841e-02 1.01125538e-01 1.41101801e+00
-1.32459268e-01 5.04207611e-01 5.80833331e-02 -7.27166414e-01
-7.29141176e-01 8.69556308e-01 2.63514072e-01 1.10368140e-01
-4.47671205e-01 8.92581642e-01 -5.24047464e-02 8.06619897e-02
3.97798210e-01 -5.37003875e-01 9.31578279e-02 2.45387927e-01
8.27141702e-01 8.19426596e-01 -9.55431070e-03 -8.02591145e-01
-4.73832190e-01 9.04224217e-01 3.94735374e-02 -9.48792160e-01
1.56302595e+00 -2.96090156e-01 -2.30764955e-01 1.11425078e+00
1.86967516e+00 4.24192965e-01 -1.12106645e+00 -3.10329676e-01
1.20173134e-01 -1.10900082e-01 -2.25951806e-01 4.04637121e-02
-8.53523910e-01 1.01115191e+00 5.48873365e-01 6.10958636e-01
9.04401720e-01 6.18475616e-01 7.58570910e-01 5.09712756e-01
6.65114745e-02 -8.85404646e-01 -2.64242977e-01 5.09449422e-01
6.08936965e-01 -8.13369393e-01 -5.19837797e-01 1.90862119e-02
-2.37539694e-01 1.13529074e+00 -1.93785280e-01 -5.04785478e-01
1.25040340e+00 5.89744508e-01 -1.20655969e-01 -1.78150862e-01
-3.75632018e-01 -1.95157498e-01 1.31020620e-01 6.08180642e-01
2.19240800e-01 -6.60869032e-02 -8.23826119e-02 7.66199946e-01
-4.75178361e-01 -2.14134559e-01 6.41914308e-01 5.26046276e-01
-6.16739333e-01 -8.86448681e-01 -4.18855816e-01 4.99679893e-01
-5.19349456e-01 -2.53111776e-02 -1.83636144e-01 3.43452811e-01
-1.09010659e-01 8.24358284e-01 1.09106287e-01 -2.52028555e-01
1.52817279e-01 6.23140037e-01 7.67130315e-01 -4.57697541e-01
5.61561882e-01 5.54261386e-01 -5.84071279e-01 -5.11857688e-01
-4.55602944e-01 -1.17243397e+00 -1.06337738e+00 -4.80118483e-01
-1.43022984e-01 6.45426810e-01 1.05209732e+00 1.14193690e+00
1.93886250e-01 -1.09178290e-01 6.47891581e-01 -1.05981851e+00
-7.40941167e-01 -8.25882614e-01 -1.20602536e+00 1.30062932e-02
7.69596815e-01 -5.75121939e-01 -1.01907659e+00 -2.48416588e-02] | [7.193763256072998, 4.1971306800842285] |
d701cd03-ddc9-42d2-8c82-86180742aec9 | gfie-a-dataset-and-baseline-for-gaze | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Hu_GFIE_A_Dataset_and_Baseline_for_Gaze-Following_From_2D_to_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Hu_GFIE_A_Dataset_and_Baseline_for_Gaze-Following_From_2D_to_CVPR_2023_paper.pdf | GFIE: A Dataset and Baseline for Gaze-Following From 2D to 3D in Indoor Environments | Gaze-following is a kind of research that requires locating where the person in the scene is looking automatically under the topic of gaze estimation. It is an important clue for understanding human intention, such as identifying objects or regions of interest to humans. However, a survey of datasets used for gaze-following tasks reveals defects in the way they collect gaze point labels. Manual labeling may introduce subjective bias and is labor-intensive, while automatic labeling with an eye-tracking device would alter the person's appearance. In this work, we introduce GFIE, a novel dataset recorded by a gaze data collection system we developed. The system is constructed with two devices, an Azure Kinect and a laser rangefinder, which generate the laser spot to steer the subject's attention as they perform in front of the camera. And an algorithm is developed to locate laser spots in images for annotating 2D/3D gaze targets and removing ground truth introduced by the spots. The whole procedure of collecting gaze behavior allows us to obtain unbiased labels in unconstrained environments semi-automatically. We also propose a baseline method with stereo field-of-view (FoV) perception for establishing a 2D/3D gaze-following benchmark on the GFIE dataset. Project page: https://sites.google.com/view/gfie. | ['Jingtai Liu', 'Bohan Zhou', 'Dingye Yang', 'Xiaolin Zhai', 'Yuxue Yang', 'Zhengxi Hu'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['gaze-estimation'] | ['computer-vision'] | [ 1.01650052e-01 6.19574599e-02 -1.29623398e-01 -6.05229139e-01
-6.84915334e-02 -5.77209294e-01 3.25081944e-01 -2.76026964e-01
-5.44257283e-01 3.72092307e-01 5.45174778e-02 -9.80272964e-02
1.99567363e-01 -2.06608504e-01 -6.27104998e-01 -8.14381421e-01
4.63503063e-01 8.61300826e-02 1.34234980e-01 -1.96352210e-02
7.58311749e-01 4.21481162e-01 -2.24295473e+00 -2.35378325e-01
4.23152298e-01 9.45299029e-01 3.71208847e-01 5.66192567e-01
1.93642065e-01 5.43102264e-01 -5.35032392e-01 -5.07590473e-02
2.57367820e-01 -4.77053612e-01 -6.67700350e-01 2.38878012e-01
9.13602829e-01 -4.06308740e-01 3.68064374e-01 1.05703151e+00
4.90228504e-01 9.09228399e-02 4.32545930e-01 -1.70539761e+00
-5.32737195e-01 -2.43355751e-01 -1.03516579e+00 3.50461394e-01
1.14010584e+00 3.19717735e-01 5.91233552e-01 -7.03368306e-01
5.47339678e-01 9.68572319e-01 4.14302677e-01 1.01694226e+00
-8.95411611e-01 -8.09489608e-01 -1.40792474e-01 2.70055920e-01
-1.59249258e+00 -8.99536967e-01 8.59365046e-01 -7.83458471e-01
4.42507893e-01 3.74971062e-01 3.25191200e-01 1.09460902e+00
1.42759457e-01 3.63762170e-01 1.30167842e+00 -8.12307000e-01
1.01883285e-01 5.49202442e-01 3.36187929e-01 6.03414237e-01
1.84203416e-01 7.38604218e-02 -1.04361665e+00 1.36616439e-01
5.46148837e-01 3.36682826e-01 -6.99770987e-01 -4.14481014e-01
-1.00529802e+00 3.56531084e-01 6.03065133e-01 2.03615576e-01
-2.87396818e-01 -1.83258533e-01 -2.22229838e-01 -3.54626253e-02
5.02060831e-01 2.00090677e-01 -1.47037104e-01 -2.13573545e-01
-8.38025331e-01 5.04096001e-02 3.10073793e-01 1.11275339e+00
9.33993638e-01 -8.21786165e-01 -1.26689509e-01 3.19260061e-01
8.24815094e-01 7.96909153e-01 5.08882642e-01 -8.42144668e-01
2.32659400e-01 1.03679144e+00 5.29370427e-01 -8.43517482e-01
-4.42607194e-01 3.72316629e-01 -4.52417694e-02 7.07121432e-01
7.30679572e-01 -1.00140601e-01 -6.84892356e-01 1.50829720e+00
7.57276535e-01 -2.06896693e-01 -5.84794462e-01 1.41628802e+00
7.50933290e-01 1.11926317e-01 -1.40824869e-01 -4.00855273e-01
1.81370485e+00 -7.29536176e-01 -1.11767423e+00 -2.33389169e-01
6.77480578e-01 -7.98646152e-01 1.48624384e+00 3.59624058e-01
-8.21099758e-01 -5.95480621e-01 -7.65237391e-01 -2.96936512e-01
-3.89035672e-01 2.88648635e-01 2.06365094e-01 9.33730304e-01
-1.12439990e+00 -4.72209007e-02 -7.09780574e-01 -8.05476964e-01
2.18811646e-01 4.24720883e-01 -3.94737750e-01 2.49763831e-01
-6.34862959e-01 8.92039001e-01 -1.51903510e-01 3.02693009e-01
-4.23921198e-01 -1.22797847e-01 -9.18779612e-01 -1.96492329e-01
4.90753055e-01 -3.78161073e-01 1.30983031e+00 -1.05451989e+00
-1.55659378e+00 1.47077119e+00 -1.08274925e+00 1.20982349e-01
2.59735107e-01 -3.78498763e-01 -4.00279999e-01 -2.90437089e-03
3.76909673e-01 6.05248928e-01 1.04729295e+00 -1.19719946e+00
-8.19352090e-01 -1.08637440e+00 -1.58378407e-02 3.64595205e-01
-1.43630467e-02 4.20683175e-01 -4.23120558e-01 1.95302129e-01
1.14927344e-01 -1.07382011e+00 4.59865957e-01 1.06205299e-01
-5.72597265e-01 -7.09399760e-01 9.95982528e-01 -3.16064268e-01
1.41701305e+00 -2.10363793e+00 -2.30206877e-01 2.94697192e-03
6.21080101e-01 2.73615345e-02 6.27153873e-01 1.27289191e-01
-1.44331902e-01 3.58082093e-02 2.66150892e-01 -7.01413572e-01
-9.56438556e-02 -2.88937539e-01 5.37976548e-02 8.11927736e-01
-4.27150130e-01 5.98511517e-01 -7.42909372e-01 -6.82622492e-01
2.74919003e-01 2.63342500e-01 -3.27558927e-02 4.32885051e-01
9.55046117e-02 7.04403162e-01 -3.56315345e-01 8.28248322e-01
7.32788205e-01 -3.84229422e-01 -4.55505222e-01 -2.08391268e-02
-6.23304904e-01 2.55064309e-01 -9.09379244e-01 1.67551291e+00
-1.12091735e-01 9.88841772e-01 2.01763939e-02 -1.16469311e-02
6.99002683e-01 1.70039430e-01 1.33396506e-01 -7.61398315e-01
5.71679771e-01 -1.00193940e-01 -3.20808589e-01 -1.07416952e+00
5.77552974e-01 1.98355675e-01 1.65421367e-01 9.56468284e-01
-2.41450444e-01 4.45536494e-01 4.09194678e-02 -2.04889402e-01
6.36822581e-01 3.27325106e-01 3.81108373e-01 -1.56916946e-01
6.48674726e-01 -8.10048580e-02 1.39146343e-01 2.64283925e-01
-7.17840135e-01 7.20048666e-01 9.29935351e-02 -4.36500937e-01
-6.74287498e-01 -6.15023911e-01 -1.15932420e-01 1.31326544e+00
5.21854818e-01 -1.29962772e-01 -1.23988104e+00 -6.82608247e-01
-3.78223240e-01 8.27628136e-01 -9.76062477e-01 1.44656256e-01
-1.45969138e-01 -6.97740167e-02 1.53709248e-01 2.34974759e-05
4.30403084e-01 -1.16804421e+00 -1.46202135e+00 -7.29820073e-01
-3.63862872e-01 -7.13557184e-01 -7.48581588e-01 -2.84731209e-01
-3.68899882e-01 -1.45105684e+00 -6.36801958e-01 -4.78412151e-01
1.02683377e+00 7.63354897e-01 9.26819324e-01 -8.16916581e-03
5.37609570e-02 4.60538119e-01 -3.88659090e-01 -8.04883420e-01
1.24752373e-01 -2.50694528e-02 2.86684602e-01 3.06943208e-01
1.33763266e+00 1.42331779e-01 -8.96379352e-01 6.97227180e-01
-2.82728523e-01 -9.79982689e-02 -6.91000605e-04 1.19475275e-01
2.36258939e-01 -3.21934849e-01 -2.42208242e-01 -6.68930233e-01
5.13601542e-01 -2.31273860e-01 -7.97977388e-01 2.41223320e-01
-6.75768137e-01 -1.76891774e-01 -2.67196923e-01 -7.09437951e-02
-9.55077648e-01 1.37127936e-01 4.10435081e-01 -5.60948670e-01
-8.19966078e-01 -2.88010359e-01 -1.22596204e-01 -8.95295367e-02
1.15119731e+00 -1.27263531e-01 2.64784724e-01 -4.33023453e-01
-6.95321783e-02 1.28283465e+00 2.85945177e-01 5.10423370e-02
5.79498827e-01 7.64554858e-01 -2.31147528e-01 -6.66059494e-01
-1.12624717e+00 -1.00930202e+00 -9.22338188e-01 -6.93814635e-01
9.03962851e-01 -7.58524954e-01 -1.64989901e+00 5.98543644e-01
-1.04203248e+00 4.00662199e-02 1.41883582e-01 4.51369613e-01
-3.70298415e-01 -3.46888453e-02 2.50481308e-01 -1.04604149e+00
-3.11231881e-01 -1.07034266e+00 1.56728363e+00 7.55331159e-01
-3.81086349e-01 -6.94008172e-01 2.64346272e-01 6.83071613e-01
5.80629855e-02 1.33745313e-01 -1.49037957e-01 -1.05584696e-01
-3.39412242e-01 -2.83473972e-02 -3.33133116e-02 -4.85034706e-03
3.49220514e-01 4.05444689e-02 -1.58114469e+00 -5.38662039e-02
3.21571648e-01 -7.74608329e-02 1.06819786e-01 6.42505646e-01
8.51605356e-01 -3.75003479e-02 -7.63993561e-01 4.83103365e-01
1.17202330e+00 2.66991228e-01 6.35025799e-01 5.73576272e-01
8.47692668e-01 9.66157377e-01 9.71402109e-01 2.75491208e-01
4.96875912e-01 9.33112383e-01 4.50764269e-01 1.95361711e-02
4.66444306e-02 -8.45441669e-02 2.81904191e-01 -1.30197063e-01
-4.82897490e-01 -4.76918995e-01 -1.30568171e+00 3.57810378e-01
-1.54185641e+00 -9.31323946e-01 -5.90226352e-01 2.43604898e+00
6.82672560e-01 -2.25776836e-01 3.89745921e-01 1.80741042e-01
1.08054090e+00 -3.08052868e-01 -5.13453543e-01 -1.18893772e-01
5.40948033e-01 -3.45156640e-01 6.59067333e-01 4.08323824e-01
-8.14248562e-01 6.88737869e-01 5.44390011e+00 1.12670302e-01
-1.46234000e+00 1.49086505e-01 2.58153468e-01 -3.71063918e-01
3.12837005e-01 -3.28189403e-01 -1.17067444e+00 7.96674848e-01
8.79699051e-01 9.79215801e-02 2.90088952e-01 7.37864196e-01
7.05350280e-01 -8.11984003e-01 -1.41576171e+00 1.39339602e+00
4.82549936e-01 -5.64412653e-01 -7.50370324e-01 4.33473498e-01
1.49985909e-01 -5.75707220e-02 5.01271151e-02 -4.74580258e-01
-3.61987710e-01 -9.14743304e-01 8.01187456e-01 8.19853127e-01
1.16177630e+00 -2.87158698e-01 5.33565521e-01 4.76698250e-01
-7.65406311e-01 1.33504689e-01 4.21541408e-02 -4.48966920e-01
1.17030472e-01 2.06553824e-02 -1.31033385e+00 -1.29038632e-01
1.11022472e+00 5.81179142e-01 -9.54321265e-01 9.83999729e-01
-3.58542085e-01 3.65910500e-01 -3.36789101e-01 -2.62791723e-01
-2.83069015e-01 -2.49285430e-01 5.65951765e-01 4.96747404e-01
3.44744712e-01 -3.21951718e-03 -5.25285423e-01 9.48944569e-01
1.04501516e-01 -3.42797399e-01 -8.51217210e-01 5.97609699e-01
5.11409461e-01 1.12580466e+00 -6.94419026e-01 1.82334900e-01
-1.86022267e-01 8.43389809e-01 2.15299707e-02 4.57804471e-01
-8.00988376e-01 -4.06424075e-01 5.74446261e-01 7.56499350e-01
-9.92319733e-02 3.48121859e-02 -3.67611349e-01 -9.34147239e-01
2.95014560e-01 -5.24004221e-01 3.68885649e-03 -1.54409528e+00
-6.98635042e-01 5.54239273e-01 2.27410253e-02 -1.58015788e+00
-3.32773298e-01 -5.09458005e-01 -3.75658691e-01 1.17947257e+00
-1.24118590e+00 -9.10416126e-01 -1.08947670e+00 1.01408553e+00
3.10110033e-01 8.02388489e-02 6.98936939e-01 1.67348370e-01
-6.53498411e-01 5.67492604e-01 -3.00837994e-01 -4.54892851e-02
1.11125791e+00 -1.15769768e+00 -1.74198002e-02 8.02383065e-01
-4.76837233e-02 9.05491710e-01 9.46933806e-01 -3.67539257e-01
-1.06604004e+00 -4.39250648e-01 1.22443390e+00 -1.23107004e+00
2.15356976e-01 -5.39460540e-01 -6.21883929e-01 8.75590980e-01
3.66889834e-01 -1.92952424e-01 7.24190176e-01 2.94516176e-01
1.28652185e-01 -1.41312420e-01 -1.22474790e+00 4.72782522e-01
1.07176614e+00 -5.90704918e-01 -7.57016420e-01 2.18615234e-01
1.82108283e-01 -5.96268594e-01 -2.26430416e-01 -1.94952145e-01
6.74735844e-01 -1.34412766e+00 5.41822255e-01 1.01095811e-01
7.80034289e-02 -6.56431675e-01 3.85944396e-01 -9.09959257e-01
-2.29276018e-03 -6.83974743e-01 1.16434380e-01 1.26584995e+00
8.34017023e-02 -7.19932199e-01 8.05401921e-01 9.17320788e-01
2.15691298e-01 -2.26182953e-01 -6.42203927e-01 -1.99801460e-01
-8.32822919e-01 -2.58043081e-01 5.27676404e-01 6.97928607e-01
2.35703573e-01 5.39321721e-01 -8.66006687e-02 2.33924180e-01
6.77411556e-01 -1.50800258e-01 1.22313941e+00 -1.42574000e+00
6.28111780e-01 -1.49278060e-01 -4.48158562e-01 -1.03872430e+00
-8.07006061e-02 -1.50992453e-01 1.31740794e-01 -1.13984036e+00
-2.01754738e-02 -2.00116709e-01 1.20482482e-01 3.39097589e-01
-1.08516753e-01 2.54711062e-01 -8.35511833e-02 7.67079234e-01
-7.34310448e-01 1.33991927e-01 1.22010076e+00 4.17046905e-01
-3.15924644e-01 3.76817375e-01 -6.44331276e-01 8.81396592e-01
7.55381048e-01 -6.00055456e-01 -3.91882777e-01 -4.02818650e-01
5.09242892e-01 -3.04259747e-01 6.23685718e-01 -9.44853127e-01
7.93883204e-01 -9.20515731e-02 2.32315347e-01 -9.09648120e-01
3.04203451e-01 -1.17058015e+00 -6.79126531e-02 9.47825387e-02
-1.17966786e-01 7.58300275e-02 -6.62614927e-02 3.60172570e-01
-2.37021279e-02 -5.36934674e-01 6.01508081e-01 -7.04984814e-02
-6.34286225e-01 4.62537482e-02 -1.07248209e-01 -1.65619031e-01
1.40130925e+00 -7.66292155e-01 -5.30287743e-01 -1.75286576e-01
-5.53566396e-01 2.41412431e-01 1.09962654e+00 4.45306093e-01
3.47802162e-01 -1.03402781e+00 -1.38975456e-01 5.16494274e-01
4.79020178e-01 1.53698668e-01 6.58347383e-02 9.57751930e-01
-5.40515125e-01 6.56797528e-01 -3.36886764e-01 -1.18255246e+00
-1.71188462e+00 5.93292654e-01 3.99673671e-01 5.76975763e-01
-1.23323739e-01 9.46522593e-01 3.29813629e-01 -1.18748389e-01
3.22327465e-01 -3.35276753e-01 -6.77732587e-01 2.21564591e-01
9.18672562e-01 5.78520119e-01 -9.83748212e-03 -1.15303433e+00
-5.41837633e-01 9.75336015e-01 1.77389473e-01 2.81800032e-02
9.15145636e-01 -1.09828329e+00 -1.95597753e-01 8.34997058e-01
8.74758542e-01 2.29123026e-01 -1.12441254e+00 -2.15666220e-01
-1.41058415e-01 -8.64297926e-01 -1.53011447e-02 -5.81920683e-01
-6.07128143e-01 8.85621667e-01 9.74556923e-01 5.21961689e-01
1.27243567e+00 1.96945310e-01 2.09915638e-01 2.12572873e-01
5.66448748e-01 -1.01435697e+00 -3.08662266e-01 1.52644560e-01
7.59090006e-01 -1.79248726e+00 -7.38778859e-02 -3.20562452e-01
-6.41386092e-01 8.37131202e-01 8.06017756e-01 3.86137605e-01
7.06586421e-01 -1.40531242e-01 4.56508189e-01 -8.71025681e-01
-3.18265349e-01 -4.66585010e-01 4.33468670e-01 6.37984753e-01
4.14283603e-01 -4.56523746e-02 -2.29539908e-02 3.33869010e-02
-5.44853568e-01 2.27873117e-01 5.52402854e-01 8.27196479e-01
-4.33170408e-01 -5.27852416e-01 -9.02413726e-01 1.48740545e-01
-4.07003939e-01 1.86478376e-01 -4.86404181e-01 6.76856816e-01
3.91374439e-01 1.40457892e+00 3.49040866e-01 -1.63521633e-01
3.78012657e-01 1.04241848e-01 3.48177791e-01 -7.56186247e-01
-3.40297788e-01 -6.14755973e-02 -3.38339508e-01 -8.54859412e-01
-1.07135963e+00 -8.91957223e-01 -8.64623547e-01 -3.68495226e-01
-5.83584726e-01 -7.20165893e-02 8.31378102e-01 8.08156192e-01
4.82216567e-01 6.69803768e-02 6.29092574e-01 -1.19371521e+00
2.23477855e-01 -1.27923691e+00 -6.98076487e-01 4.75167394e-01
7.56466925e-01 -1.08901107e+00 -6.17731392e-01 6.33573592e-01] | [14.110374450683594, 0.11671550571918488] |
d1682a4b-7b2a-49ad-9ac6-2fef8bc64e54 | stereo-matching-with-color-weighted | 1708.07987 | null | http://arxiv.org/abs/1708.07987v2 | http://arxiv.org/pdf/1708.07987v2.pdf | Stereo Matching With Color-Weighted Correlation, Hierarchical Belief Propagation And Occlusion Handling | In this paper, we contrive a stereo matching algorithm with careful handling
of disparity, discontinuity and occlusion. This algorithm works a worldwide
matching stereo model which is based on minimization of energy. The global
energy comprises two terms, firstly the data term and secondly the smoothness
term. The data term is approximated by a color-weighted correlation, then
refined in obstruct and low-texture areas in many applications of hierarchical
loopy belief propagation algorithm. The results during the experiment are
evaluated on the Middlebury data sets, showing that out algorithm is the top
performer among all the algorithms listed there | ['Vamshhi Pavan Kumar Varma Vegeshna'] | 2017-08-26 | null | null | null | null | ['occlusion-handling'] | ['computer-vision'] | [ 1.08102486e-01 -3.21887314e-01 -1.36022661e-02 -6.07104182e-01
-2.58658171e-01 5.81593849e-02 7.44784534e-01 -1.08047398e-02
-6.61407888e-01 8.37426484e-01 2.93490320e-01 9.34440494e-02
-7.07301944e-02 -7.83157229e-01 -3.64119112e-01 -7.89602339e-01
-2.09198566e-03 6.65711999e-01 8.00421894e-01 -3.40163559e-01
9.09091651e-01 5.95097959e-01 -2.05183816e+00 3.66256386e-01
8.42694819e-01 8.19851398e-01 5.03742695e-02 5.20864367e-01
-1.03674062e-01 8.52029324e-01 5.97208142e-02 -4.14323658e-01
4.76688474e-01 -2.21650466e-01 -7.81659544e-01 2.34506026e-01
8.58913004e-01 -1.59653544e-01 -3.34690541e-01 1.45778358e+00
6.37576759e-01 2.81318903e-01 6.79319501e-01 -1.01295984e+00
-2.31499467e-02 -1.66478857e-01 -7.57627308e-01 3.15379322e-01
6.79740787e-01 -1.00519925e-01 6.63420737e-01 -9.97567832e-01
8.56350660e-01 1.49278533e+00 9.18670774e-01 1.74960345e-01
-1.34387255e+00 -1.89968988e-01 -2.85936326e-01 5.39519310e-01
-1.23003399e+00 -3.60236496e-01 7.81923056e-01 -6.50519788e-01
1.05255795e+00 3.54546219e-01 7.95479238e-01 2.29771882e-01
4.72101688e-01 5.77501178e-01 1.59741402e+00 -7.07342863e-01
2.50549197e-01 -8.43378305e-02 4.02689129e-01 9.13435400e-01
5.91749735e-02 9.43822563e-01 -7.01710463e-01 -4.99879807e-01
8.70956540e-01 -4.06349860e-02 -2.95467764e-01 -6.17885530e-01
-7.43219137e-01 8.50231647e-01 3.35942715e-01 2.84679621e-01
-3.39479715e-01 -9.49733555e-02 -1.73154566e-02 1.34914592e-01
5.96532941e-01 -2.99428016e-01 -1.60676435e-01 2.24739403e-01
-1.05556571e+00 2.80894905e-01 7.32670486e-01 7.68575490e-01
1.07028806e+00 -1.95638925e-01 1.12283073e-01 1.07942986e+00
6.77715600e-01 3.57073694e-01 2.92143166e-01 -1.00767803e+00
2.06487104e-01 6.59243107e-01 -2.46511046e-02 -1.29197645e+00
-3.19079757e-01 6.97636828e-02 -7.97196448e-01 1.00948143e+00
2.66804874e-01 8.94808471e-02 -1.16784573e+00 1.11716652e+00
4.99165952e-01 1.15651768e-02 -3.06241959e-01 9.64758933e-01
9.83258605e-01 5.51603913e-01 -8.05617645e-02 -4.87856232e-02
1.03098774e+00 -1.01433611e+00 -6.49918079e-01 -1.42867966e-02
-3.53175029e-02 -1.19084716e+00 1.14546694e-01 5.14626741e-01
-1.26566613e+00 -8.97722900e-01 -1.07175052e+00 -1.84373975e-01
-3.36207092e-01 -5.87662160e-01 6.85699403e-01 5.44921100e-01
-1.59485960e+00 8.70435297e-01 -4.04836506e-01 -5.57367384e-01
-1.73613846e-01 3.91485304e-01 -5.00962734e-01 -4.78519835e-02
-8.99578393e-01 1.15579987e+00 4.53530014e-01 2.95727789e-01
-2.18383849e-01 -1.07960366e-02 -8.28653932e-01 -5.26582658e-01
-5.06525218e-01 -9.65651989e-01 7.02282250e-01 -1.01591086e+00
-1.49277997e+00 1.50597167e+00 -5.11134565e-01 -1.67353258e-01
9.60595906e-01 -9.43550840e-02 -2.76445150e-01 1.76735800e-02
-1.27831414e-01 7.73782849e-01 9.31113780e-01 -1.61333096e+00
-1.06595719e+00 -4.70450252e-01 -2.18282700e-01 4.27324086e-01
4.45589364e-01 -2.01819479e-01 -7.76915014e-01 -5.50651610e-01
8.81663561e-01 -6.56765878e-01 -3.31161588e-01 -6.50740564e-02
-1.58828840e-01 -3.62320729e-02 5.76536655e-01 -7.95495093e-01
1.18184519e+00 -2.04909778e+00 1.08780645e-01 6.93055391e-01
8.93343091e-02 -2.48742312e-01 4.03367013e-01 4.01288599e-01
-1.61057085e-01 -4.54525501e-01 -2.29926318e-01 -8.67652819e-02
-2.06901357e-01 2.65201807e-01 2.70201027e-01 7.82619417e-01
-3.90570104e-01 2.39077136e-01 -7.10892498e-01 -1.09429562e+00
6.58038318e-01 4.89527166e-01 -5.53583324e-01 2.38018870e-01
2.80871540e-01 2.88158089e-01 -2.28517011e-01 7.33260155e-01
1.14113092e+00 2.35700056e-01 -1.69608891e-01 -4.27949220e-01
-5.43383598e-01 -2.82515645e-01 -1.67733765e+00 1.58635676e+00
3.50209251e-02 8.10205162e-01 2.25339010e-01 -5.61217070e-01
1.05148721e+00 5.96185997e-02 3.98718566e-01 -9.65186894e-01
1.15482979e-01 2.30403408e-01 -3.78436834e-01 -6.63483858e-01
6.58028603e-01 -3.05920541e-01 6.27897739e-01 -1.31783977e-01
-1.58268973e-01 -3.52451652e-01 1.27560839e-01 -2.29271427e-01
6.47857547e-01 4.63627189e-01 2.54035443e-01 -8.54238868e-01
8.80560458e-01 3.63185912e-01 5.49848437e-01 7.38949418e-01
-3.71908337e-01 8.93208802e-01 -1.45504311e-01 -8.13308597e-01
-1.07503521e+00 -1.04944372e+00 -6.13334656e-01 8.00788879e-01
8.37290347e-01 2.88361967e-01 -4.90102530e-01 -4.15638648e-02
2.42112786e-01 1.75101846e-01 -3.52690101e-01 4.34574753e-01
-6.03076339e-01 -7.08833337e-01 -5.31297810e-02 1.49609938e-01
9.10404265e-01 -9.81550634e-01 -5.40484846e-01 2.38954246e-01
-2.56545901e-01 -6.94738507e-01 -2.39482716e-01 9.58438143e-02
-1.31337547e+00 -1.01767480e+00 -7.58660376e-01 -1.19479990e+00
7.19425261e-01 2.30020523e-01 1.49693096e+00 4.07691747e-01
-2.31910408e-01 1.93795398e-01 2.11578589e-02 6.38633519e-02
-1.44565552e-01 -5.50633669e-01 -4.29255158e-01 -1.42834261e-02
6.85468137e-01 -4.31195378e-01 -6.74489439e-01 4.41579461e-01
-3.83667767e-01 3.19262475e-01 2.64852703e-01 9.60837662e-01
8.78409803e-01 7.86158293e-02 -7.81610548e-01 -7.47472644e-01
2.21872464e-01 2.98089553e-02 -1.09613323e+00 -4.71722670e-02
-5.52420199e-01 1.57682240e-01 -1.86503917e-01 1.44041389e-01
-1.32406569e+00 1.80637255e-01 -2.61665821e-01 1.18961483e-01
-4.06209528e-01 3.36561650e-02 1.77467689e-01 -5.69997251e-01
4.69106644e-01 1.07354425e-01 -1.80045396e-01 -4.96070474e-01
4.18557785e-02 5.41659057e-01 6.34748697e-01 -4.73126441e-01
7.08413482e-01 9.69427943e-01 1.58343971e-01 -1.14251852e+00
-1.39619276e-01 -9.41823483e-01 -6.68167889e-01 -6.63310945e-01
1.08124888e+00 -7.21392751e-01 -4.84755576e-01 9.69219506e-01
-1.08504736e+00 -2.57396661e-02 1.32637560e-01 7.11830676e-01
-6.70237958e-01 6.50249302e-01 -8.77176583e-01 -8.20667803e-01
-1.61521912e-01 -9.28991199e-01 9.52689350e-01 3.49511862e-01
1.29881486e-01 -1.29510975e+00 7.37014651e-01 4.11727905e-01
3.57767791e-01 2.65144527e-01 8.36810291e-01 1.99438203e-02
-7.10792005e-01 -1.54422700e-01 -4.13703322e-01 1.99411526e-01
-1.88986212e-01 2.33732849e-01 -1.27529490e+00 -5.38334697e-02
1.42683253e-01 -8.61301497e-02 1.01472509e+00 7.53366888e-01
6.12132847e-01 5.04186571e-01 -1.70874670e-01 9.40476358e-01
2.16500783e+00 4.75409865e-01 9.81235206e-01 7.72962689e-01
5.65835655e-01 7.53771842e-01 6.02175951e-01 2.09679902e-01
3.22507471e-01 6.22310460e-01 4.17982519e-01 -5.81941068e-01
-1.62218720e-01 -1.43858287e-02 -5.12237959e-02 7.30222404e-01
-6.67939901e-01 8.48326087e-02 -9.24046516e-01 2.49555543e-01
-2.04124546e+00 -1.10934997e+00 -8.11384559e-01 2.28572726e+00
5.46843171e-01 3.51314783e-01 -1.10008828e-01 2.56426543e-01
9.80322063e-01 1.29171878e-01 3.03467996e-02 -7.07220316e-01
-2.80164480e-01 1.51342094e-01 7.50734031e-01 1.19047832e+00
-1.37416875e+00 6.47938788e-01 7.49647856e+00 7.34134674e-01
-8.23463500e-01 -1.23552352e-01 4.45751637e-01 7.56467223e-01
-2.09864885e-01 1.83314055e-01 -6.52548969e-01 5.12998641e-01
2.10604236e-01 2.57718295e-01 1.82241410e-01 5.89191496e-01
1.79406315e-01 -9.26729739e-01 -9.48644698e-01 1.22544146e+00
1.01408593e-01 -8.79706740e-01 -4.87277396e-02 -1.05048642e-02
9.23833668e-01 8.19231644e-02 -3.21364194e-01 -1.61352545e-01
3.04455787e-01 -9.26882982e-01 7.50478327e-01 8.00331950e-01
7.84839690e-02 -5.06989300e-01 1.05634451e+00 3.91674452e-02
-1.15021789e+00 4.80128795e-01 -4.89044905e-01 -2.77562857e-01
1.77528381e-01 7.48032153e-01 1.44251864e-02 5.38405001e-01
8.50015402e-01 6.53167307e-01 -2.81848967e-01 1.66730380e+00
-7.62936994e-02 -1.66909337e-01 -4.49664146e-01 1.61191985e-01
3.15435022e-01 -8.27865481e-01 5.20431578e-01 1.39878559e+00
9.63103250e-02 1.27263412e-01 2.94354975e-01 2.78451115e-01
5.88793993e-01 1.61621317e-01 -6.27217114e-01 9.98970866e-01
1.21982969e-01 7.85945058e-01 -6.09405398e-01 -4.09469932e-01
-5.48329592e-01 1.09224224e+00 6.01016507e-02 3.80591750e-01
-5.25036752e-01 -5.71534745e-02 2.28354886e-01 4.68693022e-03
1.53871536e-01 1.01824487e-02 -5.38368165e-01 -9.29744542e-01
-8.00920203e-02 -4.89625186e-01 5.56579590e-01 -9.33625638e-01
-1.42669725e+00 5.47512710e-01 1.07840553e-01 -1.34323919e+00
-1.01225354e-01 -6.02499127e-01 -4.80777800e-01 1.17140245e+00
-1.75199628e+00 -7.65612483e-01 -5.15460610e-01 9.28157508e-01
5.07051289e-01 -1.05425887e-01 6.08006239e-01 6.31269813e-01
-1.77576780e-01 -1.34014338e-01 1.39812410e-01 -2.51348525e-01
4.95213330e-01 -1.36206317e+00 -7.68768564e-02 8.46915007e-01
-2.75021195e-01 1.56253755e-01 1.09271944e+00 -6.12044275e-01
-9.62832153e-01 -2.98342764e-01 1.28035736e+00 -1.10205330e-01
3.29401165e-01 3.85336041e-01 -8.12315345e-01 2.96362400e-01
5.18555939e-01 -2.38720343e-01 2.75536478e-01 -6.21779896e-02
-2.11888831e-02 -2.56891418e-02 -1.39403784e+00 1.66356415e-01
9.15300310e-01 -5.32784402e-01 -1.02246559e+00 9.15982351e-02
-4.05620426e-01 -5.63822746e-01 -7.53066361e-01 4.93278980e-01
6.30279958e-01 -1.90381563e+00 1.02058971e+00 -3.71114835e-02
1.65455580e-01 -5.14313340e-01 -3.23231339e-01 -9.41050768e-01
-7.30537355e-01 -3.29405218e-01 3.95412117e-01 1.01229298e+00
2.51173705e-01 -5.56480229e-01 1.01372027e+00 4.96658593e-01
1.51699290e-01 -1.99760422e-01 -1.05853415e+00 -5.57585239e-01
-3.23006004e-01 -2.34009549e-01 -5.88287227e-03 9.02244329e-01
-1.49282247e-01 7.96262827e-03 -5.21913648e-01 9.52567253e-03
1.34932745e+00 2.12328091e-01 5.23970187e-01 -1.38102806e+00
-2.88916916e-01 -5.39679408e-01 -8.79677117e-01 -1.04608274e+00
-3.45899433e-01 -4.07460898e-01 3.11766952e-01 -1.55035293e+00
4.46223825e-01 -4.11297292e-01 -1.36270970e-01 -2.91177660e-01
-2.88654923e-01 3.65543127e-01 -7.83625618e-02 2.98997134e-01
-2.34867126e-01 1.58961236e-01 1.12872279e+00 -1.18220441e-01
-1.68934748e-01 -2.86382716e-02 3.63825977e-01 1.30815852e+00
5.57051778e-01 -3.76947701e-01 -3.41032855e-02 -5.24798334e-01
-7.53036737e-02 1.45160155e-02 2.71047741e-01 -1.16495299e+00
4.67650741e-01 -1.29570752e-01 3.67967755e-01 -1.07951713e+00
4.69902635e-01 -1.27815509e+00 5.05210102e-01 7.62006879e-01
1.34966686e-01 3.15029770e-01 -9.44955572e-02 3.81415099e-01
-7.91190386e-01 -3.05362433e-01 1.27176142e+00 -2.41996571e-01
-1.61745954e+00 8.16083848e-02 -1.38443023e-01 -1.11609593e-01
8.39685619e-01 -9.96559143e-01 2.21029092e-02 -2.73931473e-01
-8.03681672e-01 1.79122478e-01 7.16978908e-01 3.33513245e-02
6.88277960e-01 -1.48244381e+00 -7.88245261e-01 5.03213763e-01
-7.60354241e-03 -3.83521587e-01 2.52224833e-01 7.62191117e-01
-1.38549185e+00 1.05197795e-01 -5.74761212e-01 -1.02559912e+00
-1.56063545e+00 3.97733897e-01 5.60726345e-01 -1.31881222e-01
-6.00813150e-01 1.01050639e+00 5.07267676e-02 -1.34915590e-01
5.96763611e-01 -2.50236522e-02 -5.95583141e-01 -1.20363757e-02
1.64782479e-01 8.56575012e-01 -1.35187781e-03 -9.56200838e-01
-4.32922632e-01 1.33863926e+00 3.88209909e-01 -2.55659193e-01
9.73889649e-01 -3.50293696e-01 -4.84362960e-01 3.04205239e-01
1.06445777e+00 -9.25656930e-02 -1.08349037e+00 -5.35103008e-02
2.02463642e-01 -9.74785745e-01 4.63822395e-01 -3.86327267e-01
-1.00139415e+00 6.47522748e-01 1.33720040e+00 2.98901588e-01
1.15571773e+00 -5.02057076e-01 4.83443618e-01 -1.23134121e-01
7.08114564e-01 -1.43919396e+00 -5.90638518e-01 5.33162177e-01
6.41066611e-01 -1.48229957e+00 2.04461947e-01 -5.80289900e-01
-3.75017434e-01 1.33017075e+00 3.86111766e-01 -4.53611881e-01
1.17619681e+00 2.57233202e-01 3.22249144e-01 -2.71188438e-01
-3.81934017e-01 -4.43990678e-01 5.01931131e-01 6.77464068e-01
5.55013895e-01 -1.51399925e-01 -8.26404870e-01 -7.01929748e-01
1.31418839e-01 1.82561442e-01 -4.82877120e-02 1.15062094e+00
-9.76621866e-01 -9.75638092e-01 -8.35358202e-01 9.59972963e-02
-3.43454123e-01 -1.67747542e-01 -1.24220386e-01 8.21571589e-01
5.40668368e-01 1.05244565e+00 1.68793887e-01 -3.44284058e-01
5.67890108e-01 -2.53781766e-01 5.33686399e-01 3.62710468e-02
-6.57428443e-01 4.05491710e-01 2.88511634e-01 -8.85883629e-01
-1.12864506e+00 -5.37285686e-01 -8.24501693e-01 -7.93972611e-01
-2.18809560e-01 9.96268168e-02 6.21372402e-01 7.38027334e-01
-4.34915185e-01 -6.11145422e-02 5.76306343e-01 -1.34397399e+00
5.04935794e-02 -6.06688738e-01 -7.60885715e-01 6.91124439e-01
3.79528493e-01 -6.07461452e-01 -5.15792966e-01 2.70572931e-01] | [9.092700958251953, -2.4267141819000244] |
a1982bbe-46c2-45ee-8a1f-63032c4c9695 | 190807816 | 1908.07816 | null | https://arxiv.org/abs/1908.07816v3 | https://arxiv.org/pdf/1908.07816v3.pdf | A Multi-Turn Emotionally Engaging Dialog Model | Open-domain dialog systems (also known as chatbots) have increasingly drawn attention in natural language processing. Some of the recent work aims at incorporating affect information into sequence-to-sequence neural dialog modeling, making the response emotionally richer, while others use hand-crafted rules to determine the desired emotion response. However, they do not explicitly learn the subtle emotional interactions captured in human dialogs. In this paper, we propose a multi-turn dialog system aimed at learning and generating emotional responses that so far only humans know how to do. Compared with two baseline models, offline experiments show that our method performs the best in perplexity scores. Further human evaluations confirm that our chatbot can keep track of the conversation context and generate emotionally more appropriate responses while performing equally well on grammar. | ['Yubo Xie', 'Pearl Pu', 'Ekaterina Svikhnushina'] | 2019-08-15 | null | null | null | null | ['open-domain-dialog'] | ['natural-language-processing'] | [-2.44093835e-01 6.12979472e-01 1.49697915e-01 -7.66268015e-01
-4.12461162e-01 -6.98366106e-01 5.69263756e-01 -1.47831067e-01
-2.54476756e-01 9.22530949e-01 4.91239756e-01 -4.11272943e-02
7.07667053e-01 -6.68641269e-01 1.23003628e-02 -3.10262918e-01
2.92816460e-01 7.70372927e-01 -3.04202527e-01 -9.67139423e-01
7.30786175e-02 4.37454991e-02 -8.38367462e-01 4.97273356e-01
9.02501941e-01 6.45838380e-01 -5.50646596e-02 1.03025150e+00
-4.48922753e-01 1.67343462e+00 -8.56756091e-01 -8.21161211e-01
-2.29747280e-01 -8.55357528e-01 -1.43707955e+00 -1.08219430e-01
-3.38496059e-01 -4.33143288e-01 -9.89892855e-02 7.69101083e-01
4.96746689e-01 5.28032899e-01 5.01158535e-01 -1.13113534e+00
-7.86157250e-01 9.57069278e-01 7.70511255e-02 -3.08619440e-01
6.86845481e-01 5.47851503e-01 1.17897415e+00 -5.41476667e-01
6.89554155e-01 1.62950933e+00 3.84080827e-01 1.32078207e+00
-1.09934723e+00 -4.94354725e-01 1.96401492e-01 1.07215166e-01
-6.98384345e-01 -3.76998007e-01 9.88470256e-01 -3.31154019e-01
1.36059403e+00 3.74843687e-01 4.06568855e-01 1.65541124e+00
7.18986019e-02 9.28833961e-01 1.19542277e+00 -3.61873955e-01
2.00379491e-01 4.77402240e-01 3.15872639e-01 4.67876285e-01
-8.49272072e-01 -3.36900838e-02 -3.96554291e-01 -2.65141249e-01
3.90482396e-01 -3.55178088e-01 -7.49585405e-02 2.39785597e-01
-8.61941099e-01 1.34782863e+00 2.31650651e-01 5.38107455e-01
-5.15497565e-01 -5.00558503e-02 6.02327049e-01 6.49644196e-01
6.84165835e-01 1.03126442e+00 -4.74643677e-01 -5.82208276e-01
-2.66418785e-01 2.92301238e-01 1.54008710e+00 7.85678625e-01
7.77229726e-01 1.18128106e-01 -5.44739962e-01 1.19831002e+00
1.63938981e-02 2.76077986e-01 4.56431985e-01 -1.33433628e+00
1.43498361e-01 5.82645655e-01 3.79447401e-01 -6.47306859e-01
-5.38355291e-01 4.53614593e-01 -5.92738628e-01 5.18807843e-02
5.14330089e-01 -1.01443148e+00 -2.61350513e-01 1.91548741e+00
1.96149901e-01 -4.36140090e-01 4.02890056e-01 9.66985226e-01
8.54514182e-01 9.12266433e-01 4.37210649e-01 -3.38779926e-01
1.21048260e+00 -1.11002159e+00 -1.13962162e+00 -5.07930279e-01
7.12290585e-01 -5.92439532e-01 1.23818922e+00 4.89063859e-01
-1.15573752e+00 -2.29325235e-01 -5.18010497e-01 -6.57416061e-02
-9.00184512e-02 -1.80492803e-01 9.16033864e-01 5.62417626e-01
-1.07438302e+00 2.54186809e-01 -3.83050978e-01 -2.53987104e-01
-1.90287456e-01 2.86863416e-01 4.50658463e-02 3.99452925e-01
-1.75454652e+00 1.42054594e+00 7.79363289e-02 -2.02346340e-01
-5.10011852e-01 -2.75130659e-01 -9.22901154e-01 2.12149307e-01
3.75804007e-01 -3.57311726e-01 2.04699326e+00 -1.36492932e+00
-2.53281927e+00 7.73325086e-01 -1.06700614e-01 -5.07913589e-01
2.55742580e-01 -1.27162516e-01 -1.16033696e-01 8.67636874e-02
-5.86458445e-01 1.15036929e+00 3.64022464e-01 -1.07067740e+00
-2.03869656e-01 2.12695934e-02 5.25539577e-01 2.98504084e-01
-3.78828377e-01 5.72811365e-01 1.06705084e-01 -3.20532113e-01
-7.47560859e-01 -1.11181140e+00 -5.35513401e-01 -5.35735488e-01
-2.29585499e-01 -7.89189994e-01 4.38361287e-01 -6.70782626e-01
1.07075381e+00 -1.65119910e+00 2.51952738e-01 -1.88206300e-01
1.30655557e-01 2.48045608e-01 -2.04539686e-01 8.12758923e-01
1.93163499e-01 2.17440166e-02 6.20889328e-02 -2.48976812e-01
4.27419841e-01 3.49935442e-01 -3.63141954e-01 -2.96712279e-01
4.85110521e-01 1.12719488e+00 -9.47857082e-01 -5.19964099e-01
2.21620664e-01 2.66951621e-01 -7.60354161e-01 1.00261009e+00
-7.87654877e-01 7.73820698e-01 -5.24188757e-01 1.00785591e-01
2.17613220e-01 -1.08717494e-02 5.27297437e-01 5.17056346e-01
1.03546798e-01 4.38620359e-01 -3.56027126e-01 1.40214777e+00
-1.03368080e+00 4.36492831e-01 2.62120515e-01 -4.81153101e-01
1.29518449e+00 7.44874358e-01 2.15579212e-01 -5.45362532e-01
4.88969594e-01 -2.76791155e-01 3.49649668e-01 -6.70074761e-01
6.35317385e-01 -5.96838295e-01 -5.65945029e-01 8.40143144e-01
1.00746930e-01 -3.76679182e-01 5.31139132e-03 4.33715552e-01
8.98549557e-01 -7.12815076e-02 4.92720723e-01 1.04991682e-01
5.44309378e-01 1.61066018e-02 4.22020942e-01 5.04580736e-01
-4.97277886e-01 1.04734950e-01 8.05558860e-01 -4.22958940e-01
-6.48762763e-01 -4.22356397e-01 6.92185402e-01 1.59279668e+00
-1.09179750e-01 -2.08238780e-01 -1.12538266e+00 -6.19820595e-01
-5.61321437e-01 1.04751778e+00 -4.76124763e-01 -2.18603268e-01
-4.93460566e-01 -3.72388959e-01 8.46330345e-01 3.43980521e-01
3.21615726e-01 -1.79308629e+00 -5.66092014e-01 4.38753456e-01
-8.41634274e-01 -1.08814454e+00 -5.00512362e-01 1.48802608e-01
-4.36232954e-01 -5.19071400e-01 -3.74314308e-01 -6.87648237e-01
1.88469946e-01 -2.87714452e-01 1.30036211e+00 9.51413438e-02
3.18204686e-02 2.99613744e-01 -6.02600038e-01 -5.29278815e-01
-1.09913802e+00 4.39696796e-02 -2.24366352e-01 -9.74065959e-02
7.78455734e-01 -4.64201331e-01 -2.17273504e-01 3.70183706e-01
-4.47147220e-01 7.18983635e-02 4.19680446e-01 8.97182822e-01
-5.20032287e-01 -8.32385838e-01 8.87041926e-01 -9.54823613e-01
1.57242012e+00 -4.28673863e-01 5.68484627e-02 2.62113452e-01
-3.01562309e-01 2.87708908e-01 9.63969946e-01 -6.13837600e-01
-1.80422425e+00 8.54193941e-02 -4.22104925e-01 -1.88764721e-01
-6.33508801e-01 1.94703624e-01 -1.19608216e-01 1.64245397e-01
7.81917989e-01 -2.04438373e-01 1.30777761e-01 -6.37678951e-02
6.36311889e-01 9.96436536e-01 3.25512052e-01 -8.23117495e-01
2.40123957e-01 -1.60009414e-01 -7.45258808e-01 -5.20576179e-01
-8.18692625e-01 -3.26774925e-01 -2.67058522e-01 -5.55635214e-01
1.10227728e+00 -6.81066573e-01 -1.40442932e+00 3.82155806e-01
-1.65214288e+00 -8.79699886e-01 5.90881146e-02 1.35341451e-01
-6.67118847e-01 2.22149715e-01 -1.33974683e+00 -1.38048983e+00
-5.91551065e-01 -8.61920238e-01 7.89220035e-01 3.26592386e-01
-1.22764802e+00 -1.02820659e+00 4.21836197e-01 5.53534150e-01
5.31209111e-01 3.85809653e-02 7.84653366e-01 -1.00839818e+00
5.08613288e-02 1.11026250e-01 2.31072128e-01 3.31239700e-01
-9.35863554e-02 -1.56802729e-01 -1.13861692e+00 3.73136699e-01
2.79005557e-01 -1.15209806e+00 2.57991940e-01 -4.75420095e-02
7.10187376e-01 -7.65714288e-01 2.54864693e-01 -1.24727048e-01
5.74234784e-01 5.87684393e-01 6.50308609e-01 -2.75370508e-01
2.00181901e-01 1.20646775e+00 8.56207550e-01 6.94259405e-01
6.02744341e-01 6.29243433e-01 1.92635342e-01 -3.33470069e-02
4.27103192e-01 -1.66008204e-01 7.42097080e-01 6.85605168e-01
6.58197654e-03 -4.58319724e-01 -6.90486491e-01 3.27662766e-01
-1.93651628e+00 -1.13452733e+00 -1.32390559e-01 1.38471234e+00
1.44247985e+00 -2.02870473e-01 1.97659716e-01 -4.24443811e-01
6.04266763e-01 2.80203253e-01 -4.66971755e-01 -1.42715359e+00
2.69290917e-02 4.97944474e-01 -2.04377443e-01 9.79023933e-01
-6.54004455e-01 1.43880665e+00 6.20281219e+00 3.66526812e-01
-1.08383787e+00 5.93850203e-02 9.27532732e-01 -1.10201716e-01
-1.60825923e-01 -2.81717181e-02 -2.82538146e-01 1.28273562e-01
1.19407558e+00 -2.41687540e-02 7.12090552e-01 9.09011364e-01
4.65464115e-01 2.63491441e-02 -1.28580165e+00 6.62146151e-01
-1.45813264e-02 -8.04164469e-01 -3.50972772e-01 -2.84654379e-01
4.18521523e-01 -4.59213525e-01 -3.37786943e-01 9.66861308e-01
1.04773939e+00 -1.09169412e+00 3.65540028e-01 4.49254870e-01
2.49411836e-01 -8.07137311e-01 7.31650591e-01 6.46555901e-01
-4.54981089e-01 -6.24992289e-02 -1.15320876e-01 -7.55371273e-01
3.51586193e-01 1.57713797e-02 -1.41744506e+00 7.19242468e-02
2.69662470e-01 1.54927701e-01 6.09077571e-04 -4.38800268e-02
-6.65234745e-01 8.09755087e-01 8.71941149e-02 -8.48565042e-01
2.79146522e-01 -3.09120029e-01 3.35832179e-01 1.44358182e+00
-2.14622587e-01 6.78844452e-01 3.25311154e-01 1.05026305e+00
-2.06977546e-01 2.85002947e-01 -4.48424220e-01 -2.50665188e-01
2.99179018e-01 1.49006045e+00 -1.94634110e-01 -4.56300646e-01
-2.28162661e-01 1.10308766e+00 5.47499299e-01 1.83478653e-01
-8.90954494e-01 -3.36648166e-01 8.22418869e-01 -5.99355519e-01
-1.51832849e-01 -3.90020832e-02 -2.85340250e-01 -1.02470243e+00
-3.43906701e-01 -1.25435579e+00 2.38941938e-01 -1.05288434e+00
-1.58691955e+00 8.22900891e-01 -2.73893148e-01 -5.13644755e-01
-1.02252948e+00 -4.80754375e-01 -1.05116093e+00 8.72301996e-01
-9.55976069e-01 -9.78322387e-01 -1.59039527e-01 4.29757297e-01
8.93335044e-01 1.92851171e-01 1.22344935e+00 -1.66904286e-01
-4.14931208e-01 5.12972593e-01 -6.11606002e-01 3.88831049e-01
1.10791361e+00 -1.28528094e+00 2.21331269e-01 1.69554606e-01
-2.42478102e-01 6.06259882e-01 1.10711515e+00 -3.74593914e-01
-1.18097281e+00 -6.12188876e-01 1.12121212e+00 -4.73380893e-01
7.48396039e-01 -5.98087609e-01 -9.89436030e-01 6.09610558e-01
1.06143689e+00 -7.69420743e-01 8.61402810e-01 4.28817272e-01
-2.53118396e-01 4.29569274e-01 -1.24043858e+00 7.51638234e-01
6.71101689e-01 -5.21193564e-01 -9.87643182e-01 3.36740553e-01
8.40470970e-01 -3.41411084e-01 -7.73075283e-01 -1.22329406e-01
3.24918628e-01 -9.89532769e-01 5.15238166e-01 -1.09189630e+00
8.30051184e-01 4.11979765e-01 3.16171646e-01 -1.79007804e+00
-4.84333634e-02 -1.15972006e+00 6.42883554e-02 1.35813534e+00
3.93882692e-01 -5.36579072e-01 5.73485792e-01 1.31413615e+00
-2.76275147e-02 -6.38227940e-01 -4.70995188e-01 -3.60328853e-01
3.57242346e-01 -1.58516705e-01 4.38309103e-01 1.13600993e+00
1.17039275e+00 1.13706148e+00 -8.68131459e-01 -3.92170608e-01
-1.96285531e-01 3.04470718e-01 9.51416552e-01 -1.22614515e+00
-4.91536707e-01 -4.43237692e-01 2.94176638e-01 -1.12698364e+00
7.94390917e-01 -5.31354547e-01 5.69922805e-01 -1.19641089e+00
5.52965514e-02 -9.27270502e-02 2.23636344e-01 7.33342528e-01
-3.94908130e-01 -2.12232366e-01 2.42345259e-01 -3.59737098e-01
-6.77126169e-01 6.97572768e-01 1.12585783e+00 -8.16469193e-02
-4.86195058e-01 -5.74961863e-02 -9.39437091e-01 6.65593028e-01
1.00461709e+00 -2.84423023e-01 -3.01113337e-01 -9.23807845e-02
1.67586014e-01 6.19867504e-01 -2.33614445e-02 -4.00650412e-01
1.41004220e-01 -7.24503994e-01 -2.92143077e-01 -1.46309203e-02
5.02740145e-01 -4.17610884e-01 -1.94772944e-01 1.75953716e-01
-1.03767705e+00 -1.91949904e-01 4.48415279e-02 2.08296984e-01
-3.32640946e-01 -3.00088555e-01 8.22298467e-01 -3.98450881e-01
-4.94694442e-01 -1.01238392e-01 -1.11666632e+00 1.11796990e-01
9.92533326e-01 2.13381663e-01 -1.97103575e-01 -1.38925302e+00
-7.36980855e-01 5.98600328e-01 2.54125446e-01 6.63692176e-01
3.07917029e-01 -9.62834477e-01 -6.51463687e-01 -5.20747244e-01
-5.46454685e-03 -3.86332899e-01 3.72975320e-01 5.29803216e-01
-6.03952482e-02 4.59857643e-01 -3.53789210e-01 -1.39405251e-01
-1.55904698e+00 4.06331241e-01 4.84158665e-01 -7.82318413e-01
-1.65870428e-01 9.85346615e-01 2.89680719e-01 -8.46688688e-01
2.11501747e-01 -6.69129863e-02 -3.24368387e-01 6.49652630e-02
4.56227481e-01 -5.12949042e-02 -2.04695761e-01 -3.05927157e-01
1.36534479e-02 -2.03089327e-01 -1.66915320e-02 -5.35310805e-01
1.26235664e+00 -2.68842429e-02 -2.54329741e-01 4.76624578e-01
8.02776515e-01 -4.93788123e-02 -1.14411807e+00 -1.15880489e-01
9.87583920e-02 -1.65855954e-03 -5.89676321e-01 -1.19713628e+00
-5.70966184e-01 1.09522641e+00 -1.51698306e-01 3.76127571e-01
9.13933337e-01 -1.49256960e-01 1.13341010e+00 9.93506312e-01
2.62952745e-01 -1.45840788e+00 5.71442425e-01 1.30232048e+00
8.60938489e-01 -1.32673538e+00 -7.16054082e-01 -3.48439544e-01
-1.69305086e+00 1.12168682e+00 1.18567252e+00 -7.69142713e-03
-5.54339290e-02 4.09493864e-01 6.74112558e-01 -6.96046948e-02
-1.52945960e+00 -1.18825911e-02 -3.07632446e-01 5.49964011e-01
8.18041682e-01 1.84986234e-01 -3.83555830e-01 9.69153285e-01
-3.72496158e-01 -1.31151170e-01 7.24196076e-01 7.26052284e-01
-4.40473080e-01 -1.41075718e+00 -2.64771312e-01 2.05655470e-02
-3.90635967e-01 -2.04780083e-02 -1.50493991e+00 3.80671501e-01
-5.11037350e-01 1.55126929e+00 -2.16709629e-01 -4.65308905e-01
2.80926228e-01 8.15772295e-01 2.77420580e-01 -6.09958231e-01
-1.56264079e+00 -2.86423475e-01 7.71132469e-01 -4.73726898e-01
-2.25548729e-01 -3.09018850e-01 -1.54985905e+00 -2.58497894e-01
-1.75667912e-01 6.25554025e-01 3.22791159e-01 1.02149987e+00
1.92360312e-01 2.67787129e-01 9.79925811e-01 -7.11962104e-01
-8.33869994e-01 -1.46529472e+00 -1.59742609e-02 7.25051224e-01
-7.03098848e-02 -1.26851410e-01 -1.55236587e-01 1.35367345e-02] | [13.086043357849121, 7.701837539672852] |
5636bf14-bb60-4c02-8ff3-28bd97295c22 | efficient-localness-transformer-for-smart | 2203.16537 | null | https://arxiv.org/abs/2203.16537v1 | https://arxiv.org/pdf/2203.16537v1.pdf | Efficient Localness Transformer for Smart Sensor-Based Energy Disaggregation | Modern smart sensor-based energy management systems leverage non-intrusive load monitoring (NILM) to predict and optimize appliance load distribution in real-time. NILM, or energy disaggregation, refers to the decomposition of electricity usage conditioned on the aggregated power signals (i.e., smart sensor on the main channel). Based on real-time appliance power prediction using sensory technology, energy disaggregation has great potential to increase electricity efficiency and reduce energy expenditure. With the introduction of transformer models, NILM has achieved significant improvements in predicting device power readings. Nevertheless, transformers are less efficient due to O(l^2) complexity w.r.t. sequence length l. Moreover, transformers can fail to capture local signal patterns in sequence-to-point settings due to the lack of inductive bias in local context. In this work, we propose an efficient localness transformer for non-intrusive load monitoring (ELTransformer). Specifically, we leverage normalization functions and switch the order of matrix multiplication to approximate self-attention and reduce computational complexity. Additionally, we introduce localness modeling with sparse local attention heads and relative position encodings to enhance the model capacity in extracting short-term local patterns. To the best of our knowledge, ELTransformer is the first NILM model that addresses computational complexity and localness modeling in NILM. With extensive experiments and quantitative analyses, we demonstrate the efficiency and effectiveness of the the proposed ELTransformer with considerable improvements compared to state-of-the-art baselines. | ['Dong Wang', 'Lanyu Shang', 'Ziyi Kou', 'Huimin Zeng', 'Zhenrui Yue'] | 2022-03-29 | null | null | null | null | ['non-intrusive-load-monitoring', 'non-intrusive-load-monitoring', 'non-intrusive-load-monitoring'] | ['knowledge-base', 'miscellaneous', 'time-series'] | [ 3.71629685e-01 -2.45153040e-01 -5.52945435e-01 -3.39095533e-01
-9.67294812e-01 -5.13254583e-01 4.95510578e-01 8.93364772e-02
1.63276456e-02 4.64540809e-01 5.42793751e-01 -1.04977451e-01
-2.58051157e-01 -8.03620756e-01 -6.87126577e-01 -7.63189435e-01
-6.49546972e-03 1.59005076e-01 -7.16997981e-01 1.92562919e-02
-2.49819189e-01 2.65611380e-01 -1.28413296e+00 2.15046778e-01
8.62152517e-01 1.42932796e+00 2.24018142e-01 5.25250137e-01
9.22897309e-02 1.13888741e+00 -5.42210579e-01 -7.25245625e-02
8.80780220e-02 -4.37693387e-01 -6.32031024e-01 -3.68789583e-01
-3.37590650e-02 -6.00361943e-01 -4.51301962e-01 1.15047705e+00
8.78231227e-01 -3.12577076e-02 3.94913733e-01 -1.47343946e+00
-6.81921959e-01 1.33430457e+00 -6.14614308e-01 3.61061931e-01
4.41017777e-01 3.58809739e-01 1.26055205e+00 -4.16970760e-01
-1.88245803e-01 7.02274323e-01 8.07206571e-01 1.13231331e-01
-1.50431633e+00 -7.95307815e-01 2.00704068e-01 7.75766730e-01
-1.37984264e+00 -5.67881584e-01 1.20814943e+00 5.54492064e-02
1.80380332e+00 6.97941899e-01 5.88856876e-01 1.43022633e+00
2.24753261e-01 1.03676140e+00 9.44369018e-01 -1.30825758e-01
4.60545570e-01 -3.40920419e-01 7.24346116e-02 2.33863637e-01
4.04205546e-02 -2.39463802e-03 -6.78005040e-01 -1.61903307e-01
3.79797071e-02 2.31318697e-01 -3.74859542e-01 1.86480552e-01
-8.96715999e-01 6.26034796e-01 5.12522280e-01 5.55408299e-01
-7.57418633e-01 3.74650210e-01 5.56860328e-01 -1.05930157e-01
4.39762264e-01 3.47305924e-01 -7.17416465e-01 -6.16690814e-01
-1.12170696e+00 -4.49928671e-01 7.63124585e-01 8.32745016e-01
4.58353490e-01 3.63073647e-01 -4.22946930e-01 6.73569322e-01
7.14575639e-04 8.51530313e-01 8.76239121e-01 -7.79235423e-01
6.00919187e-01 6.22348905e-01 -2.92474955e-01 -6.13503456e-01
-8.93890381e-01 -4.63014275e-01 -1.33682704e+00 -3.72440040e-01
-2.81683713e-01 -1.29052237e-01 -5.66190064e-01 1.96036315e+00
-1.13985598e-01 4.55682516e-01 -1.95233494e-01 4.91585553e-01
4.11129326e-01 9.47672307e-01 2.56724566e-01 -7.18999803e-01
1.32172787e+00 -6.34277999e-01 -1.18183994e+00 -1.43248990e-01
4.31931525e-01 -2.18236476e-01 1.14117336e+00 2.76670545e-01
-9.63074327e-01 -2.60789037e-01 -1.15503979e+00 -7.89299160e-02
-4.58061755e-01 8.75941943e-03 4.66082752e-01 5.92371702e-01
-6.18701518e-01 6.56922817e-01 -1.27899194e+00 -2.27957606e-01
6.46273077e-01 3.52388620e-01 2.44130030e-01 1.50154397e-01
-1.22384620e+00 6.76166952e-01 2.94588149e-01 1.83039829e-01
-6.66991770e-01 -1.13616514e+00 -8.64802837e-01 4.24395204e-01
3.30178678e-01 -4.67624426e-01 1.26619911e+00 -4.52572763e-01
-1.59708393e+00 2.03074887e-02 -2.72914708e-01 -8.23638320e-01
8.47752616e-02 -3.23690176e-01 -7.36719072e-01 -2.01524347e-01
6.43307418e-02 1.30617455e-01 5.54412425e-01 -7.61167765e-01
-5.92723429e-01 -5.01647949e-01 -3.80740404e-01 -6.25711828e-02
-6.88916266e-01 -5.05197048e-01 -2.14616284e-01 -7.85156429e-01
-2.44206503e-01 -6.81270123e-01 8.16700310e-02 -8.74476969e-01
-7.80616581e-01 -2.69663006e-01 9.55280006e-01 -8.59130263e-01
1.80397117e+00 -2.03103900e+00 -1.86820284e-01 2.94907779e-01
1.01732917e-01 9.21959355e-02 -1.55798689e-01 3.45205903e-01
-1.81112096e-01 -7.65513852e-02 -2.95068949e-01 -5.06940961e-01
6.88392103e-01 3.46870571e-01 -3.12474638e-01 4.66905206e-01
-2.31155813e-01 1.77436864e+00 -7.38518655e-01 -1.61048397e-01
7.15502620e-01 6.96464658e-01 -1.72878593e-01 1.46621689e-01
-2.04460979e-01 8.42855871e-02 -2.02915952e-01 7.43357301e-01
3.71834248e-01 -5.40632069e-01 3.07552785e-01 -9.80324388e-01
1.66155443e-01 6.97333813e-01 -7.20699787e-01 1.77556694e+00
-1.03239524e+00 5.33687592e-01 -7.56199956e-02 -1.02142525e+00
2.47211501e-01 3.77389967e-01 1.24521244e+00 -1.26635337e+00
3.57652962e-01 -1.15587726e-01 -3.73753667e-01 -3.12333226e-01
2.80799717e-01 2.27961585e-01 -3.46683592e-01 4.66045678e-01
-7.03358278e-02 4.10943091e-01 7.67123401e-02 -1.04645140e-01
1.48330820e+00 -8.53736475e-02 5.50630689e-01 -2.25836605e-01
2.65454262e-01 -4.88183349e-01 6.94488645e-01 5.90110183e-01
-1.25760004e-01 8.17740988e-03 6.32160576e-03 -1.25073269e-01
-6.39119446e-01 -9.53155816e-01 -1.45314366e-01 1.22426307e+00
-1.30481988e-01 -6.74080849e-01 -6.61692977e-01 -5.86006820e-01
1.34806678e-01 1.30099523e+00 -4.79306191e-01 -3.19179326e-01
-6.52837217e-01 -1.30601990e+00 5.16910493e-01 9.00822163e-01
6.29636228e-01 -9.60453033e-01 -8.80932748e-01 6.27641201e-01
-4.46542680e-01 -1.17164636e+00 -8.31990778e-01 7.63713658e-01
-4.73169267e-01 -8.37307274e-01 -1.65873632e-01 -8.59556720e-02
1.46550447e-01 -3.99374664e-02 1.40612388e+00 -6.76987708e-01
-3.18108857e-01 4.43492740e-01 -1.90768436e-01 -3.69009018e-01
2.07066480e-02 4.24170911e-01 8.44056755e-02 7.38392845e-02
8.87174070e-01 -1.17415559e+00 -7.43124127e-01 4.27778885e-02
-7.39473104e-01 -2.09079355e-01 6.64725661e-01 5.51007986e-01
8.66857111e-01 5.48546791e-01 5.71202219e-01 -6.77063227e-01
6.81001306e-01 -6.58427596e-01 -6.21339142e-01 7.94529170e-02
-1.16465890e+00 1.17836706e-01 1.01462281e+00 -3.90567809e-01
-6.94403946e-01 4.16047499e-02 -2.28857309e-01 -5.27558029e-01
3.32709104e-02 2.42302313e-01 -6.20266855e-01 3.12546849e-01
1.64393768e-01 4.89542305e-01 -7.28496194e-01 -6.63893163e-01
4.35420007e-01 5.97117126e-01 7.79134870e-01 -2.33582661e-01
5.99311054e-01 2.07367092e-01 5.99214658e-02 -5.84456623e-01
-6.92838252e-01 -2.67697841e-01 -4.35949303e-02 1.91766366e-01
8.36796582e-01 -9.98419762e-01 -1.37362766e+00 4.14272070e-01
-8.25258613e-01 -4.65762347e-01 -8.55171323e-01 2.82056689e-01
-5.40773690e-01 1.30004391e-01 -6.35407984e-01 -9.63916004e-01
-8.93640876e-01 -1.00110495e+00 1.27311373e+00 -1.06921174e-01
-5.51725864e-01 -8.06786001e-01 -2.05044940e-01 2.71053821e-01
6.73071265e-01 3.73077542e-01 8.70270848e-01 -5.55332303e-01
-4.62241650e-01 -9.99534130e-02 7.43577182e-02 2.01625884e-01
5.71313620e-01 -9.26787734e-01 -1.21196628e+00 -3.84757787e-01
3.14323664e-01 -3.87475230e-02 6.09836698e-01 6.32714212e-01
1.62621665e+00 -9.06566024e-01 -3.73516560e-01 7.38042116e-01
1.53798461e+00 2.68600881e-01 4.49958891e-01 -1.40880451e-01
9.33223844e-01 -1.79752350e-01 -3.87789831e-02 7.94345260e-01
6.65526688e-01 7.50454843e-01 3.46946090e-01 -9.57178101e-02
-1.35226697e-01 -3.87167871e-01 5.97469091e-01 1.19491541e+00
4.56699818e-01 -5.20657539e-01 -5.20267487e-01 6.00353301e-01
-1.76261783e+00 -1.03268063e+00 2.64408082e-01 1.87837374e+00
9.19892132e-01 8.47487822e-02 1.04810253e-01 4.69974607e-01
1.55935556e-01 4.27172929e-01 -1.26138985e+00 -2.11339712e-01
-2.22632572e-01 2.09989443e-01 8.56925964e-01 4.47111040e-01
-1.01425242e+00 1.00048050e-01 5.55911493e+00 1.06907046e+00
-8.66340458e-01 6.15695059e-01 5.35759091e-01 -6.31126225e-01
-3.16015720e-01 -7.72665083e-01 -6.69471264e-01 1.26495910e+00
1.40342510e+00 3.01902965e-02 1.10021889e+00 5.90214074e-01
4.65166330e-01 8.83475095e-02 -1.48480797e+00 1.51567566e+00
3.29970010e-02 -9.51923966e-01 -4.10195291e-01 1.57006070e-01
7.38745987e-01 5.37334442e-01 -7.05620050e-02 2.37119332e-01
4.68296260e-01 -1.07392561e+00 4.54223335e-01 5.49254894e-01
5.80290258e-01 -8.09156001e-01 8.15218031e-01 3.11144173e-01
-1.81611907e+00 -5.54431260e-01 2.71065176e-01 1.74539149e-01
4.13238615e-01 1.05198216e+00 -4.65465605e-01 7.39678383e-01
1.02450228e+00 7.83348739e-01 -3.05038005e-01 1.96032986e-01
1.22234896e-02 9.59239542e-01 -9.48758543e-01 4.97577786e-02
-7.69333243e-02 -1.08122140e-01 3.43194902e-01 1.09261394e+00
4.06813115e-01 1.23830147e-01 1.89483300e-01 8.63779604e-01
-3.04470897e-01 -2.07608253e-01 -3.97510827e-01 -4.46641259e-02
7.38797069e-01 1.10855925e+00 -8.29852670e-02 -2.08978727e-01
-6.05822325e-01 1.13686240e+00 -1.76198721e-01 4.26065534e-01
-1.04511321e+00 -2.29490325e-01 9.43622887e-01 -3.39654461e-03
4.52156842e-01 2.21118510e-01 -5.98132968e-01 -1.05681229e+00
2.01138362e-01 -8.13812375e-01 5.16421199e-01 -6.34156942e-01
-1.62067044e+00 3.39725971e-01 -9.34977010e-02 -8.63000154e-01
-5.33590496e-01 -1.66009422e-02 -4.83038038e-01 6.31979644e-01
-1.40558708e+00 -1.24055207e+00 -3.03222507e-01 6.73694551e-01
6.46066546e-01 1.21905006e-01 8.08237374e-01 5.13779700e-01
-8.70401442e-01 8.89674544e-01 4.07378137e-01 -2.12878454e-03
4.77982759e-02 -1.43423212e+00 5.12356400e-01 8.03444564e-01
2.73558885e-01 8.46422017e-02 6.52094841e-01 -4.12916720e-01
-1.90731907e+00 -1.48391545e+00 6.00296855e-01 -5.73189378e-01
7.37447023e-01 -6.04515910e-01 -6.38144970e-01 6.69375300e-01
5.42665124e-01 -9.13878158e-02 8.39925110e-01 -4.54042852e-02
-1.66012794e-01 -7.56555080e-01 -1.21924305e+00 2.59032756e-01
1.01414919e+00 -8.46217215e-01 -2.49373361e-01 3.32865179e-01
5.07110775e-01 6.12514606e-03 -1.42246068e+00 5.92074215e-01
4.56945479e-01 -5.64926267e-01 9.82581496e-01 3.51966843e-02
-1.40807420e-01 -4.01060969e-01 -6.87138438e-01 -1.24759662e+00
-5.93490422e-01 -1.00926232e+00 -1.16178310e+00 1.71755493e+00
2.27984399e-01 -8.26588571e-01 5.36965370e-01 6.21905684e-01
5.37302680e-02 -8.61996949e-01 -1.19422579e+00 -7.45153189e-01
-2.68118918e-01 -9.05858338e-01 1.36064053e+00 6.88254118e-01
3.09200585e-01 4.70644474e-01 -5.19844532e-01 3.49420875e-01
6.98096633e-01 3.58869284e-01 1.90885767e-01 -8.49922776e-01
-5.48249245e-01 -5.24556398e-01 -2.27423668e-01 -1.15629804e+00
2.43470654e-01 -9.08673406e-01 1.54565006e-01 -1.35057092e+00
1.57119021e-01 1.34222537e-01 -9.05102551e-01 9.49331045e-01
1.89424921e-02 4.89792883e-01 2.40185857e-01 2.22067800e-04
-5.47946453e-01 7.82277048e-01 2.96831042e-01 -7.20002711e-01
-3.46744686e-01 -2.00391620e-01 -8.71353805e-01 4.85034108e-01
1.18105996e+00 -2.98008054e-01 -6.05884910e-01 -2.15208545e-01
1.46083266e-01 -2.82710880e-01 1.39561042e-01 -9.62677956e-01
2.07699016e-01 1.25281438e-01 5.89815259e-01 -8.69120002e-01
3.76704514e-01 -1.24896216e+00 4.81739640e-01 3.78586709e-01
-1.28203362e-01 2.16179773e-01 1.94396213e-01 5.77200174e-01
2.75092751e-01 3.97532731e-01 4.22471792e-01 2.49012873e-01
-5.79592228e-01 3.64823997e-01 -2.85946548e-01 -1.47520021e-01
7.99801171e-01 2.22323537e-01 -4.04355526e-01 -3.15671325e-01
-3.86480570e-01 5.31390131e-01 1.02407828e-01 4.94085073e-01
4.25819233e-02 -1.80496299e+00 -3.77751648e-01 2.81292647e-01
-3.35902050e-02 -2.83377200e-01 3.35839689e-01 1.06629980e+00
3.64722937e-01 7.20068574e-01 5.05282283e-01 -5.78571439e-01
-8.57523382e-01 8.41799617e-01 3.47100288e-01 -5.97822130e-01
-8.08016360e-01 2.02351421e-01 6.71230927e-02 1.77104045e-02
3.46442878e-01 -8.55427086e-01 1.21908456e-01 1.20681442e-01
5.81164956e-01 8.27772081e-01 4.94561821e-01 -5.80496371e-01
-5.74709177e-01 6.04493260e-01 2.73138314e-01 3.87948304e-01
1.35238588e+00 -4.73971486e-01 -3.38824354e-02 5.75695455e-01
1.46296442e+00 -1.73686117e-01 -1.11797297e+00 -1.16324939e-01
-1.12879679e-01 1.46658327e-02 6.48512602e-01 -1.28704977e+00
-1.54629612e+00 4.68780041e-01 1.04397416e+00 5.41926205e-01
1.99410307e+00 -2.17723139e-02 1.41985261e+00 2.29291126e-01
3.07689428e-01 -1.28699589e+00 -6.21801615e-01 -5.15152439e-02
6.16481900e-01 -9.63663936e-01 9.28731170e-03 1.72121286e-01
-2.08531693e-01 3.41427475e-01 2.04278573e-01 4.88012552e-01
8.92998636e-01 8.99181247e-01 -2.17177510e-01 -5.68736251e-03
-7.95095026e-01 -2.20637441e-01 2.20924526e-01 5.50627768e-01
5.27480766e-02 4.28949803e-01 9.74777266e-02 1.03954947e+00
-2.65196770e-01 1.58863649e-01 -1.26317918e-01 7.17822194e-01
9.46689770e-02 -7.83858478e-01 -2.45698959e-01 8.67648005e-01
-5.64399600e-01 -2.31990516e-01 8.78834203e-02 1.73557431e-01
1.76982433e-01 1.38457727e+00 8.85066763e-02 -5.40195644e-01
6.01618707e-01 2.63068557e-01 3.27948332e-01 1.91715956e-01
-5.99133015e-01 1.40243560e-01 -1.27337486e-01 -1.10517490e+00
-2.48827174e-01 -5.64734578e-01 -9.94560242e-01 -5.36439478e-01
-1.81047678e-01 -1.88154131e-01 4.53041613e-01 9.56422865e-01
7.52325475e-01 1.09232342e+00 9.42885399e-01 -9.48966205e-01
-7.00698674e-01 -1.11099148e+00 -7.11955786e-01 4.67837214e-01
4.92598891e-01 -2.52441615e-01 -1.51182458e-01 -1.05795518e-01] | [16.069480895996094, 7.582119464874268] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.